What software (& other things) should I get for a first attempt at speaker design

[JonW]

Well, it looks like my first speaker project, the Modula MT’s, went pretty well. I’m now intrigued by the idea of trying to design a speaker from scratch. This might just be a passing phase, but it sounds appealing at the moment. Maybe start with a simple MT design to see if I can get something sounding OK. If that goes well, later design a dream speaker (or just dream about doing that. ) Pick a couple drivers that look good from testing people have posted. I’m not looking to save money, just learn. And get some great sounding speakers out of it. So no need to think in terms of ‘bang for the buck.’ Rather, just think in terms of what will sound great. Perhaps, to start, I’ll stick with a Parts Express cabinet because, for me at least, making the cabinets was 90% of the work. So how to design a crossover?

I see various software packages that people use- LspCAD (~$200), LEAP (~$1500), etc. Some are cheap and others are pricey. I’m fine to spend $200 on software, but I hesitate at $1500. So what software is good to consider, that will actually give a decent chance of things coming out well? (There isn’t any decent Mac software, right? PC is OK, I guess.)

How tough is it to actually design a crossover? How much time and computer work does it take? Anything else I should think about ahead of time?

Thanks!

-Jon

[Jed]

If you haven't been to Roman's site, you should do some reading there first because it outlines a lot of what you will need. Also, he has links to other sources that will help make it easier to chose which direction you would like to take.

Jed

https://www.rjbaudio.com

[Dennis H]

Jeff Bagby's free Passive Crossover Designer for Excel is easier to use than any of them. I don't know if it will run on the Mac version of Excel.

[JonW]

Thanks guys- nice suggestions. :T
The Passive Crossover Designer didn't work on this Mac but I could try a PC at home, later. Might be interesting to play around with.
I just read Roman's walk through of designing a speaker- an excellent tutorial.

So if I use some of these basic tools to design a speaker... what are the odds of designing something that actually sounds really, really good? Rather than just OK. Assuming I start with good drivers and build a really solid cabinet of appropriate dimensions. It seems to me all you need are a really good, well built cabinet (I can do that), excellent drivers (I could pick from some of the favorites out there), and then a great crossover. Then tweak the built crossover to taste.

I suspect there is the science, but the art probably comes into play at some point. For the Modula MT's, I built them from a solid design and then tweaked (again- thanks Dennis) to suit my ears. Is that it or is there more to it? I'd think there has to be more.

[TurboFC3S]

I did my first DIY totally with the free FRD tools. Started off with the SPL Tracer to make my .frd and .zma files ... then import those into Frequency Response Combiner to model the response on my baffle ... and last into PCD for crossover design. It worked well, but was still somewhat off in a final result. If I didn't have a measurement setup, I probably would be chasing my tail for months trying to figure out just exactly how it was off.

[Jed]

Thanks guys- nice suggestions. :T
The Passive Crossover Designer didn't work on this Mac but I could try a PC at home, later. Might be interesting to play around with.
I just read Roman's walk through of designing a speaker- an excellent tutorial.

So if I use some of these basic tools to design a speaker... what are the odds of designing something that actually sounds really, really good? Rather than just OK. Assuming I start with good drivers and build a really solid cabinet of appropriate dimensions. It seems to me all you need are a really good, well built cabinet (I can do that), excellent drivers (I could pick from some of the favorites out there), and then a great crossover. Then tweak the built crossover to taste.

I suspect there is the science, but the art probably comes into play at some point. For the Modula MT's, I built them from a solid design and then tweaked (again- thanks Dennis) to suit my ears. Is that it or is there more to it? I'd think there has to be more.

They key is to get accurate manufacture's data. Without that the simulations are not going to yield correct summed responses. My suggestion is to just practice with a variety of designs. I personally like speaker workshop. Some people think its hard but like anything once you get used to something its not too bad.

edit: Start with a 2-way!

[Dennis H]

Jon, make sure you have the Analysis Toolpack and Visual Basic (VBA) installed with Excel. They don't install by default and, without them, most of the FRD tools won't run on a PC. No experience with Excel on a Mac. As well, the tools are a bit sensitive to Excel version. I think some of them use routines added with Excel 2000. I'm still running the 97 version (ain't broke, why fix it) but I have had minor problems with some of the FRD stuff and Unibox won't run at all. Maybe it's time I brought my software into the 21st century, sigh.

[Chris7]

You may consider starting with measurement before you get into speaker design. You can pick up a Behringer ECM8000 measurement mic and a mic preamp with phantom power for about $100 total. Add another $40 if you want it calibrated, though the stock unit is pretty good. You can compare your Modula with your Totems, see how they behave off-axis, both horizontally and vertically, how they interact with your room, etc., as well as see how your resistor changes affect the FR of the speaker. Great place to start.

[Dennis H]

Yeah, what Chris said. As I recall, Chuck made a subtle suggestion a few times to buy the mic as well.

Good measurements are a skill (art) all their own, one you really need to master before you can take the next step to designing a crossover. You'll get to thoughly explore the wonderful ramifications of FFT theory -- short window on the impulse response, cancel out the room reflections but lousy low frequency accuracy; long window, good LF but you're measuring the room as much as the speaker. Learning how to balance all that really is an art although it's one based in science (math).

[chasw98]

Hmmm.................. What is being born here before our eyes?

Give a man a fish and you feed him for a day. Teach a man to fish and you feed him for a lifetime.

Chuck

[Paul H]

JonW,

I've been very happy with my purchase of:

Speaker Building 201
Testing Loudspeakers
LSPCad
Behringer mic/preamp

The first 2 (books) are very informative and between them cover a lot of theory and practical application.

LSPCad is a very good piece of software with measurement and design capabilities. As far as I'm concerned, it paid for itself during the design of the 4-ways in my avatar (side note-don't start with 4-ways ).

The behringer mic and preamp give you all the testing accuracy you'll ever need as a diy'er.

[JonW]

Thanks for all the tips, folks. :T

Yes, definitely start with a 2 way. I’ve already picked up the ECM8000 microphone and Mobilepre USB soundcard. I’ve got Speaker Building 201 and the Loudspeaker Design Cookbook. If I’m really serious about this, I’ll need to reread both books and try to take more in.

Hmmm.................. What is being born here before our eyes?

Give a man a fish and you feed him for a day. Teach a man to fish and you feed him for a lifetime.

Eh, what can I say? I like seafood.

Although this is all just an idea at the moment, it seems really appealing. So I’ll start learning these things, slowly, and see where I end up. Maybe begin using the FRD tools and the procedure Roman outlined on his excellent web site. My only hesitation at this point is from A) all the time this will take (free time is scarce for me) and B) all the time using computers that will be required (not much of a break from work, you know)

So to start thinking about things, based on Zaph’s testing, maybe a nice 2 way could be had from the Peerless 810921 ($59) tweeter and the Usher 8945P ($110) mid. I’ll have to look and see if they play nice together. (And, more importantly, see if I can figure out what playing nice together even means!)

Is picking 2 drivers a good place to start? And really, how do you pick your drivers? (Let's avoid price as a restraint- the point is to make amazing speakers, not save money.) People like MarkK and Zaph have lots of good test data out there. What am I really looking for from the tests?

After picking drivers, should I begin with measurements, before doing anything else? Given that I have no idea what that involves, should I just buy the “Testing Loudspeakers” book and go from there?

[cjd]

My suggestion would be to work with the drivers you have now. Or at least work with drivers that people are both familiar with and may have access to other measurements. Getting a crossover design to work right can be tricky - I don't know how many I've done that failed miserably. We won't mention the ones that used to sound good to me but make me cringe now. Learning to measure well is a game of its own, which is where known drivers can be useful as folks can look at samples and see problems (or not) as they may exist.

Then too, it's nice to have existing designs to compare against - model yours, model theirs, compare.

While the FRD tools are superb (I do most of my "sketching" with these), for some things like the comparisons I really like Speaker Workshop. I also use it for measuring. And it's a free Windows program.

The software you buy is awesome, but expensive. And it doesn't help you understand what's going on any better than the free stuff.

So, really, dive in and learn, but do so with the goal of keeping budget under control and not worrying about "the best ever" results. I started with $0.99 buyouts and cheap tweeters. Sooo glad I did. I learned a ton with them.

C

[SteveCallas]

I know Jon wants to do everything from scratch, but does anyone have experience with allowing Madisound to create a crossover for your desired speaker? Jon had mentioned that they will make you one for $50. Do they have a reputation of - at the very least - giving you something solid to start from?

[ssabripo]

Madisound built a custom XO for the Seas Excel MTM CC that I recommended for a friend, and he is happy with it. How great and accurate they are, remains, as I doubt they actually built the cabinet size I gave them and tweaked the XO, but nonetheless, the results seem to be good.

However, building XOs is the heart of a speaker design, and without measurements of the actual design, complete with baffle, position, etc, is hard to get perfect. Certain drivers are specially sensitive to the XO, like the Seas, as you can royally scew the sound up by not getting the XO done right.

I'm actually reading the Cookbook now, and since Chuck got the measuring equipment ( :lol: ), it wont be long before we get our hands dirty I'm sure.

I gotta admit though, it's hard to pass some of the great designs and work of some of the guys here like JonMarsh, ThomasW, Evil-twin, and many others....some of these guys have tons of experience, and there is never substitute for that in this field.

[ThomasW]

I know Jon wants to do everything from scratch, but does anyone have experience with allowing Madisound to create a crossover for your desired speaker? Jon had mentioned that they will make you one for $50. Do they have a reputation of - at the very least - giving you something solid to start from?

Unless one has in-baffle measurements that can be sent to Madisound, their LEAP XO designs are taken from whatever published plots are available. (usually the mfgr's data).

The LEAP program doesn't factor in things like 1st breakup modes and other important design considerations. So it's certainly better than a generic XO design, but it's not a fully optimized design that takes all the performance criteria into consideration.

[JonW]

Chris-
Thanks for the perspectives. I realize this could be a complicated and lengthy undertaking. Sounds like fun, but time is always the problem. So we’ll see.

I see your point about sticking with established drivers and designs. I don’t have any drivers that aren’t in use, so I’ll have to buy some new ones anyways. Might as well pick some good ones to increase the slim odds of my first design not being offensive to the ears, as long as there is sufficient ‘support’ for them out there.

The LEAP program doesn't factor in things like 1st breakup modes and other important design considerations. So it's certainly better than a generic XO design, but it's not a fully optimized design that takes all the performance criteria into consideration.

So how does one take all that into account? Or is that a complicated topic best left for after I’ve designed my 20th speaker?

Unless one has in-baffle measurements …

So would this be a good progression of events for me to shoot for…?
-Pick a tweeter and a mid based on Zaph, MarkK, etc. testing
-Buy/make a cabinet, mount the drivers in there
-Take measurements of the actual drivers in the actual cabinet
-Import all the measurement data into the suite of software programs
-Then start with the crossover design

Do I need anything for the measurements other than the microphone I have? If I want to do anything for measurements it sounds like I need to get that book first. That's if measurements are really where I should start.

[Paul H]

...

So would this be a good progression of events for me to shoot for…?
-Pick a tweeter and a mid based on Zaph, MarkK, etc. testing
-Buy/make a cabinet, mount the drivers in there
-Take measurements of the actual drivers in the actual cabinet
-Import all the measurement data into the suite of software programs
-Then start with the crossover design
...

Good list but I'll suggest adding one step. After you pick a tweeter and mid, do some simulation using one of the baffle simulation programs like "Edge" (free download) - it will help with picking an appropriate baffle size and driver location.

http://www.tolvan.com/edge/

Edit by Mr Moderator to add link

[JonW]

Good list but I'll suggest adding one step. After you pick a tweeter and mid, do some simulation using one of the baffle simulation programs like "Edge" (free download) - it will help with picking an appropriate baffle size and driver location.

http://www.tolvan.com/edge/

Edit by Mr Moderator to add link

Thanks Paul (and Mr. Moderator).
My cabinets took me so long to build, I was thinking about using the PE pre-made ones for this exercise. Just get the large 0.75 cf MT ones, I'd think. In which case I don't really have much flexibility with the baffle size. But I could, of course, place the drivers anywhere on the baffle. Is this an OK approach or am I asking for trouble starting with a fixed cabinet?

[ThomasW]

The PE boxes are a great deal. If that's the size you want to use for your project, no reason not to use them.

[JonW]

The PE boxes are especially a great deal now- on sale and free shipping. And they will make this process much faster for me. I'll be less invested if the speakers sound lousy.

Given that I should start easy, with an MT design, I'd think I should just get the same 0.75 cf boxes of the Modula MT's (17" x 10" x 12.5"). Sound OK?

http://www.partsexpress.com/pe/pshowdetl.cfm?&Partnumber=302-732&CFID=18754&CFTOKEN=55953283

You know, if I actually buy something, I'd be committed (in general and to this project, etc.) What am I getting myself in for?

[ThomasW]

MT or MTM not a whole lot of difference. If you buy the 1 cu ft box you can experiment with either a MT or MTM. And you can recycle the box since they sell replacement baffles.

What am I getting myself in for?

Sleepless nights, premature gray hair, premature baldness, ulcers, and on and on..... :B

[JonW]

Thanks again. Yeah, the replacement baffles and ability to recycle the cabinets are nice touches.

I'm leaning toward the 0.75 MT cabinets because I'll just want to try a simple MT design at first. And that would also allow me to use them for 2 more Modula MT's to make 4 for the rear/surrounds, if my own speakers don't come out so great. I'd better order them soon, before the sale and free shipping ends.

So far, I have not been able to talk myself out of this. Uh oh... :??

Sleepless nights, premature gray hair, premature baldness, ulcers, and on and on..... :B

Sounds like my kind of crowd.

[JonMarsh]

The PE boxes are especially a great deal now- on sale and free shipping. And they will make this process much faster for me. I'll be less invested if the speakers sound lousy.

Given that I should start easy, with an MT design, I'd think I should just get the same 0.75 cf boxes of the Modula MT's (17" x 10" x 12.5"). Sound OK?

http://www.partsexpress.com/pe/pshowdetl.cfm?&Partnumber=302-732&CFID=18754&CFTOKEN=55953283

You know, if I actually buy something, I'd be committed (in general and to this project, etc.) What am I getting myself in for?

What ThomasW said... worst case, you could end up like me, waking up in the middle of the night on international busines trips thinking about equations and acoustic transfer functions (instead of visions of sugar plums and fairies in your head at Xmas time), trying to type Xmas and getting Xmax instead, etc. It's a sad sight....

beware the dark side.... worst case, you might wind up like Evil Twin.

[chasw98]

What ThomasW said... worst case, you could end up like me, waking up in the middle of the night on international busines trips thinking about equations and acoustic transfer functions (instead of visions of sugar plums and fairies in your head at Xmas time), trying to type Xmas and getting Xmax instead, etc. It's a sad sight....

beware the dark side.... worst case, you might wind up like Evil Twin.

This is only the first step. Pretty soon you will find yourself up late at night prowling the back alleys of eBay looking for great used Tektronix, Fluke, Wavetek, and HP test gear. Then you will descend into the lower reaches fishing for esoteric parts to build your creations and soon you will wake up one day and say "How did I get here!" (compliments to David Byrne).

[JonW]

Well guys, I appreciate your, ummm, encouragement.
And it’s kind of odd how I’m starting to see themes in the personality traits of this crowd (myself included). :??

This will be great fun. I’m busy so it will take me a long time to get a finished design. But hey, part of it is the journey.

Here’s the plan: Design an MT. Buy the PE cabinets now. (I’m still thinking 0.75 cf unless there’s a good reason to go 1 cf) Read everything I can find about driver testing results on the web. Pick the best tweeter and the best mid pair I can find that might play nice together. Read that book on driver testing and actually measure the drivers in the cabinets. Then start with all the software programs, edge diffraction, xover design, etc. Actual construction should hardly take any time with the PE cabinets, so this will mostly be computer time. And we’ll see if I can make a decent MT speaker. I’ll be able to compare my first outing with the Modula MT’s and my Totem’s (also an MT).

[Hank]

Yes, you could get really messed up. You could be doing a cabinet design, getting your son to put it into AutoCAD, thinking about an access panel design instead of going to sleep...you could be at December 20th and having bought only ONE Christmas present and comitted to designing an open baffle for another customer, and building surrounds for another...yes, you could be fairly well messed up.

How 'bout this: buy a pair each of the PE .75 and 1.0 ft^3 cabinets on sale. Start with the .75's and design/build your two-way, tweak until it sounds great, and then give them to someone who will appreciate them. Then, tackle an MTM with the bigger cabinets. I haven't bought any PE cabs yet, but since they're on sale, I'm thinking about some in Dark Side gloss black.

[Evil Twin]

Well guys, I appreciate your, ummm, encouragement.
And it’s kind of odd how I’m starting to see themes in the personality traits of this crowd (myself included). :??

This will be great fun. I’m busy so it will take me a long time to get a finished design. But hey, part of it is the journey.

Here’s the plan: Design an MT. Buy the PE cabinets now. (I’m still thinking 0.75 cf unless there’s a good reason to go 1 cf) Read everything I can find about driver testing results on the web. Pick the best tweeter and the best mid pair I can find that might play nice together. Read that book on driver testing and actually measure the drivers in the cabinets. Then start with all the software programs, edge diffraction, xover design, etc. Actual construction should hardly take any time with the PE cabinets, so this will mostly be computer time. And we’ll see if I can make a decent MT speaker. I’ll be able to compare my first outing with the Modula MT’s and my Totem’s (also an MT).

Ah, the Force is strong in this one...

Trust your feelings! Join me, young Jon, and together we can rule this forum!











Wait, wrong movie....







Just one point- as soon as you become more than a Padawan apprentice, I don't want to hear any of this "Once I was the pupil, now I am the master" line of reasoning.


That's MY schtick!

[kingpin]

You guys are great comedy relief after dealing with customers in the reatail business all day.

Although, I do think some of you have inhaled too much fibreglass or something over the years. :B :B :B

Mike

[blueenergy]

Wow what a great discussion. There are many good points here on how to get into "design". From a pure learning experience I can see why you should start simple and cheap then focus on learning the tools (many are available free) and build your skill set.

Getting a crossover design to work right can be tricky - I don't know how many I've done that failed miserably. . . Learning to measure well is a game of its own, which is where known drivers can be useful as folks can look at samples and see problems (or not) as they may exist.

So, really, dive in and learn, but do so with the goal of keeping budget under control and not worrying about "the best ever" results. I started with $0.99 buyouts and cheap tweeters. Sooo glad I did. I learned a ton with them.
C

For a study on the effects of where drives are place on the baffle a cabinet like the PE can be purchased or like the Microbe design on rjb audio lets you keep reusing a cabinet with mutiple baffles and easy access to swapping/changing crossovers. Maybe I could find some closeout specials and build some "computer" speaker for the office.

[nerd of nerds]

Ah, the Force is strong in this one...

Trust your feelings! Join me, young Jon, and together we can rule this forum!











Wait, wrong movie....







Just one point- as soon as you become more than a Padawan apprentice, I don't want to hear any of this "Once I was the pupil, now I am the master" line of reasoning.


That's MY schtick!

Evil twin, you seriously crack me up! Every time I see your avatar I know some hilarious star wars jabber awaits me and it about kills me every time.

:rofl:

[Brian Bunge]

Hank,

At least it's nice to see that you are building again! I'm in the process of working on several subwoofers myself.

[cjd]

For a study on the effects of where drives are place on the baffle a cabinet

Far simpler is to grab either the BDS over at the FRD Consortium (requires Windows and MS Excel) or the program The Edge. Takes a while to grasp but very powerful once you start to understand what you see.

C

[JonW]

From a pure learning experience I can see why you should start simple and cheap then focus on learning the tools (many are available free) and build your skill set.

Makes sense. For me, time is my most precious commodity. I ‘cut my teeth’ on the Modula MT’s. And that took me months. So I’m not going to try and save any money on the next project. If some expensive parts increase the odds that the speakers come out sounding good, I’ll just buy them. I’m especially thinking about this in terms of drivers and trying to find a tweeter and mid that play well together. But I will keep things simple and just do an MT design to start with.

Yes, you could get really messed up. .

Just one point- as soon as you become more than a Padawan apprentice, I don't want to hear any of this "Once I was the pupil, now I am the master" line of reasoning.


That's MY schtick!

OK, I can’t say I wasn’t warned.

Seriously, this looks like it will be great fun. :T

Actually, I’m curious to see how poorly my speakers end up sounding when compared to ones designed by a master (Jedi, Evil, or otherwise). Part of why this looks like an appealing experience is because there is a firm ‘science’ basis to it all, but it seems to quickly get into the ‘art’ range once you have the fundamentals down. I’m now reading everything I can about drivers to pick a promising tweeter & mid combo. Holiday travel will slow things down, but I’ll pick it up again next month.

[nerd of nerds]

Makes sense. For me, time is my most precious commodity. I ‘cut my teeth’ on the Modula MT’s. And that took me months. So I’m not going to try and save any money on the next project. If some expensive parts increase the odds that the speakers come out sounding good, I’ll just buy them. I’m especially thinking about this in terms of drivers and trying to find a tweeter and mid that play well together. But I will keep things simple and just do an MT design to start with.










OK, I can’t say I wasn’t warned.

Seriously, this looks like it will be great fun. :T

Actually, I’m curious to see how poorly my speakers end up sounding when compared to ones designed by a master (Jedi, Evil, or otherwise). Part of why this looks like an appealing experience is because there is a firm ‘science’ basis to it all, but it seems to quickly get into the ‘art’ range once you have the fundamentals down. I’m now reading everything I can about drivers to pick a promising tweeter & mid combo. Holiday travel will slow things down, but I’ll pick it up again next month.

I'm in the same boat you are in...in a few months I hope to get all my measurment equipment so i can start taking driver measurments and get working on my own speaker...

I think working with drivers that have been used a lot before would be a great idea (seas 27tdfc, etc) as there will be plenty of measurements lying about the internet to verify your results...