Crossover Design 101

[blueenergy]

I want to start a Modula MT project soon.

Instead of simple blindly following someone else’s design I want to understand the crossover designs completely. Many of the designs offered here on HTGuide are far beyond the simple 2-way 2nd order designs with 4 componets that I have worked with in the past. Most have far more components being used. For example I would like to reverse engineer a 2-way MT design like MarkK’s Dayton RS RS225/RS28A crossover to understand what each component is for and why it was chosen. How a 4ohm tweeter and an 8ohm woofer can be used together with what resulting impedance? Then maybe I could design a new crossover from scratch and end up with the same results.
I know this site has many nuggets of wisdom buried in the thousands of posts, but I am looking for it in a more organized format. So if you can point me in right direction that would be helpful.

Is there a “really good” book on crossover designs including “notch filters” to read? If not any other resources?
During assembly what meter do you recommend for verifying values of components, validating the network is assembled properly and drivers parameters?
What free or low cost (less $100) software can be used for crossover design?

Thanks in advance for your replies.

[cjd]

Oh my goodness... you want to understand the designs completely?

It's going to be a few years before you get to listen to anything! Really.

Um... books. Speaker building 101, Measuring Loudspeakers. Check the PE link up top, they have some good starter books.

They won't explain the stuff Jon is doing these days still.

And, though a bit esoteric, Speaker Workshop covers most your needs and is free. The FRD Consortium also has some superb tools (also free) if you have Windows and MS Excel.

C

[ThomasW]

They won't explain the stuff Jon is doing these days still.

Correct...

You're not going to find the Modula MT XO design in a book. It's a Cauer-Elliptical filter placed inside a 4th order L/R topology. I'm not aware of any freeware or low cost modeling program that will allow you to model this crossover, since it's pretty much defining state-of-the-art, as opposed to following in someone's footsteps.

[blueenergy]

Oh my goodness... you want to understand the designs completely?

It's going to be a few years before you get to listen to anything! Really.

No I don't want to wait too long. Basically to be able to read a XO schematic and understand it. Example when I see a simple coil and capacitor connected to a driver then I know that is 2nd order HP XO. But when there is 3 coils 3 capacitors and bunch of resistors I have no idea what is what. Let alone how to change the HP filter from 1200Hz to 1400Hz.

Um... books. Speaker building 101, Measuring Loudspeakers. Check the PE link up top, they have some good starter books.

They won't explain the stuff Jon is doing these days still.

I understand the basics ideas that I am sure those 101 books offer. I am looking for better resource in learning the "current" designs offered here. I have read Loud Speaker Cookbook back in the 90's and might be find that dust it off.

[cjd]

Hmm. I haven't read the 101 book, but I'm pretty sure it DOES answer some of the questions you have.

However, the Modula MT is a hybrid-series type crossover and is NOT going to be explained anywhere. I'm not sure I can tell you how it works, entirely.

Changing the filters is never what you think it is, because textbook crossovers rarely work anyhow, not just because impedance doesn't work that way, but neither does phase. What works at 1200 may not work 1400.

So, I'm back to: you'll be learning for a while, or you'll decide you'll build and listen and enjoy while you continue to learn, or you'll give up on learning. But the learning will be a while if you follow that road.

It's a bit faster if you've got the calculus down pat. Dig up some EE texts, they'll start to explain more of what's going on.

C

[blueenergy]

Correct...

You're not going to find the Modula MT XO design in a book. It's a Cauer-Elliptical filter placed inside a 4th order L/R topology. I'm not aware of any freeware or low cost modeling program that will allow you to model this crossover, since it's pretty much defining state-of-the-art, as opposed to following in someone's footsteps.

Exactly my fear. 8O I don't even know what Modula stands for or even a "Cauer-Elliptical filter placed inside a 4th order L/R topology".

So what is a better process for learning this stuff? Buying a good design application that is capable to modeling these types of designs, do some tweaks to the model and study expected result and learning from that?

[Dennis H]

Probably the best way to learn is to model some existing designs in software and fiddle with removing/adding/changing components.

[digital desire]

Probably the best way to learn is to model some existing designs in software and fiddle with removing/adding/changing components.

Even that would be emulating more that anything else.
To DESIGN you will need some EE under your belt.

[ThomasW]

'Modula' doesn't stand for or mean anything, it's just a name Jon and I have used for 20 yrs for a particular type of loudspeaker design.

[JorgenMan]

It's a bit faster if you've got the calculus down pat. Dig up some EE texts, they'll start to explain more of what's going on.

I agree. You'll want to have an understanding of why certain crossovers do what they do, as opposed to just memorizing what they do (two inductors and a capacitor and a resistor equals a such-and-such filter, etc.) Read an EE text or two (or however many it takes, I guess), so you know how to analyze a filter. Then you could start designing with at least a little confidence...

I don't do much circuit design, so the only book I know of is my college circuits textbook by Sedra & Smith. It's a good book, but probably not quite what you want. It has a lot of stuff that doesn't really have anything to do with crossover design, and assumes the reader has a basic electronics background. I'm sure there are books out there that are better suited to what you're looking for.

[JonMarsh]

The AES Anthologies of published articles in acoustics and crossover design would be good reference material for background of the development of current practices.

Another good book with some projects and practical information is "Speaker Building 201", should be available from PE and Madisound. Getting a little deeper, try "Testing Loudspeakers" by Joseph D'Apollito. Similar availability.

Make sure you understand your basic S-plane math and calcuations; the current computer programs will do the repetitive work for you, but without a basic theoretical grasp you'll not have the intuitive understanding needed for some "creativity" in the area.

Read up on passive and active filters; you should understand what the different types and orders are, though in Audio the most popular types these days are all-pass filters of various sorts. If that phrase doesn't mean anything or much to you, well, you've got an idea where to start. Butterworth filters are not in favor, especially not since people started paying attention to the radiation patterns and power response of non-coincident drivers.

Cauer-elliptic filters are most commonly used in communications applications; they are classically referred to as equal ripple filters, as the pass band and stop band ripple are equivlant in amplitude. They allow steep transition slopes with a minimum number of components, and are very useful where phsyically imperfect components are the norm (say, at microwave freqeuncies).

I first used a cauer-elliptic filter in "audio" for Class D amplifier ouptut filters back in the 80's; I presented a paper on that topic at the 91 AES conference, which was subsequently enhanced and published in the AES Journal in 92. Later I adpated the concept to loudspeaker crossover networks, but minimizing the passband and stop band ripple, and emulating the transistion characteristics of filters such as an LR-8 for the first 50 dB or so of attenuation.

I've seen some other papers published around 2000 (for crossover alignments with notch filters integrated) by Neville Theil, but they don't have as deep a stop band rejection or as low pass band ripple, and they aren't aimed at emulating other transfer functions like an LR8. Neville Theil was granted a patent in Feb 2005 for an active and passive realization of notched high and low pass filters in active and passive format, with complementary transfer function and a -6 dB crossover point to realize a two way transfer function. The combination produces an all-pass transfer function. US Patent 6854005.

As some might say, you can get almost anything approved by the patent office these days. Defending it might be another matter. It's a very competently written patent, but as the topology and filter concept is IMO just a refinement of "standard" cauer-elliptic filters (no change in topology), I'm not buying in. Plus the strict all pass implementation he uses trades off some other points that can be improved, such as keeping the stop band ripple much lower in level. Say, -50 to -60 dB in level instead of -30.

Joseph Audio uses an "infinite slope" crossover system developed by Richard Modaferri which achieves some of the same characteristics through inductive coupling to get sleep slopes and cancellations.

Now, the Modula MT and the Natalie P use a modified series crossover configuration which acoustically is roughly a 3rd order all-pass (not a popular text book configuration), with BSC, and with additional filtering to suppress the out of band hash on the midwoofer (sort of a "wideband notch filter" if you will). It uses the series connection in a way to achieve Q tuning and tracking through the crossover region with one resistor. Dissipative, yes, but then we've got efficiency to kill in the tweeter padding and BSC comp anyway, don't we? If any of those terms don't mean anything to you, start will Google- it's amazing what you can find online, including Wikipedia.

Why, you might even be able to get some good hints about why the North Koreans didn't get very good yeild on their recent efforts, and what the challenges are in that area of endeavor- one that's a lot more trickier than crossover design!

Last, there's a somewhat obscure Danish mathematician by the name of Steen Duelund who had some interesting ideas for unified multi-way crossover design; I'm looking into his three way synthesis all pass concept for my next few projects. Google, of course.

Regards,

~Jon

[Paul H]

Pretty good for a psych major ...

[JonMarsh]

Actually, it's all just a Freudian slip of the tongue...

And here on the board you guys get to listen to MY expression of Jungian archetypes... but then, where do you think Lucas got some of his ideas?

[blueenergy]

. . . Make sure you understand your basic S-plane math and calcuations; the current computer programs will do the repetitive work for you, but without a basic theoretical grasp you'll not have the intuitive understanding needed for some "creativity" in the area.

Read up on passive and active filters; you should understand what the different types and orders are, though in Audio the most popular types these days are all-pass filters of various sorts. If that phrase doesn't mean anything or much to you, well, you've got an idea where to start. Butterworth filters are not in favor, especially not since people started paying attention to the radiation patterns and power response of non-coincident drivers. . .

Thanks Jon for guidance on how to "begin" to understand the advance design shared on HTGuide by you and others. It will take me a few weeks to research what you have said here and reread it a few times. Without a Electronic Engineering or even Calculus in my background or formal studies it seems that understanding these designs enough to alter them on my own in the near future is not possible. I rather get focused on building one of them first so that I can start enjoying it.

I know that I that there is NO way I could come up with one of these brilliant design on my own. ops: But I hope that I could at least understand them enough to make small changes.

For example I may find that after some initial testing of the drivers received they are a little off spec that they may be better suited at a higher XO point. So I may want to change design from 1200 to 1400Hz. Also with the increasing cost of metal many of the components going into the XO have much higher cost than a couple of years ago. I would like to know which component (if any) I could use lower tolerance version (5-10% instead of 1%) or smaller gauge coils (18ga instead of 14ga). If cheap components are used in part of design what will be the effect. After pricing out the XO it can easily cost over $150 per speaker if you stick with the best components.

It might be best if a start a project thread related to the one design that I want to use and then inquire about how to make the changes.

Thanks everyone for your feeback so far.

[JonMarsh]

We're all just here to have fun and learn a bit from each other. Welcome to HT Guide.

~Jon

[---k---]

Jon,
I haven't been around here long enough to have read all of your posts. But so far, I've gathered that you have a degree in psychology, and not formal training in EE, but yet somehow you've had a successful professional career designing electronics for Crown, Siemens, and I'm sure others, you've been published in AES, had speaker companies rip off your ideas, and much more. If you ever feel the urge to impress us all and post a bio or your vitae, I would sure love to read it.

Same goes for you Thomas.

[JonMarsh]

Hi Ryan,

I'm not here to promote myself, but to relax, build speakers, and have fun with friends.

I do have a somewhat "checkered" past, and have had the fortune to be involved with a lot of interesting things at one time or another. I've been a partner in a high end audio store in the 70's, designed pro audio gear (in the 70's), developed mil/aerospace power supplies (an outgrowth of my interest in Class D amplification) in the early 80s at a Boulder company (have since met guys who work at current customers that joined that firm just after I left- was fun to compare notes and what happened to whom), and joined Siemens as a Power Semiconductor applications engineer in 1985- I've been with them since, including their spin-off as Infineon Technolgoies. Infineon is a low profile company with some pretty serious engineering resources; we pioneered superjunction MOSFET technology and have also done a lot of work with Silicon Carbide power semiconductors, including introducing some of the first commercial products, including the first merged structure (Schottky + PN junction) SiC rectifiers. (this is an area I work in). In 12-18 months we'll should be introducing high voltage SiC power JFETs; I've been working with them in application projects in my lab for several years. I've published about 40 papers altogether for Siemens/Infineon on a variety of topics, including simulation models, automotive smart power components, eletrothermal simulation techniques and models, and high frequency power conversion.

For the record, I'm the Applications Engineering Manager for Infineon Technologies power semiconductor products for computing and industrial segement products in North America, and the North American representative for three product segment core teams responsible for product definition and application roadmaps, so my day job has gotten a bit demanding of late. I spend far too much time in airplanes lately, about 80% in the last 7 months.

I never took an EE course in school, though I tutored the drummer in a band I played in in Boulder CO who was an EE major (I wonder if ThomasW remembers who Rich Rieback was? But that might have been before his time) (ThomasW knew me in my rock and roll days). I started getting into electronics via shortwave radio when I was 10 and eventually got a Technician class ham license when I was 13. Looking back, I realize in comparison to my daughter and our local friends I was one rather strange young guy.... When my folks made me ditch my Ham gear at the end of the eighth grade because of a move and because my dad realized it was "dangerous" and he didn't have any time to spend with me, I turned to the Dark Side- rock and roll, keyboards and electric guitars, and motorcycles. The rest, alas, is history. :rofl:

And no, I haven't figured out a way to store 1.21 gigawatts of energy to improve the transient/time domain response of my stereo, though I haven't given up on it....

[ThomasW]

I wonder if ThomasW remembers who Rich Rieback was?

As I recall a rather wild fellow who kept a full drum kit in the middle of his living-room.... :B

[---k---]

Jon,
Thanks for that bio info. I understand why you don't want to brag, and how you want to keep this a hobby. Though, I have to embarrassly admit that until recently, I dind't realize how advanced some of your cross-over designs were, and in general how knowladgable you and Thomas were. I was previsouly under the false assumption that you were just a couple of hobbiest always erroring on the side overkill.

So, anyway. Thanks for the bio. It helps me appreciate a even more the effort you put into this.

I'm still looking forward to the day that I've had my current speaker long enough in the eyes of she-who-must-be-obeyed that I can upgrade them to one of the designs here.

[chasw98]

Hi Ryan,
I started getting into electronics via shortwave radio when I was 10 and eventually got a Technician class ham license when I was 13. Looking back, I realize in comparison to my daughter and our local friends I was one rather strange young guy.... When my folks made me ditch my Ham gear at the end of the eighth grade because of a move and because my dad realized it was "dangerous" and he didn't have any time to spend with me, I turned to the Dark Side- rock and roll, keyboards and electric guitars, and motorcycles. The rest, alas, is history. :rofl:

Very interesting...... My father and I built a shortwave rig when I was 11. (He shot up rockets for NASA back when). I then got my first class. Then The Beatles came around and the rest (like you), alas, is history.

Did you and Thomas ever hear of the Ophelia Swing Band in Boulder? Or know Dan Sadowsky (AKA Pastor Mustard) who played in it?

Chuck

[JonMarsh]

Yes, I've heard of the Ophelia Swing Band, and of Dan Sadowsky, though I didn't "know" him. I had a number of friends still in the Boulder music scene in rock and jazz when I left Colorado for the silicon valley area with Siemens. Some of the other semi local people like Firefall I've actually found CD's of, but I understand the only release by Ophelia Swing band was a Japanese label back around '77.

[JonMarsh]

As I recall a rather wild fellow who kept a full drum kit in the middle of his living-room.... :B

Yeah, he also had one of the first Genral Radio caculators, thing cost about $650 (in 1973 dollars!!). Italian, but big afro like hair, somewhere I think I still have a picture we took of "Family Hand" for promotional purposes with him.

[chasw98]

Yes, I've heard of the Ophelia Swing Band, and of Dan Sadowsky, though I didn't "know" him. I had a number of friends still in the Boulder music scene in rock and jazz when I left Colorado for the silicon valley area with Siemens. Some of the other semi local people like Firefall I've actually found CD's of, but I understand the only release by Ophelia Swing band was a Japanese label back around '77.

I didn't even think Ophelia ever made a record! I have worked with Rick Roberts numerous times on the western slope, though. I came to know Dan quite well and ended up doing maybe 3 or 4 private projects with him. Even pressed some vinyl for 2 of them.