WMTMW Open Baffle project - so far so good

brad944911 · « **on:** 2 Jun 2016, 07:48 pm »

Guys,
I am working on a new project and need some hard facts here; what the heck is the right way to implement the spacing of the mids and tweet in an MTM(D'Appolito) speaker? For the life of me I cannot seem to find the no kidding answer. I've gotten so many points of view that sound good on either side, but what's the truth? I understand the <1 wavelength(someone even suggested <1/4) @ crossover freq spacing part, but between what? Between the two mids center to center or between the mid and tweeter CTC? My brain says between mids CTC, but I'm self taught and may just be completely off! The spacing I've planned is 20.75in between 10's and 6.5in between 4's. So in theory even the 1900Hz crossover to tweets will be OK even if it's mids CTC, but I'd like to take tweets x-over freq up a little higher if possible.
Just for info, the speakers I'm using are: Woofs - MCM 55-2891's (from the original Hestia...sorry Ed) / Mids - Dayton RS100-8's / Tweets - Vifa XT-19TD00-4.
The ranges I'm planning on using are (W)90 - 400 Hz (M)400 - 1900 Hz and (T) 1900 - 20K. All crossovers are active and 4th order LR (built into the amps) and can be moved around a bit if necessary to get this right.
Thanks in advance,
Brad

JohnH12 · « **Reply #1 on:** 4 Jun 2016, 01:59 am »

Brad for a conventional closed back speaker the preferred spacing is less than or equal to the wavelength of the crossover frequency between the centers of the tweeter and mid. At 1,900 that’s distance of 18 cm between the tweeter and mids (343m/s divided by 1,900 cycles/s).

This is just a guideline that results in the tweeter to mid integration occurring close to the speaker. As no one listens very close to a speaker you can expand the guideline to reasonable driver frame separations.

For the two mids use the midpoint between them to the woofers as a guide. Again the spacing is a guideline to ensure decent integration.

For open baffle the Vifa XT19 needs to reach down to the RS-100 below the first dipole null for constant directivity. The dipole null on a narrow rectangular baffle of 5 inches is about 2,300 Hz. So your baffle width OB would be considered more important for determining your x-o point than the CTC spacing.

I like to use ABC Dipole to look at driver roll off OB http://www.musicanddesign.com/A_B_C_Dipole.html The help file is very good.

John H

brad944911 · « **Reply #2 on:** 6 Jun 2016, 02:10 pm »

John,
Thanks for the link, advice and sanity check...I've been doing everything longhand, so I'll play around with this a bit.

brad944911 · « **Reply #3 on:** 14 Jun 2016, 03:30 pm »

Well I got this project up and running and so far, so good. I don't have FR plots for the speakers, but am using a DEQ2496 to equalize it to flat at about a 30deg off axis center position and so far I dig the openness of them. This being my first MTM type speaker, the controlled floor/ceiling bounce is VERY noticeable...not a bad thing. After a little more research, I ended up with the subs going up to 80Hz, the woofs from 100Hz(to help cancel a room boost around that area) to 415Hz, the mids from 415Hz to 2KHz and the tweet is from 2KHz out. Here are some pics. If anyone has any ideas or suggestions, please chime in. My ego is fully in check and have a passion to learn. Thanks Matavena for the original guidance on the Hestia and John for his input here. Brad

matevana · « **Reply #4 on:** 14 Jun 2016, 04:50 pm »

Hey Brad,

Interesting concept. I might try crossing to the XT19's much higher; something in the 6-7k range. This will likely reduce some potential straining, increase your dipolar response and benefit the overall polar pattern as well. Plus with the Behringer it's an easy change to try.

I'm guessing the frames were cut on a CNC machine. Can you describe the magnet clamp on the back brace? Have you noticed a lot of vibration between drivers? Kudos to you for taking risks in a design! Tell us more as you continue to evaluate them.

brad944911 · « **Reply #5 on:** 14 Jun 2016, 06:20 pm »

Ed,
The main concept I was trying to go for here was as much dispersion as possible--maximize that fabled dipole "figure 8" pattern of each driver and avoid them beaming. I'm afraid if I push the mids to tweet x-over that high, the mids will be in their beaming range, right? I'm gonna try it just because you suggested it, but just letting ya know my design philosophy...shoulda said that in the beginning.

That's why I used the most minimal baffle possible, to avoid any anomalies from them. I couldn't find a lot of info on using minimal or no baffles, so if you or anyone can give me some good research to digest, that would be awesome!

I wish I had a CNC...man that would've saved me a LOT of time. I used basic tools: a router with a circle jig made from 1/4" lexan, a router table, table/jig/band saw, drill press and a bench top belt sander.
As for the magnet brace/spine, its (3) 1x4's glued together(1st picture) with a cavity for the wires in the center layer. I glued/clamped them, then used furniture bolts just to make absolutely sure they were solid. I measured the magnets with a caliper then routed out about 1/64th smaller in diameter(much trial and error), then hand finished so the magnets were snug to get in place. It is all held together with a 5" stainless lag bolt(non magnetic

) that goes through the back and is tightened into the "phase plugs" in the 10's holding them tightly in place.
Vibrations were actually a huge concern since the "baffle" is only 3/4" thick. I have spacers and screws for between the baffle and spine at the ready in case. So far I have noticed minimal vibrations, but I may still use them just in case.
So far I am quite happy with the soundstage and imaging; very open and very accurate. FYI my listening area is a bit funky and NOT optimal in any way. We have a BIG 90 degree sectional in which the middle of the couch is the center of the room...imagine what home plate looks like. The back of the couch would be the point and the opposite side is where the speakers are placed. So as you move from the center of the couch you are also moving much closer to a speaker than if just moving along a parallel wall. All that to say that they may actually have a much wider sweet spot, I just don't want to tear everything apart to see how wide. I saw an improvement.....sooooooo

The RTA showed fairly flat with pink noise...some +/-5dB spots, but I can take care of that easily with the DEQ-2496. Hopefully this will help it open up some. I'll keep you guys posted with the results of any suggestions...I like tinkering!
Now to get the boss to approve a color and get them painted

Take care,
Brad

matevana · « **Reply #6 on:** 14 Jun 2016, 08:07 pm »

The Dayton RS-100 has a cone that's probably closer to 3" than 4", so beaming at these frequencies shouldn't be too much of a concern. More importantly, since you're not using a ribbon (or a 2nd rear facing dome tweeter) your current design is only dipole up to the mid drivers cutoff. By crossing at 2k currently, you are likely missing a lot of those spatial cues that reflect from the wall behind the speakers, which help them disappear as a focused point source.

Nice job on all the hand routing!

brad944911 · « **Reply #7 on:** 15 Jun 2016, 01:03 am »

Point taken...and THAT'S why I ask for input! Gonna raise it n see what happens. Cheers brother

!

JohnH12 · « **Reply #8 on:** 15 Jun 2016, 02:03 am »

Nice job, wonderful build.

When you push the mid above the dipole null you lose constant directivity. The figure 8 pattern breaks down to like petals of a flower. On axis it looks nice and flat off axis you have dips. It’s not beaming it’s the front and rear wave summation where the rear wave and front waves are out of phase with each other at the frequency of the baffle width.

Not likely to be a problem if you have one place you normally listen at.

I do agree 2K is likely to stress the XT-19. I also like the idea of higher dipole frequencies. Depending on how young you ears are, try running without the XT19, or you can chose another chambered tweeter.

However, it's a great build. I'd use it and save your ideas for something else later.

Guy 13 · « **Reply #9 on:** 15 Jun 2016, 02:42 am »

Hi Brad, nice job.

I like your comncept - design.
If it was me, but it's not,
I would worry about the unit to tip forward
since it's not very deep ? ? ?

Guy 13

matevana · « **Reply #10 on:** 15 Jun 2016, 12:36 pm »

Thought you might appreciate this thread from a few years back. Not a WMTMW, but a similar minimalist design concept.

http://www.audiocircle.com/index.php?topic=79036.0

brad944911 · « **Reply #11 on:** 15 Jun 2016, 02:25 pm »

Guys,
Thanks for the kind words and ideas so far! I really look forward to refining this with your inputs!

I may have misunderstood some things while researching this project, so if you can, help me understand a little better here. Why I designed this like I did:

X-over freq - Based on the twin mids center to center(CTC) spacing of 6-1/2 inches I chose the 2K crossover point to maintain "constant directivity"(CD) and keep them from beaming. John helped with input at the beginning, but I never really found substantiated facts on whether the CTC of the mids or the CTC of the mid to tweeter would determine the x-over point, so I went with the most common answer; mids CTC spacing. Sorry John, I didn't ignore ya, I just couldn't get hard facts, so I had to go with a majority answer. I raised the X-over last night BTW and it did seem to open the midrange a bit, but the HP x-over in my amps(Crown XLS 1002's - Got a KILLER deal on 4 of them) only go to 3KHz, so I'm kinda confined there.

Baffle diameter and thickness - As John mentioned, I was trying to avoid any nulls in the off axis response, so that's why I went with only 1/4" extra diameter outside each of the drivers and made the "baffles" 3/4" thick. From what I understood, those dimensions would keep CD well above 10KHz...right?

Guy,
They actually are a little rear biased due to the woof's heavy magnet and aluminum frames. I put spikes as far out and forward under the "outriggers" on front and as far to the rear on the spine as possible to balance them. They seem OK...I don't have small kids or big dogs

Ed,
Holy crap, I nearly doppelgangered that build, huh. Too funny!

Once again guys, none of this is me arguing my point, but looking to be corrected if I misunderstood anything. I've read almost everything I could on the interwebs, but maybe I didn't go far enough in. Your time and advice is MUCH appreciated!

Brad

matevana · « **Reply #12 on:** 15 Jun 2016, 04:44 pm »

Brad,

Have you considered running the mids and tweeters w/o the highpass filter engaged and simply using passive components for these two drivers instead? You can probably get by with a single inductor for the RS100 and a single capacitor for the XT19 (1.0 or 1.5 uF) You may even have appropriate values left over from your earlier build... at least for testing purposes. The 3khz limitation on the XLS (with a fixed slope of 24db/oct) may not let you take full advantage of the higher dipolar frequencies you are hoping to benefit from.

JohnH12 · « **Reply #13 on:** 15 Jun 2016, 06:34 pm »

Without measurement equipment we are guessing a bit but your baffle sims like a dipole null of 2,500.

The Edge program could do something close to the dumbbell shaped baffle.

brad944911 · « **Reply #14 on:** 15 Jun 2016, 06:41 pm »

Ed,
Early in the build I did look at passive, but I was still looking at crossing over at about 2-2.2KHz to meet the goal of keeping the x-over freq below the wavelength of the center to center spacing of the Dayton's, about 6.5"=2150Hz. That's what I understood MTM design practice to be at the time, but if I use the mid to tweet, it's 3-1/4---about 4KHz, I can go to the 3KHz my amp will allow and unburden the tweeter a bit...it has a low fs of around 700Hz anyway. Add to that the fact that the conventional knowledge I found on MTM's said to use a 3rd order or higher crossover and I just decided on another amp.

I guess I'm missing something all around here, so please help me understand. The Dayton's cone is around 3-1/4 diameter which means it will be full on directional at around 4100Hz, right? So by then it's no longer working in a broad, dipolar manner any more, correct?

I hear what you're saying about the dipolar information missing without a rear tweet so I am looking at PE right now to pick up another XT19 or maybe I'll just use one of the OT19's from the coax's I tried.

I can tell you feel strongly about pushing the x-over up, so I KNOW I'm missing something. Hit me with the AH HA tid bit

Brad

brad944911 · « **Reply #15 on:** 15 Jun 2016, 08:09 pm »

John,
Thanks for that. The baffle around the mids is actually 4-1/2 inches and around the tweeter it is 4 inches, so by that logic, the null in the mids will be pushed up to about 3K, right? So if I cross the mids to tweets at the 3KHz max of my amps I could add another rear facing tweeter and, with the summing of the mids and tweet, kind of overcome that null? Then to overcome the tweet's null...the first null in that response would be 6.7KHz then again at nearly 14KHz, which would require some equalizing with the Behringer. Does this sound like a better plan or am I totally missing your point?
All in all, my goal is as uniform as possible horizontal dispersion. The only way I can think to accomplish this was to ensure the drivers are operating as omnidirectionally as possible by avoiding their beaming freq.
Once again guys, I don't want you to think I'm arguing, I just want to understand your points of view so I can learn from you. I'll buy ya a beer or twelve if you're ever in Tampa

JohnH12 · « **Reply #16 on:** 16 Jun 2016, 01:49 am »

3K sounds right.

Adding a rear facing tweeter will create another dipole null at the width of the tweeter, you are not eliminating a null it’s always there. Adding the rear facing tweeter adds to the off axis issues, but you gain some reflected sound.

If we step back bit, generally constant directivity is the goal if you are building a reproducible design for others. However, you can operate above the dipole null and eq the response to sound good at your listening point in your room.

http://www.musicanddesign.com/Integrating_a_tweeter.html

You might enjoy a measurement mic more than a second tweeter

matevana · « **Reply #17 on:** 16 Jun 2016, 10:17 am »

Hey Brad,

I didn't realize you are in Tampa. I usually have at least one meeting at USF each year. I work for one of the other state universities in South Florida.

What you are saying isn't incorrect. And based on your design goals for the project, you have not deviated from any of the laws of physics that might impact your goal. When I experimented with three and four way dipoles, I came to the conclusion that I preferred the midrange driver when it operates as a true wide band, driver, and then supplemented this with a woofer and tweeter at the extremes. Of course this doesn't necessarily support your design goal. I was more curious what impact shifting the X/O point upward might have on the overall enjoyment of the design.

brad944911 · « **Reply #18 on:** 16 Jun 2016, 01:28 pm »

John,
I was looking at using the exact same vifa XT19 tweeter if I go this route, so that may make it a little more predictable. The whole Constant Directivity idea came from some experimenting with "omnidirectional" designs. Once I figured out the issues there, lower SPL, less distinct transients and a helluva lot of work for a diffuser, I decided on a 4 way that would maximize the dispersion patterns of each speaker and be open baffle. I really enjoy the open and uncolored sound. So that's how I arrived here. I already have the Behringer ECM8000 mic(solid enough?) that I use with the DEQ-2496 to help equalize the room, but no software...got a suggestion?

Ed,
Yea man, been stationed here now for 4 years and I'm home! I love this area! Give me a shout on IM next time you're in town and we'll hit Coppertail Brewing...great beers. But I digress...back to speakers...I think I kind of backed myself into a corner by using the WMTMW format...the all critical M to M spacing and baffle size kinda have me limited. I have been playing with the idea of using just the bottom WMT and seeing what happens there. Then I can push the x-over up. Right now, I think I'll just add that second tweet and see what happens.

Thanks again guys,
Brad

matevana · « **Reply #19 on:** 16 Jun 2016, 02:20 pm »

LOL...Florida Inspired, Tampa Brewed. Our Beer, Our Way.

I think the next Hestia design will incorporate Barley... somehow.