Loudspeaker engineering : How to design speaker crossovers | Tanasescu Marius | Skillshare

Loudspeaker engineering : How to design speaker crossovers

Tanasescu Marius

Play Speed
  • 0.5x
  • 1x (Normal)
  • 1.25x
  • 1.5x
  • 2x
36 Lessons (4h 52m)
    • 1. 0.1 - Promo

      1:43
    • 2. 0.2 Contents

      1:54
    • 3. 1.1 Introduction

      3:03
    • 4. 1.2 Crossover components

      3:13
    • 5. 1.3 Capacitors

      6:03
    • 6. 1.4 Inductors

      4:27
    • 7. 1.5 Objective

      5:37
    • 8. 1.6 Basic Circuit diagram

      4:12
    • 9. 2.1 Diagram basics

      11:09
    • 10. 2.2 Attenuation L pad

      5:03
    • 11. 2.3 Baffle step compensation

      5:38
    • 12. 2.4 Zobel network

      3:43
    • 13. 2.5 Series notch filter

      4:58
    • 14. 2.6 Parallel notch filter

      3:47
    • 15. 3.1 Project 1 : Presentation and speaker selection

      6:33
    • 16. 3.2 Project 1 : Enclosure description

      2:29
    • 17. 3.3 Project 1 : Measurements

      8:43
    • 18. 3.4 Project 1 : Crossover design

      15:29
    • 19. 3.5 Project 1 : Prototype crossover

      5:43
    • 20. 4.1 Project 2 : Description

      10:26
    • 21. 4.2 Project 2 : Measurements

      7:09
    • 22. 4.3 Project 2 : Crossover design

      12:47
    • 23. 4.4 Project 2 : Prototype crossover

      4:46
    • 24. 5.1 Project 3 : Enclosure presentation

      7:14
    • 25. 5.2 Project 3 : Crossover design

      19:54
    • 26. 5.3 Project 3 : Ladder delay network

      6:41
    • 27. 6.1 Project 4 : Description

      7:37
    • 28. 6.2 Project 4 : Crossover design

      21:00
    • 29. 7.1 Project 5 : Description

      14:51
    • 30. 7.2 Project 5 : Crossover design

      20:31
    • 31. 8.1 Component Voltage and Power

      5:05
    • 32. 8.2 Component values

      6:09
    • 33. 8.3 Inductor placement

      5:18
    • 34. 8.4 Crossover build bass side

      16:04
    • 35. 8.5 Crossover build tweeter side

      17:44
    • 36. 8.6 Finishing touches

      5:26

About This Class

Design you own speaker crossover

This course is the last piece in the loudspeaker design process. You will learn how to use the FRD and ZMA files (check Acoustics 201 course) to design 2-way and 3-way crossovers. These designs will be done in XSim which is a free application. Crossover design, while it does have some general guidelines, it's unique for every crossover. For this reason, 5 different types of enclosures have been built for this course : two 2-way bass reflex bookshelf speakers, a 3-way bass reflex, a sealed MTM and a 3-way bass reflex floorstanding speaker with dual woofers. This way you will get a better understanding on how the process works. It takes not only knowledge, but also intuition and experience when designing a crossover.

Specific circuits for crossover design

Besides the basic filtering for speakers (1st order, 2nd order, etc), there are other circuits which are implemented to correct the frequency and the phase response of the system. Since we have 5 examples to play with, we have enough crossover projects to go through most of these in a practical way. To mention some of the circuits : impedance equalization circuit, attenuation pad, baffle step compensation, ladder delay network, notch filter, and more.

Testing the new crossovers

Designing a crossover on your computer is convenient but we will also test it out. I have a large amount of electrical components, capacitors, inductors and resistors, from small to large values. After we design the crossover in XSim, we will assemble a prototype crossover and check how it measures and how it sounds. Some tips and tricks on how to make your life easier and what to avoid when building a prototype.

Building a crossover network

Building a crossover network will require some handy work. You will need to be aware of the size and material of the board, how and where you place your components. What will you use to fix the components to the board. How to make sure the components won't rattle about. Unavoidably you will acquire some basic soldering skills as well.

At the end of this course, and by taking the previous 2 courses as well, you will be able to fully design from scratch your own multi-way loudspeaker.

Transcripts

1. 0.1 - Promo: Hello and welcome to the third course in the acoustic Siri's, which will be about crossovers. I'm Matias. You should know me from acoustics one on one and acoustic store. One, of course, is if not, then you should take those courses beforehand because otherwise you will have difficulties in the lessons which are about to follow. So in this course we will learn how to design a crossover network. And since crossover design is unique to every project, I will try to cover all the possible scenarios by showing you five examples. So we have ah to bookshelf speakers, which will be two way. We will have Ah, three way speaker on a sealed MTM in a three way based based reflects ah, floor standing speaker with dual wolfers. Now, besides the basic filtering from the crossover design, we will also implement different circuits to reach our goal. And our goal is to have a linear frequency response and the good face coherence between the drivers. And you will see something like attenuation pads notch filters to leather delay networks on DMA anymore. Now, after you finish this course since acoustics, one on one and 21 courses are mandatory. Then it's safe to say that you can design your own loudspeaker projects from scratch. So in that case, I wish you good luck and have fun with your lot speaker projects. 2. 0.2 Contents: First of all, let's try to summarize the structure of this course. In the beginning, we will talk about the passive electrical components, the resistors, capacitors in doctors. I will describe how they function, what to look for in which one you should big for audio applications. In the next section, we will take a look at different circuits, which we will implement. Besides the filters, these are used to better shape the frequency response. So, for example, toe bring the output off speaker lower eso. It matches a second speaker dame a peak in the frequency response. So the overall response is more linear and so on. After that, we will take a look at our first speaker project. First, I will give my insight on how I chose the speakers. Secondly, we will look at the enclosure. Then we will discuss the speaker measurements and how we are going to use them. Finally, we will design the crossover indexing. When that is finished, a prototype crossover will be made just for test. We will take a measurement and compare how the measurement looks, in contrast with the calculation from except four more projects will follow in a similar fashion, this time with more complex speakers, including on MTM and a three way for standing speaker with dual wolfers. At the end, we will actually start building across a report. We will follow the circuit diagram and actually assemble the cross of our network. I will show you how to place the components on the board and what to look after, how to solder properly. And finally, I will show you how to fix the board to the enclosure. So now that you know what to expect, let's start with the first step. 3. 1.1 Introduction: in this first video, it will be like an introduction for the things which are about to follow. So, uh, if you remember from acoustics one no. One at the end of the course, I talked about the basics off how to design some crossovers. So we learn about the different that apologies. And, ah, names like a link quits Riley or Butterworth. And you can you can safely ignore all of that because those use formulas and consider the speaker like a resistor. Resistor has fixed the impedance. But as we know the driver, the speaker driver has a variable impedance, and it's all over the place. You have the spiking impedance at the resonant frequency, and then you have the rising impedance as the frequency increases. So, uh, all of those formulas are basically incorrect. So, um, to get around this, we use software and, um, you still here? Ah, these names. So if you, um if someone says I designed the second or third language, really, you basically know that the responses has have a summed up flat. So if the crossover point is six decibels below the linear frequency response, then you can consider that a link quits right when you say liquids. Riley were Butterworth you are actually referring toe the shape of the knee. But in the rial case scenario, you look at where the crossover point is compared toa the linear frequency response. So if you see a bump, then that there should correspond toe a Butterworth response because the summation adds two plus three db over the linear response now. Ah, In this first part, we're also going to talk about the passive components. I'm going to give you some basic description off each component and what should you choose for audio application? And also, we will take a bit. We will take a look at Exim and uh um, I will make a short briefing about the electrical circuits, and many of you might skip this because it's really, really basic stuff. But I have to cover it, since there might be someone who doesn't know the basic off electrical circuits. And that should be. Then let's move on 4. 1.2 Crossover components: Now let's talk about the components you will find in a crossover network. Since ah, this crossover, it's called a passive crossover. You will find some passive electrical components, and we will use resistors, capacitors and in doctors, and we will use them in different positions. Toe. Achieve our goal. Now, if we start with the capacitor, the capacitor filters low frequencies, so you will find the capacitors in series with a Twitter, for example, and depending off. Depending on the value of the capacitors, which is measured in the micro Farage's and the impedance off the tweeter, you can determine which frequencies are rolled off, so these are the two factors in determining this. And since the impedance is not a flat line, is not a fixed number. It's a big complicated, and that's why we need the help off Exim. Our software, which we're going to use if we talk about in doctors. They are the other way around, so they feel they're high frequencies and let the low frequencies pass. They are measured in merely Henry's and depending on the on the value of the doctor and the impedance of the driver, some frequencies are him rejected and some not. And the last component, the resistor, um, the resistor filters all frequencies and the the amount off off. How much reduction in output you get depends on the value of the resistor. And ah, since the resistors are not a fancy component, let's talk more about them. Besides the rating, which is measured in homes, you will find the power power handling rating, which is measuring watts. Some might be five watts. Then what's 25 watts You can see like big ones with, He thinks on them toe resist high power demands. And ah, it's good because we can check this power handling in Exim. So when we finish when we finish a circuit diagram, we can check various components from the circuit, and then we see the power demand for each component depending on the frequency, because not or not all frequencies are treated equal. So let's say you have a high power demand that 40 hertz, but like nothing for ah 1000 hertz for the same resistor, so that that should cover all about the resistors and we'll talk more about in doctors and capacitors and what the types of them we should use for audio applications 5. 1.3 Capacitors: No, let's talk about the pastors. So we said at the beginning that capacitors filter low frequencies, so they are used for to protect the tweeters. But how does the capacitor look inside? So if you take the basics of a court off capacitor, you will see they have two plates which conduct electricity and sandwich between them. You will find the insulating material which here blocks current flow. So in a set up where you have direct current D. C, you will find that the positive charge will align on one plate and the negative charge will align on the other plate. And they cannot past one another because we have that insulating material between the plates. If you take that capacitor away from the from from the source, you still have charged the capacitor. That's why you hear. Oh, you're gonna zap yourself if you're gonna touch a capacitor, even if the power supplies not on. That happens Onley in D. C in a sea. However, depending on the frequency and on the value of the capacitor, some frequencies do cross over, so you have the east rating material. But despite of that, some frequencies pass over to the other side and some do not. So that is the filtering mechanism of the capacitor. Now we have different types of capacity. We have many types of capacitors. But for all your applications, I suggest you use Onley. You basically choose between the cease to you have electrolytic away and ah, film film capacitors So you can see here I have ah film capacitor and, um, these are considered superior in audio quality compared toa electrolytic and ah, they have ah many advantages as well, but they also have disadvantages compared to the inferior counterpart. So, uh, the production off this capacity where you basically have a sheet off conductive material and a sheet off insulating material and you roll it off and basically make this capacitor and as the capacities value increases, you will find that this capacitor can reach really high, uh, sizes. So this is a 39 micro fair of the capacitor and is considered quite big. Imagine if this capacitor would have bean 100 the microphones. So this is in ah, electric capacitor. And this is a 220 micro fair, and this is a 39 so you can understand the size difference. So size is an issue here. Prices also when you should, because this is way more expensive than on electrolytic. And another thing that you should taking into consideration is that these capacitor has a high voltage rating. So it's not uncommon to see 400 volts like on this one or lesser ones might have 250 but that is also sufficient. If you look at the elect, the electrolytic, you will find that this is rated at 60 free votes, which is pretty low. You might find us some which I rated higher. Maybe it 100 volts, some lower, maybe had 50 volts. So you have to take into consideration the voltage rating as well. We will check that way, will check that in Exim so we can see if this capacitor is suitable for our application. If not, then we have toe bump upto something which is more stronger if the voltage rating is exceeded. So if we take, uh, let's see the D. C. Example, we have the plates and the charges which which are attracted to one another but can't pass for the insulating material if you exceed the voltage rating So you bump up the power supply, you will find that people rip for the insulating material, and it will basically ruin the capacitor. So just remember, these are considered superior, as in audio quality. But don't get toe hung up on this because both are very good. If you're not designing really, really high in the speakers, then both should be fun. You have to consider the prices Well, because as the capacitance value increases, the price goes up considerably. And these are for electrolytic ALS. Uh, you will find that the even if you increase the value very much, the price doesn't go up. Uh, that much so you have to take all factors into consideration, so that should be it for the capacitors. 6. 1.4 Inductors: Finally, we're gonna talk about conductors and the types we use for audio applications. We have, ah, Iran core. We have Air Corps and for really fancy fancy, we have, ah, foil foil in doctors, which are really, really expensive, and we're not going to use them. But let's talk about the Iran core I have here one which is Solent cylindrical. But most of them are have a square core and the wind ings over the Iran core. And, ah, the qualities off these type of in doctor is that it's very cheap because since it uses this core it the has lesser wind ing's. So when you using an air court, it doesn't have nothing in the middle, you have toe make more wind ings to reach the same induct inst level. Also, because it has lesser wind ings, it has a lower resistance value on in doctors. Since it uses this wire, it has a certain resistance, which at which adds up to the crossover circuit, and you have to take in consideration because it does matter. Actually, every component has an a serious resistance, but when you consider a capacitor, it has really low, serious resistance and basically you disregarded. But it is there. But when we're talking about the in doctors, the resistance value is critical. So if we look at this in doctor, we see it's a 4.7 million Henry, and it has a 0.31 home. And if we think the same, the same in Doctor, This is an Air Corps. AmeriCorps has nothing in the middle of the says. It's basically a plastic casing, but this has no property at all. It's only aesthetic, and it keeps it the altogether. But I can show you another one, which has absolutely nothing. This is also Air Corps, but I'm showing you this one because it has the same inducted. Its value. So 4.7 million Henry's exactly like the Iran core, but they have different resistance values, so this has 0.31 this one has 1.25 homes. So 1.25 soames is really high, really high, its considerable enough that it will affect the euro frequency response. So let's say you want to use an Air Corps because these are considered superior as audio quality compared to the Iran core But let's say the serious resistance is too high and you need a lower one. So you either use Iran core in doctor or you use on air Corps. But you have to increase the gauge of the wire so the thickness of the wire determines the resistance as well. So this have these two in doctors have the same, uh uh, a wire thickness. But if you use ah higher, thicker wire than the resistance will go them, so you have to check all of your parameters. And, of course, price is a n'importe factor. Here is well, because Air course can get the really expensive not to mention foil in doctors. Those are exaggerated, really, really expensive, but they are considered to be the highest quality available. So when it comes to conductors, you have Iran Corps and Air Corps. Of course, you have many other types of in doctors, but when you're designing a crossover, I suggest you check on Lee. These two types 7. 1.5 Objective: what is the objective of this course? So we want to learn how to design speaker crossovers. And when we were designing a crossovers, we are looking for two things. We want to have a very flat frequency response. So the linearity off the frequency responses one goal and the other one. We want good phase coherence between the drivers. So this means that the speakers are in face. Since we know that the speakers have different sizes and the acoustical centers off, the speakers are different, we have toe match them accordingly. So considering that the phase the phase alignment of the speakers are different, we have to match them at the crossover frequency so so that we achieve these goal toe alter the phase responsible speaker you have tow, add different components. So we we learn, we learn from, uh, that parting the acoustics one on one course that let's say, a first order fielder will phase shift by next degrees, a second order by 100 and 80 degrees and and so on, so increasing the complexity off the crossover will disrupt the face. So we have to take this thing into consideration when we are designed to cross over. So we do this to our advantage. Also adding different circuitry to the The crossover design will also introduce some face shift. So we have to take this into consideration because when we are achieving the frequency Reese passed a linear frequency response, we we might reach a really good linear response. But when you look at the face response, it might not be that great. Sometimes. If your project is a cheap one and you don't want to add too many components to your coarse salt crossover, you might accept you might accept that your frequency response looks great, but your face response isn't perfect. And that is OK when the speakers are not correctly aligned in face. You might have some issues with the sounds which are very transient. So, like snares or something like that, that required a really fast movement. Since they are delayed by somewhat in not, uh, the sound doesn't go at the same time. Since the sound is very fast, you might pick up this phase mismatch. So when we are designed to cross over to check if the speakers are aligned perfectly, you reverse the polarity of one speaker and take a frequency response measurement since they are perfectly in phase. When you reverse the polarity, they are perfectly out of phase and you will see a huge deep at the cross over at the crossover frequency. So this phase alignment is important at the crossover frequency because that is the part where both speakers play the same frequency bandwidth. So if you are talking about, let's say, 50,000 hertz, you are not interested that the mid base is in phase with the Twitter at that frequency because they do not overlap at that respective frequency. But at the crossover point, all the frequencies around the crossover point around the crossover point are played by both speakers, so we have to check that both that both speakers are in face. This isn't also an important characteristic when designing a crossover another another important factories that you should have the least amount of components. This none only reduces the cost, but it also improves the audio quality because each component in the signal path alters the signal and audio. We do not want to do that. So having said that, that we should the go ahead and start building our projects. And now, since each crossover design project is unique, there are some basic guidelines. So from the electrical engineering standpoint, But you have to take into consideration that each project is unique and you will have to, ah, take your intuition and experiences or so a big factor. So that's why I taken five example off crossover designs, and by the end of that, you should get a feel off how to design a crossover. 8. 1.6 Basic Circuit diagram: Now, before we get into the actual crossover design, I want to talk about the basics off. Ah, electrical diagrams. So my advice is to skip this part altogether because this is really, really basic stuff. So most of most of you will know this, but I have to cover this anyway, so I'm going to work directly in Exim. So when you use the diagram, you have these three components. This is an in doctor which looks like a coil of wire. This is a resistor, which is a rectangle. And this is the capacitor, which we talked about, which looks like two plates and the insulating material between the two plates. So these our he are our three components and the the speaker driver, of course, but this is specific to our design. These electrical components you will see in any other electrical diagram. So let's, uh, delete this. So when we wire different components together, uh, let's say we want toe add to, uh, components in Siris. If we add them up in Siris, Wait. Go from a positive positive to negative positive to negative. And so these are wired in Siris. And this should go to the other side, so these components are wired in series. If you if you want wired thes two components in parallel, you have to match their polarity signs so minus the minus and plus two plus. So basically, I wired these two components in Paris and when we were designing crossovers, there are two types of crossovers. So we have serious crossovers and parallel crossovers. Uh, Siri's crossovers are really, really rare, and you won't find many of them. And we were going to talk on Lee about Parral crossovers. So we take the signal from the amplifier and we split them up into Let's say we have a two we design, So we have two speakers. So basically, we wired these two speakers in parallel. And now, um, you will find the designs where the minus goes to Ah, the minus off the the source. But ah, I think it's more ah, more elegant to use this ground I can over here. So this means that the the ground is common, so each year component will have a negative terminal which goes to the ground. So instead of having all these wires all over the place, we simply use this ground Icahn, which, uh, indicates that that negative thermal goes to ground. So these are all the same spot. So I think that this should ah, cover some some some basics. I was afraid that the people won't know that this brown signed, but you never know. So Okay, that's that. Let's go to the actual crossover design. 9. 2.1 Diagram basics: So before we get to the actual enclosures and ah, designing the crossovers for them, I want to show some basic stuff about the filters. And then we will talk about different the circuitry which will complement the filters to achieve the linear frequency response. So when you are starting a project, the Exim should look like this. You have the drivers. In this case, we have a three way design. So we have the Twitter up top in the mid in the mid. We have the mid range, and at the bottom we have the buffer. So if we check them, we can see we have the Frd and Zahra May file. And the Hilbert both transformation applied like we talked about in acoustics to one. The Twitter doesn't have a delay, but if you check the mid range in, uh, you can see a small delay for the mid range and a larger one for the wolf, for this is to align them acoustically. I showed you how to do this in acoustics toe a one with a toy design. But for the three way design, I will show you in this course it will be somewhat similar Now, if we look at the frequency response charge BC that we have the three drivers, so s one is to in its three which stands for a speaker Want to entry on and you have the notes over here. So this one is the tweeter and the corresponding s three is the wolf for and ah, we can see their frequency response curves and the blue one is the combined curve. Now, when you start to design the crossover first you will begin by adding the filter sections. So if we're talking about a wolf for you will add Ah, a low best filter so low pass filter is basically undocked. Er, if you wire on in doctor in Siris with the wolf for this is basically a first order filter and you can go from there. So this is the first order. If you want to go higher in order than you add a capacitor in parallel. If you want to goto third order, you add another in doctor in Siris. If you want to go even higher in order, you go for a 44 order and so on. So this is a 44 order for the wolf for in the same fashion for the Twitter. For example, if you want a first order fielder, you start with the capacitor and then in an alternating fashion. If you want to increase the order of the filter, you go with the in doctor in parallel. This is ah, second order. If you go for 1/3 order unique in another capacity re Siri's. And for a four foreigner, you need another in doctor in parallel. And this is a 44 Therefore, the tweeter. Now we'll for the mid range you need the band pass filter. So this is basically ah, low pass filter and a high pass filter wired in Siris. So first, I usually start with the high pass filter. So this means we start with the pastor and then, uh, if you want to increase the order of the crossover, you start adding components in the Fashion week, Dr. Earlier. So this is ah, second order. This is the high pass section. And then if you want to separate by a piece of wire toe, I know which one is which. Er, then you start adding the components for the low pass section, so starting with the doctor and let's say we want a second order. This will mean another capacitor in parallel, so we have a four footer at the top. A second order band passed, so we have ah, second order high pass filter and a second order low pass filter. And together it's actually a second order band pass filter and at the bottom before the wolf, where we have a 44 order low pass filter. So, as you can see, we added these components in the circuit diagram. But if we look at the frequency response chart, we can see that it's all over the place. So frightful screen this. We can see that this looks horrendous. But if you start feeling with the ah values off the crossover, then you will get the right curves. Some people like toe use the formulas, which I showed you in acoustics one a one and have a base start from there so we could try this. We're not gonna use it when we are actually designed the crossover in the five examples in this course. But I'm going to shoot. I'm going to show you how to do that just for for the example sake, because I think it's a waste of time. So let's ah, start with the two way fielder for all sections. So let's delete. Let's delete three extra components. So instead of 44 order, we will have, ah, second order low pass filter and let's do the same for the Twitter section. Let's delete the extra components so we have only a second order filter for ah, every section off the crossover design. So if we start our Excel spreadsheet from the acoustics, wanna one here we can input the the crossover frequency and the nominal impedance off the drivers that this Excel spreadsheet the only calculates for two way filters. But if you have, ah, three way in, like in our case, we can do a neat trick so you can consider the wolf for in the mid range a two way in the mid range and the tweeter another two way. And since you are adding, uh, and since the mid range is in the first case, um ah, hi best fielder and in the second case, a low past here, a low pass filter. When you combine them, you will have a band pass filter, so the impedance off the offer is eight homes. The mid range is also ate OEMs, and the Twitter is forums. So let's start with the wolf for mid range combination. So the wool for is eight homes and the Twitter, and in our case, the mid range is also freedoms. And let's choose the crossover frequency somewhere about the 900 yards and let's go for a second order liquids. Riley and we have the values off the components right over here. So we have, ah, the capacitors at the 11.5 and the in doctors for 2.83. So let's so let's start tuning the components. We have the capacitor at uh 11.5 and the induct ER at through 0.83 million Henry's. And since they have the same impedance value, the components on the mid range side is the same Onley for ah, the high pass section. So the capacitor is you live in 0.5 and the induct ER is 2.3 million Henry's, and now we are going to, uh, check the values for the rest of the components. So let's head back to the Excel spreadsheet. So now the Twitter is the forums, which is the actual Twitter. In the mid base is the mid range and let's increase the crossover frequency two or 3000 hertz. So we look at the liquids Riley section. So the Twitter is 6.63 for the capacitor and 0.42 for the in doctor. So six points. 63. Let's close this sub zero 42 for the conductor. And for the low best section we have, ah, 0.85 3.30 0.85 and 3.32. And as we look at the frequency response chart, we can see we did absolutely nothing. We can see that the slopes are more defined, but still the mid range, which is the yellow A graph we can see. It has a band pass shape. This is a low, best shape, and this is a high pass shape. So the filters did what they need to do. But the problem is, you can see that the levels aren't match. The mid range is way higher than the basis section, so this needs an attenuation pad to lower the level so it matches the based side. The Twitter has adequate ah, level compared to the mid range, but not compare to the Bayside. So again, the twitter needs on attenuation also. So this is why I tell this is a waste of time so you can basically start with of fiddling the components right away. There is no need toe use these basic formulas to calculate. Ah, a starting point. So people considered. Hey, this is a starting point and then you start from here. But I consider the waste of time because it achieves absolutely nothing. So I suggests that you use the default values from Exim, which is 10 microphones for the capacitor and one million for the doctor, and started, stand start right away from there and fiddle with the values. So this is the introductory part for the filters. And now we're gonna discuss about different circuits, and then we're going toe dig deep into the actual crossover design 10. 2.2 Attenuation L pad: now the first thing we're gonna talk about these on a 10 waiter. So basically, when you have a speaker which is louder than the other, so if we look in our chart, we can see that the midrange driver is louder than the base side. And also the tweeter, while it's on a par with the mid range, is also louder than the basic. So in these situations toe equalize the levels. You cannot increase the level off the base. Instead, you have to lower the output off the loudest speaker. So in our case, let's remove the twitter from our circuit. So if I delete this branch over here, the twitter will disappear. So we have the yellow curve, which is the mid range and the black one, which is the base. So the easiest way to attenuate to reduce the output off for the louder driver is to put the resistor in series. So if I delete this section over here and place a resistor, let's see what happens. And if we tune it so if I increase the level of the resistor, you can see the output dropping. But ah, in our case, we can see we get this weird hump over here So sometimes, uh, one syriza resistors. One series resistor is sufficient. But the in this case where we get this were hump we have to use on Anel pad. This means we have to use another resistor in a power parallel l pads stands for this l shape. And ah, this time, if we fiddle around with the values you can see, we get a much flatter response so you can play with the both resistors. So you will find formulas on the Internet on how to calculate these attenuation pads these l pads. But I always do this. You just slap him in the circuit diagram and just play with the numbers and reach, uh, the the sheep who you want because like we saw on the the other lesson the previous Listen , if we follow the formulas, we will get basically nowhere because the drivers are doesn't have ah, fixed resistance. The resistance is variable with frequency, So basically, this looks good. You can see I reversed the polarity of the mid range because in they are out of phase in, cancel each other up. But this is not our folks right now we're forcing on the L pad. Sometimes people use the resisters in front off the filter. So this is on on the speaker side. How how they call it. And you can use it on the amplifier side as well. And you will get ah, somewhat different response. Uh, when you do these crossover networks, you have to try all the possible scenarios and see which one fits you the best. So if we delete all this and try this scenario, so let me move these components over here, so I make room for another resistor. So let's place a resistor here and see how the response looks like this time. As we can see, we didn't get that weird hump. So you have to play with the different positions and see which one works the best. Because, like I said, there is no, uh, there is no textbooks solution. You have to try out and see which works best for your project. So basically here we covered how to attenuate a driver which is louder in output than another driver. So if we hook up the twitter over here, we have to use another attenuation l pad for the twitter as well, because it's also ah louder than the Bayside 11. 2.3 Baffle step compensation: the next simple circuit I want to talk about is the baffle step compensation. So well, we know about the baffle diffraction effect. So when you place a speaker on a baffle, some frequencies get boosted by the baffle, and some do not. So in this case, we have the peerless before and we can check the frequency response on the right. But let's look at the manufacturer spec sheet then then And let's check the that frequency response, because that is measured in an infinite baffle set up. So no diffraction issues. So, Aziz, we follow this chart, we can see that the response starts to roll off somewhere about 150 hertz. And then it goes quite linear until it reaches about the 1 1500 hertz. And then it has a dip until about the three K, and then we see a big rise which suggests the cone break up. So the response is pretty linear. Oppa until about the 1005 100. So now let's get back to our chart, and we can see that this is a big difference. So we're going to ignore from 200 hertz and below because this is a gated far field measurement. But we see that, like in the manufacturer graph from 200 hertz and up, it should be linear. But starting with about 500 we see a noticeable bump that is the diffraction effect of the baffle. And then we can see that it's these peak. He's maintained to about 1500 hertz, and then it has that deep up until three K and the cone break up after that. So, uh, if we look at this example, to be honest, this is, ah, mild example off off baffle step gain. But in some cases, the gain might be significant. So ah, to compensate for this gain, we integrate a bevel step compensation circuit, so that is pretty simple. It consists in a nen doctor in a resistor in peril. So let's delete this in place in doctor and a resistor. Let's wire everything up. So basically, in this circuit already, it looks pretty good by now. But in this circuit, the induct er dictates the frequency from which on the attenuation happens and the resistor determines the amount off attenuation. So to see this effect, let me increase the value of the resistor. And, uh, if I move, if I reduce the value of the induct er, it will shift the frequency which the frequencies which are affected by this circuit to the right. So Onley higher frequencies are affected by this baffle step conversation. Second, if I increase the value, it'll go lower. So try to you can see if I increase lower and lower frequencies are are affected from. So I'm lowering the point from which the frequencies are affected by this circuit. So if I reduce it, you can see the lower frequencies are reaching their normal point. So now we have to find the point from which we get the that bumping frequencies. So that will be around, Ah, 405 100 hertz. So I'm increasing the value of the doctor right now, I feel that I'm exceeding that point. So I'm gonna lower this by just a bit. This seems to buy about right. But the value of the resisters the resistor is clearly high because I intentionally did that. So let's lower this. That should look about right so we can see that now the the response is much more linear from 200 to 1 K and still like in the chart from the manufacturer. We get that a slight dip happening from, ah, 1005 100 hertz. And then at three k, we get, uh, this rise. This is spike because of the cone break up. So this is the baffle step compensation circuit. 12. 2.4 Zobel network: Let's talk about impedance equalization circuits. These are also called Zobel Networks. So when would you would you want toe integrate such a circuit in your design? So this is a tweeter. We have ah Tweeter on the circuit diagram and we are focusing on the impedance chart. So Tweeters exhibit this behaviour with ah rising frequency, you'll also get rising impedance. And sometimes when you are doing a crossover network, you will find that the this behavior will not work to your advantage. So if you would flatten this frequency response I mean the impedance response. If you flatten the impedance response, then it will get closer to a resistor value. So this is the time where our equations from the Excel spreadsheet work. In conclusion, if you're having a hard time finding a correct filter for your Twitter, you might want to implement a Zobel network and then try again, see if that helps you out. So, basically, a global network is ah, capacitor and resistor in parallel with the speaker, so you probably could pass their here and resistor. This is basically a Zobel network, so a zoo we can see the impedance shifted quite a bit but this doesn't look very good. So this is, um, a similar to the Bethel step of compensation circuit where the capacitor dictates from which frequency on the impedance is affected. And the resistor that term is the amount off attenuation on Lee. This is reversed. If I lower the value of the resistor, uh, we will get, um or, uh, Attenuation. So let's increase it because it's kind of too much. And let's lower the value of the compressor because this doesn't look all right. Let's lower the value of the resistor. Let's fill it with the good pastor. Once again, that's lower value. There we go. There we go. So we see that the impedance more like a flat line, which is comparable in level to the impedance from the left, off the resonant frequency which is from the impedance bike. So when you're playing with the crossover, you are more interested in this part off the graph, because the tweeter will not play at the resonant frequency. Although there might be some cases where you want to flatten this impedance speak, while this one is not very not very large, Sometimes you might want toe reduce the impedance peak. And that is for a future lesson. Well, we have to do a notch filter, but in this case, this is the global network. This is how it works. The impedance is flattened, and this might help you in the overall crossover design. 13. 2.5 Series notch filter: let's talk about, See reason, notches. So first of all, when would you want toe integrate the Siri's notch? So, um, if you look in this example, we have the before Twitter And if you check the impedance chart, you will see that at the resonant frequency, it has a really sharp high peak. Now, although this is that around 600 herds or something like that, this will not be in the playable range. So you want to cross over the Tweeter way before this frequency. So although this is partially filtered by the crossover, sometimes when you have ah, piece Scharping Peten speak. Uh, the twitter might ring at the resonant frequency. So the theme the attenuation from the crossover my might not be enough toe to silence the tweeter at the resonant frequency if you hear audible distortion when testing the cross or this might be the cause and to solve this we need to tame a the impedance speak by using a Siri's notch filter. So basically we wire in the peril with the speaker on your doctor, a capacitor and a resistor. This is also called on lcr lcr hell stands for in Dr C for capacitance and our for resistor . So that place that resistor in the ground And there we have it. This is a series notch filter and again we're going toe fiddle around with the components. So ah, the in doctor and capacitor dictates from which frequency toe which the impedance is affected and that the value there is the value of the resistor dictates the amount off continuation. So, in this case, if we fiddle around with the in, Doctor, uh, let me reduce the value off the resistor to increase the attenuation so we can see better what we're doing. So if I increase the value of the in Dr were see, we're going towards the peak because we want to attenuate it. But like you see here, no matter how I try to a match the frequency on the on the peak all I'm doing is a 10 reading the middle of the peak and creating two smaller peaks and this is no good. So what we need to do is broaden this notch filter. So we wanted to affect ah broader frequency range. So So right now, the notch filter is concentrated in the middle of the peak and the result is not satisfactory. So we have to broaden the frequency range. In that case, we try to look at this frequency. We're going to somewhere higher than that. So maybe it you can see we're reaching somewhere 2000 hertz. And this should be okay. And now if we increase the value off, the capacitor were basically going somewhere over here. So we're affecting all of this. All of these frequency. And since the range has broaden, all that stands between them will be attenuated, so we can reduce that impedance speak. So let's increase the value of the capacitor and see what we're doing. Who they have no more impeding speak. And then you can fiddle around with the resistor. But it looks about right four point 4.3. So, as you can see, the impedance peak is flattened out. So if you if you would have ah, a ringing problem and the reason frequency. And now this should solve fish 14. 2.6 Parallel notch filter: Now it's focused on a different type of notch filter the parallel notch filter. So, um, compared to the Siri's one, Usually the Siri's notch fielder is used outside the playable band with of the speaker. So, like I showed you there at the reason frequency of Twitter, that is not in the playable range that either will never play at that frequency. Another example might be to tame down the cone break up, which is also not the playable rage. But in this case, at when we are using parallel notch filters, we are trying to model the frequency response in a playable rage. So where the speaker actually plays, So in this example, we have a two we design with the second order filter for the well for and a second order filter for the Tweeter. Now, if you if we look at the frequency responses, it's absolutely horrendous. But that is not the point. We are going to show what a parallel notch fielder does in. Uh, if we look at around the what is this? 6 6000 her said 6000 400 hertz. We have a big peak over here. So what? We're trying to do is to flatten that peak so we can do this with the parallel notch. Filter it. The consists in the same components on in Dr Capacitor and resistor. But, uh, they are old wired imperil between them. So we have the in doctor, we have the capacitor and we have the resistance. And let's wire them up together and hook it up to the speaker so we can already see that the response curb has been modified and now we basically do the same. We alter the value off the in doctor and the capacitor to change which frequencies are affected and how broad. Uh, the frequency span eyes affected, and the valley of the resistor determines the amount off attenuation. So let's ah fit around with the doctor. So I, as you can see as I'm decreasing, I'm moving to the right. Here we go, Uh, basically, then it. So I just fiddle around the value of the injector and you can see how flatten this response has gotten. But just to show, if I move the it's basically like in the series Notch further on. Lee with were playing with the frequencies, Boss Now and the resistor dictates the amount of attenuation. So if I lower the value, you can see the response is getting higher because is it's ah less attenuation. And if we increase it, we can see the peak dropping down. So you've got to play with these values and balance each other out, so you get the desired frequency response. 15. 3.1 Project 1 : Presentation and speaker selection: Now for our first the crossover design project. We're going to start with something simple. So we are going to design a two way crossover for a bookshelf. Speaker. Ah, base reflects bookshelf speaker. And as for the drivers I chose in a peerless the mid bass driver and a more rail tweeter Now, ah, the driver selection is a process off its own. So you have to understand that I have a bunch of speakers on have to mix him around to create these crossovers. So, uh, we will see what we have to look for when we are picking our drivers. Because some of these selections are not cherry picked. They are chosen toe just to mix them around and show different crossovers. But you will understand what to look for when you are choosing your drivers from all of these five examples. So if we start with our meat based driver, the first thing we need to look at is to look at the frequency response chart so we can determine what is an appropriate crossover point. So if we look at this chart, we can see that the response is linear, starting from 150 hertz and all the way to about somewhere around 2000 Hertz and then from 3000 and up, we can see a big rise that is a cone break up. That is a clear indication that the cone is breaking up. So when you are choosing the crossover point, you have to make sure that the crossover point is nowhere near this 3000 hertz. So a basic rule is to take the cross more point, it least one octave away from it. So since the since the cone breakup starts at the three K, we should we will choose the crossover point at 1.5 k So we have established that the crossover point for the meat based driver is at 1.5 kilohertz, or somewhere below that. So now let's switch to the Tweeter when you are trying toe, interpret the frequency response chart for a tweeter, and to determine the corresponding crossover point, you have to look at the frequency response where it starts to roll off so we can see the response starts to roll off somewhere, roughly at 1700 hertz. So if we use the same formula, we can deduce that the cross more point will be at 3400 hertz or above, and you can see. First off, we have a problem. So the crossover point for the mid bass driver needs to be from 1500 or below, and the crossover point for the Tweeter needs to be at 3400 or above. So we we have no overlapping of the two to set a crossover point, which is good for both of the drivers. So this is a good example off bad choice for drivers. But, uh, we're not going to scrap this project. We're gonna continue with it. So even though we have made a poor choice in driver selection, I'm gonna show you what you can do. Toe actually do something with this project. So first of all, let's start with the Twitter. We can see that the resonant frequency is at 780 hertz. Let's round this upto 800 one rule of thumb is that you never cross a Twitter any lower than the double of its resonant frequency. So in this case, if we apply this formula, the conclusion is we cannot cross over any lower than 1600 hertz, which is double off 800 so we can see we went even closer toe to its limits. But we're not going to stress that this Twitter that much. So now we head back to the Middle East River and and try to squeeze more frequencies from this part of the spectrum. So we know that the corn breakup starts at the 3000 Hertz so we can push this. Let's say we can push this driver all the way to 2000 hertz, so I don't want to cross that tweeter right at the limit. So at 1600 Hertz, let's say a good middle round will be 2000 hertz. But if we do this, we have to take into consideration that we are stressing the Tweeter. The meat base might not sound right, but the tweeter you can burn it up if you don't protect it correctly. So if we choose this 2000 these 2000 Hertz crossover point, we have to realize that we are stressing the Twitter, and in this case, we have toe but to choose ah, high order crossover. So the slope of the Rohloff needs to be very steep. So the crossover point it, it started 2000 Hertz and then it will have toe abruptly reject the lower frequencies, so we need to choose. At least the three order recommended would be a 44 order filter for the Tweeter. So now that we establish the crossover point and something about the slopes for the Twitter , let's take a look at the enclosure. 16. 3.2 Project 1 : Enclosure description: Now let's talk about the enclosure. So we have Ah, to a bookshelf speaker. This is about 18 liters. It's a base reflects enclosure, which is June toe about, um 51 hurts, so you can see that I used some crimping some cream terminals for easy connection to the speaker and tow wire the wires outside of the box so you can measure the drivers and test your crossover. You can route them for the base triplex sport, but I usually like to use the holes from the for the binding posts. So these are binding posts. I think they use seven millimeter hole, so simply drilled two holes for the biting post, and I simply robbed the wires through them and use some silicone to seal the extra space. Um, around the wire. After I finished the design of the gross over, you can pull the cables out. You force them out to rip the silicone and then insert your binding posts in the hole. Now, there is nothing fancy about this enclosure because this is a test enclosure. So I haven't used any recess step for the, uh either for the Twitter or for the meat base first of all because it will take extra time . And it's I don't care about the aesthetic aspect because, like I said, it's it's just a test enclosure. And secondly, if you look at the box with the speaker inside, you can see that the purest driver is not around driver. So it's rounded one in, and then it's straight. This shape is it's good because you can cramp Maura Mead. Base driver's close to one another. But if you want to do this, recess steps off. The speaker is flush with baffle. You will have a really hard time. If not, I think there are some techniques to do this, but most of them you will need a CNC machine to cut that precise shape. So this is the speaker. Now let's get to the measurements and to the crossover design 17. 3.3 Project 1 : Measurements: Now, let's take a look at the necessary measurements. We need to design this two way crossover. So we learn in acoustics to a one ho to do these measurements. So we need the frequency response of the feeder and the frequency response off the mid base now in a kristic store. One I showed you how to make the full frequency response of the mid base. But in this case, although I did the full frequency response, uh, we need to combine the near field with the far field toe, get that complete frequency response. But, uh, when we're going to design across our, we're going to use only the far field response. So you simply do a far field measurement. And that is the measurement we're going to use in the cross of our design application. What does this mean? This means we don't have the lower part off the frequency spectrum. Some say, somewhere below 250 hertz. We were missing that frequency part. Ah, this This is not an issue because the the bass response off the speaker is given by the enclosure. So no matter what crossover you are designing, you are not affecting the lower octaves. So the crossover is somewhere in the mid range in the high frequencies. That is where it operates. So, um, off course, we're going to apply the Hilbert boat transformation and get the calculated full frequency response or so to say. But we are not using the near field response in the crossover design application because when you are splicing the near field with the far field, you are getting some phase errors. So simply use Onley, the far field measurements. So in this case, you guess he we have the Let me tickle these that I have the near field measurements. I've done all these measurements and I'm going to show you why. But in the actual crossover application, we're going to use only the far field measurements. So you can see I have the meat based the far field for the mid base, which is the red one, Twitter, which is the Green Line and also the combined the to the fourth with measurement off the two drivers combined because we need to calculate the acoustic center, the critical difference between the two drivers. Now, I also have the near field measurements, and here is the the mid base with the complete with the with the bass response, including in the frequency chart. And you can do these measurements and it will help you determine the Hilbert. Both transformation more accurately. So after you made these measurements, you make effort the files on Lee for these three. So the mid base, the tweeter and the combined the two drivers wired in peril. Now, if I switch, if I switch it to my folder over here I have the three year 40 files. So I have the twitter of Hardy, the mid base and the two drivers wired in peril and also those automated files. And this is done with the Dayton audio device. And I always simply go to Exim and we have to load these files up, so we simply have to place to drivers. So we have we I usually place the twitter up top and the mid base just below it. So let's, uh, wear them up to the amplifier. This might take a bit. That's the okay, let's connect him to the ground. And then let's load the files. We have the treater, frd and Twitter visit me, okay? And the if, ah, ready for the main base and the meat base said to me, And now we have to apply the Hilbert boat transformation. So let's derive this. Let's see the higher tail at 20 kilohertz. That looks correct. I think this will be a steeper slope like these, like 24 decibels for Octave. And since this is the far field measurement, let's increase this to more than 250 toe. Let's say 300 and let's adjust the slope. This looks about right, nine decibels productive. So this is the Treacher with the Hilbert born transformation. And now let's tune the meat base. Now the higher tail is always easier for the mid base. Just go a bit lower in frequency and match the slope. This looks about right. And now when we go for the lower tail, you go just a bit above from 250 hertz because that is from where the far field measurements start being accurate. And now you have to sit a very low slope. So if if you don't, um if you having a hard time determining the slope, you measure the mid base with the complete response, and then you take a look at that. So let's go toe Roma que Wizard. And if we look at just the mid base So if we look at this chart, let's see, we take a look at 2 250 hertz where we're going toe set the tail. We have 66.7 decibels and let's say we go two octaves lower than this. So 125 and somewhere about 60. So two octaves below 250 60 hertz somewhere around 60 hertz again. See, at 60 hertz, we have like, uh, 63 decibels, so this is a two octaves lower. So, uh, it is roughly 1.5 disables loss per octave. So if we switch back to the Exim, we already have this choice 1.5 decibels. Uh, so you can leave it like that? We have the lower tail from 259 hertz with 1.5 decibels per active slope. So now we apply the hill Bourbeau transformation for the mid base. Now all we have to do is ah said the delay for the Bayside. So we have to load our we have to load the our file with the combined response. There it is. So now all we have to do is manipulate the delay and, uh, let's match the curves. But before we do that, let's, uh, raise this up so we can see better what we are doing. That's much better. And now let's kun the delay. I simply increased this until you have matching slopes. So sometimes I like to full screen this so I can see better and try to do the matching as best began, because you are not going to get the one on one match. So somewhere 7 74 uh, or 73 I think 74 looks best. So this is the, uh, acoustical delay set for the mid bass driver, and now everything is set up for the actual crossover design. So in the next lesson, we are effectively starting the crossover design for this two way bookshelf. Speaker 18. 3.4 Project 1 : Crossover design: So now let's start designing the crossover. When we are designing this crossover, we're trying to achieve a linear frequency response. The good face coherence between the drivers and we also have toe keep an eye on the impedance chart because you don't want it to deep below four arms. So if we look at the graph right now, that impedance chart is terrible. But when we're going to add components, it will start to to rise up. So most of the amplifiers are having trouble with. Impedance is which are below for alms. But if here, if you have some slight dips at higher frequencies, that might not be so bad, and most of the amplifiers will handle that unless they're very bad amplifiers. So first thing I like to do is to add a second order filters for each site. So let's start with the well for an ad Annan doctor and capacitor. Maybe they may be the Bayside. Sometimes one in doctor is sufficient. So let's let's delete the capacitor and just leave this in, Doctor, because we want to cross over to be as simple as possible. Now, uh, normally I would I would add a second order filter for the Tweeter, never at the first order because even though some manufacturers say that their tweeters are designed for first order filters, I highly advise that you increased the order of the filter to protect the tweeters. Most of the time, first water is really not enough. But in our case, we when we chosen drivers. Since this is not the most fortunate pick for the Tweeter, we've already established that this will be a high order filter. So from the beginning, I'm gonna sit directly 1/4 order filter because we need to protect this treater. So let's have a capacitor on in, Doctor in parallel and another capacitor and another in doctor. Now, most of the time, the Twitter will be hiring output than the mid base. But let's check this out. So we have this weird cancellation over here. If you have, you can always play with the phase by inverting the polarity off the way. It doesn't matter which one you revert, so if I invite this one, it's it's the same if I do it for the Tweeter. So first of all, it's the curves are not showing up, so we need the curves. So this is the mid base and this is the tweeter, and we can see the the cancel where they overlap. So, ah, if you invert the polarity, you face shift by 180 degrees and you get the opposite effect. So it doesn't matter how you play with the polarity of the speakers as long as you get the desired frequency response and the desired phrase response. So this is our starting point. So as again, see, the Twitter is a bit higher in amplitude compared to the mid base. So for sure, we are going to use resistor here. So let me who are up a resistor. And now we are going to fiddle with the components. So simply, let's start with the mid base because this is a first order filter and there is not much to do. Then increase the variable. The increase, the value of the in doctor. Because as we increase the value, we feel their mawr and mawr high frequencies and we can see that this mid bass driver goes all the way to like six K seven K. And this is no good. So let's increase the value of this in doctor and see what it does now. Ah, as you can see, as I'm increasing the value of this in, Doctor, uh, I'm blowing this portion of the graph is well, and now it's linear and on par with the Twitter. But I need a steeper roll off. And if I increase this, I destroyed the linearity. So I have to reduce the the induct INTs for this in doctor and add a steeper roll off. So I have to increase the order off the crossover. So I need to add another capacitor and see how that plays out. As you can see, we have a huge cancellation over here, which is a good and a bad thing. It the my suggests that the drivers are in face. So if I do the normal popularity now, the drivers are in face. But I have to set up this curve correctly. So let me play around with this capacitor so you can see I'm ah, the slope is now more, uh, mawr aligned correctly. And then we're going to play with the treater. So you basically play with these components. The first capacitor is is the most important one, but, uh, try to understand that the other components shape the curve. A swell. So let's fiddle with this capacitor is well, again. We see that cancellation. So let's revert to the polarity. And now it's that seems about right. Let's decrease the value of this resistor to get the response higher. Yeah, I think is better this way. And let's play with this capacitor. Now I think we're going. I think that we have to play with the MiG base more. Let's let's, uh, tune disk a pastor because we have to match. So we said the crossover point around 2000 hertz. So the curves needs to match somewhere in that area. So my starting point was wrong. So let me tune these capacitor so I can reach that. Uh, so I reached that frequency I want. So now you can see they're overlapping in the in the area I'm interested in. Let me try to increase this, okay? No, it's it seems toe look better now so you can see that the response is quite linear. Now, this is one of the first the aspect we are searching for, but we also need to have a good face coherence meeting between drivers and to and to do that when toe check for that, you have to give her the polarity of one speaker. So let's say the twitter and you have to see a big dip. Now, in this case, we do see a dip. But this is not a big one. So the phase when you convert to the polarity you have 100 degrees out of phase between the drivers. But since this is not a perfect cancellation, probably is around, I don't know, 120 degrees out of phase or something like that. But we are searching for the frequencies around the cross of our point to be perfectly out of face. And the bigger the how do it. The deeper this deep is so the march this is a much If, if it's deep, then the cancellation is perfect. If it's shallow like this one, then that that is not a perfect cancellation. So now, uh, in this case, we play with the components. Would we start fiddling with the components and try not to ruin the frequency response or we start adding more components so we add more complexity to the crossover in hoping to achieve these face response that we want and at the same time to not ruin the frequency response because this looks decent enough. So let me play. Let's way haven't figured with this girl component much. Let's so as you can see now I'm I'm leaving the inverted polarity so I can have a closer look toe the notch. And as we can see, if I increase this capacitor, we get kid bigger and bigger notch. So we know we are in the right direction and you also inverted to see Do I ruin the frequency response and we can see it's still linear, which is good. So let's increase it even more. We can see as we increase it. We do not ruin the frequency response, but instead we are achieving that phase response that we're after. So let's increase this until now. This is silly. We already reached 110 parents, and this is a way big way to bake a pastor and not much benefit we get from it. So you have to take these factors into consideration as well, so we can see you have ah 51 micro fired capacitor. And if we increase the value, we can start. The notch goes deeper, but Onley by an insignificant amount. So you have to weigh that 36 36 micro fair. It's now let me in verse the parole polarity so it can take a look at the frequency response. And it looks pretty OK is increased. This I see the response. I see the response getting linear, but I don't know what what it does the face response. So, as you can see, as I increase thes the notch gets bigger. But again, we don't want to exaggerate the value of this and dr because, ah, 4.3 million Henry in Doctor is quite expensive. So we have to weigh this somewhere around three million Henry's Ah, shouldn't be should be fine. Now, let me see if I play around this capacity. If I get the notch human deeper Well, it seems that I do. Let's, uh, raised this chart. But im ah skeptical that I keep that the frequency response. So let's see what it does. Oh, it looks even better now. As you can see, if I lower the value, I increase the frequency, the crossover frequency. So we stated that we wanted it to be at 2000 Hertz. Now it said, uh, somewhere close toe 2000 almost 2200. Now, this is better for the Tweeter, but I guess it's okay. Let's, uh let's try it lower. Maybe it 200. 2100. Let's see how the face response looks. Now, this this looks like a pretty decent Nagy. This is even lower. So I think this is Ah, this looks about right with this crossover design. So if we invert this, we can see we have a bit of a home here, but, uh, sometimes you don't want to overcomplicate things. This is just the three db hump, which is fine. We can insert the parallel notch filter here in the same this down, but we're not going to do this. We're going to keep this simple because the responses linear enough, we don't want to add any more complex into the crossover because it will increase the cost and adding more components will deteriorate the sound. So I think this is linear enough. You don't want something like a ruler flat because ah, though those speakers, like I say, don't have personality. So this looks correct. The face response is also great. So this is it. This is the crossover designed to Waycross of our design. For our first, our first build, our 1st 2 way bookshelf for speaker. 19. 3.5 Project 1 : Prototype crossover: now that we finished designing the crossover, I like to do a prototype crossover, so I actually have, ah, a bunch of spare in doctors and capacitors of different values. And I'm just gonna test the crossover and see if it works correctly. Now, this is not mandatory, but I usually like to do this so I can listen to the speaker and then decide if something it's tweaking or if it works correctly or not. So when I do this prototype, I usually place a chair behind the speaker, and I plays the components on the chair and connected the toe the speaker. And then I do the measurements. Now, of course, they're not going to solve their components together or something like that. But I people usually like to use alligator clips. Now I have some advice here. When you are using alligator clips, some are good quality. Some are bad ones, so it's always good to take your meter out in the measure, each alligator clip. In my case, I like to make some custom ones, so use thicker wire and custom clips, so I know these are high quality. So if I take my multi meter and measure the The test leads because those have, ah resistance as well. We can see the measure at 0.1 homes, and when I measured my custom alligator clip, they measure 0.2. So if I subtract the resistance of the test leads, my alligator clips are measuring at 0.1 homes. So ah, negligible resistance. But if I measure the alligator clip, I can see it. Zero. It's a 1.4 hopes, which is huge. Or basically it's a net off a 1.3, but still it's It's a very large resistance value. It's like having a next for a resistor in your circuit. And you don't want that because it will. It will shift your frequency response by by quite some margin, so you have to take this in account. You can buy some alligator clips, measure correctly. Some do not. So watch out to the quality off your alligator clips because resistance is, ah, a big factor in your crossover design. So try to use these clips as ah less it's possible. So I'm talking about the cables you can use how many single clips as you want to put your components in place. And after I finished the cross before I did that. A measurement in room e que wizard. So you can see this is the measurement with the crossover in phase. And then I speak to the polarity of the twitter and do another measurement and you can see the big notch over there exactly, like in Exim. And now we can go to X him and I exported that these measurements in 40 files and now we can switch back to accept, and we can upload these curves into the frequency response graph. So you important the file. I have it over here crossover in phase. And since I did this measurement in another day, I, uh you can see they're not perfectly aligned because I used different amplifier levels. It was not made in the same day, so it didn't kept the the same output settings. But as that isn't not a problem, because we are interested in the relative response. And we can align this this curve by adjusting the the wattage rating on the amplifier. So I have to go down. So you see, if I go down these start toe overlap, let me full screen this. So if I go to 750 they're ago. That is, that looks like a good match. So you can see we did all these measurements with the oldest wicking in Exim. And look how hope perfectly thes curves bunch up. So, uh, it's nice to see that all we did was correct. And the measurement, the actual measurement of the crossover matches with the calculations from Exim and ah, we can, um, remove this curve and add the one out of phase so we can see how that looks like. Let's invert the political twitter, and we can see there is a slight mismatch here, but anyway, it looks well on top of one another. So, uh, that concludes the crossover designed for our first enclosure, and it looks pretty good by now. So let's ah, doom or example to gain more experience in this crossover design 20. 4.1 Project 2 : Description: the next speaker. We're going to talk about Eason MTM now on MTM stands for meat base Tweeter mid base, and this is, ah, standard position for, ah, for a speaker, which has to meet based drivers and the Tweeter. So when you are seeing a speaker with Tweeter up top and to meet bases down below No, that that design is not a two way. When we're talking about a MTM, the crossover is a two way you can you can not designed to weigh crossover if you place the Twitter up top and the two mate meet basis down below. Why is that? Because the first the crossover frequency will be too high. Let's say, somewhere around 2000 Hertz, 2500 hertz and the distance between the Twitter and the second meat based driver is too large, and it will create the sum of phase issues. When you do actually see this design, it is either a three way or ah ah, more popular choices, a 2.5 way crossover. So what a 2.5 way crossover is. Basically the 1st 2 are two way, a normal to away and the second mid bass driver takes over and reinforces the Bayside. So let's say the tweeter is from 2500 all the way up to 20,000. The first meet base is from 50 hertz toe, 2500 the second, and the 2nd 1 is like from 50 hertz toe 500 Lexi. So it reinforces Onley the base section, because that the that lower octaves have very long wavelengths and do not interfere with with the Tweeter. So the crossover, the frequencies near the crossover point of the Tweeter. Now let's switch over to our MTM design. So what are the particularities, often MTM. Besides having the extra efficiency from to mid bass drivers, there's a certain trade for using this design. Let's look at the polar plot for a conventional two way speaker. So when I say polar plot, I mean, how does the sound ready through the room? And we are talking about the vertical dispersion. If we look at this representation, we see that the distance between the Twitter and the meat base is equal toe one wavelength off the crossover frequency, and this is just a reference point because we are going toe, Compare it to the MTM design where this rule is more important. But, ah, let's focus on how the sound radiates. So those lobes you see are the vertical dispersion. So on access, you see a little listener over their own access. You get the full, the full frequency response. But if you move your head up or down, so if you go go up or crouch down, you will experience parts off lower intensity and then higher intensity. This is called a comb filter effect. Now, when you switch over to Owen MTM design, we keep the same rules. So the distance between the Twitter and the mid base is, ah, one wavelength off the crossover frequency. And another thing we have to take into consideration is that the filter needs to be 1/3 order. So if we look at the dispersion chart, we can see that if we use a second order fielder, we just get a big global on the own access. But if we use 1/3 order or a combined second order with third order, we get a linear dispersion of sound, regardless off where we are on the vertical axis. So If we stand up or crouched down, we have the same sound experience. So when people designed these empty and boxes, not only are they looking to get a better efficiency from the mid base sections because you have a one additional speaker, but also regarding the vertical sound radiation. And now that we talked about the characteristics off this MTM design, let's look at the enclosure. So basically we have, ah, the to mid basis, which are symmetrical from the Twitter, and the twitter is not in the center of the baffle. It's slightly offset. This is toe reduce some baffle diffraction issues. Also, you can see that the ages of the baffle around it actually for for all the enclosures off this course, the edges around it to reduce the fraction issues we have to brace is on the upper and on the lower part off the tweeter. And this is a sealed enclosure, so we don't have the opportunity to route the cables through the base replace sport. But I did my trick where I route the cables through the holes off the binding post and uses I'm silicone toe seal everything up. Then I placed the speakers into the enclosure and again using those spade the cream per connectors toe for easy connection with the speakers. And now we can talk about the driver selection. So here I'm using the same Morial treater, but instead the off the peerless mid base. Now I'm using sass mid base and, ah, before we talk about the frequency response and where to choose the crossover point, we need to talk a bit about efficiency. So when you having two drivers, you have to take this into consideration. So if we look at the efficiency off the mid base, we can see that it's 88.5 decibels measured at 2.83 volts. But since this is an eight on speaker, that is the same as one. What, one meter? But because we have two speakers, we need to add six disciples. So actually, the efficiency of the mid bass drivers is 94.5, and now let's switch to this toe the tweeter. We can see that the efficiency is measured at one what one meter and the efficiency is 89 decibels. Now this is lower than the mid base, but since the Twitter usually gets the full effect off the baffle. So plus 60 b we have on 95 decibels. Efficiency rating. When you have multiple drivers, you have to check these efficiency ratings beforehand because since you have to meet based drivers, you have the options tow, wire them in parallel and we gain this six decibels off increase output. But if it's too much, you can wire them in Siris and not get this benefit. So depending on the project, you have this decision to make. Now that we talked about the efficiency leads move on to the crossover point. So we already talked about where to set up the crossover point for the Tweeter in the previous project. So let's look at the new mid based driver. So as you can see these drivers way more friendly because it doesn't have that weird calling break up, this will be a better match with our Twitter. So if this would have bean ah, normal two way, like in the last project, this would be a better match compared to the peerless driver and since this MTM is basically two way also, this is a great match for our project. Now we can see that this driver plays linearly up until four K 4.5 key. But we're not going across this driver that high, and we know that the optimal point for the Tweeter is 3400. But when you're designed the cross over this, when you actually doing the designing, your a crossover point might shift maybe a couple 100 hertz up or down. Ah, in ah, sometimes you decide the crossover point on the spot. Now we know that the Twitter will have ah third order filter toe to keep the desire vertical sound radiation and the mid bass driver can have also 1/3 order. If not, a second order will result the same benefit in a sound radiation. So we know that the Peter will have, ah, slightly higher order filter so it will have adequate protection. So if we go to like 2500 it'll be fine. So now that we established where a rough estimate where our crossover point will be, let's move on, toe the measurements and let's set up everything to do the actual crossover design 21. 4.2 Project 2 : Measurements: Now let's say a few words about the measurements like we did before. We are going to use only the far field measurements so you can see up top that I have the mid base and the Tweeter and also the combined for the acoustical center. But some people get confused in this, and since you have two drivers to mid bass drivers, some people like to measure just one driver and place two drivers in Exim. I find this that sometimes these this method of doing things will you will encounter errors . And maybe in this case you can measure each driver individually and place them in parallel in exim. But sometimes let's say you have a base reflects enclosure. This is a sealed enclosure when you have a base reflects enclosure. If you measured one driver, far field, you will get the output off the port as well. And when you measure the second driver, you will also get the output of the port. So basically you are getting to port outputs and you have only one. So you understand that is some scenarios. You have no way to do the measurement accurately. But if you wire both drivers in parallel or in serious, however you want and do one single measurement. And even you have even if you have two drivers. Three drivers just place one speaker on the Exim circuit diagram. So for the Bayside, even if you have two drivers, three drivers just place one speaker. But when we are doing the measurement, measure, all the drivers at once. So, like I said, we have the mid base, the tweeters, the Twitter, that single Twitter and the combined. If our these are old far field. And of course, I have the near field measurements. This is the near field and the near field with diffraction applied. And here we have the four response off the mid base. Now, in this case, when you are doing this measurement on a tell you just to make a point. Since you have two drivers, you are probably inclined Toe measure one driver near field and add six decibels, so you get the output off to drivers. But that is simply a waste because when you are matching the output for the far field, so you have to scale the near field far field anyway, so you are basically doing nothing by adding six decibels because we are changing the output anyway. Toe match with the far field. Now these are the measurements, and I have them all bunched up in this folder. So we have the Tweeter and Meat Base. If RG and the combined the party all far field measurements and this enemy filed, we created using the deaths. And now let's load them up in Exim. So we have to do is place to drivers. This is the Twitter, and this second driver represents the to mid bass drivers. Let's wear them up together. Let's place the grounds and let's load the files. Preacher if RG creatures that it may and the same for the meat based driver basic 40 and said at me Now, let's quickly apply the hill burble transformation. So we look at this tweeter. Let's assume the the role office a bit steeper. 24 disciples prerogative and let's go to 250 hertz. Maybe a bit more. 300 match the curve looks oh, on par with the slope, so I think to have their supposed proactive is okay, so let's finish this and let's go to the mid bass and do the same. And let's decrease the frequency. Okay and increase the slope. That seems about right. Let's go to 250 hertz and the slope. Let's check. Let's check the four frequency response if you have any adults. So this is the full frequency response. If we go to 250 hertz, we see that we have, ah, 62.5 decibels if we go to octaves lower. So at the 62 or something 62 we have 52. So we lost 10 decibels. That means, like five decibels per octave loss. So let's head back to exam and we have a 4.5 disciples per octave loss. Let's hit. Okay, and now let's find the acoustical center. Let's rise this and place the combined. If our d we said that that they pretty much match. But let's get it done correctly. - I think this is the closest match, so 0.13 inches. So now everything is set up. Let's move on to crossover design 22. 4.3 Project 2 : Crossover design: Now let's begin with the crossover design. So what we established beforehand so normally, like I said, I would add to order fielders for each section off the crossover. But since this is an MTM design, we need 1/3 order filters for at least one branch. So since the Twitter is an issue here, we know that for the Twitter to be comfortable, it needs to be crossed closer to 3000 hertz. But this will vary a lot when you actually make the crossover. And my guess it will be, ah, 2000 plus crossover rather than closer to 3000. So for this reason, we know for sure that the filter for the Twitter will be 1/3 water. And regarding the Bayside, that will be either a second order or 1/3 order. So let's start by adding 1/3 order filter for the for the Twitter and for the mid base, I'm gonna add a second order filter because I'm want to ensure we have a minimal component count. If we will see that we can't get the phase right, then I will add another in doctor to increase the order off the filter and maybe we get the face to align correctly. So let's add 1/3 order for the Tweeter and a second order for the meat base. So we have a capacitor on a doctor in parallel and another capacitor. This is 1/3 order. We shall see if we need the resistor or not. Because remember, we have to meet based drivers so the efficiency might be quite on the spot. So let's connect these two and at the second order filter for the base height. So on in doctor and a capacitor, let's wire everything up and connect this to the ground. Let's remove the combined curve and add the two drivers so we can see what we are doing. Okay, so let's start with the easy part. So as we can see from this, the levels are quite match. So there might be a high chance that we will not be the Siri's resistor for the Tweeter. So first of all, it's, uh, let's play around with the mid bass and sea and let's lower the crossover frequency because we can see it's quite high. So this is somewhere around above two K and that now let's move on toe the tweeter and see and play with these components. I think this in doctor is way too. Highlands lower this more comfortable level. - It seemed that it seems like this where the value of these capacitor went way too high. There we go. Now it seems like we're getting somewhere. Now we have this weird hump over here. So now this is a perfect scenario toe implement a parallel notch filter. So we are gonna tame these frequencies down in hopes that we will get, um, or clean your response. So we have to add the three components the healthy are like they say than in Dr Capacitor and resistor. You can put him in any order you want. Just make sure you wear them, you wire imperil between them. So there we go. This is our parallel notch filter. And now we have to target these frequencies. The frequencies around this area, probably. Now it's around over here because I get that this dip, This is the thing is we're toe highlights, huh? I think we're getting there. Let's lower the value of the resistor because we have too much stimulation. We see we're right on the spot we Nice. Yeah. So this looks Ah, much more linear. Let's help the face responses. I m not that great. We'll see what that we can do about this. Listen, verse the face back, huh? Knowledge lets the response now is quite linear. So let's work on the face now and see what we can do about it. Maybe work on the base side. Who? Oh, I think we ruined the frequency response. Now, let's let's let's convert. Yeah, The mid base is extending too much in this area, so we have to lower. We have to increased this capacitor. We know that if we, uh, if we decrease the valley, we get the face response, but we don't want to ruin the frequency. So we have to get a spot where we are comfortable with both. Now, when you get that, these, uh, these in doctors with the value is higher than one million. Henry, you have to know that they have, ah, resistance off the room. So if these are Air Corps, if you tune it, you see that you have ah, a serious resistance here, which is 0.35 bombs, 350 millions So, uh, I'm pretty confident that this number is higher for a 1.5 million Henry. So at, uh, three million Henry's. You probably have one home for, ah, one millimeter wire section. So I'm gonna go to, like, 600 to match what's in reality. Sometimes this shift the response quite a big. So you have to take this into consideration. When you have big in doctors, you have to check their resistance as well. So you have toe introduce it in X Emma to get the correct frequency response. Now, let's June this in, doctor, over here. But first, I want to change the value of this capacitor because I don't know. It's why it's 12.1 because, uh, this is not the usual capacitor value capacitors Commons in standard values. Usually they come from 10 12 then 15 12.1. You won't find something like this. So let me change it to 12 like, uh, because that is the correct value. So now let's fiddle with this in doctor and see how we can change the face response and we can see that we're getting in the right place. But also we don't want toe ruin the frequency response. So let's invert the polarity and see what we're doing. Like we see, we get a big hump. So as we increase it, we are losing the phase response The desire for his response No s as we as we increase it with regained the phase response that we want. But we got this little hump over here so we don't want to exaggerated by too much. So let's say that here, at 0.22 is a nice sweet spot for both the face response and the frequency response. So we see we get this big notch over here so we know that our drivers are aligned correctly . And if we invert the the polarity, we get a decent enough frequency response now. Like I said, people always try to get those who ruler flat frequency responses. But I say that if you keep the minimal component down and have a pretty decent the frequency response with the variation off, like two decibels, three decibels from linear, it should sound great. And I can tell you that I already made that this speaker in its in sounds awesome, I think, because it has so few. It only has this resisted. It doesn't impede the current flow to the speaker debt much, usually Speaker have a lot of resistors and stuff and try to match the levels. But now the levels of the speakers are somewhat the match match, the naturally and it sounds really great. 23. 4.4 Project 2 : Prototype crossover: Now it's time to do the prototype crossover and is the last time I showed this, because there is no point in showing you for each enclosure how to do the prototype, because is the same thing for every scenario. So I put the chair behind the speaker and I set up the components on the chair and connect them to the speaker using custom made the alligator clips. So if we look at the close up, we can see better what I'm doing. So I'm using the input I. I also made some custom cables that go to the amplifier, which have a red and black alligator clip. These are the positive and negative connections to the amplifier, and like I said, try to use the least amount off cables as you can. If you watch closely this prototype crossover, you will see that I'm using cables on Lee to hook up to the speakers and Onley to connect to the ground all the connections, all the rest of the connections. I connected the components directly and clamp them together, using some simple alligator clips without cables. But you can do this however you want. Just make sure you don't abuse these cables because you are increasing the resistance off the circuit, and that is no good. So now let's switch it to room, make you wizard and you can see here I have met made the measurement. Let's antique all the measurements besides the crossover. So here we have the crossover with the speakers in phase and, of course, with the speakers out of phase. And we get that huge notch at the crossover frequency. So we export the these files into a $45 you freeze which over two exim we can start loading the files. So let's load the file, which has the drivers in face. This is the crossover response. If our defile and we can see it's a lot higher. This is because in the same fashion, I did the measurement in another day, and I didn't kept all the settings the same. So this is a bit hard, but there's no problem. We're gonna June this this amplifier over here. Let me four screen this and we try toe, match the curves and let's go for 45 watts. Now this looks about right so you can see the curves are overlapping almost perfectly way have to look at the blue line and the green line. We can close the other two so we can have a better if you add it. We can see that they overlap pretty nicely, almost perfectly here, a bit of a gap here. But you have to understand that in order to do this, you have tow measure each component, so I haven't measured each component individually. I just went with the values the manufacturer quoted. So maybe I placed the 1.5 million doctor and in reality, it's some like something like 1.45 I don't know. So slight mismatches also caused these slight deviations. But I trust the, uh this crossover is the designer exactly like in this diagram. And we can also choose the curve with the reverse, the polarity of the speaker and we can see that they match up pretty good slipped gap between the two graphs over here. But I'm quite satisfied with the result again. We also have to look at the impedance chart so everything if the impedance is above four homes, then it's perfect. We can see a slight dip over here, but like I told you, if you have the small dips below for alms, then that should be OK. It's not OK if you find them in this section of the graph. So for low frequencies, if you find your impedance Schardt dipping below four rooms, that's no good. But here in the four K area, that is no problem for the amplifier. So this is our second crossover design. Let's head toe the next one. 24. 5.1 Project 3 : Enclosure presentation: Let's take a look to our third enclosure. So this is ah to a designer based reflects enclosure and we are going to use for Twitter and Assayas mid base. This is about 16 liter enclosure and the tuning frequency off the enclosure is at 51 hurts . And this is an enclosure, which I did the wild back so you can see that it already has the binding posts fixed in. So I had to run the cables through the base, reflect sport. And, um, it's funny enough because I made this enclosure a wild bag. So I had my variation off crossover design back then, and I did it again now, and the crossover design is different and we shall see my new approach, and we will take a look at my first approach also because it's interesting. What I'm trying to say is that as you play with the crossover design, you see which components do what, and you will get more experience and have have a feel off how how the circuitry goes around . So a particularity about the speaker is that I offset the tweeter and you're probably wondering why, where first of all this is to solve some diffraction issues. But let's look on a specific chart, so you understand what I'm saying. So this is Ah, modeled diffraction response using sound easy. If you have something easy, you can do this by herself. But if not, then let me explain this graph and you might find this solution helpful and implemented in your future projects. So this is the model diffraction response for the Tweeter in the center of the enclosure and for the tweeter offset Tow us toe. Want either side left or right? It doesn't matter. What we can tell from this graph is that the diffraction response is not ah, a simple curve. So you see the linear game and it reaches the level which is actually higher than six decibels. And then it goes down a bit and then little ripples until it stabilizes toe a flat line. Now, the first dip after the plus 60 B peak has been reached. That is significant if you can see it's somewhere in the round the two K three K area, and that usually corresponds with the crossover point in a two way system. And since the mid range driver is in the center of the bear. Phyllis. Well, that dip will overlap with the deep from the tweeter. So you will, because of off this diffraction, you will have two overlapping dips, which will create an even bigger dip in the overall response. So correcting a huge dip in the crossover is a massive challenge because usually it's easier to tame a peak. So if frequency is too high, you can tame it down. But if a frequency is too low compared to the others, you cannot bring it up. So you have to find creative solutions. Tow this problem so you can see if, if I offset the Twitter, you can see how the green line the diffraction response goes. It does not overlap in that frequency range, so it like counter balances the diffraction issue off the mid base so you can see this will help us out in making our lives easier in design. The crossover. Now let's go in choose the crossover point. And this time I'm not going to use the frequency response charts. I'm going to use the actual measurements in room accuser. So here we have the far field meat based the far felt either, and the combined the response of the two with the two drivers, where the parallel and far from measurement for both of them, we have the near field measurements and combined with the diffraction. And this is the the mid base, the complete response of the mid base. But we are going to look at the mid base and at the distortion shot so we can find a good cross of our boy. So as we can see, we have in percentage. So we're looking toe. So we are looking where the distortion exceeds 1% or 2%. So, actually, we're not really interested in the lower section off the graph because we are going to cross it with the Twitter somewhere in the two K three K area somewhere over here so we can see that the distortion is under 1% all the way to like six K over here. So this means if we go on active lower to three k, it should be perfect. So somewhere around the three k area is a good cross 4.2 k 3 k. So we can see starts to rise at one K, but it's still well under 22% which is also good. So anything in the three K area should be a nice crossover point. Now if you go to the Twitter because if we don't choose the correct crossover point we might damage Tweeter, we can see that the distortion is under under 1% all the way toe one k. So if we go one octave about that two K, we can safely say that this Twitter can be crossed from two K or above. Now that we have established that the crossover point will be somewhere in the two K three K area. Now we can hit the word toe crossover design. We will go directly to the designing part. I'm going to skip that phase where I load the files and apply the Hill Bourbeau transformation, and I will skip the prototype crossover as well. I'm just go. I'm just going to show you the designing off this crossover and the variant, which A did a long time ago because I used a leather delay network and I want to show you how to implement one, because you might do it as well in your future projects. 25. 5.2 Project 3 : Crossover design: Now everything is set up for the crossover design. So let's begin. So if we compared this project with the other two, we have no rules here. Basically, because for the empty em, we had to have at least one branch with three order and the 2nd 1 with either second order or 1/3 order crossover and, uh, at the other one, we need high order for the Tweeter to protect it. But here we don't have any of those restrictions. So to begin with, we will use a first order for the mid base. So let's place an in doctor over here and for the Tweeter. We will use a second further because the first order is and not enough for for Peter. So this is the second or a filter. So this is where we will begin. We will add components if we need to, because we want toe maintain the cross over a simple it's possible. So let's look at our chart. And, uh, first of all, we see that the mid bass driver is extending way too much, So we need to increase the value of the conductor to restrict the higher frequencies from the mid base. So let's increase this. I'm gonna do it until the mid basis around flat. This looks about right. It still extends way too much, but it's at a lower level. Maybe it will be good. Maybe not. Let's see, Let's do something for the Twitter because the level is way too high for the Twitter. And we knew. And we need to, uh, implement on attenuation pad. So let's see if one resistor is okay. Live. Replace the bed. Try to match the levels more ignorantly, it seems like cannot get rid of this weird hump over here. Usually for the second order in Doctor one million. Henry's usually wait to my So let's lower. This seems like we're getting we're getting there. So it seems to get rid of that hump. I need to decrease the value of the first off the first capacitor by too much. So now the crossover is four key or something like that. So, uh, I'm gonna add another capacitor here to increase the order of the filter and see if that helps me in any way. Let's play around with these components. No, no, let's rise This. Let's increase this So we increase the the crossover point, Let's play with this one as well. It looks like now we're getting more to the desired response. But let's take a look at the phase. So we see that the drivers are in cancellation somewhat, and we see we have double dips. And this is because the mid base is extending way too much in into the response of the Tweeter. And since it's close enough to the toe tweeter output, you will see a cancellation here is well because for all of these section, they're out of face. It seems that here they're not quite helpful faith. But we're interested in in closer to the crossover area. So let's see if we can, uh, make this face response a bit better. No, this looks a bit better. We can see the notch going down. Hope I didn't ruin the frequency response. I just got a weird hump over here, but I'm gonna show you how we can fix that. So let's see if we can improve the phase response. So now we've got this deep notch over here. So that's inverters. We see we get this this weird hump over here so When you get this hump, you can wire another resistor with the component, which is in parallel with the speaker, in our case, the doctor. So here, if I place a resistor, we can see what can fix that hump. But now, with the value of the resisters way too high. So let's reduce it. We can see that we flatten that that peak, which was right before the roll off. But we have to look at the phases. Well, so now we fixed the frequency response Slight, slight bump over here. But what kind of ruined the the freeze response that I don't like how the face response looks over here. So I think we have to do something more complicated. Let me eso. This is the actual crossover design part. You have toe do trial and error. You have to try different components in different places. You have to increase the complexity off the crossover. Then you try to reduce it to some way and see if you can reach the desired response, the desired frequency and face response. Because in our case over here, I'm not quite satisfied with this response. So, uh, we so we could get rid of this resistor over here. So if I short this out, basically it's not not even there. We can see that the face response is good, but we get this weird hump which I don't like, so let's do something else. Let's increase the complexity off the base I so we can shift the face off the mid base and see if that helps up. That helps us in any way. So let's put a capacitor over here and let's connect it to the ground. Now, since we have ah capacity in place, we can play with the slow pull off the mid base. Even better. So in that case, let me reduce the value of this in doctor and tried to fiddle with the capacitor is well, since I went this high with the mid base, you can see I was losing quite a bit of output from the mid base. In that case, we need to rise the the Twitter response because we have a new attenuation pad in place. So let's play with these resistors used. Seems like we're not getting where we want toe. Let's for L A series resistor with this comm pastor sometimes if you if you do this, you can shape the response quite well near where the response starts to roll off. So even if the twitter is a bit lower compared to the mid base, you consider the mid bases extending in the Twitter responses well and embalming, and it's bumping up the overall response. Let's take a look at the we can see that we have these Tween twin dips. So basically, I'm moving the crossover point a bit higher. So we're now in the 3500 range and see if I can get a good the phase response. Think we've found something so as we can see them frequency responsible square. Good. Hope we're not ruining the frequency response. Let me Chicken crossover point is somewhere it 2800. This is no good where we're ruining the phase response. So this is the best for his response. So I think we need to extend that twitter more into this area and have, uh, steeper slope. So let me add another in doctor. So I increased the order of the filter of the tweeter. So this will basically a 44 door now and see if we can do a bit better. We can see we're on the were on the right track. So if you remember this, we can revert back toe this design because we have a huge notch. So we're different. Definitely in the in the right direction. So, uh, let me increase this because I want to go a bit further, but with a steeper slope. So let's check the face response. Not looking that great. We need to extend more. Something like this. Were in tow. The 2400 which is good. The frequency response looks. It looks pretty nice living lower this value. This doesn't much. Maybe does something to the phase hip Hope I didn't ruin the frequency response. Let's reduce this. Let's change the frequency. No, we can see that the linearity and I was pretty good. Now let's see what we can do about the face. So if I draw a line between 70 and 65 you can see that the variation is just a couple of disciples. So now let's see what we can do about this. We have this hump over here. Maybe we tuned this resistor. Yep. Maybe the face response gets better as well. Nice. They look at the frequency response. So I'm quite happy with this response so we can see that we got a huge notch now. So now the drivers are clearly in phase, and the response now is is quite linear. But as this advantage we have 44 for the Twitter, which is Ah, high component count and the second order for the mid Pacey. I guess it's OK, but considering how good the frequency response and face response looks, I think this is a nice compromise. So let's look at the face response again. Yep, this is pretty good. This is way better than our first, the alteration. So you have to play around with the components and see what's what tried different combinations at complexity. And then maybe, maybe now I should try to reduce the complexity because now I know I reach my goal. But maybe I can do a bit better, who knows? But right now, I'm gonna stop at this design, and then I will show you how I did it with the leather lane network because that might be something which could help you out there. You know your projects 26. 5.3 Project 3 : Ladder delay network: Now let's talk about the leather delay network. Now, this is a circuit I should talked about in the first part of the course, but I kept it for this later stage because it's a bit complicated. And it This serves well in this example because we already made the crossover without the leather delayed network. And then we shall see how we can do with the aid off one. So s u K C. From this diagram, we have the same V for Twitter and assessment base, and you can see how simple this crossover is. We have a first order for the mid base and a simple second order with the syriza resistor for the Tweeter. And the response looks quite lier. But if we reverse the face, we can see that it's nothing that resembles, like a phase alignment. So if I returned to the original state, what a leather delayed network does is that it keeps the same frequency response, but it only others the face. So if you find yourself in a point that hey, I don't know what to do, I cannot align these drivers phase wise. I tried everything. I added more components I made the crossover so complex and I still didn't manage to do the face right. Then it's time to implement a leather delayed network. And the leather delay networks are off two types. There are first order and second order, but we will talk only about the first order because the 2nd 1 used to so many components, it's just a waste. But it does add a bigger delay. If you would need a bigger delay. You will have to implement the second order leather delay neighbor. But we're going on Lee going to talk about first order. There are many ways you can draw it in the circuit diagram, but I'm gonna use the way it is described in sound easy. So I'm gonna take this resistor here, and, uh, I'm gonna delete this and it looks something like this. And then another one by these two then connected the speaker like this and then place a ground over here. So don't be fooled. The this this thesis part does not connect to the ground. It connects toe this node over here. So when you are actually building the crossover, take note of this. This does not go to the ground. It goes to the node over here and the resistor. We're gonna place it over here in Siris with the speaker, the symbol of the polarity. And now the the rules you need toe keep in check is that the's in Doctors must have the same value, and these capacitors must have the same value. So if you tuned is to like 6.8, then you have to do in this to 6.8 as well to keep the effects off the leather delay network. And then you have to play all day with these components until you reach your face response . So you against you were kind of getting there. I'm just gonna input the components, the value of the components, which I know 0.8 to end this one 5.6. It's just a merely coincidence that it's that they are the same values of the second order filled their components so you can see it's the same 5.65 points. Expect more six and zero pointed toe for the three in doctors is just the coincidence. But you can see we get the notch we wanted. We got this little deep over here. But it doesn't matter when we evert the face, we can see you have that. We still have that flat response like it does quite slightly changed the response. But, um, the letter delay network focuses on Lee on the phase response. So if you find yourself that you cannot manage to get the face right, then you might as well implement a leather delay network. Now, the whole point of this lesson is to show you how leather delay network works. Don't judge this exact crossover design because it's clearly inferior to our previous one. If we look at the component count, we can see that they're basically the same. But if we look at what the crossover looks like so the crossover point here is that 1.8 k compared toa way higher than that. We know that the twitter must be crossed somewhere at two K and higher, and this is a second order for their, so the actual slope is not that high. So in this case, the protection for the Twitter is not the best. So if we compared to the 1st 1 this one is clearly inferior. Also the face of Lyman isn't not that good compared to the first iteration. So don't judge the actual crossover design, but focus on the leather delayed network and how it works, because that is the purpose off this lesson. 27. 6.1 Project 4 : Description: Now it's finally time to move to a more complicated design, So we are going to design a three way crossover. So if we look at this enclosure, we have like, ah, bigger bookshelf. This is not the floor standing speaker, but it's a three way. So we get the V for Twitter. We got a focal mid range. We haven't used this driver before and the peerless based site, and if you look at the back of the enclosure, we can see it's a based reflects enclosure. The bass driver is in a separate chamber compared to the midrange and tweeter, and for that reason we I routed the cables for the Tweeter in the mid range. Through the holes off the binding posts and the base, I routed the cables through the bass reflex port. Now, if we switch over toe the measurements, so let's go toe room into wizard. You can see we got a lot of measurements now, so we got the base Far field. We have the mid range Far Field and the Tweeter Far Field. We need these for the X seemed a crossover design application and, of course, the combined frequency response now you have to hook all three drivers in parallel and do a far field measurement to get the correct combined response. These this will we will use to determine the acoustical centre. And I'm gonna show this process because it's a bit different, since we have three drivers. Not only, uh, the bass driver needs alignment, but the mid range as well. Ah, I also have the near field measurements. So to set up the hill Bourbeau transformation correctly, I need the complete response for the for the base. We have it over here and the mid range. Also, because I did a near field measurement for the mid range is, well, a simple near field is enough because there's no based reflects port for the mid range. And since it these are the measurements. Let's take a look at the at the mid range specifications sheet. This is a pretty old driver, but it has a nice frequency response when you are designing a three way crossover. Determining the crossover point is the simpler task because now you have way too much slack where you can place the actual crossover points. You have to cross over points, so you can see the driver roll offs at about 200 hertz. Sometimes people like to leave the mid range alone and just place it in a sealed enclosure . The acoustical damping off the enclosure itself will increase the resonant frequency off the driver and shift the roll off a bit higher. If you do this, you will have less components in your crossover, but you will limit the power handling off the speaker. In our case, we will do Ah band pass filter so we're will not going to let the driver roll off. Naturally, we're going toe set a crossover point, which is a bit higher somewhere around 1000 hertz like 807 100 or maybe 1.1 kilohertz. That should be fine, and we can see the driver extends well into the high frequencies. And it's better for the twitter if we choose a higher crossover points or, like three K four k tau eliminate any stress from the Twitter. So, like I said, this is the easy part in determining the crossover point because there is no, uh, fixed positions. We have to said that the points we have, ah, way too much room to set the crossover point in different places. Now let's switch over to Exim, and I'm gonna show you how to determine the acoustical center. Now all the drivers are set up in the Hilbert board. Transformation applied, and we need to determine the acoustical centres. So we applied the correct delay for each driver. Now let's load the combined the, uh, the far field with the combined drivers. So here we go, and we need to match the curve as usual. When you have a three way design, not only you will have to add a delay to the base driver, but you will also have toe added delay to the mid range driver. Depending on how you place the microphone, let's say you place the microphone on axis with the mid range. Sometimes the treater sound will get will arrive later compared to the mid range, and the twitter will need a delay, so you have to check each driver individually. The midway's will surely needed lay, but normally the mid range will need a small delay. But sometimes the tweeter needs also, you have to check. You have tow fiddle around with the settings and see the best way to match the two curves. So let's start by tuning the mid based driver. Let's four screen this away, see it better. So let's increase this and see where we are going. Probably will not going going to get a perfect match because we need to tune the other drivers as well. So let's see. So now, as you can see, is I increase it. I'm actually going the wrong way. So let's find a middle ground and then start fiddling with the other drivers. So let's close this up. And first of all, I like to tune the Twitter to see if we're doing more bad than good. Because then I know for sure that the Twitter does not need adding a delay so we can see if I start to add a dilated the Tweeter like I expected, I'm doing more bad than good. So we're going to leave the Twitter. We're going to leave the Twitter alone and start fiddling with the mid range. So here's the mid range, and as you can see, we're going towards uh, so we have to add a certain delay here and maybe fiddle around with the Bayside again. and see if we're getting closer. As you can see, we are getting a better shape. Is the increasing its bad? So I think this is the right spot for the for the Bayside, and probably this is the best I can get for the Yeah, see, it's not doing any good. I think this is the best part, the the best we can do for the matching curves. So as you can see, we found that the critical center for the speakers and we said that the appropriate delay for the Bayside and the mid range side and we left the Twitter alone. What follows next is the more complicated part where we actually have to design the crossover. When you have a three way crossover, you will see that the complexity goes up by quite a bit. So let's see how well we fare to this challenge. 28. 6.2 Project 4 : Crossover design: Now it's time to do our 1st 3 way crossover, and this will be a bit complicated complaint compared to our previous ones. But don't get discouraged if you actually think about it. You are practically designing 22 way crossovers, so we have, ah, to weigh between the base and the mid range and another two way between the mid range and the Twitter. You can look it like that. Basically, you can look it like that. So let's start with the usual Let's add a second order filter. But first, let's let's show the curves because we cannot do anything without the curves and let's bump this up so we can see what we are doing. So again we can see the responses are all over the place, and we need to restrict the band with off each a speaker separately. So let's I usually start in the same fashion. I add a second order filter for each branch, so let's do this right now. So let's have the second order filter for the base side, which is an in doctor and a capacitor and, uh, ground. Now let's connected speaker and for the mid range driver, we have, Ah, second order High pass in a second order low pass. So we get that band pass effect and I usually place that the high past first. So a capacitor and conductor and then we have, ah, the low pass section which is the other way around an in doctor and, uh, capacitor. It's places correctly and it's connected to the ground. And this is the band pass filter. And now for the twitter, we have ah ah, second or high pass. So basically a capacitor and an in doctor in a barrel with the speaker. And this is our starting position. And, uh, because the values are basically incorrect, you can see that in our frequency response chart, everything is going crazy, but we will fix that in a bit. So now we are going to focus on the base and on the mid range, and we are going to fiddle with the components off the Bayside and Onley the high pass section off the mid range. So basically the crossover point between the base and the mid range, and after that we shall take a look at that. The other side where the frequencies cross between the midrange and Tweeter, so we have to cross over points. First, we're gonna take care off the one on the lower frequencies. So let's adjust the components on the Bayside and see if we get something to resemble a nice roll off curve. So let's increase the value of this in, Doctor. So we bring the response toe Amora manageable state. Let's increase this. So now we got the response off the base side to look to look like a normal, uh, normal roll off slope. And as we can see, I think the mid range is hiring. Output can really tell because we have this huge peak over here. But I think the mid range is higher in, uh in output than the Bayside, and we probably need to implement an attenuation pad for the mid range. But first, let's, uh, I think I'm gonna work with these components because I hit that peak. I It kind of distracts me. So let me fix that just a bit. There we go. And now I'm not going to touch these components anymore because we were not focusing on that part right now. But let's focus on this section, which is the high pass section and see if we can get it to a match with the base side. So let's increase this capacitor to bring the response over to to the base side and who, right off the bed, we can tell that the mid range is higher output than than the base side. So let's implementing Attenuation pad toe, bring the levels to the correct values. So one resistor in series and one in parallel. And let's who kept roughing up. And I think we're getting there. So let's leave it like that for a moment and see what we can do about it Now, Aziz, you can see the Tweeter extents over here. So this ah, disrupt our our crossover point between the mid range in the base. And I said that we're not gonna feel with the treater until we get to that specific station . But we have to do it now because the tweeter extends below 1000 hertz in it. Um, it introduces errors into our summation off frequency responses between the drivers. So we have to increase the decrease this capacitor, so we don't see that the response anymore in our graph. Okay, so Basically the Twitter is out of the way. So now we can work on Lee with the mid range and the Bayside, and we can see that the response goes down a bit. This means that they are somewhat out of phase. So, um, we shall see what we can do about that. So if I invert the phase, we can see the response goes up. So let's work with these components as we as we can see right now, we got a good frequency response between the two. But the problem is that, um the face responses isn't isn't great. So at this point, I'm thinking off shifting the phase off the base side. So heading another in, doctor, over here. Let's see if that helps us in any way. Now, don't be afraid to add complexity to the crossover because you can always remove them afterwards. So we see we have a this additional in doctor now, and you can always check between second order and third order. We have third or now, but if I short this out so we have deception to short this out. Basically, the in doctor is not there, and we can compare between two responses and see if this in doctor is a good choice in that particular spot or if we can remove it altogether. So let's, uh, take it to the normal state and see what we can do about this crossover. So let's tune this, and we're starting to get that the weird hump. And, ah, like I told you how to remove that, we can add the resistor in series with the component, which is in peril but the speaker. But then homes is way too much. We should go to something like one home. There we go attained that peak. Let's see how the you can see we're still not getting that phase response that we wanted. So now when you find yourself in this situation, you can do all sorts of thoughts of you can still play around with the components, which you have on the circuit diagram. Or you can increase the complexity off the the the filter, either on the Bayside or the midrange side. And just because we are playing around with things and experimenting, let's add another capacity to the Bayside and see if we bring this toe 1/4 quarter. What will happen. So if we get ah, nice face response so we can see the frequency responses, it's it's OK. It could be better. But the first response is, uh, uh is definitely not good. It's funny cause they're on. The sweeter side is good, but the frequencies passes is it's terrible. So let's play around with these with these components. See if we can change the slope, toe a certain toe, a certain angle where the phase alliance to the correct spot. So let's bring with this capacitor. So as you can see, if I increase this the value of this capacitor, we get, uh, into the right spot. Let's see, let's work with this. Let's let the mid range like this for now. And let's work on the on the base side. So let's work on the on the oversee. We got a little dip over here. Let's see what happens when I think if I increase this year, if I increase this, I get I get a bigger dip. But the frequency response is not that great. No, no good. I think that's a great looking dip. Let's, uh, bring this up. Let's take a look at the frequency response now, which is okay, ish. So this is the point where so we see we got a nice flat line over here. The dip. And now let's check the frequency response again. Now, this looks pretty flat to me. So we got a nice, um, we got a nice phase alignment and the frequency response is pretty flat. Now. We finished on the first parts of the cross over, so we got the crossover between the base in the mid range, right? And now we basically have to fix that twitter because it's all over the place. So let's scroll up. You can always use the function, right? Look and fit. All So we got all the components on one page and, uh, first of all, I'm going toe adding a thing vision pad because we have to attain that twitter down. The output is way too high. So let me had a resistor and another resistor. I'm not going to mess around with the attenuation pad on the mid range side because you can match the levels between the twitter in the mid range by fiddling with the intimation plant of the mid range as well but that will disrupt our alignment with the Bayside. So I'm gonna leave the attenuation pad on the MENA region alone and focus only on the one which is on the Twitter side. So let's, uh, let's work around with this first, let's let's work on the known of filter because it looks pretty bad. Now it looks better. No, let's check the phrase response. Who? It's already good. So basically, we have a second order at 4.8, which is pretty good. I can't believe this is, uh this went so so smoothly. We basically not even fiddle with the components on the Middle East side. So, uh, so there is a nice stroke of luck. Let's see. Yeah, this frequency response looks pretty linear in the face of response. Looks good. This well might be. Maybe there is a slight, but this is the overall frequency response of the mid range, and it doesn't sum up were with the responses off the twitter and or the base I This is actually the response of the mid range alone. So I want to point out there's something over here. I already did this cross over a PSA test before I filmed it. And the fact of the matter is it was a different crossover than this one. It was actually more complex than this one. And I will show you right now what I did before. And, uh, the point I want to stress out is that crossover design is so unique that you have to try different components, different values. And if you put your effort into it, you can do a simpler design or better designs if you do trial and error stuff. So in my case, you can see I have a second order filter on the sweeter side. Let me show you what was my initial design. So I had a 44 there on the Tweeter. So what I'm gonna do is I'm gonna save this frequency response, and this is the the crossover with a similar design. And then I'm gonna show you my first iteration, and then we are going to compare the two and see if we got slightly better response for the added components or not. So you can see that if you try different combinations, you may end up with with a simpler design and which is better for the overall audio quality and it keeps the cost of the crossover down. So let me show you my first iteration. So I had a 44 door on the Tweeter branch, not a second order. So let me delete this and let's had a capacitor and on in doctor the help that stays the same. So I had here on 8.2 and here I also had resistor. Let's have the resistor and this this waas zero point 0 56 and this was 0.6 home. So with 0.6 Okay, this capacitor was 4.7 and this in doctor was exactly one million for Prince Evident then Holmes And on the midrange society this was 0.82 and this was three micro Farid capacitor with resistor also which was three Rome resistor. So, as you can see, this was the This was my first direction. As you can see, the crossover is has two additional components not counting the Australian sisters because resistors r really cheap components But it's a simpler design nevertheless, So the extra complexity is not a good thing. And if I reverse the phase, you can see, we got the notches in both places, so I guess the response looks basically the same. This is a 4.7 crossover point, which is has basically the in the same spot. So my conclusion is, don't get discouraged by crossover design. Try different combinations and add complexity the crossover, and then try to simplify it and remove components and get to the spot you want, where the frequency response is to your liking and the face response is good. You have that deep notch when you invert the polarity of the speaker. So if we look at the the impedance response, we can see we've got a dip below for alms somewhere in the higher frequencies. Like I said, the this shouldn't be a problem for most amplifiers. So this is our 1st 3 way crossover. Didn't look that complicated. We got it fixed in the end. So, uh, let's take a look at another three way cross over now for a floor standing speaker 29. 7.1 Project 5 : Description: Now it's finally time to talk about our last project, which will be a floor standing speaker with the dual woofers and the base reflect sport. This is, Ah, three way design. So we have the Twitter, the mid range and the dual wolfers. Now, this is a huge enclosure because because the base drivers have ah, really high via is number 150 litres each. So, uh, actually these enclosures is quite quite a bit of compromise. I managed to squeeze and 135 liters off net volume for the base side. And the tuning frequency is not that greatest somewhere around 23 3 hurts. But since then, this enclosure is so big, uh, you have to put the really small port decide. So basically, I just made the four inch port with an inner and outer flair and the part between the flares I did it as minimalist possible. And this led toe Ah, resonant frequency off 23 hurts. I used them a Mastic sealant. And then I used the screws to fit that the port to the enclosure like we have seen in other projects. I route the mid range and the Twitter cables through the holes off the binding posts and used some silicon sealant, toe patch everything up and the bass drivers I wrote the cable through the base reflect sport. So let's look at the measurements because there is something interesting here. So I have here all the measurements the far field measurements, which is the base, the midrange and tweeter and the combined response off all the speakers wired in parallel for determining the critical center. And then I have the near field measurements to get the full frequency response of the base and off the mid range to aid us in playing the labor board transformation. But let's stop and look at the far field based response. When I first looked at this response, you can clearly understand me why I start scratching my head. There is a big no in the frequency response somewhere around the 1.75 K and usually you start, think about maybe there's a cancellation. But if there were any cancellation, uh, especially the low frequencies will be affected. Not nothing. The high frequency region. But then I realized this is basically a sub before, so the part over one k. It shouldn't be playable for a sub over anyway. So actually, what happens here is that there response starts to roll off naturally, starting with this frequency like 6600 hertz. So from 600 hertz, it starts toe roll off naturally, and then is thirst to rise. Why is this happening? So this is actually the cone breaking up. The cone no longer acts like a rigid piston, and it starts toe wobble around. So the part the part inside the the cone itself transforms into mini speakers which are very efficient at high frequencies. So this is not the not the actual response of the speaker itself. It's just a cone breaking up and producing. We can call it distortion because this is not the natural response of the speaker, but the actual cone breaking up. So that's why you see this rise in high frequencies and then they naturally roll off. So even if you look at the frequency response chart coated by the manufacturer, you can actually clearly see this null over there. So there is nothing to worry about. We're going to use this response as is if you look at the mid range. There is all very normal. And if you look at the Twitter, we can see a dip over here, which is not natural to the response of this. Peter and I attribute this to the baffle itself. Somehow the the beautiful it's determining this dip, maybe the positioning off the other speakers because they are not the flush with the bathroom. Maybe they are producing this artifact and we can see this is at the 4.2. Usually, I tried to see if this happens in multipliers. Oh, for 4.2. So I know there is something with the baffle. So 4.2, its 8.4. So nothing wrong over here, 12.6. So here we have another dip. So 6th 16.8 So we can see that another deeper happens at a multiplier frequency. So that might suggest that, uh, it could be a problem with the baffle. Some some other Bethel dimensions might be more fortunate. So this is the response of the Tweeter. And of course, when we are going to choose the crossover point, probably we will start with the Twitter rolling off from here to choose the crossover point somewhere around four K g five k to get rid of this dip. Because this is a problem. We cannot choose the crossover point here because we have issues with this. This effects of the linearity quite a bit. So you understand why we have to choose the crossover point in this area to mask this deep with the response of the mid range. So if I go to the base, let's here. Of course, we have to stay away of this deep and try to choose the crossover point somewhere in the 400 range, Toby in the safe zone. So actually, we are going to stress the mid range quite a bit. Well, actually not. I'm sure this mid range can handle 400 hertz. No problem. And, of course, going into the upper frequency because we need to mask that deep flow from the tweeter. Now, before we go to X Emma, where I want to show how to determine the acoustical center. Let's try and take a look at the impulse response. So if you look at the impose response, there is nothing out of the ordinary for the mid range. Nothing out of the ordinary for the Tweeter, the big spike, the DK and then the room response. This is the the room response picked up by the microphone. But if you look at the base, we can see the this is did. This does not look that that great. You have to realize that when you're measuring floor standing speaker with dual wolfers and the base reflects sport on the front baffle. So if the port would have bean on the back panel, that would imply more accurate measurements. But since we have that port over there, you can see that if you place the speaker directly on the floor because it's high enough that the microphone would be at Twitter level somewhere around 1.1 meters, 1.2 meters because the speaker is very tall. But the problem is the port is very close to the floor and you can understand you will have very early reflections from the port when we are measuring the base speakers Far field. So in that case, I had to place the the the speaker on that is basically support for enclosure, but you can use any things, books or something like that toe Elevate the speaker somewhat normally in in my room in adequate conditions. When we are measuring a bookshelf speaker, for example, I can go and quickly to about four point six milliseconds. So right about here. But as you can see in this case, 3.6 milliseconds is all that I got from raising the speaker, uh, above the floor. And you can see this even though I did this. You cannot clearly identify the room response from the from the actual director response. So you have to take into consideration when you are doing this kind of measurements. The Bayside will not going to be that a neck read mission. So if you look at the Twitter, you can see the response has shifted quite a bit toe 3.7 milliseconds because I have to raise the speaker up from the floor. The twitter is not, is not at the halfway distance from the ceiling and from the floor. And since I raised the speaker, now that Twitter is closer to the ceiling and it creates on early reflection and we can see it over here now, the gate window for the Twitter is shorter. If I go for the mid range, you can see that this is this is better. This is that for milliseconds because it's not that close from the ceiling and the window is larger. But when you're doing the combined response, of course you are going to take the the shortest the window for ALS, the all the drivers, the wired in parallel. Now that I talked about the measurements and showed you the differences in these kind of speakers, let's head on toe x mnc. What we can do about the acoustical center because this is this is will be a bit challenging for, ah, for standing speaker with so many drivers now, here we have all the drivers in place with the Hilbert board transformation apply and the effort against the Emmy files loaded up. And the first thing we have to do is low the combined curve. Okay, now here is the combined cuff and we can clearly see Let's raise this up a bit, okay? And we can clearly see a definite gap between the two curves. So we know that we have quite a bit off acoustical offset. And, of course, like I said before, remember that these measurements when you have, Ah, the speaker so close to the floor, you have to understand that the accuracy of the mission, it is not that great. So you have to accept some kind off every percentage. But in we'll start to with the Bayside, because clearly this will have a big delay. Let's full screen this. So let's, uh though this bit by bit so we can see the response shifting in our favor. And there we have. If I increase this, you can see this matching better over here. But this is going the wrong way. But remember, we can Do you know, the the delay for the mid range is well and for the Tweeter as well. So we can see if we shift this. We get it closer to our There we go. So this is kind of matching. This is matching Good. Good. Now here, this is bad. What? We can go ahead and fiddle with the Tweeter as well. So let's increase. This is as against he were going toe the right direction. Now, this is the part where I talked about where, um, you have to accept that that's light error because you can see we don't have these squiggly lines. Here we have, ah, line that crosses through. And there is a slight gap on each of these measurements. So let's go to the base site, see if we can do anything better. Actually not. This is doing things worse. So let's stop here. Let's try to tune the mid range again. Nope. This is good. So when you have a delay for each driver individually, there is no point. You can leave it as is. But normally there is no point in having delays for all of the drivers. You have to take a reference point for one driver. So all right, this has the least amount of delay. So this is the closest with a microphone. So we have 0.17. I'm gonna take this 20 and the subtracted 17 from this self. 0.6 47 minus 0.17 equals zero point 30. And subtract another 17 from this guy. We're here. So one point five. And as you can see, the curves match exactly the skin exactly the same. Because the critical difference is the same. So now we have everything in place and we have to start with the actual crossover design 30. 7.2 Project 5 : Crossover design: Now it's time to do the crossover design for our last project, the floor standing speaker. And from the previous lesson, we have everything set up with the acoustical offset in place. And let's raise this curve of it so we can see what we're doing and check all the drivers, the Tweeter, the mid range in the base and, um like we know in the three way design, we have a bit more leeway with the actual crossover points because we have on additional driver. And, uh, we're not stressing each year driver that much. But in this case, I said the in the previous lesson that we will choose the crossover point for the Bayside. So between the base and the mid range, somewhere around the here, around 405 100 area. So we stay away from this big dip over here and we are in the comfort zone of the mid range as well. But I already designed the crossover with this crossover point and the advantage off having set up approach that cross over. So I actually got the chance to listen to it. The problem is, the midrange sounded the muddy I don't know how to explain this. But the problem is when we have this dual first set up and since the wolfers air big their 10 inch woofers, the distance between the second wolf for so the one down low and the mid range is pretty large, so that distance is considerable. So we know that when you are designing a speaker, you have tow place the speakers as close as possible toe one another. So you can understand that that this gap is not favorable to our and go. So usually there's a rule of thumb that you have to place the speakers very close to one another, and when you have to place them somewhat at a distance, you have to choose the crossover point so that half of wavelength off that particular frequency. So the the crossover point frequency does not. That distance does not exceed that half wavelength off the crossover frequency. So if we choose a growth of four point, which is lower in frequency, we get a higher wavelength, a longer wavelength, we get a longer wavelength and that permits us toe place the speakers further apart. So in our case, I find that that we need to choose a crossover points somewhere lower. We are going to stress the mid range a bit, but I think it can handle it. So somewhere around this area, about 300 hertz, maybe even over, let's see what we can do with that. And I'm gonna test that and see how it sounds. So it So it's always best to check how the crossover actually sounds, because doing it in a circuit diagram and looking at a frequency response chart will not tell you how the speaker actually sounds. So for the bullets do the usual, you know, three way design. Um, let's start with the second order filters for each branch. So let's start with the Bayside and place on in Doctor and Pastor, That's where this to the ground. That's the Bayside. And now let's place a band Best filter for the mid range. This means the high pass and low pass second order. Of course, that's gonna everything up great. And the high pass filter for the Tweeter? No. Okay, so as well as usual, let's start with the the lower part. So the crossover between the base and the mid range, so we're actually gonna play with these components and the high pass section off the mid range. So, like I said, I want a low cross over point. So that means I'm going to increase this in Dr to a lot, okay? And this capacitor also. So I get this is so much. Okay. Eso usually when you will have ah, fine doctor like this or 4.6, actually, that's in a real value 4.7 a normal in Dr Value. And usually when you have such high value the resistance, let's say you have on Air Corp the one millimeter, uh, thickness wire. You will probably have here, like 1.5 if you have on Iran core or ah, higher gauge wire than this will go down. But normally you get a one millimetre wire Air Corps, and this will have some slightly above one home in resistance. So you have to take that in the consideration. No, no, no, no. Let me tune Disk a. Pass it, er no. Okay, so let's leave it like that for now. And let's play with the midrange because it's looking like a big mountain over here. So we're going to cross this driver really low, Like we said somewhere around 203 100 Hertz area. This means that this capacitor will have a really high value so you can see the response shifting to the left somewhere over here, let's say a bit higher compared to the base. Oh, he's on. Now let's play with this in, Doctor. Oh, that's good. Sometimes usually like to play with the resisting and says, Well, so see what that does. So we're getting, like, a freeze a light month if I, uh So you see, we got a nice flat response here, but the resistance will not be that high for this. In Dr Usually for a two million, Henry. Ah, the doctor will be like no 0.5. But we can place a Siri's resistant with this with this in Doctor and we can change the value of the resistor. So it was like five bones. Let's look at the response. And so now we see were about 270 hurts. Well, she would got a nice flat line. The phrase response could be better. I guess so. Let's see if we can do something about that. This shift this aligns the face better, but we don't want the crossover to be that low. There. Go. And now that's reverse the phase and we can see that the responses Priti, Priti Lee! Yeah, See if I increase this. Nobody in ruin the face. Yes, I did. Okay, So let's live this like that for now, Let's play on the Twitter side. So the twitter is playing we over here. So let's decrease this capacity in this in that there is well, it's really do is just be extend. That's and I resistor here to fix that hump because we need the midrange to extend in tow The higher frequencies on this capacity that you see this resistor is way too big. So let's reduce the value. Looks like we need to place in the thing we should pattern treat around something like that . But that's it looks about right. So let's place in Attenuation pad, see what we can do next What? So that twitter is way too low? It's increased the level Well, I don't want to go too high, so we see this high frequency response is on the lower side Now this isn't the problem. Most of the time. So when you see the response over 15,000 hertz that is bigger than the rest of the frequency response so or like in this case is somewhat lower than the rest of the frequencies. You shouldn't be concerned with this part of the chart that much. Some guys might, uh some guys might put an emphasis on this part of the frequency spectrum. But bear in mind that Aziz, you age, you really cannot hear much over 16,000 hertz, even from a young age. So if you're like 30 or more, you're gonna lose that part of the spectrum. Your years will not be able to pick up those high frequencies anyway, So let's see what we done. Oh, so the phrase response looks good. Let me see if I can get better linearity. So, like I said, this is such a small town in Dr Will not have this kind of resistance somewhere like 0.2. So let's place ah Siri's resistor with this in Doctor so we can play as we like. Uh huh. So you see, as the Peter goes down, we we can see it actually cancels the response of this meet range. So in this case, I might try to put 1/3 order filter for the Tweeter and see how that plays out. Starting to better already. So we see we start to rise. This which is good. Actually, this resistor is no doing any any good here. So we're gonna show a short this out. So let's say we place this at one home and look at the frequency response. When we shorted, we get Ah, better, more Lena response over here, which is great. So let's look at the please is looking good. Let me tweet this. Nope. This is clearly better. That doesn't have much their fees. So this is linear. Maybe we can get it linear more in this Bayside. So let me let me see what we can do about that. I think that is a bit better. Yeah, things that that looks more lean year. So the crossover points somewhere around 250. That is a beautiful push for the mid range, but, uh ah, I think it it can handle it because it's ah, five inch driver. So let's check the face again. Everything looking good. So now if we look at the the Beaton's chart. We can see this dip over here, which is which probably isn't that good. So I think we have to fix this. Maybe if you can do a perfect cross over, you can check it out and see if your amplifier is struggling or it, uh, goes into protection. But if you have to fix this, let's think about it. How we can fix this. This is on the Bayside. So to raise the impedance off this, we have to place a Siri's resistor with on on this part of the circuit. So let's look at this And doctor, if we raise the resistance off this, the graph over here should shift up. But also, when we're doing this, the frequency response will, uh, alter its well, so you have to balance everything up. So let's increase this, see what happens. So as you can see, the chart shifted up. But as you guess, you were losing over here so you can see that we are losing, uh, some low frequencies. 2.5 were barely on the 40 mark. So we keep our based response and the impedance sharp start to look good. So basically, this is our third order, a crossover designed for our floor standing speaker 31. 8.1 Component Voltage and Power: Now, before you start choosing your components for your crossover axe, Emma has a need little function where we can check out the voltage and power requirements for your components. So if you will have, ah, high voltage treating where a capacitor, then you should look a tow, a film work a pastor. If not, then you can go for a cheaper alternative so far on electrolytic. Now, if you go to add graph, we can and click more. Let's go for a component voltage. So we have this little graph over here right now. It doesn't show anything because we need to check the curves so the let's go for the capacitors. So let's check older capacitors. So we have them on the little chart over here and to see their their ah voltages requirements so we can see that all of the capacitors have, ah, a rating below 10 volts, and that doesn't seem to be a problem. But the issues here the number is so low because the amplifier power is said toe one what? But we need to simulate a scenario where the speaker is at its highest vote. So this crossover design is for the three way speaker, not the floor standing one, the smaller one. So those drivers, they won't exceed like 50 watts off power. So if you have ah, 50 watt amplifier, most likely you can drive them to their full capacity. So if we increase the power off the amplifier to 50 watts, we can see the chart going up. So let's stop at 50 watts on this closest up, and let's full screen the chart, and we can see that the highest value is founded. The blue curve, which is the see one capacitor one, which barely exceeds 50 volts somewhere around the one que Hurt frequents. So this means that we can choose electrolytic capacitors for all off our capacitors from this crossover design, because an electrical will have 50 volts rating. 63 8100 though those, uh, if you have ah, better one. But usually they're not below 50 volts, so any capacitor will do now if we talk about the induct er's. Usually if you have one millimeter thickness where you won't have any problems with the induct Er's either air corps or Iran core usually should take a look at the resistance and don't worry about the doctors heating up, because if you do find that you're in, doctors are heating up, you probably have other problems as well. So let's says, Keep the in Dr Part and let's focus on the resistors. So now we need to check the power and link. So let's go for a component power dissipation and let's check all the resisters So we have . Our three are for our six r five R one and r two, and you can see that the power handling power dissipation is off the chart. And let's skill our chart. Let's go to 10 watts and here we can see that the blue and which is our three. It exceeds 60 watts. So you know you need a beefy resistor here, the ones with the hit things on them because you will have a lot of a lot off power dissipated into this resistor and our three. We can find it on the Twitter in series with the Tweeter, and you need the serious resistor there. So if you look at the other ones like the Red Line are four or the black one, which is our five something like 20 watts 25 watts will be sufficient. So for the red one are for where we confined our four. We can find our four over here. This one then what resistor will do. So this will help you out in determining the actual components. And in the next lesson, we will focus on the values because these components have, ah, pretty standard Hooiveld. So let's see what that is all about. 32. 8.2 Component values: Now it's time to build an actual crossover. So this is the design from the MTM speaker. It's the crossover. We designed a few lessons back, So the first thing we will do is check the the voltage and power. Anything's just like we did in the previous lesson. So let me do this here really quick. That's so like the voltage. And, uh, check all the pastors so we see so we can see if we can choose electrolytic alone or, ah, film capacitors. So let's join this up to, like 50 waas. You can even go toe something higher just to be on the safe side, and we can see that allow capacitors have low voltage readings. So even this one that yellow on the C one, which is the 1st 1 in series with the Tweeter, has a rating off peak rating off 16 volts. So any electrolytic capacitor will do. But I'm probably going to use film work. A pastor's. It depends how you want to make your project because budget is an important factor. So if you wanted to be low budget, then you will clearly use electrolytic capacitors unless the voltage rating requires otherwise, So for the in doctors, we have a 0.10 point with the to. For these two, there is no point in using something else other than air course. Because these are low value in doctors and are cheap anyway. And for this 11.5 million Henry a decent Their court shouldn't cost that much. The only component which which looks expensive is this capacitor over here. So 24 microphone with a capacitor Ah, film foil capacitor will be decently expensive. So you might, uh, exchange this one for electrolytic If you want to keep the budget down and let's check this resistor also and see, I see the wattage. Oops. So let's selling her one and we can see it So it's over the chart. Let's kill this down. 50 watts, 50 watts. Let me turn it down a bit like 50 together. Reasonable number. So yeah, something like a 25 watt, maybe higher. To be on a safe zone, get a resistor which can handle a lot of wattage. And now that we've finished with this part, you have to set up the values of the components. What went me if you June. A certain component, you can see that it says 12 micro fire. It's if I go up one match. It goes to 13 this fortunate coincidence, but now you can see it's 15 not 14 16 18 2022. These numbers are standard numbers when you when you look at what the manufacturer produces in terms off values, off capacitors or in doctors, because they're not in the increments off. One. So 12 13 14 15 They're not like that. So I devised an Excel spreadsheet blitz, which he over to that Excel spreadsheet you can find here the most common values for ah, capacitor and foreign doctor so you can check your when you are using exam. You can introduce whatever value want, but it's better to use the standard values. So let's say, for example, you need a 1.2 Farid capacitor. You don't have it over here, but you can. You can combine them. So let's say you combine one microphone with 0.22 so you get a 1.22 micro Farid capacitor, and here you find a little calculators. So if you wire capacitors in series, I will show on the screen the formula, because it's a more complex formula, but for parallel you simply add them up. So if I say one to you can see we have the summer off all these components when you wire the capacitors in parallel. But if you wire them in Siris, you get, like I said, the formula, which I mentioned previously. So you get, uh, some awkward number to combine them in different position. Maybe you get the value you want. So if you need a higher ah, higher capacitance capacitor and you don't have it, you can wire Maurin parallel and get the desired result. And for the in doctors is the other way around. When you wear them, Siri's, you simply add them up so we can see each number is adding up way. Here we get the total, and for the parallel, it's exactly like on the capacitors are wired in series. So you have this formula, which I show you on the screen and then when you wire in doctors in parallel, we get something like this if you apply the foreman. So when you are choosing your component values, you can check this table to see the most common values, and if you don't find them in the stable, then you might combine them in different ways in series or in parallel to get the values that you need, or maybe that you don't have. But you can work around by combining different components and getting the value that you want. 33. 8.3 Inductor placement: Now, before we get into the actual building off the cross overboard, I want to talk about in Dr Positioning. So why is this relevant? Well, you can find this. This is a conductor schematic online. I don't know which the original sources, so I cannot quote the original source because it you will find it over the internet. This in Dr Positioning when you have the board ready. So here is my example. I have a board which is made out of indie. If we have, ah, four millimeter and give. But I stacked two boards and glue them together. Also, I drill the for pilot holes for the screws, so the holes are champ erred. So we get this cruise nice and flush with the board. I will show you how we will screw these to the enclosure later on. But now let's focus on the board and own in Dr Positioning. So if we look on the schematic, we can see that we have to place the in doctors far apart as possible. So in our case, I have to in doctors, to a recording doctors. And if we place him like that flat on the board we don't want Oh, stepped in close together. You want toe, take them further apart as the board allows. So this is, ah, scenario, that might be plausible. But if you really want to get better with positioning, you have to imagine that somehow they're shooting something from the hole and you don't want to intercept those lines. So something like that will be really bad. Something like this will be better. Something like this will be even even better than than that. So you have to spread them as far as possible from each other and try to not intersect the lines which come through through their holes. So you can see on the last scenario this is the, like, terrible Don't don't do this. So why do we have to spread the day in doctors apart? It's because they generate a magnetic field. And if we want him up together, those fields interact with each other and the value of the in Dr changes. So you since we have a three in doctors in our crossover design, we have to take this into consideration when we space them out, so we find a good position so they don't interact which with each other negatively. So since since we have finished with this positioning the size of the board is also important because you have to make it as biggest possible. So you have room adequate room for all of your components. But you have to make it small enough so it you find a place for it inside the enclosure. Usually you place it on the back panel or on or on the bottom panel. And also in our case, since the enclosures are already made, you have to make it small enough So it fits for the biggest hole. So the mid bass hole, you have to check this in through the enclosure, so it has to fit through that hole. So let me get the enclosure so I can show you what I'm on about. So now we can see the mid base cut out the whole which here the mid base will go. So if I take the board, you can see it goes inside pretty nicely and you can fight a spot for it down there, for example. Now if I take it out if I place it on the long side, it doesn't fit on Lee on this side. So? So before you start assembling, you have to check for an additional factor, which I will show you right now if you have a tolling doctor like this one, This is a tolling, doctor. If you when you try to space the conductor's apart, you're probably like to place one on the side. So something like this place it on the side, so you have enough room to place another in doctor on the other side. But watch what happens when you have a tolling doctor like this. When you try to to place the board inside the enclosure, you will find that the in Dr hits the side of the hole so you cannot place this board inside the enclosure once, although components are in place. So you have t to check for this and see if you have a tolling doctor. If you do have one, try to place it on the middle, so if you place it on the middle, then when the board goes inside a little bit difficult for me to hold it. But when the board goes inside, you can see now being doctor has adequate room to go inside. No problem. So we have to check for this for this particularity also. So now that we covered with the in Dr Positioning and the board, we should go ahead and start building the crossover. 34. 8.4 Crossover build bass side: Now it's time to build our actual cross overboard. So we have the board over here with the four holes that we pre drilled with a sham furred. So, uh, the screw goes in flush with the board way, have our soldering station and just spray some water in the sponge over here so we can clean our iron tip. This is for cleaning the iron tip. Also, we have some Salter over here, and this is Ah, something with about 220 degrees melting point or something like that. The lower the point, the better it is. More easier is toe solder. Thies is a pair of helping hands Be my needed. We might not. And here we have our components. We have the four capacitors three and doctors in the resistor. Here we have the binding posts and we will use thes. So I have these wires over here to read one black one in a red one, and I cream some term knows which go directly into the binding post. Usually these guns with their own on their own termination here, that goes between these two nuts. But I somehow lost them, so I got these standard, uh, standard crimped terminals that go on these binding posts. So, uh, this will go to the when we look to the cross over diagram, this will go to the plus side off the amplifier and this on the negative side. Now, to place that the components on on the board, you can use either zip ties, which we will use today. Or you can use hot glue or whatever. Uh, you might see fit. So to use our zip ties, we have to drill holes. 34 These the small zip ties three millimeter holes should be enough. And we will look, the zip ties in Fix that the components to the board. Now, before we start fixing the components to the board, I want to mention something about the in doctors. So we have these air corps in doctors. But if you use a moon door for air court, you see, they have this plastic casing and sometimes people Ah, they see this hole through the middle, and they think, Hey, I'm going to stick this on the board and I'm gonna screw something in this hole, and I'll stick it to the board like that don't do that because the screw is made of Ira and it will shift the the inductions off this in Dr. Because you're basically taking this air corps and saw how making an iron core hybrid or something like that you will. It will not have this rated value anymore. So you have to use something that doesn't affect the inductions off Thean doctor. So zip ties are plastic and they are pretty five. So now let's go and look at the cross of our schematic and see how we will place the components. So for a first timer, when you see you an electrical diagram, you somehow I think that when you see these lines connecting the components that you have to place a wire there and while it will work, you have to use actually if you can no wire it all. But sometimes when components are spaced too much apart and you have to go to the when you when you connect to the ground, sometimes you have to use wire. But if you can, don't use any wire a tall, so you have to use your imagination. And when you see this first diagram, you can see that you can. You have to look it like notes. So, uh, which component connected the plus side of the amplifier? So we see a wire which goes to the right and goes them to the wall for side. But actually, if you think about it, you have the capacitor C one and the in Dr L two which connect to the same spot in the plus side of the amplifier. If you subtract all the wiring, so that's what we will do right now we are going to, uh, fixed these three parts into a node and then we will continue with the rest of the components. So we have the conductor l two, which is the 1.5 million Henry in Doctor, and we also have. So we're going to face it to the board, something like that. And we also have the C one capacitor, which is a 12 12 microphone with a capacitor, and we have the plus at which goes to the amplifier, which is this wire over here. This will connect to the binding post, and the binding post will connect to the amplifier. And now all we have to do is drill some holes to look the zip ties and, ah, figure out how we will place these components. So, uh, in our case, we when you see a ground symbol, so we have 12 free for five connection to the ground, all those connection converge into one single spot. So I'm gonna try to place the components on the edges of the board and try toe, connect the ground somewhere in the in the middle of the books, have all the components like this. And when we have something going to the ground, all shifted to the middle of the board like that, going to the middle of the board. So in our case, we have our base side, which is that the in doctor and then we will have a capacitor which goes to the ground. And then I will go with another capacitor which go to the Peter side over here. So we have adequate spacing between the components so we can fit all of them into this small board. So right now I'm going to drill some holes to to fix this components into the board. So I'm going to use this small drilling and I would you see that the doctor is fixed to the board? No problem. And we and we used this sniffer toe. Cut the excess zip tie. Don't worry about this protruding on the back because I will show you how we can fix this issue and fix the board to the enclosure so it doesn't rattle about now. We focus on the components and how we fix them to the board and and sold them, or on connect them to each other and solar him up. So now we have to place a disco pastor, which goes to the Twitter side over this side. So we have to drill another two additional holes toe hold his capacitor in place. You know, we got the capacitor snugly in place like this. So now we have to connect this wire over here, which is the red one, which goes to the amplifier in it. And we have to make a three way connection here. Between this in Dr Disk, a pastor and this piece of wire that will go to the amplifier. So right over here, we need to make a soldier point and connect these three components so And what I like to do . Is this just somehow strangled the wire through the for the two wires of the two components . So they have a nice good the connection, and then I'm going toe solder them up. So let's start our soldering station and wait for the Iran tip to hit up. Now, I'm not expert in salt ring, but I can give you a few tips that I know off. So before you do anything, you have to clean your tip. So we're gonna use this sponge over here. We change. Use some straight, some water on it, and use this also and clean any residue that might be on the end of the tip. So you can see. Let me see if it'll focus. You can see that the tip is now a nice and shiny. That's it is how we wanted. Because if it has any imperfections and erase that used, then it will not make perfect contact, and it will not heat up your components as you would like to. So the basic principles when you are doing some soldering job, you want to heat, you want to heat on the one on the one side and place the solar on the other side because solder is going towards the heat. And, uh, you don't want to do something like places on the tip and then over here, because that is not a good connection. You have to place the heat source on the on the one side and place the soldier on the other side, and the soldier will go through the wires and the the the ends of these components and goes towards the tip. And then the seller goes through everything. And when it when it hardens, it's it's a good connection. You might want to apply just a bit off soldier on the iron tip and then place it like I like. I told you, - There we go. We've got a nice good connection over there, and now we can sleep the excess of wire. So that is our first connection. Let me stop the heating station. So that is our first connection. Now we're going to finish with the Bayside, so I'm going to add the second toe of my preferred capacitor. And again my orientation will be towards the centre because that is where I will. Our ground point will be. So I'm going to place the capacitor some something like that toe oriented, this this part towards the center. So here we have. If you look at the circuit diagram, we have another three way connections. So we have between this and Dr this capacitor, which go to the ground and the positive side of the speaker. Right now, we're not going to connect the speaker. I don't want to clutter this board too much for right now because those are long wires and will make our life a little bit difficult. So I'm going just I'm not going to make the center point. I'm just going to fix this capacity, the board using the zip ties. So now we fixed that this the capacitor to the board as well, and we made we made this connection over here. We still have to place the wire that goes to the plus side of the mid base. But we're not gonna do it right now, which is gonna live it like this. And we'll do this solar point a to the end. So let's leave this right now. Let's leave this like this for now, and let's continue with the Twitter side because these two are all of the components for the basis 35. 8.5 Crossover build tweeter side: So now if you look in our circuit diagram, we can see that the next note is between capacitor C one in dr L one and capacitor C two. So we need this small in Dr Off. 0.22 million Henry's and the big capacitor the 24 27 micro fire in the past. And now, like we learned in the previous lesson, we have to take into consideration the doctor positioning. So the best way to place that. So since we have this flat on the board, this will be in the same position is the 1st 1 But but the further apart, that's possible. However, this will be an even better solution. So we have a different orientation from their centers. So we're gonna try to place it like this so this one will go to the ground. So I'm going to, uh, the second end of the doctor I'm going toe orientated towards the middle off the board, and we have to find some room for the second capacitor because this is a big one, and it's going to be something like this. So I'm going just going to screw now. I'm going to drill the holes so I can feed that these components to the board case . And now we got the conductor in places. Well, so we got it in the correct orientation so we can see it's going into, uh, the other the other way. I could have placed it a bit more to the end of the board, but I think this will be easier for me to make these connections. I got a place, this base it a bit more. It would have been better if it was a little bit further up, but I guess it's OK over here as well. So now we get to make this solder point over here into these three components, and this will go to the ground. But we will make this connection after all of the all of the components that go to the ground into one big seller point. So I just twist. I just creased these wires to get her, and now we're going to make a night a nice ah solar point over here. So let's start up our soldering station and wait for it to heat up. No, it's close to Solar Station and we can see We made a nice solder joint over here and I want to sleep the excess wires with this trimmer and there we have it. Now, if you look good on our circuit diagram, we have the the notch filter and we need to connect three components in Paris. So we have this resistor this capacity in this in doctor. Now we know what the resistor should be like a 20 watt resistor. But if you have, for example, only 10 What resistors and you want toe be able to handle a 20 watt. So for heat dissipation, you can, since in our case, we have with 3.3, you can wire to 6.6 resistors in parallel to get the same resistance, but with a double word with the double the power handling. So you can take that into consideration if you don't have the adequate power handling resistors. So now we have to find the additional space on the board to hook these three components. And we have to be wary about the induct replacement because we have three of them now. So try to imagine where we should fit this in, doctor so that it's adequately place. Now the correct orientation will be something like this. So let me see if you can see it, right. Places like this, Because this way, like I like to say none of them none of the in doctors looks the one another. Just so now you can see a fix the capacitor in the resistor off the notch filter, and I slip the the ends over decide. Because here we have to make a four way connection point. We still have to place the in doctor. We have this capacitor, this one and the resistor and the in doctor, and I'm going to place it. Ah, like we mentioned like this and tried to brought with the end over here so we can make the four way connection point over here. And we have to make the three way connection point, uh, down over here. So I think this is an adequate position regarding our needs and, of course, the conductor positioning as well. According toe the other two case I know the doctor is fixed in place, and now we have to make the four way connection over here and then make a seller point and three way Actually, the here is the four way, because we have to connect the plus side of the off the tweeter as well. So, as you can see, the needle nose pliers is really good. Toe bunch up thes wires altogether. And now we're going toe assaulted them over here in Sleeper, the ends. So, like I said, your place just a little bit of solid on the tip, and then place the tip on the other side where you're applying solder. And then I started making the connection. Like I said, the Sutter will go to the heating point so you can see made a nice connection over there. And now we're gonna stream the ends like So Now we have to make a four week connection over here, but we have to get our cables. So this will go to the blast side of the Twitter and here will go to the plus site off the wolf for So let me get that. Those cables. These are exactly the same cables I used to make the test enclosure. You can remember I sell them up, are using silicone through the binding whole posts. But you simply rip them apart and the silicone goes away. And, uh, this, uh, we have two sets. This is the mid based site, so we have two sets that goes to the mid base because we have an empty em with the dual meet based drivers. So here is the positive that goes to the mid bass and here the negative, which will go to the ground point. So basically, we need to connect the plus side to the Bayside over here. So this, uh, this capacitor doctor the three way point on the the three way connection point in the mid base of branch. So the induct er the l to induct er the C three capacitor and the plus side off the mid base. We have to connect them all in this boat. Basically, just find a way toe, make the cable stay in place until you make a starter. The starter joint. Because you don't want them moving around while you you are soldering the wires together. What? So try toe. I like to strangle the wire between the wires off the components because those are easier to work with. This is not so flexible. So playing the tip like we did before a place him sorter and then heat everything up. Oh, you can see this. Got a nice joint over there, and I'm gonna sleep the ends. Now. I have to do the same to this section over here only we have to connect the Tweeter here. So we have a separate a pair of cables, this one with the with the smaller crimped terminals. So we have to connect the plus side toe a bunch of these three together and hooked this Peter positive wire as well. See you. We got this done as well. And now if we look at the crossover diagram, we can see that all it there is left to do. We have to connect the the ground point. So the ground point we have this one of the negative site which goes to the, uh the amplifier, the negative side which goes to the Tweeter and the negative, which goes to the mid base and also these two components in the middle. So we have in dr L one and capacitor C three. So we have to make a big blob in the middle. So let's see how we can do this. I just twisted these wires into, ah, big bunch over here and see how I can fix them in place using the wires off the components . So I just made a big mess of wires over here, so Ah, and we're gonna play a bunch more Southern. So we can, um I'm not gonna close it up to the camera because I'm afraid that the wires will fall off , so they have to maintain their positioning. Why? I make the solder joint, so let's clean the tip and apply sufficient soldiers. So your fingers connected correctly. Now that looks like a messy solar point. But that's how it is. Sometimes you can use, uh, if you have too many connection, you can use a no bar, something like a bar with you when you can make individual connections and hook every wire to the same spot. But, uh, in this case, I did. This is not the most elegant solution, but it does the trick. So basically, this is our crossover boy. Our finished the crossover board. And now let me show you what to do about the backside. So we have all the wires here that go to the way have the plus and minus that go to the binding post also to the amplifier. And here we have the mid based connections, both of them. And here we have the Twitter connections, all going to this same cross overboard. So for the backside, let me show you what can be done. 36. 8.6 Finishing touches: now the crossover is complete. All we have to do is fix it to the board, and we can see we have dismiss over the back side, so we can't really fix it directly. But I can show you what we will do to make spacing between this and the actual back of the enclosure. And also to, um, add some damping because we don't want this thing to rattle about the enclosure so you can use the spongy with adhesive announced on one side. But I have some some rubber grommet over here, and we are going to place this on the back side and, uh, like this and had a sufficient robber, uh, on, uh, on every available space on the back side. Okay, so now we based the on a fair amount off rubber on the other side of the cross overboard, and now you simply have to place it inside the enclosure. So let me get a testing closure, so normally you will have to stick it in inside through the hole like this. We're not gonna do it right now because you won't be able to see what I'm doing. And simply use four screws toe. Screw the crossover onto the enclosure. So let me get this into frame so you can see. Now I'm showing you this on the side of the enclosure. I'm not gonna You gonna scrap this enclosure anyway, so I don't care? Fight, Drill a hole through it. But you will have to take the cross overboard and stick it inside. And you have to do this inside the enclosures who have to place it either on the bottom panel or on the back panel. But I'm gonna show you here on the exterior so you can see what I'm doing. So basically, we have us these small screws which go in each individual hole, and these are small enough. So even if you tied them real good, you have to understand that you will have some spacing here because of the the rubber on the bottom of the board. But even if you do tight them really good, you will not have them passing. Proved the wall of the quarter. Considering this is 18 millimeter India. So while you are inside the enclosure with this cross overboard, you will not have room to drill pilot holes or such so you will have to screw the screws directly, but it will go. It will not go that far into the MDF panel that it will introduce the cracks or something like that. So simply streak. Take your power drill and drill the the screws inside, - and I can see the board is fixed to the enclosure. Did this rock solid and it bold rectal about because it have. It has that rubber between the crossover board and the inclusion, and you can use sponge as well sponge with adhesive on one side. That is also fine. Now we can also place hot glue between the components and the board, and maybe you can space them mawr correctly. Then this, because I have to do this while filming and show you everything I do. And if you take your time, you can do a great crossover board that looks good and it's functional and it doesn't rattle about. The components are fixed, corrected to the board, and this is just just one way you can do this. You can design in many other ways the point where you go where you connect all the ground points. You can have different connections This is not a very elegant one, but you get the idea. This is a beginner crossover course. Maybe we will have more advanced ones, but for now, this will ensure you that you will build your own crossover for your own multi way speaker .