DIY MTMWW Questions

[SteveCallas]

Since I had such great success with my DIY sonosub, largely in part because of all the great help I received on the audio forums, I am now interested in building some DIY mains and a center. I realize building a successful speaker will be a lot tougher than a sonosub, but I am in no immediate hurry and will do as much reading and researching as I have to. I was pretty set on auditioning some Ascend 340s, but I am wondering if I can get the same performance out of DIY speakers that I did out of my sub if I use a similar "no compromise" approach. I plan on these speakers being as large or as extreme in design as they need to be. No WAF or much of aesthetic concerns for me - purely performance.

The design I have in mind is a MTMWW similar to the one Brian has going. The main design difference is that I plan on isolating each driver in its own sonotube enclosure. The tubes will be mounted in an MDF shell that will consist only of an exterior and some braces inside to support the tubes. This cabinet will probably be very deep in order to supply enough length to the tubes for the woofers to come out with a pretty optimized volume. Without yet knowing much about how to build speakers, I assume this design would instantly eliminate several of the common problems associated with cabinets, volume, and driver interaction.

I am pretty set on using the Vifa Ring tweeter, as I have heard nothing but good thinsg about it. For the rest of the drivers, I am not so sure. Since my sub can handle everything below 100hz with power and precision, I don't need deep extension. Based on the limited researching I have done so far, I am thinking of using a pair of Hi-Vi F5 midranges and a pair of Dayton RS270s. I believe in Brian's thread, cjd suggested the Dayton Reference Subwoofer drivers might have been a better choice.

Also, since I am trying to go strictly for performance, is MTMWW not a good option? Would an open baffle or something else be much better?

For a crossover, since I only have very limited circuit knowledge/experience, active crossovers might be a very real choice.

Basically, I was hoping you guys could point me in the right direction to start learning about driver layouts, baffles, impedances, crossovers, etc., all the basics I would need to know to start building a set of speakers. I'd rather not buy a book because I'm sure all of this info can be found online, but if I had to, I would. Also, what programs can I use to start playing around with deisgn parameters and get predicted results?

I would appreciate any and all help. If I were to optimize this design as much as possible, I am assuming a speaker based off of the Vifa Ring tweeter and some high quality woofers could outperform the 340s? I ask because this project will likely cost me around $1200, about twice the price of the 340s.

[ThomasW]

I advise that you copy one of the existing designs we have posted on this forum. Designing a fullrange loudspeaker from scratch is not for the beginner or anyone without a significant amount of test equipment and knowledge.

The original Arvo dipole was an MTMWW that used the XT25. We also did a box design that was a MTM using the XT25, all you'd need to do is a bass bin for the woofers.

Note that the XT25 is a good tweeter, but there are others available that are as good if not better and cost less. Look at Jon's Modula MTM designs for guidance

Finally I see no benefit to using tubes for the enclosures then making a cosmetic cover for them. That's like building 2 sets of enclosures when only one is needed

[SteveCallas]

Finally I see no benefit to using tubes for the enclosures then making a cosmetic cover for them. That's like building 2 sets of enclosures when only one is needed

What about:
-no cabinet resonances or standing waves
-no backwave from one driver interacting with any of the other drivers
-optimized volume for each specific driver
-no need for extensive bracing
-no need to make driver cabinets asymetrical
-ability to make a seperate baffle for the tweeter, mids, and woofers very easily and best suited to their performance

If I'm way off on these, let me know, but when I read about building DIY speakers, these are often compromise areas. I realize that drivers used and crossovers affect sound much more than any of these areas do, but if I'm trying to get the best performance I can, using two enclosures isn't much of a task.

Also, if I just copied an already established MTM, I might as well just buy the 340s.

[cjd]

Choosing drivers is half the fun, and sometimes gets you into more trouble than you think.

You are not going to solve any volume issues with Sonotube - for most folks it's a matter of how big they won't go. Any well designed box will not have driver interaction issues either. If you look at the internal views of the box I did, there is no driver interaction unless you count both mids interacting or both woofers interacting: but they DO interact electrically whether you isolate them physically or not, and I'm not really sure that the physical interaction is anything bad or good - I think it's simply a non-issue. My other concern with sonotube is whether it will allow you to put together a properly spaced baffle with respect to C2C.

If you're wanting to go with a 100Hz crossover, consider something like the Modula MTM over a 3-way. I just don't see the gains in the expense of a 3-way (crossover is more than twice as expensive as a 2-way plus the extra drivers and box materials).

Regarding the RS sub drivers in the 3-way I'm working on: they would give you better bass extension but would require at least bi-amping OR 2ohm (or lower) impedance through the bass section. Their sensitivity is low comparatively. 10's could be ported to <20Hz response in the specific box I did. 12's would also drop in sealed quite happily and have room for a L-T circuit to again give tremendous bass response.

However!

Were I to approach a boxed 3-way I think I would listen to myself and not Brian in choosing the drivers. I would go with an MTM using RS150's for the midrange and up. I would probably bi-amp at the least and go with the 10" RS subs.

The Vifa does get mostly excellent reports, some less so. I would have no hesitation recommending it.

I think my favorite in my price range is still the 27TDFC. The RS28A is very very and may grow on me as I tweak the crossover I have on it, but it almost seems too brittle for my ear. That, of course, may not have anything to do with accuracy and everything to do with personal preferences.

Man, it has taken me a while to write this up! Amazing what actually having work will do to your pass-time activities.

C

[JonMarsh]

Steve, I think Thomas had some very good points. Considering the kind of physical effort and expense you're talking about, I'd consider pretty carefully what tweeter to use, and what midwoofers to match up with it.

The problem with your component choice, even if you go wtih the Scan Speak version of the ring tweeter, is that MTM's are problematic in performance UNLESS you can have the crossover frequency low enough so that at least the center to center distance on the woofers (if not the edge to edge radiating area) is within 1/2 wavelength at the crossover frequency, AND you have moderately steep crossovers. Otherwise, most of the finger pointing done at MTM's in many other forums is mostly valid, IMO.

Regarding the overall MTMWW, been there, done that.


X1 SLAMM CIOWIN Grand Prix

Now, Wilson audio has a new top of the line MTMWW, the Alexandria X1. They also have the Maxx and other similar systems. The Alexandria DOES use a ring radiator, but only as the ambience tweeter.










The ambience tweeter is the high end Scanspeak ring radiator; this only runs above ~ 8-10 kHz, where the distortion of the ring radiator is reasonably low, and dispersion is not such an issue, as it's designed to "bounce" off the rear wall. The main tweeter is still an inverted dome Focal.



OK, something you may or may not realize- a speaker design like this makes a sub superfluous. Or, looking at it another way, maybe you don't need something quite this large for your mains. HOWEVER, don't assume that a smaller system with sub will have the same impact- one of the benefits of the X1 and X2 configuration is the large vertical source in the 100 Hz - 300 Hz region, which provides great room drive in the power region (100 - 250 Hz) and minimizes floor bounce nulls. Just putting an MTM on stands and connecting to sub doesn't do the same thing....


OK, last, I'm going to make a more radical suggestion for your consideration- think outside the box. I.E., don't use a box. One of the "features" of the big Wilson systems is that they have a factory trained tech from the dealer come in and carefully evaluate your room and speaker placement, and work on setting them up to get the best overall performance, considering how much a large box speaker like this interacts with all the room boundaries in both the bass and midrange. Solution? Get rid of the box.

Go dipole.

A lot less work and expense to build, easier to setup in a variety of rooms and get good to great results, and weigh much less when you have to move them around. Especially if you have a monpole sub you can use in the below 35-40 Hz region, this is, IMO, the best solution.

That's what ThomasW has.

So do I.








There are many other approaches, too, such as Monte Kay's Dayton RS based dipoles.







~Jon

[cjd]

-no cabinet resonances or standing waves

Not accurate.

-no backwave from one driver interacting with any of the other drivers

A well designed box will elminate this where it matters.

-optimized volume for each specific driver

A well designed box will do this.

-no need for extensive bracing

This one may be true. Where do you mount your crossovers though?

-no need to make driver cabinets asymetrical

What? You don't need to. If you do, it has everything to do with diffraction, nothing to do with construction imitations.

-ability to make a seperate baffle for the tweeter, mids, and woofers very easily and best suited to their performance

Can be done. Maybe easier if you find boxes difficult to make?

Also, if I just copied an already established MTM, I might as well just buy the 340s.

Um. No. You would have everything you would get if you manage an outstanding design, except you wouldn't be having to learn all the ins and outs of design. I started with drivers I picked up for dirt cheap. My first handful of designs sounded horrible. I keep getting better, learning new things. The design I finally settled on for that first project now has very obvious shortcomings that I was NOT aware of at the time. If you want to learn design, dig in. It's a long but fun road.

If you want to build yourself some top notch speakers you do not need to learn design. Every day, I consider whether I am wasting my time (and money) trying to push my understanding of design further.

C

[Brian Bunge]

Hey don't blame me! I'm blaming it all on Jon. I didn't realize there'd be issues adding dual 10's to the RS 7's. I figured they'd integrate better with the 10's than the RS 6's would.

Concerning an established design vs. the Ascends, well to me the whole point is to get a much better speaker. The Ascends look like good speakers for the money, maybe even great speakers for the money, but I'm not fond of these aesthetics of them and I don't really care for 5/8" MDF enclosures. They may be fine for this design, but I like really sturdy, heavily braced cabinets. I'd expect Jon's Modula MTM's to blow them away in many ways. Plus, building them is half the fun!

[JonMarsh]

Re the Ascend 340's, there price is very reasonable, considering they use some decent Audax Aerogel midwoofers, but I question whether they would compare well with a good DIY design, even a "budget" oriented one like the Natalie P.

By there own specs, enclosure is built with 5/8" MDF, not 3/4" or 1".

The FR and off axis response are better than many commercial designs I've tested or seen tested, but have a rather large BBC dip, and not a very smooth one at that.




The cumulative spectral decay they publish is fairly hashy in the 400 Hz to 1 kHz region, which points to cone energy storage in the midwoofers, or underdamped rear wave in the cabinet.



The Audax Aerogel woofers tend to be a little short on Xmax, too; even the 6.5 versions are typically only 3-4 mm, depending on model.


It probably isn't realisitc to think that you'll build a "high end" DIY MTMWW system for no more than $1200. Check the threads on the Modula and some of the variants like the Natalie P. A three way active or three way passive crossover itself that's well designed will set you back some serious change.

The least expensive implementation that might work well would be the MTWW that Dawaro is working on, with the crossover designed by Dennis Murphy- a tower version of that center channel design could work well. (RS28a, RS150, 2x RS225).


On my first post, I hadn't picked up on your budget goal of $1200.

Just some things to think about.

[SteveCallas]

Ok, thanks for all the helpful replies. I guess my best option would be to state what I want to get out of these speakers and then hopefully one of you guys can point me in the right direction to a premade design.

My sub is already capable of reference level bass from 100hz down into the low teens. I would want solid performance from the speakers down to maybe 60hz, so I could crossover at 80hz and get the extra clarity, dynamics, and slam (which Jon touched upon) in the upper and middle bass regions. I want room filling sound that is detailed and dynamic - good for music, movies, and an occassional video game. Probably 50% movies, 40% music, 10% video games. My sub is capable of so much and can do it so cleanly that now I want speakers that can do their job just as well.

Size, looks, and weight aren't much of an issue at all for me, as long as it can fit through a door. Extreme designs are not a problem either. I'd like to keep the cost for a pair of mains under $1500. A couple hundred of that will probably have to go towards a custom wood shop to cut the MDF, as I am in an apartment and don't have access to very many power tools.

Also, whatever design I go with, I want to be able to build a matching center shortly thereafter, assuming I am pleased with the results. I'd expect it to have a nearly identical sonic signature, and the height would become the only concern. I don't expect to stack it on a tv, but even with it on a stand or if it had it's own stand, I don't want to have to position my tv so that it would be ridiculously high off the ground.

I've read plenty about the Ascends and they seem to be great speakers. I just don't think they could be dynamic enough or fill the room enough to match the sub with only two 6.5" woofers. I'm sure the upgrade bug will hit eventually, as I'm only 22, but I would like whatever I build to satisfy me for a while and be something I can rely on.

Again, all help is truly appreciated.

[cjd]

Hey don't blame me! I'm blaming it all on Jon.

Heh. The issues are tweeter/mid not mid/woofer. I know we discussed RS150 vs. RS180 - you gain some and you lose some. And I may get the RS180's figured out yet. Have a couple tweaks in mind I want to try specifically. That's for the other thread. But the flexibility to cross a bit higher suits a broader range of tweeters. Also crimps Center to Center. So, always trade-offs.

One thing that may not have been said yet: If a fully DIY design is the goal here, figure out your assumptions about why things are done this way or that, and then throw them away and ASK why. You will get different answers, but in the process you will learn the ins and outs - no-one is likely wrong in why they do one thing or another, just different reasons, different compromises they prefer over other compromises.

C

[Brian Bunge]

Chris,

Let's hope they're simple tweaks with existing parts or limited change in parts!

[cjd]

To comment on your most recent post (which went up while I was typing that last):

Goals are excellent.

What is your subwoofer?

The big RS 3-way towers have an F3 of about 50Hz or something. In-room, it's probably closer to 40Hz. If someone else is building cabs, *just* in your price range if you're careful and do things like assemble large cap values from the good deal GE Poly's.

I think that you may find a more moderately sized setup with RS225's/RS150's would do everything you want though. The one Jon alluded to is certainly a good option, particularly as he's starting with the center channel. Long ago I made a pass at such a setup and worked with Robert early on, but realized (as with the big 3-ways) that live measurements really were a necessity, which is when Dennis got involved. I trust his work.

A "simple" two-way such as the Modula may also do everything you want it to do. The benefits here are in cost and complexity. Gives you the opportunity to really see what such projects are capable of and since this particular one is produced with a ready-made cabinet it would solve that issue for you handily.

I'm not sure that dynamics are what would lack if room-filling-ness is a concern for you: rather, I think it may be related to presence, that potentially elusive quality that I personally (and others) feel is related to driver size compared to the frequency range. It's one thing I like a lot MORE about my big 3-ways compared to the MTM's I did with the RS150. Which would suggest that you may be on the right track in considering a 3-way. Of course, it could be an issue of cone material too: poly cones in particular often seem "mushy" to my ear, and that softens the resultant presence significantly.

See, SOOOO many things to consider, argue about, get entirely depressed because you don't know WHAT to do, pick a road, walk it, find it's not right, wander around aimlessly for a while, and maybe, hopefully, find yourself where you want to be at the end.

C

[cjd]

Chris,

Let's hope they're simple tweaks with existing parts or limited change in parts!

erhmmm... suuure. Sorta. Limited change, probably so. Hopefully quite reasonably test-able. Really don't know, just some thoughts in the back of my mind that aren't fully there yet, so I'm letting them whisper in conference and they'll send me a memo with the results when they get there. Then maybe I'll try it.

[Brian Bunge]

I believe Steve has a BIG ported enclosure with an Avalanche 15 in it! So I'm guessing he wants something that'll keep up, SPL-wise.

[cjd]

I believe Steve has a BIG ported enclosure with an Avalanche 15 in it! So I'm guessing he wants something that'll keep up, SPL-wise.

If it were sealed I might be worried.

I would, however, aim for a system that will cross 50-60Hz, 80Hz at the highest.

C

[SteveCallas]

I have a 700 liter ported sonosub using an Avalanche 18 - port is 8" diameter and tuned to 14.5hz. Programs predict 115db solid to 15hz - I moved it out into the middle of my living room, 8' from the nearest wall and measured +/- 4db from 100-10hz at 95db. Didn't want to disturb the neighbors TOO much by testing at a louder level. It is super clean and tight with music, and handles all the low stuff in movies with ease. The scary thing is that I am building a second - so I will probably never upgrade my bass needs again.

Here is a link Sonosub

Image not available

I think "presence" is a better word to describe what I am after with this project. I really don't think a MTM with 6.5" drivers will give me what I am after.

So does it seem like a MTWW dipole would be a good option? What makes the MT better than TM? I would then assume the center would be MT vertically with a W on each side?

[Dotay]

Something else to consider if you want a big 3-way system is what you'll be driving it with. I know from following the Dayton RS WWMTM thread that cjd has had some serious amp issues trying to power his set. This could really throw a wrench in your budget and could be another reason to try out one of the simpler MTM designs like the Modula. Just a thought...

[SteveCallas]

I was planning on getting another Carvin hd1800, it's very powerful and very inexpensive. Getting the speakers at 4ohm would probably work best in my situation.

[cjd]

Something else to consider if you want a big 3-way system is what you'll be driving it with. I know from following the Dayton RS WWMTM thread that cjd has had some serious amp issues trying to power his set. This could really throw a wrench in your budget and could be another reason to try out one of the simpler MTM designs like the Modula. Just a thought...

Understand that the chip-amp cost me maybe $100 in parts and is unquestionably NOT suited to 2ohms. Its performance at lower volumes is superior to the Nakamichi I have on it at the moment. But yes, amplifier IS a good thing to note and consider.

Yet another reason to go with the RS150. IF you're talking RS drivers, but you really would be hard pressed to find a better value for the performance you get. The network I optioned for dewaro's TMWW doesn't go below 3ohm and would be just fine on a 4ohm stable amp.

C

[SteveCallas]

Ok I looked at Dawaro's thread - I don't think that is what I'm after. The sensitivity is lower than I am after, it uses two 8" woofers, and I was pretty set on the Vifa. Also there was some discussion that the design was somewhat compact which was a bit of a detriment - I really don't have any size limitations.

I guess I would be leaning towards the Arvo - though I would rather have it optimized to be sensitive instead of being able to hit below 40hz. How much room do you typically need to give a dipole like that from behind to the nearest wall, at least 4'? That might be a problem.

Any other suggestions?

[cjd]

Any other suggestions?

Do more research before making decisions.

The Vifa is a fine tweeter. I don't know anyone telling you not to use it. There are other (often better) options in its price range, some a bit cheaper. Most of the designs you are being pointed to are done without a pre-chosen tweeter, so they use what they use for whatever reasons the designer intended. So, go ahead and use the Vifa! No one here is going to tell you that you shouldn't (though some might suggest you could do better. )

Sensitivity (on dewaro's project): I do not know that anyone properly adjusted sensitivity to 2.83V/1M (or 1W/1M). The one I put out should be somewhere on the order of 90dB sensivity under such conditions. I suspect most of the rest will be in that ballpark - maybe a bit less since I see padding on the tweeter for most. Jon's should be right in there, and I have no basis for comment for the rest. Also, there is nothing preventing you from going with larger volume boxes. However, the ability to make huge boxes does not mean you must.

Generally (at least to the best of my experience) you can not optimize a given design for sensitivity (unless you're looking for an extra dB here and there). You can certainly produce high(er) sensitivity designs by using different/more drivers. I don't see much (if any) way to get significantly more output out of the M8's in this case.

I'm not sure what your information base is as far as some of the assumptions/constraints and whatever else you're making. So maybe I'm simply missing something.

Arvo's would be a stellar option though. I eye them now and again but have other dipole things in mind these days.

C

[SteveCallas]

My knowledge on building speakers is minimal. That's why originally I wanted to get pointed to some good reading to get more familiar with the basics.

Basically I was expecting to go for a design that did everything it could to get the most performance out of the drivers. Angled baffles to get the woofer closer to the listener then the tweeter was something I had in mind. I've read about dipole designs and that could be a solution for me - depends how much room you need in behind.

I would like to go back over a couple of the concepts of using sonotube again, if you don't mind. Don't take it as me being stubborn, I just wanna learn.

You said that no cabinet resonances or stadning waves was not accurate. By using tubes as the enclosure for each driver, then sliding them into braces with holes cut out to fit the tubes, with the inside edges lined with weatherstrip or rubber, how would my cabinet have any resonances? The drivers would basically be floating in air. And isn't it true that all box enclosures have standing waves and some type of fill is needed to tame or reduce them? A cylindrical enclosure does not suffer from this, or at least at a much reduced level.

As for the physical interaction between two of the same drivers, I'm not really sure how detrimental it is, but I would have to assume that optimally each driver in it's own enclsoure would be best.

And for making driver enclsoures with asymetrical dimensions, doesn't this help eliminate standing waves? Paradigm mentions it.

http://www.paradigm.com/Website/RandD/RandD/DesignComponents/RefBassDrivers.html

And finally, I would think the more volume you could give a sealed driver, the flatter the reponse will be.

As for getting more sensitivity instead of deeper response, I guess sensitivity was the wrong word. I should have said allow the speakers to play louder in the 80hz and up range. By getting them to play flat down to below 40hz as opposed to say 60hz, aren't you giving up some spl?

[JonMarsh]

The sensitivity of DAWARO's project is not "revealed" in those plots, because Dennis didn't normalize the data to a reference level. Some of the guys attenuated the mid. My design didn't, while still incorporating 4-5 dB baffle step. The actual 2.83 VRMS sensitivity with the crossover design I did (which certainly wasn't the cheap one) is about 89-90 dB. Note that there's a lot of inflation in most so called specs on sentivity. Any speaker with two 7" midwoofer in an MTM with BSC comp at best will only do a real sensitivity about equal to the IEC baffle rated sensitivity (that's the figure you see published for the drivers); this is becuase of crossover insertion loss and BSC comp.

My philosophy is to research, design, simulate, and ponder many, many times; build only once or twice for a given project. The research phase involves a lot less cold hard cash, so the better you do at that point, the better the value proposition.

Note that my 8" two way system published in AudioXpress took four prototypes befor I was satisfied more or less iwth the design, and the current version is 4.2, if you consider the changes in tweeters.

PS. Check out MarkK's distortion plots on the Vifa XT25 before you consider running it low in a two way. Just my 0.02.

And just keep posting and researching what's here and other locations online- eventually a clear picture of what you want to do to meet YOUR needs will coalesce.

~Jon

[SteveCallas]

I would be fine with 89-90db sensitivity. In the thread, I believe I had read an 84db sensitivity.

And I am not in a big hurry to make these - I was planning on having to do enough reading and modeling to not even begin ordering parts or begin construction until January 1, 2006.

I still want a three way, and I was thinking the Vi-Fi F5 would be a good midrange because it goes higher than any of the Dayton References. Without knowing too much about crossovers yet, I was thinking of crossing over to the mids at around 3khz and then to the woofers at around 500hz.

[cjd]

A cylinder will have two frequencies at which it might develop a standing wave (though the diameter dimension may be less strong than in a recangular enclosure? Not entirely sure on standing waves in this respect). A rectangle will have three assuming height, width, and depth are all different. Fundamentally, that is. Resonances are a different issue: At low frequencies the thinner walls of a sonotube will be fine since the resonant frequency is much higher than the output frequency range.

I mean, a chime is a resonant pipe.

What round does do well is eliminate more need for extensive bracing, particularly useful for a sub where size means weight, and things get big and heavy awfully fast. Even force distribution and all that.

Also, a round enclosure is *further* from asymmetric - it is symmetric any way you cut it on one axis, and in that respect less ideal. This is not to say that it is a bad idea, just that I'm not sure your WHY you might use it are on the right track.

And assuming anything is always risky IMO. You probably DO gain some measure of optimization if you have a separate enclosure for each driver. As long as this does not limit volume or proper center-to-center spacing there is no LOSS from it. There may also be no gain from separating drivers operating on the same leg of a crossover (i.e. operating as a single driver for all intents and purposes). Simply put, it is not enough of an issue for it to be anything most folks give more than a brief a second thought. It's far down the list of things to have as a significant issue to compromise over. If you can, great!

Re: volume (sealed) - the bigger the enclosure, the lower the system Q will be on average. There is a point at which it no longer makes sense going larger, as the gains are negligible. This coming from someone that tends to have folks say "oh man, those are HUUUGE!" in reference to my speaker projects. Past a certain size you have to pick your drivers with more care as well. Look at the IB sub concept. That's about as far as you can push this idea, and look at how few drivers are really labeled as suitable (though in this case it's often as much an issue of excursion, but not exclusively).

Do you have access to MS Excel? If so, the tools over at the FRD Consortium may get you going on some experimentation and discovery: particularly Unibox and the Baffle Diffraction Simulator. And give you some more specific points on which to ask questions.

Regarding your sensivity clarification: Isn't getting them to play louder the function of the volume control? What you're asking about is, I think, really an issue of baffle step. A given driver will have a specific response curve on a specific baffle - lower frequencies have a harder time below the baffle dimension(s) and this is often seen as loss in output - I like to think of it as the defining the zero point (from which we measure response being above/below when we say our response is 40-20kHz +/-3dB for example) of my response target (within reason, I'm not going to try to pad everythign in my 4" two way so it's flat from 20Hz on up). Less baffle step means you may have a higher zero point with a given driver (or array of said driver).

In the case of the Arvo, I do not see that the M8's are really being padded (though one crossover I peaked at does look like they have a little padding? 2ohm? didn't dig in TOO deep). Same goes for tweeter sometimes. In the case of dewaro's project, you may note that, at least in my proposal, the tweeter has no padding whatsoever. This means I can not get more output from this selection of drivers/baffle. Same with the big RS 3-ways (so far).

I can only speak for myself, but I use the T/S parameters and whatever else to get an idea of what the low frequency output potential of a system is going to be. That includes understanding where that zero point will get established. I try to match components so that, in a 3-way, minimal padding is required (sometimes tweeters need it, such as using a 100+dB rated ribbon in a two-way TM). If I want more bass output below, electronic boost circuits (such as the Linkwitz Transform) are used. If I'm trying to match particularly mismatched items (as I would if I put the RS subs in my big 3-way towers) I would look to multiple amps and work from that. Which is, I believe, what ThomasW's version of the Arvo's does.

Many ways to approach a project and get your desired results though, as always.

And I hope to be corrected (or challenged) if anyone has information to the contrary of what I've posted. I do not consider myself an expert, or even close. Just bored at work and *thinking* I have answers.

C

[cjd]

I still want a three way, and I was thinking the Vi-Fi F5 would be a good midrange because it goes higher than any of the Dayton References.

nah, it doesn't.

3kHz means other things entirely. ribbon or planar midrange perhaps. Probably not a cone driver. At least by Jon's standards for usable.

C

[ThomasW]

For some reason marketing types tend to focus on the concept of how well their technologies eliminate standing waves. The reality is that issues with standing wave are minimal and easily dealt with just a standard box design. As a result there's no need for exotic technologies or cab designs.

Take the time to read the first post in Jon's Modula MTM design. If you have the ability to generate the types of measurements and computer sims illustrated in that post, then you're equipped to design and engineer your own speakers.

We've played with several of the Hi-Vi drivers and the only ones we consider good enough are the M8a.

There are very few high quality drivers that can be used to create a system with a sensitivity as high as you want. Since watts are cheap these days, we design for the highest quality performance not fixating on efficiency/sensitivity

[SteveCallas]

Ok thanks for the info - I will definitely look into the Baffle Difraction Simulator - any other programs like that that would help me simulate a speaker design? I have Unibox and WinISD already. Any that allow you to see predicted results of multiple drivers working together?

Also, any way for me to take a look at some of your speaker projects cjd without having to dig through numeorus threads? Do you have them listed on a page like Thomas?

There are very few high quality drivers that can be used to create a system with a sensitivity as high as you want. Since watts are cheap these days, we design for the highest quality performance not fixating on efficiency/sensitivity

Makes sense.

[cjd]

Only other project that's probably worth looking at is my MTM.

I'm a relative greenhorn in these parts. I have too much time on my hands at work to read and ponder, and I have this knack for picking things up by reading and managing to do them (stuff like ice skating when I was a wee twirp, where I watched everyone else make mistakes for a while, and as folks started getting the hang of it, I went out and got it... without making any of the mistakes they did.) Plus a tremendous ego. And occasional helping souls.

Of course, Jon and Thomas and co. have a good number of years on me.

C

[ThomasW]

Of course, Jon and Thomas and co. have a good number of years on me

Hey! we're still young at heart ... :roflmao:

[SteveCallas]

Ok, after thinking about the Arvo design some more, I'm not sure it would suit my needs because a center channel would have to eventually be in the plans, and with an entertainment center or some kind of stand to hold the tv and equipment right behind it and to the sides, I would have to think that would affect the sound negatively of a dipole.

A three way design like Brians would probably be my best option, using two 150s, two 270s, and I guess the RS28as, since many here seem to favor its performance. I'll model up the 270 to see what the sealed response looks like compared to ported. I still like the idea of dedicated tube enclosures, so if the 270s look good sealed, I would probably go all tubes. If they look better ported, I would go tubes for the tweeter and mids, and then allow the woofers to use up all the rest of the cabinet volume with a port below the second woofer.

How does this sound? Would it be close enough to his design to use a similar crossover?

Driver cost would be around $500 with shipping. As long as the crossover wouldn't be over $350, I'd be in good shape. Would an active crossover be an even better idea?

[ThomasW]

There is a center channel design for the Arvo. I haven't built mine yet, so nothing has been posted.

One thing you should realize that the drivers for the current version of the Arvo aren't shielded.

I strongly suggest you consider the Modula MTM then get a low buck active crossover and build a bass bin using a couple of 10" or 12"s. This design is shielded and has a center channel.

BTW, another reason for using standard box construction (over tubes) is that when drivers are sharing a box, like in a MTM, the performance is improved over them being mounted in individual enclosures.

[Brian Bunge]

Well, porting dual 270's gets big fast. I calculate 200L tuned to 28Hz. My cabinets are roughly 125L sealed. Based on the XO design that Chris came up with, he spent around $320 for parts.

I'm not quite sure what your fixation is with tubes unless you're not going to build boxed enclosures at all. Otherwise, it just seems pointless to me.

[SteveCallas]

There is a center channel design for the Arvo. I haven't built mine yet, so nothing has been posted

It's not that there isn't a center design, it's that with a center dipole, there is going to be something right behind it - and a dipole needs open space behind it doesn't it?

I strongly suggest you consider the Modula MTM then get a low buck active crossover and build a bass bin using a couple of 10" or 12"s. This design is shielded and has a center channel.

Would I be able to use this active crossover for the center as well with two RS270s outside of the mids?

BTW, another reason for using standard box construction (over tubes) is that when drivers are sharing a box, like in a MTM, the performance is improved over them being mounted in individual enclosures

How so? I would be really interested in learning how/why. I really haven't been pointed to any reading I can do to learn more about this subject. I truly appreciate the help and opinions, and I understand it is probably best to go with a proven design, but is there no room for fresh eyes or creativity?

Well, porting dual 270's gets big fast. I calculate 200L tuned to 28Hz

Size really won't be a problem. One of the big benefits of tubes in this instance is the greatly reduced weight in regards to reduced bracing and not having to make double thick walls. Again, drivers will essentially be floating in air, not physically coupled to the outer enclosure. Until I can figure out any ways in which tubes would be a BAD idea, I see no reason not to go for it.

[SteveCallas]

Ok I made up two drawing real quick. The first one shows just how little bracing would be needed. Add the outer shell and I'm done. The second one shows a sample of how I could mix and match the front baffles. All I would have to do is use a couple circle pieces of MDF the size of the tubes for the inside of the caps, then on the outer layer of end cap, it can be whatever shape I want it to be. I'm not saying this is the shape I would want, just an example if what I could do very easily. This would be a lot harder with a typical box design. After all, is it ideal for all drivers to be mounted in the same plane? Don't you want the woofers a bit closer and the tweeter a little further away?

Threw in a quick mockup of what the center layout would be.

[ThomasW]

and a dipole needs open space behind it doesn't it?

No more than any other speaker.

Would I be able to use this active crossover for the center as well with two RS270s outside of the mids?

No the passive XO has EQ (baffle step compensation) that can't be duplicated with a generic active XO

How so? I would be really interested in learning how/why...... but is there no room for fresh eyes or creativity?

When drivers see a larger box they work better. So a shared box for MTM's works better than individual enclosures.

Creativity? Speaker design is hardcore science and engineering. And generally speaking newbie ideas don't work, because they don't understanding the physics of what's happening.

After all, is it ideal for all drivers to be mounted in the same plane? Don't you want the woofers a bit closer and the tweeter a little further away?

Not in this case because that is compensated for in the passive crossover design

Threw in a quick mock up of what the center layout would be.

Your verbal description was adequate to the task. And IMO you gain nothing by doing this.

What you need to understand is that proper multi-way fullrange loudspeaker design involves the application of some very sophisticated science. And the designs Jon's posted here represent the experience of someone who's been designing speaker systems for more than 35+yrs, and is a twice published member of the AES. Basically what this means is that there are few people on the planet that have his level of knowledge on this topic....:wink:

[SteveCallas]

and a dipole needs open space behind it doesn't it?

No more than any other speaker.

Doesn't the dipole design eminate sound from the rear as well as the front? With an entertainment stand and other things directly behind it, wouldn't I be getting all kinds of reflections that would be avoided with either a selaed or front ported center speaker?

What you need to understand is that proper multi-way fullrange loudspeaker design involves the application of some very sophisticated science. And the designs Jon's posted here represent the experience of someone who's been designing speaker systems for more than 35+yrs, and is a twice published member of the AES. Basically what this means is that there are few people on the planet that have his level of knowledge on this topic....

I don't doubt that one bit, and truly I am thankful for people like him and yourself sharing your knowledge in this area. If the tone of my writing suggests otherwise, believe me, that's not the intent.

What I've gotten so far is do it our way or it will be bad - and really I don't mind that at all, but a big part of my original post was for me to also be pointed in the right direction to find information on the subject and learn some of these things for myself. Would I end up with great speakers if I just did it your way? Most likely. Would I learn anything from the experience? Probably not.

[jdybnis]

Well I'm not an expert but maybe I can comment constructively anyways.


Tweeter - most tweeters are sealed so don't need an enclosure. The cheaper ones (<$60) are often sealed with plastic. Those can benefit from being isolated from the midrange backwave. A small box less than 1cm bigger than the tweeter works well. No need to worry about standing waves there.

Woofers - standing waves aren't a problem at bass frequencies in a speaker enclosure, but stiffness matters. A sonotube could make a stiff lightweight enclosure. But keep in mind that one reason speakers are heavy is that they need the mass to counteract the moving mass of the cone. As Newton said, "every action has an equal and opposite reaction". Take that into consideration.

Midrange - This is where you hit problems with standing waves. With a cylindrical tube you will have strong standing waves develop at the wavelength twice the diameter of the tube, and then at every integral multiple above that frequency. Boxes make it easier to eliminate standing waves using stuffing inside the box.

[Brian Bunge]

Just did a quick calculation and if you wanna go ported you better use some BIG tubes. I calculated for a 14" internal diameter tube (since my cabs are 14"W) and determined that you'd need a 39" long tube per 10" driver, not including the volume taken up by the port or driver. That shows 2 disadvantages to using tubes. For a given internal width, a rectangular enclosure is going to have more internal volume than a tube. The other disadvantage is that you can't use the volume left over behind a mid-enclosure since they'll all be in separate tubes.

[ThomasW]

Doesn't the dipole design eminate sound from the rear as well as the front

No, there's a null at the sides of the baffle where the front and rear waves meet.

Would I learn anything from the experience? Probably not.

That's a strange comment. A few year back Jon and I each decided to 'klone' a design from a famous mfgr. Jon chose the Wilson X1-SLAMM I chose the Legacy Whisper. Both of us learned a lot from copying someone elses design.

I guess the question should be, do you want to scratch build an original design and suffer the trials, tribulations, and basic failures of a first time builder? No one is going to walk you through this step by step, because it's way too difficult and complicated to do. And most often first time designs end up in the trash, because the builder simply lacks the skill, tools and knowledge to do the job the properly.

Understand that designing/engineering and building a fullrange multi-way loudspeaker, is orders of magnitude more complicated and difficult than building a tube sub....

[SteveCallas]

With a cylindrical tube you will have strong standing waves develop at the wavelength twice the diameter of the tube, and then at every integral multiple above that frequency. Boxes make it easier to eliminate standing waves using stuffing inside the box

Ok but who said it's against the rules to stuff the tube?

The other disadvantage is that you can't use the volume left over behind a mid-enclosure since they'll all be in separate tubes

I had originally thought the benefits would have outweighed this disadvantage, but aparantly not. I guess it seems everybody is against the tube idea, I'll probably just let it die then.

No, there's a null at the sides of the baffle where the front and rear waves meet.

Wouldn't this null happen as a result of only the initial sound waves coming from the front and rear of the drivers that meet at the baffle edge, not the reflections? There would be almost no way to predict how the reflections of sound on objects behind the rear drivers would interact with the waves coming off the front of the drivers and when they would reach your ears. Is it assumed that the energy in the rear sound waves will have died down enough after reflecting a few times so that if it does cross that baffle it is largely inaudible?

Understand that designing/engineering and building a fullrange multi-way loudspeaker, is orders of magnitude more complicated and difficult than building a tube sub....

I have realized and expressed this from the start. :T

[cjd]

What I've gotten so far is do it our way or it will be bad - and really I don't mind that at all, but a big part of my original post was for me to also be pointed in the right direction to find information on the subject and learn some of these things for myself.

I've not noticed this though I suppose it can be read that way. I've simply noticed your attachment with your idea, and a few of us pointing out that it would be workable but that the things you think are advantages are, in fact, not really so, and it would introduce some design restrictions that might actually leave you with a less desireable end result (baffle layout restrictions being foremost in my mind). You also get insanely deep cabinets due to the lack of the volume from the "corners" (were you to use a cylinder the same width as a cube).

Also, the RS270's are best suited to a sealed enclosure. It's not just a volume thing, it is also the nature of the driver's performance characteristics. Unibox is usually pretty good in picking a recommended enclosure for what that's worth. In this case, you could go deep, but you would run out of excursion darned fast. Similar reasons the RS270 isn't really well suited to dipole: It has the potential frequency extension but can't manage it at a decent output level. I remember Monte commenting that the RS270's were a limitation in his dipole (pic posted in this thread). But I really am not sure why you're considering ported given your sub situation. Sealed RS270's will cross *very* well at 50-60Hz to a sub.

I think I commented on this before, but really grasping all the different things to consider such that you understand why and can make choices beyond taking a stab in the dark takes time. The amount of time we've been at this discussion is definitely a bit on the short side. If I were teaching, I'm not sure we have covered more than half an hour of class including some discussion points. A bullet point or two on my teaching notes. Though it's been a little "all over" at times.

So, keep askin questions, keep challenging, sit back and enjoy the ride.

It may help to tackle only one or two things at a time and research them. Say, baffle diffraction and baffle step.

C

[JonMarsh]

Understand that designing/engineering and building a fullrange multi-way loudspeaker, is orders of magnitude more complicated and difficult than building a tube sub....

Not that it can't be a lot of fun, but it WILL be a lot of work, especially with a steep learning curve if you're not an EE or physicist.

For starters, read everything on this site related to speaker design and construction. It's really unfortunate that Madisound and Parts Express Tech Talk use software that's so unfriendly to searching and accessing archives, or I'd suggest doing that, too. That's a lot of reading, but it's a lot faster and cheaper way to get your "education" than building and discarding your first design, especially a complicated project.

For doing an initial design, it's widely recommended that you start with a simple project, i.e., a two way in a box, and get the hang of doing good measurements, working with design software, getting some experience with power tools, and designing and actually building the crossovers, then measuring and tweaking the combo after it's assembled (crossovers outside the box).

In other words, we're just trying to recommend you learn to walk well before you try running a Marathon.

Just for fun, I'm going to try to remember all the HiFi speaker systems I've built over the last 3-1/2 decades--- and keep in mind that I just designed my best ever two way systems (IMO) the beginning of this year...

1. High efficiency monitor, built with Atltec 451H in folded horn, 511B horn, JBL compression driver, crossed at 800 Hz, not the usual 500 Hz for 511B horn becuase of the ripples in bottom end of the horn response.

2. 12" three way system, sealed, CTS woofer, Fostex midrange, Long Engineering dome tweeter, given to girl friend I was living with.

3. AR LST Klone #1

4. AR LST Klone #2

5. Transmission line speaker using long throw CTS 10" driver, Foxtex Midrange, Long Engineering dome tweeter. Over 5 feet tall, with folded line, fiberglass damped. Very strong output to 25 Hz, but took a Phase Linear 700 to really bring out the most of it (a Phase 400 wasn't quite up to the job).

6. TL speaker based on Philips woofer, Celestion dome midrange, Long Engineering dome tweeter.

7. Dual 12" sub woofer for Magneplanr MG2.

Started using B&K mic and HP mic preamp for measurements - all other design aspects by hand.

8. Three way sealed system using B&W DM6 components- woofer, midrange, tweeter, but with custom designed crossover. Mostly time aligned, but not evaluated for impulse response.

9. Three way system for apt with Audax tweeter, Vifa mid, Audax woofer; stolen about 1 month after I built it... 1977

10. Time aligned 3 way using Audax 8" woofer, MB dome tweeter, and I dont' remember brand of dome midrange- would reproduce 1 msec pulse with good waveform fidelity. Yep, time aligned transient perfect speakers in the late 70's.

11. Time aligned 4 way using 10" CTS woofer, 6-1/2" Audax midwoofer, dome midrange, and MB 3/4" hard plastic dome tweeter. Reproduced very high fidelity 1 msec pulses. 1978-79.

12. Slot loaded subwoofer design in ultra compat box using Audax 15" HiFi Pro woofer (Qts 0.18, in box, 0.55)

13. "Black Monoliths" -Reference design with 2" thick cabinet walls with celotex constrained layer damping, faceted baffle design, dual 10" woofers, 6-1/2" midrange, dome mid, MB 3/4 dome tweeter. Donated to Chas Hansen, who influenced by it's design, later founded Avalon Acoustics (now he's CEO and principle engineer at Ayre Acoustics).



At this time is when I started using very early versions of AutoCAD for drafiting and cabinet design, and SPICE for modeling electrical impedance of drivers and simulating crossover network performance. PWL data files used to input measured driver performance, and combined with filter response.



14. Modula 1 System, woodstyle sub cabinets with 13" Seas, stacked custom Woodstyle narrow enclosures with Seas 6-1/2" to implment line array with MB dome mids, Panasonic Leaf tweeters in solid oak frame (ca. 1984). Yep, line arrays 20 years ago.

15. DAS Mk2 Seas 8" woofer, Pro Audax mid, Panasonic Leaf (2)

16. "AudioWorx" series CM-1 12" four way in Woodstyle cabinet, 6-1/2" mid, 1.25" Audax midtweeter, Seas 225 super tweeter.

17. "AudioWorx" CM-2 8" three way in Woodstyle cabinet.

18. "AudioWorx" 6-1/2" CM-3 two way, Seas components, dome tweeter

19. JKA series- P21REX, Audax dome mid, Panasonic leaf ribbons

20. JKA 1 Surround speaker using Seas 6-1/2" midwoofer and dual Panasonic leaf tweeters

21. JKA Monitor 3, Seas 13', Modula 2, Seas 13", 6-1/2, Audax dome mid, four Panasonic leaf ribbon tweeters. (1985)

22. DAS Model 4 Sat/Monitor 2.5 way TMM with Seas 5" PP midwoofer, Seas tweeter, custom cabinet, aluminum front panel.

23. "AudioWorx" M-10, Woodstyle cabinet, Gold sound 10", two Seas 5" midwoofers, MB Quart titanium tweeter (MTMW).



At this time I started using DOS based CACD for optimized crossover design, and MathCAD model for boundary loading conditions in room.




24. JKA 2M6, MTM, Woodstyle box, dual 6-1/2" Seas, MB Titanium tweeter

25. Custom TS, 4 way D'Apollito (Dunlavy clone); ScanSpeak 10's, SS Kevlar 7", MB domes mid, MB dome tweeter.



At this time I started using an early version of SoundEasy, and acquired CLIO DOS.



26. "AudioWorx" Modula 3.0 - WMTMW in Woodstyle tower box, Seas 8" woofers, MB 2" dome mid, MB Titanium dome tweeter

27. "AudioWorx" Modula 3.1 - WMTMW In woodstyle tower box, ScanSpeak 8" woofers, same dome mids and tweeters.

28. "AudioWorx" SS two way, 7" Kevlar Scanspeak, Seas H400

29. Legacy Whisper Klones (first dipole) - Pioneer 15" woofers, Eton 370, Focal TC120dx2

30. Wilson X1 SLAMM Klones (combo passive and active crossover) Audax Pro 13" and 15" woofers, Eton 370's, Focal Tc120dx2



During the development of the X1 SLAMM Klones and whispers I transistioned from SoundEasy to LspCAD. Acquired CLIOWIN, but too buggy and incomplete in version 4.


31. MTM in woodstyle box using Eton 370 and Focal TC120dx2

32. Legacy center channel clone (similar to Whisper) dipole

33 Sealed WMTW Center channel, Eton 8-800, MB dome mid, MB Titanium tweeter. Custom one off.

34. AS-15 Aerial Stryke 15 Sub (ported HE-15 tuned to 18 Hz, "box" sub).

35. M8 Mk1 - Focal 8" woofer, MB titanium tweeter; tone target was Avalon Eclipse - first cauer-elliptic crossover

36. M8 MkII - Eton 8-800 midwoofer, Accuton C23-6, CE network

37. M8 MkIII - HiVi M8a, Focal Tc120dx2, CE network

38. M8 MkIV - HiVi M8a, Vifa XT25, CE network

39. M8 MTM - HiVi M8a, Vifa XT25

40. Arvo Part MkI - Dayton Titanix MkII woofers, dual HiVi M8a, Vifa XT25 - all passive dipole

41. Elaine Marie MT, HiVi D6.8, Vifa XT in PE enclosure, quasi series network - designed with SPICE simulation, not possible to model in LspCAD5

42. M8 MkIV.1 - HiVi Mi8a, ScanSpeak SS98000, CE network

43. M8ta - Mk1 HiVi 8" woofer, Hales Transcendence tweeter CE network


Transistioned to LspCAD 6, CLIO 6 WIN.



44. M8ta - MkII Dayton RS225 Woofer, Seas Excel Millinium or Seas 27TDFC or Seas H1212 CE network

45. Elaine Marie MkII, Peerless 850439, Seas 27TDFC, quasi series network

46. Elaine Marie MkIII, Dayton RS180, RS28a, quasi series network.

45. Arvo Part Mk II TC2+, dual HiVi M8a, SS98000, all passive dipole

47. Modula Mk IV - MTM with RS180's, Seas 27TDFC or H1212

48. Modula Mk IV.1 - MTM, Dayton RS180's, Dayton RS28a,

49. Natalie P MTM, Dayton RS180, Dayton RS28a tweeter, PE enclosure, quasi series network.

50. Xover Design for DAWARO's TMWW Center channel (Dual RS225, RS150, RS28a)




A tip of the hat to ThomasW for helping me with the timeline and documentation archives for names and model numbers.

Seriously, we all wish you the best of success in your upcoming project, we're just trying to give you some background information and guidance and to help you steer clear of the shoals we've already visited.

PS, my recommendation for the money you want to spend (less than $1200) would be the RS28a version of the Modula's or the H1212, with the capacitor upgrades to the tweeter network. With a good crossover to your subs, would work quite well, and be esthetically OK, not too big, either. Any of the more ambitious things you're talking about will set you back more money, IMO, to execute properly.

~Jon

[SteveCallas]

Ok first off thanks to everyone for replying. I guess I did come off pretty pushy. I am aware that I have about 1% of the knowledge and experience most of you guys do in this area, and again I apprecaite the time you guys take to share your info and designs. I'll just keep my eyes and ears open and go along for the ride.

I never really said I needed to stay below $1200, that was just what I figured my initia idea would have cost. My budget is around the $1500 mark (not too big a difference). I don't doubt that the Modula's are a superb speaker, but in my mind, if I'm going to go DIY, I want to go for something more. Are a pair of Arvos able to be built for around $1500? I'm guessing something like that would be my best option so far, though I'm still unsure how a matching center would play out in my setup.

[ThomasW]

I'll add that the first summer Jon and I met, we built a new loudspeaker design every weekend for the entire summer. Those designs weren't documented, and parts were recycled from one project to the next......

Also there have been a 1/2 dozen or more one-off full blown custom designs for individuals. Unfortunately we've lost or misplaced most of the documentation for those designs.

Finally, not included in any list, are all the multiple prototype line arrays built and tested. The data from those will eventually lead to the 'Saint-Saens' design......