How on earth do you pick the right crossover for a DIY speaker?

[JonW]

Warning: Newbie question ahead.

So I’m thinking about trying to make my first set of speakers. I might have some idea of what drivers to use (see below). The cabinet dimensions will be obtained from some calculations, I’d guess. But what about a good crossover? I’d think this has a tremendous impact on the sound of the speaker. So what do you do, make some crossover that looks good on paper, put it in the speaker and just hope for the best? That can’t have high odds of things turning out well, can they? Heck, speakers of similar design by different companies sound so very different. Or maybe there is less variation from the crossover than I’m thinking.

For what it’s worth, I’ve got a pair of Totem Forests- lovely speakers that I really like a lot. I could add two rear speakers and a center channel to make a movie setup. So maybe try and have things sound, more or less, the same as the Forests. Start with similar drivers, build the cabinets according to what a program tells me I need. But what to do for a crossover? I bought a book (Speaker Building 201) that I will plow through soon. So maybe that will provide me with some perspective. Just curious about this in the mean time. Thanks.

[KeithM]

To do it properly, you test the frequency, T/S parameters, distortion, etc, but that may be past your budget. Try to find drivers already tested and get Vance Dickason's new book here
Not only does it have supplement info to your book, but it has free software to help you build the crossover. Or you could download speaker workshop for free.

Either way, it's best to do it with help from a computer.

[cjd]

Where in Indiana are you?

If you're near Chicago, I'm not too far out of the way if you get stuck on measurements for systems that don't exist.

But, in my experience there are rules of thumb to look at and consider when picking a crossover starting point, but from there it's often a tweak-and-see game to really find the point at which synergy exists. Driver Fs, breakup nodes, the point at which off-axis starts to differ from on-axis response, driver diameter, orientation, all go into crossover points.

And, even with boxes you can pick a baffle shape, or you can design a baffle shape and layout.

I will say the Totems you have probably did not bother designing the baffle to optimize response. It's not a problem, just something additional to deal with in the crossover and perhaps some response anomalies you just get to live with (often no trouble at all).

I tried to kind-of outline the steps in the RS 3-way thread (stickied) as far as some of my process, but it's different for everyone.

Folks like Jon probably close their eyes, put on their ruby shoes, click their heels three times while chanting and the crossover will appear before their eyes.

C

[JonMarsh]

Usually in the shower. It's kind of disappointing, because when I was younger, I'd get visions of attractive women in the shower, or at least someone interesting looking. Now I have to settle for crossovers. Hmmmph.

[gimpy]

Well, Jon, that's not necessarily all bad, is it? Especially at my age.
Frank

[JonW]

To do it properly, you test the frequency, T/S parameters, distortion, etc, but that may be past your budget. Try to find drivers already tested and get Vance Dickason's new book here
Not only does it have supplement info to your book, but it has free software to help you build the crossover. Or you could download speaker workshop for free.

Either way, it's best to do it with help from a computer.

OK, thanks. I'll try to get through the book I have now (Speaker Building 201) and then maybe get teh one you mentioned. I'm guessing some computer modeling will be needed. I guess I don't know (among many other things) what resulting model will sound best to my ears. We'll see...

[JonW]

Where in Indiana are you?

If you're near Chicago, I'm not too far out of the way if you get stuck on measurements for systems that don't exist.

But, in my experience there are rules of thumb to look at and consider when picking a crossover starting point, but from there it's often a tweak-and-see game to really find the point at which synergy exists. Driver Fs, breakup nodes, the point at which off-axis starts to differ from on-axis response, driver diameter, orientation, all go into crossover points.

And, even with boxes you can pick a baffle shape, or you can design a baffle shape and layout.

I will say the Totems you have probably did not bother designing the baffle to optimize response. It's not a problem, just something additional to deal with in the crossover and perhaps some response anomalies you just get to live with (often no trouble at all).

I tried to kind-of outline the steps in the RS 3-way thread (stickied) as far as some of my process, but it's different for everyone.

Folks like Jon probably close their eyes, put on their ruby shoes, click their heels three times while chanting and the crossover will appear before their eyes.

C

I'm in Lafayette. I get up to Chicago now and then, maybe 2-3.5 hours away, depending upon where and traffic.

Tweak and see, eh? I guess that's what I figured. I'll read through that long RS 3-way thread. I'm sure to learn something. I'm starting with about zero knowledge right now.

I shopped around for a long time before finding speakers that I really like. Every speaker sounded sovery different. It was amazing, really. Maybe having a pair of speakers at home that I like will be helpful. For my first attempt at making speakers, I can try to stay somewhat close to the Totems. And I'll have something for comparisons right there.

[JonW]

Usually in the shower. It's kind of disappointing, because when I was younger, I'd get visions of attractive women in the shower, or at least someone interesting looking. Now I have to settle for crossovers. Hmmmph.

Oooooohhhhh nnnnnoooooo... Maybe I'd better stop this idea of building speakers right now. I certainly don't want to end up like you, with decreased visions of the pretty ladies... :P

[JonMarsh]

Well, Jon, that's not necessarily all bad, is it? Especially at my age.
Frank

Well, Frank, we're pushing a pretty similar number of years, '51 is when I was brought into the world kicking and screaming. I still haven't adjusted to the fact that I'm not 35 anymore, and I can tell I may be overdue for a midlife crisis by the interest I've had lately in picking up a re-issue Les Paul SG Custom, which is the SG equivalent of a '57 Black Beauty Les Paul. I last had that SG model back in my early 20's.

Image not available

That, and a Hammond XK3, and some time off to play them might make me forget for a while that I'm in my 50's.

~Jon

[wrz0170]

Hi there,

Great post! After a lot of reading, as a newbie, I've come to the conclusion that making a great XO seems to be 1 part science, 1 part art and 1 part luck.

Do software programs like Bassbox pro and xover pro really work and work well? The simple premise of the program(s)s is that you input speaker parameters and it spits out cabinent dimensions and with x-over pro, the schematics you need to build your XO. I am sure it's a bit more than that. Unless you know your schematics and electricity, I would take the results at face value.

Just curious if those programs are worth it??

William

[gimpy]

Well, JonMarsh, speaking of a "midlife crisis", I, just last year, have gotten into DIY speaker building. The LOML (for the last 30 yrs) commented that she thought I was going through a midlife crisis. My response to her was, "at least it is speakers and not women and cars." She had no response to that.
Sorry, JonW, to get off track here.
Frank

[JonMarsh]

Well, motorcycles, fast cars, and speaker building wouldn't qualify for me as "mid life crisis" material as I've been doing most of those for decades- bikes and bands since mid teens, the cars a bit later. My 21 year old daughter just figures "that's me". That and computers. I'm not sure what I would have to do to have a behavior change clearly identifiable as an uncharacteristic "mid life crisis". Going back to music might count. (last time I played publicly was 1986). Taking up options trading and futures on the stock market would certainly qualify, too.

Hi there,

Great post! After a lot of reading, as a newbie, I've come to the conclusion that making a great XO seems to be 1 part science, 1 part art and 1 part luck.

Do software programs like Bassbox pro and xover pro really work and work well? The simple premise of the program(s)s is that you input speaker parameters and it spits out cabinent dimensions and with x-over pro, the schematics you need to build your XO. I am sure it's a bit more than that. Unless you know your schematics and electricity, I would take the results at face value.

Just curious if those programs are worth it??

William

Programs like you're talking about are a bit too simplistic in their approach to get very good results.

The problem is that loudspeaker impedance varies with frequencies, unless you go fairly overboard connecting zobel networks up which flatten the impedance. I often do that, but folks often complain about how much the crossovers I design cost.

Second, drivers aren't "flat", especially not after you mount them on a baffle which modifies their frequency response as a function of the baffle size and shape, and the driver positioning.

So, for a given set of drivers and enclosure or baffle, you really have to have good measured data for the driver performance as mounted, including impedance, and then use a program like LspCAD or Soundeasy or SpeakerWorkshop to work out the details in the crossover. That requires some insight into what networks work best with what driver characteristics or behavior, and how to tune those networks manually or with the optimizer program usually built into this software (which optimizes selected component values (the ones you select) to hit a target transfer function.

If this sounds complicated, realize that I've given you the very much shortened readers digest version of the process.

Search the forum for the topic "crossover design" and I'm sure you'll get lots of material to read about and think about. There's a fair amount of acoustics and physics involved in selecting drivers, the cabinet design, and developing a good crossover with a reasonable cost/performance tradeoff.

Now, if you're NOT aware or cognizant of these issues, then yeah, there would have to be a fair amount of luck involved to get a good design. As Clint Eastwood would say, "Are you feeling lucky today?"

The key is to start with a small project if you want to learn design. But expect to need some measurement gear, design software, a few reference books, and some time and patience.

If you haven't built any speakers at all, the first step might be finding an existing kit design for sale or projects that appeal to you like those on this and other forums/sites, and try your hand at constructing one of those. Then all you're having to cope with is learning construction skills and techniques.

But above all, have fun!

~Jon

[JonW]

Programs like you're talking about are a bit too simplistic in their approach to get very good results.

The problem is that loudspeaker impedance varies with frequencies, unless you go fairly overboard connecting zobel networks up which flatten the impedance. I often do that, but folks often complain about how much the crossovers I design cost.

Second, drivers aren't "flat", especially not after you mount them on a baffle which modifies their frequency response as a function of the baffle size and shape, and the driver positioning.

So, for a given set of drivers and enclosure or baffle, you really have to have good measured data for the driver performance as mounted, including impedance, and then use a program like LspCAD or Soundeasy or SpeakerWorkshop to work out the details in the crossover. That requires some insight into what networks work best with what driver characteristics or behavior, and how to tune those networks manually or with the optimizer program usually built into this software (which optimizes selected component values (the ones you select) to hit a target transfer function.

If this sounds complicated, realize that I've given you the very much shortened readers digest version of the process.

Search the forum for the topic "crossover design" and I'm sure you'll get lots of material to read about and think about. There's a fair amount of acoustics and physics involved in selecting drivers, the cabinet design, and developing a good crossover with a reasonable cost/performance tradeoff.

Now, if you're NOT aware or cognizant of these issues, then yeah, there would have to be a fair amount of luck involved to get a good design. As Clint Eastwood would say, "Are you feeling lucky today?"

The key is to start with a small project if you want to learn design. But expect to need some measurement gear, design software, a few reference books, and some time and patience.

If you haven't built any speakers at all, the first step might be finding an existing kit design for sale or projects that appeal to you like those on this and other forums/sites, and try your hand at constructing one of those. Then all you're having to cope with is learning construction skills and techniques.

But above all, have fun!

~Jon

Jon-

Thanks for that info. I guess I don't really have any perspective on what you said. If I were to embark on my first speaker building endeavor, what are the odds I'd get something that sounded OK? I realize from what you say it won't sound perfect. But will it sound simply awful, like cars crashing and finger nails down the blackboard? Or just not as good as it possibly could be?

Now, I don't know what I'm talking about. But I'm starting to do some reading (books and web). Here is what I've got in mind so far. Maybe you can tell me if there is any chance this would sound OK or terrible: Start with copying an exisitng design to some degree. More or less try and clone the Totem Forest. I've got a pair of the real ones for comparison when it's all done. Use the same 2 drivers. Use some software to get cabinet dimensions, port info, etc. And then design a crossover somehow (no idea on that part yet). Put it together and see how it sounds. If the crossover parts aren't expensive, the whole outing would be a fun experiment, I'd learn something, I'd have 2 more speakers, and it wouldn't be expensive. (The 2 tweeters and 2 mids would only be about $200 total. Although I'm conveniently not factoring in the table saw and router I'd have to buy. 8) ).

A worthwhile idea or am I living in Fantasyland? Thanks.

-Jon

[JonW]

Sorry, JonW, to get off track here.

No problem. In the warm months, I spend weekends driving around race tracks. And this stereo/speaker building/woodworking is something to distract me for the winter months. I'm 36, but with the fast cars and such, I'll also have no way of seeing when a midlife hits. 8)


-Jon

[ThomasW]

I'm 36, but with the fast cars and such, I'll also have no way of seeing when a midlife hits.

Boy are you in for a rude awakening....

[JonW]

Boy are you in for a rude awakening....

Oh my. So, umm, what do I have to look forward to or dread in the coming years? h:

[ThomasW]

Oh my. So, umm, what do I have to look forward to or dread in the coming years?

You really don't want me to answer that question. Sometimes ignorance is indeed bliss.....

[JonW]

You really don't want me to answer that question. Sometimes ignorance is indeed bliss.....

Very funny. And fair enough. There may be some major changes in my life on the horizon (e.g., wife, kids), so really, I don't dare predict what life will be like 10 years from now.

[JonMarsh]

Jon-

Thanks for that info. I guess I don't really have any perspective on what you said. If I were to embark on my first speaker building endeavor, what are the odds I'd get something that sounded OK? I realize from what you say it won't sound perfect. But will it sound simply awful, like cars crashing and finger nails down the blackboard? Or just not as good as it possibly could be?

Now, I don't know what I'm talking about. But I'm starting to do some reading (books and web). Here is what I've got in mind so far. Maybe you can tell me if there is any chance this would sound OK or terrible: Start with copying an exisitng design to some degree. More or less try and clone the Totem Forest. I've got a pair of the real ones for comparison when it's all done. Use the same 2 drivers. Use some software to get cabinet dimensions, port info, etc. And then design a crossover somehow (no idea on that part yet). Put it together and see how it sounds. If the crossover parts aren't expensive, the whole outing would be a fun experiment, I'd learn something, I'd have 2 more speakers, and it wouldn't be expensive. (The 2 tweeters and 2 mids would only be about $200 total. Although I'm conveniently not factoring in the table saw and router I'd have to buy. 8) ).

A worthwhile idea or am I living in Fantasyland? Thanks.

-Jon

Well, you're not living in Fantasy land, IF one of your goals is to get some hands on experience with this whole process. IF it's a one shot deal, I would question the wisdom, because there is a learning curve for all things (how much experience do you have woodworking, for example?)

I'm not the person to talk rationally to about this stuff- I've been building musical instrument and HiFi speakers since I was about 20, and I'm an EE, too, which gives you an idea of my background for electronics design.

If you were a friend of mine here in CA or in CO where ThomasW is and I used to live, my suggestion would be to consider using the Modula MT/Natalie P crossover concept, use the HiVi D6.8 midwoofers (available from Parts Express), use the Seas H1212 tweeter (available from Madisound), build the cabinets per the dimensioning of your Totems if that's what you prefer, design a downward firing port using Unibox in Excel based on the internal dimensions of the cabinet (or adjust the cabinet slightly to optimize response for extension and a slight downwards slope in response from 100 Hz so that it isn't too heavy in room).

The crossover component values would have to be adjusted a bit because the Totem Forest has a narrower baffle than the Modula MT, so the baffle step compensation will be different (that is, the drive will measure differently in this cabinet, and to get a target smooth response, the crossover would have to be adjusted). Also, the HiVi D6.8 has different impedance characteristics and doesn't have the upper frequency resonance, so the crossover could probably be simplified a bit.

Re the HiVi D6.8 and the Dayton RS180 used in the Modula MT, the HiVi will have a bit more LF output, but is a bit less efficient. The RS180 will have lower midrange distortion, and more immediacy in that frequency range, but not the same "guts" or slam on the bass below 50 Hz.

If you or someone else had measured data in these cabinets, it would be easy to do. (well, at least for me it would be a small amount of work to make an updated crossover simulation in LspCAD.

Look at the Modula MT thread and you'll get a feel for what's involved. A little larger internal volume would be fine- I was working with off the shelf cabinet designs to make it easy for first constructors.

It would be fun to to a project like this, but my day job has been such a train wreck the last year that free time has dwindled to a very low level. Here it is Sunday morning and I'm working on materials (as I did yesterday) for an FAE training next week in Detroit.

Regards,

Jon

[JonW]

OK, thanks very much. That's kind of what I was wondering.

Well, you're not living in Fantasy land, IF one of your goals is to get some hands on experience with this whole process. IF it's a one shot deal, I would question the wisdom, because there is a learning curve for all things (how much experience do you have woodworking, for example?)

Yeah, that's pretty much it- try to get some experience at this whole speaker building business. See if I like it. Might as well try to match them to my Totems. But no big deal if they don't sound similar. My woodworking skills are not great, but growing.

I'm in the middle of building a subwoofer. I thought it would be interesting to do and I'd learn a few things, develop a few skills, etc. but I'd be sick of it when it was done. To my surprise, I've been enjoying this more than expected. And so I'm looking for more fun projects.

I'm not the person to talk rationally to about this stuff- I've been building musical instrument and HiFi speakers since I was about 20, and I'm an EE, too, which gives you an idea of my background for electronics design.

I'm a scientist and been getting into engineering in my research lately. It seems this place is bursting at the seams with engineers. :P

If you were a friend of mine here in CA or in CO where ThomasW is and I used to live, my suggestion would be to consider using the Modula MT/Natalie P crossover concept, use the HiVi D6.8 midwoofers (available from Parts Express), use the Seas H1212 tweeter (available from Madisound), build the cabinets per the dimensioning of your Totems if that's what you prefer, design a downward firing port using Unibox in Excel based on the internal dimensions of the cabinet (or adjust the cabinet slightly to optimize response for extension and a slight downwards slope in response from 100 Hz so that it isn't too heavy in room).

OK, those tweeters and midwoofers are cheap enough that I won't be out too much if I screw things up badly. I'll have to read through those Modula MT/Natalie P threads in detail and see what I can learn. At the moment, the circuit diagrams are over my head. But I'm willing to learn.

The cabinet doesn't have to look exactly like the Forests I have. Whatever sounds best. Although the Forests do look kinda pretty.

The crossover component values would have to be adjusted a bit because the Totem Forest has a narrower baffle than the Modula MT, so the baffle step compensation will be different (that is, the drive will measure differently in this cabinet, and to get a target smooth response, the crossover would have to be adjusted). Also, the HiVi D6.8 has different impedance characteristics and doesn't have the upper frequency resonance, so the crossover could probably be simplified a bit.

No need to keep the same baffle size. But a simpler crossover would be good for a beginner like myself. I'll have to figure out the very basics of crossovers first. I'll read through the book I have and then those threads. And see what many questions that leaves me with.

Re the HiVi D6.8 and the Dayton RS180 used in the Modula MT, the HiVi will have a bit more LF output, but is a bit less efficient. The RS180 will have lower midrange distortion, and more immediacy in that frequency range, but not the same "guts" or slam on the bass below 50 Hz.

Some people (maybe it was you?) have speculated that the Forest sound is due to a lower output in the mids. I'm not sure, but if it's true, from your description the HiVi D6.8 would be the way to go, given that I like the Forest sound so much.

If you or someone else had measured data in these cabinets, it would be easy to do. (well, at least for me it would be a small amount of work to make an updated crossover simulation in LspCAD.

Sorry, I don't have any measurement gear.

Look at the Modula MT thread and you'll get a feel for what's involved. A little larger internal volume would be fine- I was working with off the shelf cabinet designs to make it easy for first constructors.

Which thread do you mean? There's a few sticky MT threads up there.

It would be fun to to a project like this, but my day job has been such a train wreck the last year that free time has dwindled to a very low level. Here it is Sunday morning and I'm working on materials (as I did yesterday) for an FAE training next week in Detroit.

Yup, I'm here at work on Sunday as well...

-Jon

[JonMarsh]

Try to hack your way through these two threads.

Natalie P Thread

Modula MT

Both use a similar crossover concept, which is lower cost and simpler than my Modula MTM, or M8ta, or Arvo Part, but still is pretty high performance.

If you were feeling frisky, you could implement a Totem Forest Klone on steroids with two woofers in a large enclosure using an approach like the Natlie P. The Seas H1212 would work without crossover changes in place of the Seas 27TDFC versions of those crossovers. The Dayton RS28a is in the "premium" versions; it's a little more expensive. It's a very nice tweeter, doesn't sound at all like a Metal Dome in the classic negative sense that some complain about. The H1212 is quite good, though, too, with very low distortion even below 1.5 kHz.

After next week some time I could probably rough out an LF alignment and suggested enclosure approach. If you have access to Excel, I'll do that in Unibox (a free download- google it), and you can play around with the parameters to, to get a feel for it. If it can come in close to 1 cu ft, I could use an extra set of baffles and prototype it in one of my 1 cu ft MTM PE enclsoures, perhaps. They look narrow enough to be similar to the Totem Forest.

~Jon

[knifeinthesink]

Jon-

Thanks for that info. I guess I don't really have any perspective on what you said. If I were to embark on my first speaker building endeavor, what are the odds I'd get something that sounded OK? I realize from what you say it won't sound perfect. But will it sound simply awful, like cars crashing and finger nails down the blackboard? Or just not as good as it possibly could be?

-Jon

I cant give you any advise, but I'll share my own experience with you. I was in a similar situation a year and a half ago, but without the benifit of any science background (i studied art history, fun but utterly useless). I couldn't tell you what a volt was if my life depended on it. Nevertheless I somehow managed to get a chip amp together without electrocuting myself and only had one puff of smoke. I built some open baffles with some crummy drivers I found on the sidewalk (theres a kind of universal understanding in toronto that when your done with something that could concievably be of the slightest use to anyone else, then you put it on the sidewalk and the reusit gnomes come in the night and take it away) If I had a pair of good speakers to compare with I probably wouldn't have been impressed but I didn't so I was. And these things did have some merits on account of the open baffle.

I finally built a full range single driver speaker from zaphaudio.com and then added some 12" alnico magnet woofers from some speakers I bought at a garage sale. I had some inductors I salvaged from some other garbage speakers and used those on the woofer and let the 3" full range fall off naturally. These actually sounded really good. I had some acclaimed mission bookshelf speakers that a friend had paid about $500 for and left behind when he moved and the frankinspeakers absolutely blew them out of the water. I still marvel at that but somehow it worked out. Now Im ready to start using measuring equipment and building some more sophisticated projects.

So what my point? You can really get some nice sounding stuff very cheaply if you keep things fairly simple to start. For me, it was great to start with a very simple and very cheap single driver speaker, and then play around adding some bass. If i had tried to clone something I think I would have been in over my head. Something like the single driver HiVi project at www.zaphaudio.com would probably be great and John says that 3" HiVi is better then the excellent Tangband in the speaker I ended up with. Plus the HiVi is ridiculously cheap.

Like I said, Im not the person to give advise, especially in this forum, but for me it was really worth while to jump in. I'd go for it if I were you. Just dont spend a pile of money if your doing your own.

Madisound has alot of really cheap drivers in their sale section. The Peerless and the Audax look like good buys and the response doesn't look bad. I just recieved a pair of the audax and they look good for the money but I haven't tried them yet.

[Brian Bunge]

Jon (and JonW),

I loaded the D6.8 parameters in Unibox (found them at Madisound) and here's what I'm getting. If you go with a 1ft^3 enclosure you end up getting a decent hump at the tuning frequency of the enclosure (Fs is 43Hz). Tuning down around 30Hz doesn't look too bad, but you have a downward slope in the response, sort of like an EBS (Extended Bass Shelf) but then have a couple dB peak at the tuning frequency. If you stuff the box fairly heavily, the peak goes away and you do end up with an EBS alignment. I'm not sure how safe it is to tune so low below the driver's Fs, but it does look like a nice response to me.

[JonMarsh]

Thanks for running the numbers

Hello Brian,

Thanks for running the numbers. What you're describing at the end is EXACTLY the kind of alignment I usually use ported. Sounds like it will work out perfectly, then, or close enough for government work. Usually I shoot for around -4 to -5 dB from the 100 Hz point down in the 30-40 Hz region- that sounds good "in room" with articulate bass without sounding heavy or muddy. Probably "polite" sounding to some, but you know how I am about my music...

~Jon

[Brian Bunge]

Jon,

It's a little more gradual than that. It's really only about -2dB from 100Hz down to 30Hz, with the "knee" being just below 30Hz. But the "shelf" really is right in the 30-50Hz range. I believe this is sort of like the TL-type response you claim you get from your Avalon Klones (can't recall the name at the moment).

One thing is for sure, this is definitely suited as a tower design with a downward firing port. I'm showing 16.17" port length for a 3" diameter port.

[JonMarsh]

I would expect that, but the 3" port will get the most LF air motion and control velocity. With smaller enclosures you can get by with a 2" flared, but I'd normally not prefer that.

Thanks again for taking the time to confirm the concept for the D6.8.

~Jon

[Dennis H]

Some basics for JonW. It's easy to miss the basics on this forum because discussions here tend to get into the fine points of state of the art and assume the basics.

1. Distortion. Tweeters distort if you play them too low and mids distort if you play them too high. So, the first whack at determining the crossover frequency is where neither the tweeter nor the mid is distorting too badly.

2. Polar response. If the center-to-center spacing between the mid and the tweeter is more than half a wavelength at the crossover frequency, you will get uneven response between seated and standing listening positions. The sound from one driver cancels the sound from the other when you stand up. That's simply a matter of the geometry -- both drivers are playing the same frequency but the path from one to your ear is longer than the other to your ear and they cancel -- your ear hears a + from one driver and a - from the other. Bottom line, you want to keep the crossover frequency as low as possible within the constraints of #1.

3. Badly behaved drivers and the implications for how steep the crossover needs to roll off the response. Some midrange drivers, with the best distortion behavior at the frequencies where you plan on using them, act really nasty as you go higher in frequency. It's sort of a tradeoff. You can use a high-distortion driver with no bad behavior higher up or you can use a low-distortion driver and build a complex crossover that vigorously suppresses all the higher bad behavior. Jon's designs fall in the latter camp. He's picking drivers that are very low distortion at the frequencies where he's using them but they behave badly outside that range so he needs complex (and expensive) crossovers to suppress the bad behavior.

[cjd]

Ahh, Dennis... you mean to say, you can use drivers with low distortion through the region where you want to use it, and make sure that the spots where the drivers sound bad are supressed, OR you can use drivers that distort EVERYWHERE and claim it's nice sounding distortion AND you just saved money on crossovers because the really really bad stuff is muffled by the driver just like the stuff you want. :P

(for those of you that might not understand, the prior sentence-ish is intended as tongue in cheek, despite it being something I have found to be somewhat true in my limited experience. it's far from gospel though).

IMHO one of the best ways to get into crossovers is to go find some cheap buyouts that seem to work well where you want them to work, dive in and get your hands dirty, and understand that it just may take a few attempts before you don't run away crying. The first one I thought "ahh, that sounds good!" is now on my list of not-so-good early attempts. Though not bad - it's probably no more off-balance than Paradigms. :B Cheap stuff means you can experiment without feeling much guilt if you botch something or, say, short a driver and toast it, run a screwdriver through it on your fiftieth install, etc.

C

[Dotay]

No problem. In the warm months, I spend weekends driving around race tracks. And this stereo/speaker building/woodworking is something to distract me for the winter months. I'm 36, but with the fast cars and such, I'll also have no way of seeing when a midlife hits. 8)


-Jon

It's off topic for sure but where do you race? I just got into SCCA SoloII with my S2000 this past year and had a ball. I know there is at least one other former racer (cjd) on the board too. Oh, and what is your car of choice?

Oh and seeing how your from Lafayette I'll go ahead and say that the IU vs. Purdue games were so much more entertaining when both teams were battling atop the Big Ten. I miss the classic Bob Knight (his last year at IU was my last year there incidentally) Gene Keady matchups...

[cjd]

Yeah, I did SoloII and got sick of the looong days, the crappy boring drive TO the events, and most significantly, the politics. Sucked all the fun out of things. So now my poor ACR just has to deal with a 5 mile commute to the train. Needs a new clutch too. :P

If I may make a recommendation: Stay the hell away from doing ANY timing and scoring. They were trying to make me T/S chief.

C

[JonMarsh]

Ahh, Dennis... you mean to say, you can use drivers with low distortion through the region where you want to use it, and make sure that the spots where the drivers sound bad are supressed, OR you can use drivers that distort EVERYWHERE and claim it's nice sounding distortion AND you just saved money on crossovers because the really really bad stuff is muffled by the driver just like the stuff you want. :P


C

Chris,

Have you joined Evil Twin on Dark Side of Cynicism? :yesnod:

[cjd]

Have you joined Evil Twin on Dark Side of Cynicism?

Not sure I'm that far. I would say it is more likely my level of cynicism measures up almost to yours. Almost, because I don't have the hears to let it steep and mature. Yet.

C

[Paul H]

Chris I disagree - that seemed like a very mature and developed cynicism to me

Paul

[JonW]

Try to hack your way through these two threads.

Natalie P Thread

Modula MT

Both use a similar crossover concept, which is lower cost and simpler than my Modula MTM, or M8ta, or Arvo Part, but still is pretty high performance.

If you were feeling frisky, you could implement a Totem Forest Klone on steroids with two woofers in a large enclosure using an approach like the Natlie P. The Seas H1212 would work without crossover changes in place of the Seas 27TDFC versions of those crossovers. The Dayton RS28a is in the "premium" versions; it's a little more expensive. It's a very nice tweeter, doesn't sound at all like a Metal Dome in the classic negative sense that some complain about. The H1212 is quite good, though, too, with very low distortion even below 1.5 kHz.

After next week some time I could probably rough out an LF alignment and suggested enclosure approach. If you have access to Excel, I'll do that in Unibox (a free download- google it), and you can play around with the parameters to, to get a feel for it. If it can come in close to 1 cu ft, I could use an extra set of baffles and prototype it in one of my 1 cu ft MTM PE enclsoures, perhaps. They look narrow enough to be similar to the Totem Forest.

~Jon

OK, I've read the Natalie P thread. Haven't hit the other one yet. (Work does tend to get in the way...)

Right, so the 2 woofer idea... I had actually thought of that before you mentioned it. On one hand, I want to try and match the Totems somewhat, so maybe I should go with just one woofer. On the other hand, if a second woofer would improve the sound, I should do that. It'd be silly to stay beholden to an 'inferior' design. (Which makes me wonder why the Forests only have one woofer.)

I do have Excel here. But I'm a Mac guy. I haven't been able to get Unibox towork properly. I input a few numbers and then spreadsheet goes to all text or something. Grrr... (d'oh- stupid PC's!)

So maybe stick with the Seas H1212 and HiVi D6.8 to keep close to the Forest sound? Or maybe try something different?

I've been thinking about this project a lot. I think building my first speakers will be quite a challenge. But it could be very fun. And I haven't been able to talk myself out of it yet. I am wondering if I should start with something simpler, to begin with. But, eh, that would be making poor speakers from the start. That's no fun. So maybe I should go for this Forest-ish clone.

[JonW]

knifeinthesink-

Thanks for the perspective. It’s nice to see what other people went through… Before I get too far into this.

Jon (and JonW),

I loaded the D6.8 parameters in Unibox (found them at Madisound) and here's what I'm getting. If you go with a 1ft^3 enclosure you end up getting a decent hump at the tuning frequency of the enclosure (Fs is 43Hz). Tuning down around 30Hz doesn't look too bad, but you have a downward slope in the response, sort of like an EBS (Extended Bass Shelf) but then have a couple dB peak at the tuning frequency. If you stuff the box fairly heavily, the peak goes away and you do end up with an EBS alignment. I'm not sure how safe it is to tune so low below the driver's Fs, but it does look like a nice response to me.

Thanks for running the numbers. If only I knew what you are talking about.
Any reason you kept the enclosure to 1 ft^3? Or is that a standard place to begin? The Forests are floorstanders, so making a big speaker is AOK by me. Whatever would sound best, really. For what it's worth, the Forests are ported, in the back.

[JonW]

Some basics for JonW. It's easy to miss the basics on this forum because discussions here tend to get into the fine points of state of the art and assume the basics.

1. Distortion. Tweeters distort if you play them too low and mids distort if you play them too high. So, the first whack at determining the crossover frequency is where neither the tweeter nor the mid is distorting too badly.

2. Polar response. If the center-to-center spacing between the mid and the tweeter is more than half a wavelength at the crossover frequency, you will get uneven response between seated and standing listening positions. The sound from one driver cancels the sound from the other when you stand up. That's simply a matter of the geometry -- both drivers are playing the same frequency but the path from one to your ear is longer than the other to your ear and they cancel -- your ear hears a + from one driver and a - from the other. Bottom line, you want to keep the crossover frequency as low as possible within the constraints of #1.

3. Badly behaved drivers and the implications for how steep the crossover needs to roll off the response. Some midrange drivers, with the best distortion behavior at the frequencies where you plan on using them, act really nasty as you go higher in frequency. It's sort of a tradeoff. You can use a high-distortion driver with no bad behavior higher up or you can use a low-distortion driver and build a complex crossover that vigorously suppresses all the higher bad behavior. Jon's designs fall in the latter camp. He's picking drivers that are very low distortion at the frequencies where he's using them but they behave badly outside that range so he needs complex (and expensive) crossovers to suppress the bad behavior.

Thanks! I appreciate your coming down to my level. What you say all makes perfect sense. I'm now halfway through reading the one speaker book I have (Speaker Building 201) so I'm still learning such basics.

KeithM recommended getting a copy of Loudspeaker Design Cookbook. Would you all agree I should get that book to learn some more? Or might there be something better to get at this stage?

[JonW]

IMHO one of the best ways to get into crossovers is to go find some cheap buyouts that seem to work well where you want them to work, dive in and get your hands dirty, and understand that it just may take a few attempts before you don't run away crying. The first one I thought "ahh, that sounds good!" is now on my list of not-so-good early attempts. Though not bad - it's probably no more off-balance than Paradigms. :B Cheap stuff means you can experiment without feeling much guilt if you botch something or, say, short a driver and toast it, run a screwdriver through it on your fiftieth install, etc.

Good idea. Maybe I should just jump in and try something. How ambitious do I get? Hmmm... On one hand, I should try something easy and cheap. On the other hand, that will leave me with a speaker I'm sure not to like. So maybe try something more ambitious, like a Forest clone. A few hundred bucks extra isn't a problem- unless it means that in my learning I'll short out $100 of drivers rather than a 50 cent resistor. h: I'm willing to spend more money to get better sound. Just not sure how much I'll break on the way.

[Evil Twin]

OK, I've read the Natalie P thread. Haven't hit the other one yet. (Work does tend to get in the way...)

Right, so the 2 woofer idea... I had actually thought of that before you mentioned it. On one hand, I want to try and match the Totems somewhat, so maybe I should go with just one woofer. On the other hand, if a second woofer would improve the sound, I should do that. It'd be silly to stay beholden to an 'inferior' design. (Which makes me wonder why the Forests only have one woofer.)

I do have Excel here. But I'm a Mac guy. I haven't been able to get Unibox towork properly. I input a few numbers and then spreadsheet goes to all text or something. Grrr... (d'oh- stupid PC's!)

So maybe stick with the Seas H1212 and HiVi D6.8 to keep close to the Forest sound? Or maybe try something different?

I've been thinking about this project a lot. I think building my first speakers will be quite a challenge. But it could be very fun. And I haven't been able to talk myself out of it yet. I am wondering if I should start with something simpler, to begin with. But, eh, that would be making poor speakers from the start. That's no fun. So maybe I should go for this Forest-ish clone.

Ah yes... I know well of the problems with regards to running Excel sheets with Quickbasic on the Mac. Annoying, and it would be even more annoying to stooping to running PC office in emulation just to run these version- for that I have an AMD64 laptop.

There is one speaker design program for box design which runs in Classic mode. Not sure if it's really worth it, will try it myself perhaps next weekend.

Now you see the nature of the slippery slope of DIY. First, it is just building something like a commercial speaker you know or admire. The next step is deciding to "improve" on it, as commerical speakers are built to make money, while DIY speakers are built to make music. A substantial distinction, in my opinion.

If you are happy with the Toem Forest, a first step towards the dark side would be a basic "Klone", keeping open the possibility to adjust cabinet and crossover parameters somewhat to be sure they are optimum for the drivers used and achieving a tone target similar to published information (like the Sterephile review). It would probably be advisable to consider a treble balance slightly less "hot" on top than what the review shows, or at lease implement the possibility for adjustment.

Once you successfully complete your first project like this, then you will have a much clearer feel in the force of what you are willing to persue in the future.

[JonW]

It's off topic for sure but where do you race? I just got into SCCA SoloII with my S2000 this past year and had a ball. I know there is at least one other former racer (cjd) on the board too. Oh, and what is your car of choice?

Oh and seeing how your from Lafayette I'll go ahead and say that the IU vs. Purdue games were so much more entertaining when both teams were battling atop the Big Ten. I miss the classic Bob Knight (his last year at IU was my last year there incidentally) Gene Keady matchups...

Yeah, I did SoloII and got sick of the looong days, the crappy boring drive TO the events, and most significantly, the politics. Sucked all the fun out of things. So now my poor ACR just has to deal with a 5 mile commute to the train. Needs a new clutch too. :P

If I may make a recommendation: Stay the hell away from doing ANY timing and scoring. They were trying to make me T/S chief.

C

I'm just doing driver schools at this point. Mid Ohio, Putnam Park, Road America, Gingerman, Grattan. It's a real hoot. I have a '93 BMW 325is for a track car. Wonderfully neutral handling at the limit. And the BMW club events are great. I'd really like to get into competitive racing. But then I'd need a full roll cage (fine- safety is good). But then I'd need a truck and trailer to get to and from the track. Ugh. Now that's way too expensive for the likes of little ol' me.

Yeah, I keep the drives to the track to within 5 hours of home. Sometimes Sunday nights are tough, driving home after 2 exhausting days on the track, it's dark, the car is loud and bumpy...

I can't wait for it to warm up to get back out on the track.

Yup, Gene Keady and Bob Knight are certainly two huuuge personalities in college hoops. Maybe after a few years of growing up/rebuilding, things will be good again. We'll see.

[JonW]

Ah yes... I know well of the problems with regards to running Excel sheets with Quickbasic on the Mac. Annoying, and it would be even more annoying to stooping to running PC office in emulation just to run these version- for that I have an AMD64 laptop.

There is one speaker design program for box design which runs in Classic mode. Not sure if it's really worth it, will try it myself perhaps next weekend.

OK, great to know I'm not the only one having trouble with Unibox. For the sub I'm building, I've been 100% dependent upon a friend to do all the calculations for me, on his PC. Kind of makes me feel helpless. If I'm going to build speakers and such, I'd like to have a better handle on calcuations. But I guess it's not so essential that I'll go out and buy a PC.

Now you see the nature of the slippery slope of DIY. First, it is just building something like a commercial speaker you know or admire. The next step is deciding to "improve" on it, as commerical speakers are built to make money, while DIY speakers are built to make music. A substantial distinction, in my opinion.

Exactly the thought process I have arrived at.

If you are happy with the Toem Forest, a first step towards the dark side would be a basic "Klone", keeping open the possibility to adjust cabinet and crossover parameters somewhat to be sure they are optimum for the drivers used and achieving a tone target similar to published information (like the Sterephile review). It would probably be advisable to consider a treble balance slightly less "hot" on top than what the review shows, or at lease implement the possibility for adjustment.

Once you successfully complete your first project like this, then you will have a much clearer feel in the force of what you are willing to persue in the future.

My current thought, based on zero experience, is to start from a Totem Forest-ish clone. But right away add in some things to the design that seem like they might improve the speaker a bit. Like a second woofer. Or any other suggestions that seem decent. And be open to changing the crossover with time. But I'll needs lots of help from you folks to make this happen.

If you're a Mac guy, I'm amazed that you didn't take a swipe at PC's and Microsoft by using the term "dark side." See, I don't have such restraint.
:P

[Evil Twin]

Well, there is some discussione about what is the "dark side" and what is the "light side", as there are those who believe that Apple's current positive trends in the last few years are due to Steve Jobs skillful use of his powers as a covert Sithlord. The well known "Reality Distortion Field" just being a way for the unbelievers to explain Force based mind control or manipulation....

"These aren't the droids you want...." "Go about your business..."

"These are the new Mac's you want..." "and, just one more thing...."

In my day job apart from acting as the right hand man of the Emperor, my new master has recruited me as a similar right hand man, but with more restrictions placed on my behavior and public pronouncements... some of this unnatural reticence may be rubbing off on my conduct elsewhere. It does not change the basic goals or passion.

Since my company laptop has been dead since the begining of December, and due to the bumbling of our laughable excuse for an IT department, is still not replaced, it gives me considerable pleasure to show up at customer meetings, internal trainings, and management conferences with my Powerbook and wow them with what a well prepared presentation looks like on Keynote... compared with the pathetic semblance of computing power emboided in their Dell laptops running Windoze 2000 and a not even a current generation of Micro$oft Office. :P

[Brian Bunge]

Well, I mainly went with that because Jon already has a cabinet that is 1ft^3 and could take measurements for you without much trouble. Also, looking at the Forest specs on the Totem website, if I take the cabinet dimensions listed and assume 3/4" walls, the cabinet actually ends up being around 1.1ft^3. Now this does not take into account any bracing used, but I think that's close enough.

I also know that Totem uses sealed chambers at the bottom of their enclosures for sand filling in desired. I don't have any way of knowing just how large that chamber is, but if I had to guess, based on Totem's low frequency specs, I'd say that the actual enclosure is around .75ft^3 with the smaller subenclosure being about .25ft^3 for sand filling. Tuning .75ft^3 around 35Hz gives an F3 of 35Hz as well. If you didn't want to bottom port the enclosures you could build something similar and use a 2" port.

EDIT: I wanted to add that I got the opportunity to hear three of the Totem towers last year. I believe one of them was the Forest as I seem to remember noticing the HiVi driver. All three struck me as very nice sounding, being run by McIntosh amps. They also struck me as very small, all being around the same height and very narrow. I will say they all seemed a little constrained and definitely power hungry, but that's just based on a short demo. I was a little shocked at the prices they were charging as well (Forests around $2500 or so?). That's a lot of money for a small 2-way tower. I wouldn't be surprised if you actually preferred a crossover designed by Jon over the stock crossover that Totem uses.

[Evil Twin]

The HiVi D 6.8 is not very efficient per se; the longish throw VC and other paramters which work to it's advantage in the low bass also reduce sensitivity (around 86.5 dB rated). With full baffle step and crossover insertion losses, one could expect sensitivity in the 81 dB or less range. Think 50-100 watts per channel. An MTM would help quite a bit, but of necessity be much larger.

Proposals will be made after some free time for calculations.

[JonW]

Hi Again,

I'm just now back from a crushing week of work. More time to now think about building some speakers.

The HiVi D 6.8 is not very efficient per se; the longish throw VC and other paramters which work to it's advantage in the low bass also reduce sensitivity (around 86.5 dB rated). With full baffle step and crossover insertion losses, one could expect sensitivity in the 81 dB or less range. Think 50-100 watts per channel. An MTM would help quite a bit, but of necessity be much larger.

Proposals will be made after some free time for calculations.

I'll be curious if you (or anyone else) have a chance to come up with any potential plans. I don't care what size the speaker would turn out to be, so long as the sound quality is optimized. Large floorstanders are AOK by me. Although I might also want to make a matching center speaker, so that might not be able to be a florstander.

Maybe go MTM rather than the direct Forest copy? In my recent reading, I'm starting to understand the general benefits of MTM designs. Should be a fun project if I can get sufficient help from you folks with the design. Not sure if I should start by building my first speakers as this clone or try something really easy and cheap, for the many mistakes I'm likely to start out with.

[JonW]

In my day job apart from acting as the right hand man of the Emperor...

Ah, I see. You are Bill Gates' secretary.
:P