MTM is quite tricky to get right.
Lots of math if you want to do it right. Idealy crossover point is determined by
2c/3d (c speed of sound, d distance beetween acoustic centers of tweeter and woofer).
Augerpro waveguide is tricky: you need to implement paralel notch filter (LR-C-R) to smooth the bump, otherwise you will lose 5db above 10khz making the tweeter dull. With notch you can keep frequency response linear, losing overall sensitivity. Keep in mind that you will need to attenuate woofers with paralel resistor (series will change qts) to match the tweeter.I will post my soulution.
Lots of math if you want to do it right. Idealy crossover point is determined by
2c/3d (c speed of sound, d distance beetween acoustic centers of tweeter and woofer).
Augerpro waveguide is tricky: you need to implement paralel notch filter (LR-C-R) to smooth the bump, otherwise you will lose 5db above 10khz making the tweeter dull. With notch you can keep frequency response linear, losing overall sensitivity. Keep in mind that you will need to attenuate woofers with paralel resistor (series will change qts) to match the tweeter.I will post my soulution.
My first try:
Tried to keep sensitivity and preferred power response over reference angle. Tweeter is dull as you stated. I keep on learning thank you for that!
Tried to keep sensitivity and preferred power response over reference angle. Tweeter is dull as you stated. I keep on learning thank you for that!
The compromise choosen because the XO is so high. Get the XO low enuff and most of the problems go away.
dave
What would be low enough? My ctc spacing is 185mm and xo is 1750hz. Should I go even lower maybe 1,5k?
In my MTM spacing is 336mm and XO is 250 Hz.
Hard to see how you add a tweeter between 2 176mm woofers and get to only 185 mm centre to centre total. Centre-to-centre is the centre of one midwoofer to the centre of the other.
The problems go away when you get to quarter wavelength spacing, getting as close as possible would always be my goal.
dave
Hard to see how you add a tweeter between 2 176mm woofers and get to only 185 mm centre to centre total. Centre-to-centre is the centre of one midwoofer to the centre of the other.
The problems go away when you get to quarter wavelength spacing, getting as close as possible would always be my goal.
dave
Oh! Thought it was from tweeter to woofer. My bad.
I should start my own threat so I wont spoil jogunast's project.
Last edited by a moderator:
You need to connect woofers in series, and add resistor to the ground after the LC to keep it around 82db. This way (like in you sim) you will end up with <15db at 20khz at listening position. I will send you my crossover schematic. I bought 8ohm drivers thinking that 90db sb26adc could keep up. Nope.. waveguide sacrifices >6db in top octave. I dont know who can listen speaker with that much attenuation in upper freq. It sounds like cardboard.
4ohm in series with pad would be the better solution.
Last edited:
That is the compromise that works well.
Otherwise you would need 3" woofer and tiny tweeter without faceplate. Or woofers with fullrange "tweeter" that can be crossed very low.. which defeats the purpose of mtm. Whole point is to emulate the presentation of fullrange speaker, by avoiding use of fullrange driver.
Plain and simple - to avoid image distortion presented by other multi way designs.
Achiveable only by fullrange drivers that are plagued with their own unsolvable problems.
So.. by using dedicated woofer and dedicated tweeter in geometric alignment you get best behaviour trough whole frequency range without piano climbing up and down trough octaves. And by using fullrange driver.. well you are using fullrange driver. I'm yet to hear one that doesnt sound like mid-fi.
Achiveable only by fullrange drivers that are plagued with their own unsolvable problems.
So.. by using dedicated woofer and dedicated tweeter in geometric alignment you get best behaviour trough whole frequency range without piano climbing up and down trough octaves. And by using fullrange driver.. well you are using fullrange driver. I'm yet to hear one that doesnt sound like mid-fi.
So do tweeters. Probalms can only be minimized. Make your compromises where you will.
They have really low image distortion (ie they have a REALLY good soundstage/image). And no lobing.
My implemation has been well received.
dave
I agree the SB26CDC on my waveguide is a bit tricky to match sensitivities to many woofers. I may have to switch back to the 8 ohm SB17CAC for my TM with passive crossover. It is possible to use the 4 ohm with the active crossover, not because you can EQ the top octave (though that helps), but with no crossover losses, you are about 83dB >10khz, and 85dB from 8khz and down. So this matches the 4 ohm well enough I think. BUT another important tip is the phase-shield version has about 1 dB better response in that top octave. I only made the non-phase-shield version for CNCers, for anyone 3D printing I would use the phase shield version.
IME with hifi domes on waveguides, you don't want a downsloping on-axis response such as you have, or the tweeter will sound too quiet. I understand the top octave makes that tough, but your response is lower than I'm getting and I think the 4th order crossover is not helping. Try a 3rd order. I also don't know that the 5" waveguide is appropriate here, at least at 1.7khz.
I'm going to investigate strategic use of a notch filter to possibly bump up that top octave. Possibly notching the 26khz resonance may have the side effect of bumping up the octave below it, as often happens with notch filters.
IF none of these options work, you might consider the SB21SDC on my waveguides. Has good sensitivity, I think the waveguide is the same shape and should drop in, and has a shorting ring. They look great on my waveguide and I'm surprised more people haven't tried it. I'm going to use it with the "entry-level" SB drivers some time next year.
IME with hifi domes on waveguides, you don't want a downsloping on-axis response such as you have, or the tweeter will sound too quiet. I understand the top octave makes that tough, but your response is lower than I'm getting and I think the 4th order crossover is not helping. Try a 3rd order. I also don't know that the 5" waveguide is appropriate here, at least at 1.7khz.
I'm going to investigate strategic use of a notch filter to possibly bump up that top octave. Possibly notching the 26khz resonance may have the side effect of bumping up the octave below it, as often happens with notch filters.
IF none of these options work, you might consider the SB21SDC on my waveguides. Has good sensitivity, I think the waveguide is the same shape and should drop in, and has a shorting ring. They look great on my waveguide and I'm surprised more people haven't tried it. I'm going to use it with the "entry-level" SB drivers some time next year.
Wow SB26adc is really so low in level after you even out waveguide gain?
I just bought one pair to use them in 2.5 way with two 6 inchers with hope not to use any resistors for level matching...
Now it seems that they will be to quiet...
I just bought one pair to use them in 2.5 way with two 6 inchers with hope not to use any resistors for level matching...
Now it seems that they will be to quiet...
This is non calibrated SPL. Its not 80db.. my guess would be around 83-84db. But yes.. waveguide is compromise, you gain some, you lose some. Sb26adc could easily play this low without waveguide.
This way you get extremely low distortion at crossover point and more control in directivity..
This way you get extremely low distortion at crossover point and more control in directivity..
I'm away from pc for few days..
My version is:
C 10uf, L0.8mh to the ground. After first CL you put 6.6uf with paralel notch around it: 15R paralel with 0.13mh 3.9R series. After that 0.18mh to the ground.
That should give you completely flat response. For woofers connected in series: 2.5mh, 13.3uf to the ground, then 0.45mh in series, 12R to the ground. Thats how i got +/-1db from 30-20khz, 84db sensitivity.
My version is:
C 10uf, L0.8mh to the ground. After first CL you put 6.6uf with paralel notch around it: 15R paralel with 0.13mh 3.9R series. After that 0.18mh to the ground.
That should give you completely flat response. For woofers connected in series: 2.5mh, 13.3uf to the ground, then 0.45mh in series, 12R to the ground. Thats how i got +/-1db from 30-20khz, 84db sensitivity.
Ah yes, doing the same thing but with the 5khz bump instead of the 26khz breakup. I like the result!
Riffing off yours, I let the optimizer go to work. I have not standardized the values or made sure they are sensible, but it looks like a good place to start with nominal sensitivity of 85 dB. (BTW this the SB26CDC on the 6" waveguide on my TM, so not exactly the same as you have)
