Comments/guidance invited: TMMMM sealed box

[Paul W]

tcpip,
Yes, the 9900 fs is 500 and that is one of the reasons I picked it. The "trouble" starts in the 7" Rev...if you look at the impedance curve on the DST site, you'll see a little wrinkle around 700-900 Hz. Whatever causes that wrinkle is ever so slightly audible (to me, but not to most people). Also, a slight breakup starts in the teens...so the 7" is absolutely supurb up to 700 and "only" excellent above that. So, I ended up pushing the 9900 down to 1500Hz which is about the limit for it.

The basic problem though is the large room...it is a "great room" of about 14,000 cu ft which means the drivers are pushed hard when we're watching movies. (Staying below xmax doesn't necessarily mean distortion doesn't change.) If I had a smaller dedicated room, this probably wouldn't have been an issue...and it would have been tougher to find an excuse to build more speakers
Paul

 

[Variac]

Don't know much about the crossovers for MTM, but in a conventional speaker, the M drivers are paralled , so are 4 ohms
when combined, which means the crossover needs to be calculated for 4 ohms instead of 8. With your one amp per driver setup I guess this isn't required. Because of how the different crossovers lobe, I think only some work well.

I think D'Apollito used a certain order at first which theorectically was most apropriate (I vaugely remember 3rd order) and later switched to I think 4th order. Maybe going to madisound.com and looking into the Thor kit (if thats the model using Seas drivers in an MTM configuration) for hints

 

[tcpip]

Don't know much about the crossovers for MTM, but in a conventional speaker, the M drivers are paralled , so are 4 ohms
when combined, which means the crossover needs to be calculated for 4 ohms instead of 8. With your one amp per driver setup I guess this isn't required.

Yes, this part is one of the many things of passive xo which becomes irrelevant with active xo and multi-amping. I love multi-amping.

Because of how the different crossovers lobe, I think only some work well.

Yes, I guess...

I think D'Apollito used a certain order at first which theorectically was most apropriate (I vaugely remember 3rd order) and later switched to I think 4th order.

Clearly I need help on this one. More reading and consulting is required. One thing that I do remember about the D'Appolito Config is that the inter-centre distance between tweeter and midbass is a function of the xo frequency... to get the prescribed polar pattern accurately, one has to have precisely the distance governed by the xo wavelength, not more not less. Am I right on this one?

Thanks a lot.

 

[tcpip]

the 7" Rev...if you look at the impedance curve on the DST site, you'll see a little wrinkle around 700-900 Hz.

Man, you must be Batman to hear that. I didn't know anyone bothered with squiggles that small. But if you can hear it, I guess it's a Sign from Above that you need to progress to a 4-way.

The basic problem though is the large room...it is a "great room" of about 14,000 cu ft

Sigh.... Just thinking of your room makes me drool. Our living room, which is quite "okayyy" by Bombay standards, is about 3000 cuft or so.

Interesting insights, thanks.

 

[navin]

I guess there's a third choice: MTMBB. Right?

given your need for SPLs I'd suggest MTBBB or TMBBB. since you are using one amp per driver we dont have to worry about impedances or sens matching.

BTW BSC is about 6db/decade. true audio has a good explanation of it on their site.

Man, you must be Batman to hear that. I didn't know anyone bothered with squiggles that small.

bet. 300 and 3k the ear is very sensitive. and i believe that women have more sensitive hearing than men. in this rage 0.5db is easily noticeable by my wife and my sister. both myself and brother in law use them for listening tests.

 

[tcpip]

Good to see you around, Navin.

given your need for SPLs I'd suggest MTBBB or TMBBB. since you are using one amp per driver we dont have to worry about impedances or sens matching.

Actually, for v1.0, I'm thinking of going with MTMMM, in a 2.5-way config as sreten had discussed earlier. That way, I'll have two drivers to handle mids (plenty SPL), and four to handle the range from BSC on down (again, more than enough). The only things I need to sort out now are the xo requirements for a good MTM.

BTW BSC is about 6db/decade. true audio has a good explanation of it on their site.

Yes, I've seen that one. It's good.

 

[tcpip]

Box dimensions I'm thinking of going with

Here is the design of the box that I'm thinking of going with. My aim here is to do 40L of sealed-box volume behind each pair of midbass drivers. I'm thinking of an MTMMM 2.5-way, if I can get a grip on the MTM xo issues by the time I start cutting the MDF. All dimensions are in cm. This means height is a shade less than four feet, depth is about 29", and front baffle width is 10". Front baffle and bottom are double-thickness.

The rear chamber behind the lower pair of midbass drivers will house the array of five Gainclone amps. The body will be built entirely out of 25mm MDF, and braced with 12mm glass sheets stuck to the inner walls with Araldite. Assuming 2600Kg/m3 as the density of glass, I'm getting 31Kg of glass alone per box. The total box weight will probably be of the order of 60-80Kg.

 

[sreten]

Hmmm.....

The glass idea in intriguing - I'd glue (using contact adhesive)
the glass sections in one sheet to either side - after using a
a roller on the back of the glass with some black paint that
sticks well to the glass, and use a full width front baffle.

Adding internal bracing is probably better than using
glass for beefing up the internals of the louspeaker.

sreten.

 

[navin]

Re: Box dimensions I'm thinking of going with

The only things I need to sort out now are the xo requirements for a good MTM.

it all depends after playing around with MTMs a bit I feel that a MT or FW (fullrange and woofer) might offer better dispersion patterns for the standard bedroom/living room.

Here is the design of the box that I'm thinking of going with. My aim here is to do 40L of sealed-box volume behind each pair of midbass drivers. I'm thinking of an MTMMM 2.5-way, if I can get a grip on the MTM xo issues by the time I start cutting the MDF. All dimensions are in cm. This means height is a shade less than four feet, depth is about 29", and front baffle width is 10". Front baffle and bottom are double-thickness..

good luck. not particularly SAF compliant though. they might end up like my B460 (2245 bass bins).

dont use glass. if you are doing this why not try marble or kadappa. or even tile. it is a lot easier to use. if i were to use tile i'd use ceramic.

this does not mean it will absolve you of bracing.

 

[tcpip]

Re: Re: Box dimensions I'm thinking of going with

it all depends after playing around with MTMs a bit I feel that a MT or FW (fullrange and woofer) might offer better dispersion patterns for the standard bedroom/living room.

It was a choice between using a readily available inexpensive well-understood driver for the midbass and adding drivers to get SPL, versus going looking for a different, probably more expensive driver. I don't know what other good midbass driver is available to Indian DIYers which can be used in just quantities of one. Would you recommend the Vifa P18s Corrson is selling?

good luck. not particularly SAF compliant though. they might end up like my B460 (2245 bass bins).

Actually, I think they'll be quite SAF compliant. Their front baffle width and height are not that much different from many commercial floorstanders. The depth is much more than usual tower speakers, but it'll be largely invisible. And good European-hardwood veneer (knotted pine perhaps?) finished with multiple coats of PU (TouchWood) should enhance SAF considerably, don't you think?

dont use glass. if you are doing this why not try marble or kadappa. or even tile. it is a lot easier to use. if i were to use tile i'd use ceramic.

Can you give me some more tips? How is it easier to use kadappa stone? Glass has been chosen here because I have some sheets of 12mm glass as "scrap", not because it was the ideal to use. If I didn't have the glass, I'd have used slate tiles (one-seventh the cost of 12mm glass sheet). But marble would be at least as bad as glass in terms of resonance properties (metamorphic rock, clean crystalline-type structures, etc), and kadappa would certainly be better, but then why not slate tiles at ten rupees a sq-ft? If I were to use tiles, I'd probably look for terracotta or rough-hewn outdoor stone tiles.

And I'd have thought bracing wouldn't be needed if somethig as heavy as glass was being fixed as inner wall lining.

 

[tcpip]

XO requirements for MTM

Dickason's book, 5/ed, has about 20 lines on the D'Appolito configuration while dealing with 3-rd order Butterworth two-ways. There he says that 3rd order Butterworth was made quite popular with this config, but designers later have shifted to LR4 for MTM, and have found no problems. He discusses the polar pattern and vertical dispersion characteristics of MTM, but does not say that there's any reason to do anything different in the xo in order to use this config.

I guess I'm all set on that front, then. I'll use my LR4 as planned, and just BSL (BSL = Build Something and Listen!)

 

[navin]

MTMs have a diufferent sound field than MTs. Generally they can sound very good only given XO freq. chosen in 6" MTMs it means there is some degree of polar cancellation and hence the impression that they are laid back.

say the woofer is XOed to the tweeter at 2k. the tweeter has a 4" dia face plate. c-c distance between the 2 woofers in a MTM is 110mm+170mm = 280mm = 11"

speed of sound 1180ft/s. 1180*12/11 = 1300Hz. that means polar interfereance starts at 1300hz.

second problem (does not apply to your design) is that MTMs mean that both woofers have to deal with the L-R BSC ckt. I have found that the large Ls used in series tend to rob a driver of it's dynamiocs (even if the L is air core and the DCR < 0.5ohms).

instead of MTM is the drivers were wired MTW where the W is BSC only the W sees the large L in series. sicne the ear is relatively less sensitive below 400hz (typical BSC point) this is easily tolerated.

Unfortunately I have NOT read Vance, Weems etc... in fact I have not read too much. I just played with a lot of drivers in the 70 and 80s so all my opinions are based on my expereinces which might be coloured by the enviroment in which i tested drivers, the music I used, my age, etc....

 

[tcpip]

MTMs have a diufferent sound field than MTs. Generally they can sound very good only given XO freq. chosen in 6" MTMs it means there is some degree of polar cancellation and hence the impression that they are laid back.

I agree that inter-centre distances will determine what is the highest practical xo freq. This is of course independent of whether you use MTM or TM.

second problem (does not apply to your design) is that MTMs mean that both woofers have to deal with the L-R BSC ckt. I have found that the large Ls used in series tend to rob a driver of it's dynamiocs (even if the L is air core and the DCR < 0.5ohms).

Yes.... Isn't it good that I don't have to worry about all these things?

Unfortunately I have NOT read Vance, Weems etc... in fact I have not read too much. I just played with a lot of drivers in the 70 and 80s so all my opinions are based on my expereinces which might be coloured by the enviroment in which i tested drivers, the music I used, my age, etc....

Modesty becomes you.

 

[navin]

I agree that inter-centre distances will determine what is the highest practical xo freq. This is of course independent of whether you use MTM or TM.

TM c-c is 85mm+55mm (half of 170mm and 110mm) = 140mm = 5.5"

1180*12/5.5=2500Hz. this is a lot better for the tweeter.

BTW the most successful design I had using MTM used 2 8" woofers and a 1" in 1987. the slopes were 6db LP and 18db HP.

i tried to keep cognizant of polar interferance, BSC etc and use as few components as possible.

still the sweet spot was very narrow.

 

[tcpip]

TM c-c is 85mm+55mm (half of 170mm and 110mm) = 140mm = 5.5"

1180*12/5.5=2500Hz. this is a lot better for the tweeter.

Thanks for the figures. I now finally understand why people crib about MTM performance... you made it all click into place. Essentially, the M-to-M distances in an MTM may be too large to avoid comb filtering, unless you really have the xo at a very low frequency, which is often not practical.

I'm still hoping that the friend for whom I'm making this will not mind... he may not be the kind who will want to sit precisely in the sweet spot always and pay attention to soundstaging. He may be just happy to get clear, even sound slightly off-axis. Then there's also the inputs from speakerbuilder.net, where the author says that he prefers to build MTMs with drivers slightly further apart. You must've seen this article about his high-end Eros MTM where he explicitly talks about this issue. So, frankly, I'm confused. And nowadays, whenever I'm confused, I fall back upon BSL!

 

[Variac]

Drivers should be spaced on center no more than one wavelength apart at the highest frequency being reproduced, but there are exceptions. The main deviation would be for an MTM configuration; in that case the outboard elements should be spaced no more than one wavelength on center at the crossover frequencywith respect to the axis of the center driver.

If this is true than it is an explaination of Why MTMs work with wider spacing than mentioned in this thread. The guy quoted
says that tthe distance from the midwoofer to the center of the TWEETER needs to be one wavelength. Suddenly 6" drivers would work well- which we know they do- there are MTMs with 6" drivers all over the place!


It is from the "driver spacing" thread going on concurrently:

Post #7

http://www.diyaudio.com/forums/showthread.php?s=&threadid=34943&perpage=15&pagenumber=1

I think that some good books need to be consulted


PS I think when responding to tcpip's design we can assume that he is using all 6" drivers and one tweeter. From what I can see, there is some confusion when people talk about mids and woofers. we throw around terms like MTM when we mean midwoofer/tweeter/midwoofer NOT midrange/tweeter/midrange
It gets confusing.. add to this that he is using so of the identical drivers for bass only, and suddenly clear comments become very confusing..

MArk

 

[navin]

oh BTW tarun (tcpip) i have not used the P18 but i have used the M18. from what i remember the p18 had a high loss surround the m18 had a low loss suround. my knowledge is emperical and so might not be wholely accurate but in my limited experience high loss surrounds tend to produce drivers with flat freq. curves or maybe I should say flatter freq. curves. low loss surrounds tend to produce drivers that are more "dynamic". I prefer the later. it is matter of flavour.

 

[navin]

BTW variac from the same page.....
http://www.speakerbuilder.net/web_files/Projects/Eros Project/eros.htm

The crossover frequency of the final version is around 1750hz and the woofers are rolled off with a 2nd order acoustic slope. The tweeters also use a 4th order acoustic slope. The relatively low crossover frequency permits a wide center to center driver spacing to improve the quality of the soundstage without succumbing to the traditional compression of the soundstage that many MTM designs are plagued with.

 

[tcpip]

I think when responding to tcpip's design we can assume that he is using all 6" drivers and one tweeter.

You're right. I'm using four of these drivers, each with outer rim diameter of 165mm, and a moving cone diameter of (only) about 115mm.

From what I can see, there is some confusion when people talk about mids and woofers. we throw around terms like MTM when we mean midwoofer/tweeter/midwoofer NOT midrange/tweeter/midrange
It gets confusing.. add to this that he is using some of the identical drivers for bass only, and suddenly clear comments become very confusing.

You're right about midrange versus midbass. I'm building my MTM to let the midbass drivers go all the way down to their lowest rolloff, so in my case I guess they're midbass, not midrange.

But tell me, in the preceding discussion about MTM and xo and comb filtering, does it matter whether the designer is doing midrange-tweeter-midrange or midbass-tweet-midbass? It's the upper limits of the midbass rolloff which causes all this comb filtering and decide the inter-centre distances, not the lower-end, right?

 

[Variac]

http://www.zalytron.com

has lots of speaker designs on thir website under the "kits"
menu. Many of them are designed by Joe D'Appolito.

While some of them appear to have been more carefully designed than others, they will give you a good feel for MTM design, because the majority of them are MTM.

The

Usually they have a crossover diagram with crossover frequency.
If not you can probably reverse engineer the freq from the crossover component specs.

I know you aren't using the same drivers, but as I said, it gives you a feel for the approach.....

the Eros that Navin mentions is a good reference too it appears, but I haven't heard it, and it bothers me that the designer can't seem to decide the precise box size

My understanding is that the MTM design is in fact using lobing to control the dispersion. So all lobing isn't bad. BUT I think the fact that the pattern is cased by lobing makes some people uncomfortable from a theorectical point of view.
There oare plen of them that sound great though...

Also, the narrower focus of MTM is claimed by some to help imaging since it cuts off more reflections from the side walls of the room, so maybe as in all speaker design, there is a trade offf? stage width for imaging?