Comments/guidance invited: TMMMM sealed box

[tcpip]

I'm trying to build a good quality sealed-box speaker which should be full-range, go up to reasonably high SPLs, and will be used for 2-ch music listening. I intend to use active xo and LT, and relatively inexpensive drivers. I am using a certain 6.5" poly-cone midbass unit made in India, and a dome tweeter (as yet undecided), xo'd at 2.5KHz. I want to use a sealed enclosure plus LT to get decent bass extension, but with one midbass unit, I'll run out of excursion. Hence I want to use four midbass units (the drivers are about USD 10.00 each, so I can afford them).

I am therefore thinking in terms of a TMMMM design. Any tweeter will do, provided it has an fs of about 1K or so and is reasonably flat and well-behaved above that. I'm using LR4 (electrical) active xo.

The background for this xo and this midbass unit has been based on a simple sealed two-way built by my friend, who used the same midbass driver, the electrical LR4 and LT for his unit. I'm using a different tweeter, but I don't see much fundamental issues for the overall design if I replace the tweeter. My friend built the speaker (a small two-way TM configuration tower) and listened to it, measured its response etc, and found that it sounds very clean, has very good crisp mids, and measures flat to about 30Hz, provided one can keep the SPL low. I have heard his speakers, and they sound very clean, detailed and musical (as in "not analytical or hyper-detailed"). I am replicating his design idea, just adding more midbass drivers.

Any comments/guidance/suggestions?

In another thread, where I had broached this idea, I had got some very useful tips from Sreten and others. In particular, Sreten had said that I should split the four midbass drivers into two pairs, and use one pair "full-range" from the bottom till the tweeter xo, and the other pair from the bottom to the BSC knee, after which I should roll them off using a first-order slope. I think that makes sense, and I'll try that.

Any other comments? I was wondering what would happen regarding comb filtering when all these M's interact with the T. (If I use Sreten's BSC idea, two of the M's will be rolled off and cut down quite a bit by the time I reach 2.5KHz, but they'll still be audible.) One idea was to use an MTMMM, where the two top M's are used full-range till 2.5KHz, while the bottom two M's are used for BSC.

On the other hand, there are speakers like the Straight 8 which seem to have eight midbass units in a row, interacting with the tweeter. What about comb filtering there?

In case you want to know, I'll be using solid state amplification for these speakers. I intend to use one amp per driver (five amps per enclosure), and the first version will use non-inv LM3875 amps.

Any comments would be useful. This is my first major speaker building project, and while I'm building on the good groundwork of my friend's design, I'm still making changes big enough to need help.

Thanks for any inputs.

 

[sreten]

Well how active is active ?

The obvious choice is using one driver for mids only,
and applying the LT to to other three bass units only.


sreten.

 

[tcpip]

The obvious choice is using one driver for mids only,
and applying the LT to to other three bass units only.

You're basically suggesting that I go three-way, not two-way at all, right? One advantage with that will be that the TM interference will be about as low as one can hope for, as clean as any two-way. The disadvantage will be that I'll lose out one one driver for driving the SPL at low freq. What do you think?

 

[sreten]

If your going active, BSC is electrically very easy, and there
is no need to maintain sensitivity matching of driver arrays.

Even five drivers is possible :

An MT mounted on the front, and 4 drivers in series/parallel
with two drivers mounted in each side in force cancellation
mode, i.e. the magnets physically linked and decoupled
mounting of the drivers.

For the above a full 3-way design is suggested, the default
response with 1st order bass / mid crossover will be very
near to including BSC.

sreten.

 

[tcpip]

If your going active, BSC is electrically very easy, and there
is no need to maintain sensitivity matching of driver arrays.

Nice to know that something is very easy for me. Sigh...

An MT mounted on the front, and 4 drivers in series/parallel
with two drivers mounted in each side in force cancellation
mode, i.e. the magnets physically linked and decoupled
mounting of the drivers.

I understood the bit about the front-facing MT... that part's just a simple two-way, probably reaching down till a few hundred Hz. But below that, the four side-firing drivers will be in push-push mode, if I've understood you right. Have I?

In that case, I'll have to bring those four drivers in at a frequency high enough to take care of BSC, don't I? This means they must reach upto at least about 400Hz or more. But at those frequencies, sound is very directional, and won't it be better ("crisper" sound and the like) if they fire forward? I'm no expert, just asking...

For the above a full 3-way design is suggested, the default
response with 1st order bass / mid crossover will be very
near to including BSC.

I don't think I understood what you mean by 1st order xo for bass-to-mid. Why should I do a 1st order? If I do a first-order, won't the side-firing speakers reach quite high and won't this muddy the midrange or something? Basically, I'm quite unclear about what happens if different speakers are firing in different directions in the mid or high ranges.

Maybe one answer to my worry about high mid-to-bass xo is to make the front baffle wide ... maybe 18" wide. Then the BSC knee can be lower, and the muddying of the midrange, if such a concern is legit, will be less of an issue.

The more I think about it, however, the more I feel that I'd probably like to build a two-way, if it can be made to work. A three-way is not totally ruled out, but I'd just like the simpler active xo of a two-way if possible.

And if I do shift to a three-way, then left to myself I'll probably go for a TMWWW, where the M is pure midrange, and all drivers are front-firing, as you'd suggested earlier. In that case, should I xo the mid-to-woofer at the BSC frequency, or should I xo at any frequency (something lower like 200Hz) and do a shelving filter on the midrange to take care of the BSC? I hope my thoughts are decipherable... if not, I'll elaborate later if you want.

 

[sreten]

For an active 3-way its best to c/o where you'd ideally like
and add a BSC shelving circuit before the bass / mid filters.

For an active two way a MTMMM with a passive 0.5 way is best.

sreten.

 

[tcpip]

For an active 3-way its best to c/o where you'd ideally like
and add a BSC shelving circuit before the bass / mid filters.

For an active two way a MTMMM with a passive 0.5 way is best.

Thanks for the patience. I'm getting the drift.

I was trying to understand why you suggest an MTMMM specifically for a two-way. I guess it's because if I do a three-way with the MTMMM, there will be only two drivers handling all the lower frequencies, which will be too little drive, and two whole drivers doing just the mids, which will be too much drive. Right? I was getting around to that view myself.

I'm now becoming clearer that for the first version of my speaker, I'll go with a two-way and keep it simple. Maybe I'll do an MTMMM, maybe a TMMMM, I don't yet know. I'd have done an MTM.. but I've read some negative remarks from others on this forum about how MTM didn't sound good and TM was much better. That's sort of put me on guard. Is there anything I should be careful about if I go MTM?

And for the BSC, I may put a passive line-level shelving xo in the path of the low-pass output of the xo. Or maybe even in the input to the xo... how does it matter? With a passive line-level xo, the circuit will be simple. Important issue for a beginner like me.

One more thought about the two-way MTMMM option. Suppose I went for this, and did a 0.5 at 500Hz, with the tweeter xo at 2.5KHz. In that case, how would it be if I did a first-order at 500Hz, and then a sixth-order or eighth-order Bessel at maybe 1.2KHz or so? That way, the two lower drivers would be all but inaudible at frequencies where the tweeter kicks in, thus totally cutting out chances of interference and combing etc. Is it worth the extra trouble? Any ideas?

 

[sreten]

Just use a LR2 or LR4 at normal crossover frequency,
as its active this will apply to all bass units.

Inteference from the passive 0.5 units would not be an issue.
Do not go for 4 speaker drive with active BSC, its pointless.

sreten.

 

[tcpip]

Just use a LR2 or LR4 at normal crossover frequency,
as its active this will apply to all bass units.

Okay.

Inteference from the passive 0.5 units would not be an issue.

Okay here too. This lays one of my questions to rest.

Do not go for 4 speaker drive with active BSC, its pointless.

Sorry, couldn't understand this one. What is the connection between passive versus active BSC and the number of drivers? When does active BSC become important?

And thanks again for all the help.

 

[Variac]

Hey Mr. Pip,

I am intrigued by your proposal. It seems to combine a lot of good ideas
floating around the forum these days.
First of all, the idea of "one driver, one chip amp" needs to be used a LOT more, IMHO.

The "pros" seem to agree that a high power amp of over 100 watts/ch is a different beast
than a lower power one, with much more serious consequenses
when things go terribly terribly wrong
So, by using a bunch of chip amps you have the equivalent of a couple hundred watts/ch
with few of the risks AND great sound quality. Making them with active crossovers,increases the efficiency so you are ahead
there too. Stir in the multiple mid/basss drivers, and you are not going to be deficient in power!!!
I think Seigfreid Linkwitz and Nelson Pass would agree as both their speakers use multiple amps.

Now a tower with multiple drivers has a lot going for it too. If it is tall and narrow,
it is appealing in a SAF way, while if it is deep enough for the internal volume required,
people can't place it so the drivers are too close to the wall
(this is the only downside of the push/ push approach)

Thinking about the Straight 8, I think that that speaker is closer to a real line source in a
typical US 8' ceilinged room compared to a 4 mid/bass speaker as you propose, so I wouldn't count
on yours working the same.

I agree vigorously with sretin that making the lower 2 drivers pretty much deal
with the lower bass only and roll off at Baffle step, is a great way to go. That way they compensate for
the baffle step and also give you excursion where you need it- lower bass- especially when you EQ to
get more bass.

I agree with you that a 2 way is much more appealing- especially if you have one to copy and
haven't designed many speakers. Many times I have preferred well designed 2-ways over 3 ways (probably because
designing the 3 ways is so much harder!!

I have to say that then having 2 drivers handling midrange as well as lower worries me a bit.
I don't know, but somehow feel that it woiuld be much better to have only one driver doing the midrange
OR!!!!!: have a MTM setup on top which kinda is the same thing!!

So I'd think either have 3 drivers hooked up for bass and one full range and the tweeter
OR 2 bass and the MTM ie MTM+MM. check out the photo link below for a speaker like this.


My vote? MTM+MM 2-way. Start rolling off the lower 2 drivers at baffle step.
For awhile I was thinking that it might be good to roll off the bass a bit on the drivers that also handle midrange, but on further thought, I think all those drivers sharing the excursion for bass will
keep down IM/Doppler distortion- especially at normal volumes!

Yes, there is a backlash against MTM it seems, but many, many, very high end models use that configuration,
In fact I'd say it is about the most popular for very high end models from well known manufacturers. Not to mentiond The Ariel by Lynn Olsen. When I discuss this with people that have misgivings about MTM they generally have a theorectical concern that it should have some downside, but generally admit that the can sound very very good, in fact say they "think " they prefer a single driver.


Check out the "Reference Speaker" thread for drivers and speaker designs that seem to be relevent to your design
in some ways. People seem to like the Vifa d 27 tweet and various 5.5 to 6.5 mid/bass drivers.
Also good info there on dealing with baffle step.


Planet 10 Dave has posted a pic on this page that is similar to what you are proposinf I think, Post #129

http://www.diyaudio.com/forums/showthread.php?s=&threadid=25590&perpage=15&pagenumber=9

This is an exciting project- especially since it appears that it is hard to find good stuff easily in India.
Perhaps one of us here can make a similar speaker with some of the drivers mentioned in the reference speaker thread

 

[sreten]

And for the BSC, I may put a passive line-level shelving xo in the path of the low-pass output of the xo. Or maybe even in the input to the xo... how does it matter? With a passive line-level xo, the circuit will be simple. Important issue for a beginner like me.

Do not go for 4 speaker drive with active BSC, its pointless.
sreten.

sreten.

 

[Variac]

It would be more helpful to explain why it is pointless, rather than just repeating the same statement. Do you mean to stick with passive? do you mean that 4 drivers would be so bad that
there is no point to bothering with BSC?

 

[Variac]

OK , I've been thinking and thinking......

Here is your proposal I think:

sealed 4 mid/woofers, one tweeter.
Individual chip amps for each driver,
active crossovers- but maybe not on BSC
active EQ on low bass.
Use the same drivers and tweet/mid crossover as your friend

Issues:

Sealed boxes have more low end excursion than vented
Eq on the low end increases the excursion even more!!!
You have lots of drivers to spread this out
BUT we are worried that even with 4 mid/bass drivers, you will have enough excursion to increase the IM distortion through the midrange.

baffle step compensation is probably required.

SO:
I was paying attention to Sretin and I have an idea - maybe what he was alluding to, maybe not . Maybe a good idea -maybe not....

I think we all feel that having a single driver for the mids would more likely result in good sound- a single source- no comb effects,
not MTM which some don't like, AND it doesn't sound harsh because the cone isn't moving so much trying to do bass.

My idea is to use ONE of the four identical mid/bass drivers as a midrange in a 3-way setup. You have the crossover between the tweeter and the one "mid" driver worked out already- so no work there...

THEN you determine the Baffle step frequency. Then you do another crossover at that frequency. NOW you have a 3-way and all the hassles that USUALLY implies, BUT: active crossovers, so no interaction between the crossovers- things are getting easier.

THIS is what I just realized: The mid driver is the same as the bass drivers, so has the SAME frequency response. Theoretically, you could cross then anywhere and they would match up perfectly. You could use first order filters and they would match up perfectly. Sure they would run in tandem for a large amount but they sound the same, so what does it matter? Of course the mid doesn't have bass EQ either, so it has the same tone and response as the other drivers, but isn't doing bass so sounds sweeeet and unstressed. First order isn't required, but it looks like this MIGHT be one of the few times it might work out fine- with the advantages of first order crossovers

The three remaining are woofers. This might work out well:
From what I have read baffle step correction is usually betweeen 3 and 6 dB. 6 is usually too much-especially in small rooms.

Having 2 bass drivers would increase the bass 3dB over one driver. 3 bass drivers would probably be about 4dB over one driver.
SO..... if you cross all 3 to the mid at the baffle step, you get effectively about a 4dB baffle step correction- probably about right. No worries about the inpedence of three drives-they are independent with separate amps.

If the bass is too much around the baffle step, then turn down the gain in the amps. If the woofers bottom out on low notes, reduce the EQ or add another driver-then turn down the gain on the woofers.

OK, now everyone can shoot me down.....

 

[tcpip]

THEN you determine the Baffle step frequency. Then you do another crossover at that frequency. NOW you have a 3-way and all the hassles that USUALLY implies, BUT: active crossovers, so no interaction between the crossovers- things are getting easier.

Yes, I too was thinking along those lines. Sreten had also suggested it as you can see in the quote below:

The obvious choice is using one driver for mids only,
and applying the LT to to other three bass units only.

But when I asked him about the feasibility of putting the xo freq at the BSC freq, he didn't seem to think it necessary.

For an active 3-way its best to c/o where you'd ideally like
and add a BSC shelving circuit before the bass / mid filters.

I too initially liked the idea of the xo at BSC frequency, so that you don't have to do a separate BSC circuit, but I am not sure I want to do it that way now. Multiple reasons:

• To keep interferences low and signals clean, steeper xo slopes might be useful. It'll certainly help eliminate interference from woofers at the tweeter xo (2.5KHz). But if I do steeper xo slopes, e.g. my beloved LR4, then that may not be right for baffle step, which needs a very gentle slope by all accounts. Even 3dB/oct may be a good idea for BSC.
• If I go three way, why not let the woofer array play below a lowish xo, say 200Hz, and keep the entire so-called "voice range" in just one midrange? That way, I get point-source coherence in the vertical dispersion pattern for the "voice range", and get the woofer array only around 200Hz or below.

Therefore, if I go the three-way route, I'll probably do an LR4 somewhere between 150-200Hz between woofer array and solitary midrange, and then do a BSC step in the midrange driver's signal. The woofer array, xo'd at 200Hz, won't need a separate BSC filter... it'll just need to be level matched appropriately with the midrange.

I too feel that I may get somewhat cleaner, low-distortion midrange if I keep one driver dedicated to it, and handle the lows elsewhere. In fact, now with all the talk of three ways, I'm tempted to try an OB mid+tweeter, and a sealed box woofer array. Basically, it'll be exactly the same three-way we've been discussing so far, except that I'll make the midrange OB and make the baffle wider. Looking at it from the front, no one will know that the woofer array is actually in a sealed box, since that box will be hidden by the wide baffle. Maybe baffle widths of 18" will be adequate for an xo of 200Hz? But I won't try this idea right now... Later, later.

 

[sreten]

It would be more helpful to explain why it is pointless, rather
than just repeating the same statement. Do you mean to stick
with passive? do you mean that 4 drivers would be so bad that
there is no point to bothering with BSC?

The point for an active 2 way had already been covered in detail
in the previous thread, and I didn't want to simply repeat that.

sreten.

 

[Variac]

I'm tempted to try an OB mid+tweeter, and a sealed box woofer array. . Maybe baffle widths of 18" will be adequate for an xo of 200Hz? But I won't try this idea right now... Later, later.

Hooo boy -great minds work alike. I had considered mentioning this, but thought that it might be best to not mention it at this stage. Er, maybe we aren't great minds, but that this is the only trendy DIY Audio idea we haven't incorporated into your design!!
Also, 18" would be considerably wider than what I think you were considering? My basszillas are 20" wide and they are huge

On my Basszilla the open baffle is 18" wide and I believe that the 8" driver starts to roll off about 180- 190 hz. Not sure BUT the intention is as you mentioned: cover
almost the entire midrange with one driver! Keep in mind that the bafle size is effectively much greater because it is on the bass box.

Maybe this is the best idea for a speaker that you will keep for yourself!

Wait!! another idea. Maybe use "wings" on each side going back-
this is generally not considered as good as a flat baffle, but it woiuld still be better than a closed box because part of hte sides, all of the back would be open. With this maybe you could reduce the tower width to about 14"

 

[tcpip]

The point for an active 2 way had already been covered in detail
in the previous thread, and I didn't want to simply repeat that.

Sorry, sreten, for being slow on the uptake, but I didn't quite understand your "pointless" comment in the earlier post. Will you please explain in a bit more detail?

Thanks again.

 

[tcpip]

Hooo boy -great minds work alike. I had considered mentioning this, but thought that it might be best to not mention it at this stage. Er, maybe we aren't great minds, but that this is the only trendy DIY Audio idea we haven't incorporated into your design!!
Also, 18" would be considerably wider than what I think you were considering? My basszillas are 20" wide and they are huge

Yes, I'd certainly not consider 18" for a fully boxed design, but if I was doing part open baffle, then I will almost certainly use a wide baffle. I too was thinking of wings. And all said and done, that Polish speaker designer's work which you'd pointed me to many moons ago also include OB mids. His baffles are not all that wide, are they? He uses short wings, though. I would too. I was thinking of using a pretty ordinary-width central panel (maybe 9") and wings of 6" on either side, at an angle of 45-deg towards the back.

I'll certainly try this once my current sealed-box project is complete and stable. I was toying with the idea of an MTM OB together with three or four midbass drivers in a tall sealed box for the lows. It'll be so tall (six 6.5" midbass drivers + one tweeter) and strange looking, most people won't even pay much attention to how it sounds!

Maybe this is the best idea for a speaker that you will keep for yourself!

Yes, such wide baffles are acceptable only by die-hard speaker lovers or recipients of free gifts.

 

[sreten]

Simply put an active 2 way with active BSC MMTMM or even worse
TMMMM will have comb filtering interference effects easily reduced
by wiring two M drivers as passive 0.5 way.

The optimal configuration is MTMMM, or TMMMM, depending on
c/o phase alignment, but the passive 0.5 way also accounts
for BSC correction, so this is not part of the active crossover.

sreten.

 

[tcpip]

Simply put an active 2 way with active BSC MMTMM or even worse
TMMMM will have comb filtering interference effects easily reduced
by wiring two M drivers as passive 0.5 way.

This far is quite clear to me. Thanks.

The optimal configuration is MTMMM, or TMMMM, depending on
c/o phase alignment, but the passive 0.5 way also accounts
for BSC correction, so this is not part of the active crossover.

I understand that bit about the MTM-MM being optimal. What I still don't understand (sorry!) is why you recommend passive BSC over active. Is this really of consequence? Let me explain one point about my intended approach to BSC. If I do BSC in this multi-midbass scenario, I'll always do it by doing a first-order low-pass on half the midbass units, not by building a shelving filter for the full set of midbass units. And left to myself, I'll do this low-pass-based approach irrespective of whether I actually implement it in active or passive line-level circuits.

Now can you tell me, does active or passive matter? Or am I missing some other easy shortcut in my approach?

Thanks again for the patience.