The Odd-1, ATC recommendations please

Current project consists of the following drivers:
ATC 12" (SB75-314SC) in BR, Hoges tuning to 35 hz
ATC 8" (SB75-234SC), closed Q approx 0,71
ATC 3" (SM75-150) - Two e.a. in MTM
Dayton AMT Pro4 (open and damped backwards)

The "singers" above will be conducted by Mini-DSP electronic crossover.
I ask for your advices for x-over and slope settings.
This speaker will be finished with oak veneer sheets.
 

Attachments

  • oddone1.jpg
    oddone1.jpg
    81.5 KB · Views: 291
  • oddone7.jpg
    oddone7.jpg
    99.1 KB · Views: 130
  • oddone6.jpg
    oddone6.jpg
    62.5 KB · Views: 98
  • oddone5.jpg
    oddone5.jpg
    85.7 KB · Views: 279
  • oddone4.jpg
    oddone4.jpg
    76.1 KB · Views: 274
  • oddone3.jpg
    oddone3.jpg
    56.6 KB · Views: 280
  • oddone2.jpg
    oddone2.jpg
    94.1 KB · Views: 283
Last edited:

eriksquires

Member
2013-05-10 4:11 pm
Usually people go nuts with DSP and try to use 4th order LR for everything, and it rarely goes well. :)

The actual driver response will be the combination of the driver + EQ. Since you have the ability to delay the drivers, you should be able to quite comfortably try to achieve 2nd order filters. If you can use that for all drivers it will allow you to keep them in the same polarity. If not, try to achieve a 4th order for the same reasons.

Your entire baffle is sloped, so make sure you measure driver delay perpendicular to the tweeter/midrange units.

As for slopes and EQ, start with the midrange, and work your way outwards from there.

When you are done I really would like to see some compression / linearity tests for that tweeter. :)

Best,


Erik
 
With dsp, you must learn to do good quality acoustic measurements! Standardize your measurement setup, speaker placement, mic placement, furniture in the room etc.
Procedure (been there, done that)
1) measure each driver's raw response in the "box" 0, 15, 30, 45, 60 and 90 degree horizontally at 1-1,5m distance. Study impulse response and use different gatings to recognize reflections in the measurement. Apply different smoothings 1/2 -1/24
2) use dsp to equalize the acoustic response for each driver at least one octave past the intended xo frequency. Make notches to resonance peaks (the 12" cone has those).
3) study what happens off-axis, reconsider xo frequencies. Check also vertical off-axis
4) apply some xo slopes, match spl and measure the system, set delays carefully (this is very difficult before you learn it)
5) study off-axis behaviour and distortion behaviour - try other xo points and slopes
6) Listen to it for some days, then fine-tune "sonic character" by adjusting eq/levels especially where directivity is changing
do 5) and 6) again and again and again...
Read this first! Prototyping a 4-way open-baffle speaker with the miniDSP 2×4
 
Last edited:
Current project consists of the following drivers:
ATC 12" (SB75-314SC) in BR, Hoges tuning to 35 hz
ATC 8" (SB75-234SC), closed Q approx 0,71
ATC 3" (SM75-150) - Two e.a. in MTM
Dayton AMT Pro4 (open and damped backwards)

The "singers" above will be conducted by Mini-DSP electronic crossover.
I ask for your advices for x-over and slope settings.
This speaker will be finished with oak veneer sheets.

This looks nice. I hadn't seen the ATC drivers before.
I like the overall cabinet shape, and the work you put into that grill, visible in photos 4 & 5. I assume that grill is removable, to allow mounting the mid range drivers.

I have more questions than advice - but how you respond might help generate better advice.

When you say "Dayton AMT Pro4 (open and damped backwards)" do you mean with the felt pads in place? I'm curious why you chose this driver, and why you put so much effort into getting a partial dipole. I'm particularly curious because I recently ordered a pair :)

Will the 8" and 12" be covering the same range (playing together up to ~400Hz) or not (e.g. 12" to 100Hz, 8" 100-400Hz)?

Given that your low frequency drivers are low efficiency ~89dB, and the others drivers are higher efficiency (93 and 94dB), why did you opt for two mids? It seems like overkill - one seems like enough, in terms of output capability.
You may find it sounds best with one of the mids operating for its lowest octave, or turned off entirely.

Is it too late to add an extra layer to the front baffle, so you can re-cut your mounting holes and offset the drivers? It seems odd to position them right in the middle, where diffraction effects will be the most pronounced.
 
Pictures are good: here is a project with a similar spacing between the midrange drivers.

134458d1243700126-10-mtm-neopro5i-ribbon-build-thread-both-speakers.jpg


http://www.diyaudio.com/forums/mult...pro5i-ribbon-build-thread-13.html#post2687123

The builder was not satisfied with a few aspects, including the nulls caused by the large spacing. "At just 15 degrees off axis, there’s a 20dB dip at 1500Hz which is completely unacceptable. Things only get worse from there, with a plethora of peaks and nulls as you move farther off axis."

If you measured a similar null, you could try running the upper ATC 3" over a limited range, maybe 400-800Hz, and see if you preferred that. Or remove the upper 3" and use it on a different project.
 
Thanks your feedback, a good way to keep on learning.
This project is made on specification from my brother who likes the Proac "punch".
The Proac Responce 4 also got an MTM arrangement.
As we wil use the MiniDSP in an active 4-way management, we are free to try different approaches for best sound - including bass overlap- etc.
The reason for dipole tweeter is to avoid the internal chamber effects (reflex)
and we will experiment with and without the rear felt sheets,
 

Attachments

  • Proac reference.jpg
    Proac reference.jpg
    41.1 KB · Views: 24
This is an interesting project with some very expensive parts. The cabinetry I must say is very well done.

I have to agree with others on your MTM arrangement. The Proac you reference, as you can see has much shorter distance between the midrange drivers. That is not to say that your arrangement can't work, but I can say it will work best in a large room with the listener a fair distance from the speaker. Due to the large distance between the mids the vertical window will be very short. Besides that I am very confident that a single SM75 is more than capable of meeting the output requirement here.
 
Speakers are finished.
LM3886 for tweeter, Lindsley Hood 15W cl.A for mids, Class D Audio for woofers.
Chain:HTPC w/ Asus Essence STX soundcard - JRiver 21 - Dirac Live software-MiniDSP 4 x 10 HD. To our nightmare there is a challenge make the chain play as the Dirac (?) seems to refuse volunteer in this context.
 

Attachments

  • odd1fin.jpeg
    odd1fin.jpeg
    97.7 KB · Views: 96
Last edited:
You have some very nice wide bandwidth drivers there. They are perfect candidates for a Harsch quasi transient perfect Xo. Basically the xo uses a 4th order Butterworth acoustic low pass and 2nd order Bessel high pass for each xo. The delay of the upper driver relative to the lower should be half the period equivalent to the xo frequency. All drivers are positive polarity. This when done right gives a near right triangle step response. Percussion sounds excellent and very realistic.

More info here.

http://www.diyaudio.com/forums/multi-way/277691-s-harsch-xo.html
 
Earlier reported problems are fixed by re-installing the Win-10 and Adobe Air.
The Harsch XO is now implemented in the MiniDSP. Main curve are (as before) adjusted by means of Dirac Live. (XO= 120-420-4200hz)
The difference is tremendous. Full live stage, no grains and superb micro/macro dynamics. Thanks a lot for the advice !
 
Last edited: