@jimbones I made some little changes on your crossover in the mid tweeter section in the file attached. Your mid-tweeter section of the crossover needs some rework, the phase tracking is not good at all.
You should work on the crossover outside the box to measure if things are right. To see if the phase is good you should invert the mid and see two big holes in the response curve.
You should work on the crossover outside the box to measure if things are right. To see if the phase is good you should invert the mid and see two big holes in the response curve.
Attachments
See if you can get the wires and x/o out and record FR, far field, same drive voltage, same mic distance:
- mic between w and mid
- woofer with x/o only; other units off
- mid with x/o only; other units off
- woofer and mid with their x/o; in phase; tw off
- woofer and mid with their x/o; out of phase; tw off
- mic between mid and tw
- tw with x/o only; other units off
- mid with x/o only; other units off
- mid and tw with their x/o; in phase; w off
- mid and tw with their x/o; out of phase; w off
@Lojzek @jerome69 My game plan is this. Make the XO changes @2200hz. Also while I have the cabinet open at that point to get the measurement Lojzek suggests. That will take some time as I will do that in between other responsibilities. In the meantime if you have any other suggestions or advice before I go into the cabinet let me know. Thank you for your help.
Give post #7 a read, it will help you with measurements.
I would postpone any mods until you get a little better measurements.
When you time lock measurements, you won't have to bother introducing Z-offset values (delay) in a simulator.
You have now 3" delay for W (no phase data) and none to other units and there should be some delay to mid as well.
https://www.diyaudio.com/community/...frequency-response-plots.425102/#post-7956335
I would postpone any mods until you get a little better measurements.
When you time lock measurements, you won't have to bother introducing Z-offset values (delay) in a simulator.
You have now 3" delay for W (no phase data) and none to other units and there should be some delay to mid as well.
https://www.diyaudio.com/community/...frequency-response-plots.425102/#post-7956335
On setting up the Xsim results window - https://www.diyaudio.com/community/threads/xsim-configuring-the-plot-window.406106/
ok so here goes, be nice 😉
Attachments
-
Measurement 6 FR Mid XO only.jpg -
Measurement 6 Imp Mid XO only.jpg -
Measurement 7 Imp M&T XO In Ph.jpg -
Measurement 7 Fr M&T XO In Ph.jpg -
Measurement 8 IMP M&T OOP.jpg -
Measurement 8 FR M&T OOP.jpg -
Measurement 5 Fr Twt XO only.jpg -
Measurement 5 Imp Twt XO only.jpg -
Measurement 4 Imp W&M XO OOP.jpg -
Measurement 4 Fr W&M XO OOP.jpg -
Measurement 3 Fr W&M XO Only In Ph.jpg -
Measurement 3 Imp W&M XO Only In Ph.jpg -
Measurement 2 Fr Mid XO Only In Ph.jpg -
Measurement 2 Imp Mid XO Only In Ph.jpg -
Measurement 1Fr Woofer XO.jpg -
Measurement 1 Imp Woofer XO.jpg -
System Measurements List 3_2025 Rev 3 Config.docx
In a JDV style and being as nice as possible.
You seem to have a reflection impulse at 3mS or so.
Maybe tightening up the window timing would give a slightly better view.
I.e. cut off the impulse by closing the window down to just before the impulse say at 2.9mS. You will of course loose low frequency information, below 350Hz or so. I am assuming your measurements were at 1 meter and that there must be a reflective surface within 3ms be it the floor or rear wall.
Well done on all the measurement's.
You seem to have a reflection impulse at 3mS or so.
Maybe tightening up the window timing would give a slightly better view.
I.e. cut off the impulse by closing the window down to just before the impulse say at 2.9mS. You will of course loose low frequency information, below 350Hz or so. I am assuming your measurements were at 1 meter and that there must be a reflective surface within 3ms be it the floor or rear wall.
Well done on all the measurement's.
@raymondj Yes I notice that as well. After doing the calculation it appears to be floor bounce. I was in "Go Mode" and did not do any critique of the measurements until I was finished. Then I noticed it, ugghh. Like you mention I am concerned with the loss of low frequency info. But I feel better that we will have both. I will go back and do more measurements. I have a busy couple of days so I wil fit in where I can. Maybe I'll just do a couple of measurements to get started. TY for your input it is appreciated.
@jimbones I look the measurements. The response above 10kHz seems not very realistic. Too lower levels.
The crossover mid-tweeter could be better 12dB acoustic for the tweeter and 24dB for the woofer. The phase tracking MT is not bad but far from a big hole and not very symetric.
Mid woofer. The phase tracking is not good, you don't see any change if mid is reverse or not. And at 1kHz the woofer has a bump, could have negative effect on the high mid, I think you have interferences with the tweeter.
You should see the big picture, the entire response on the same graph with all responses.
Measure raw response without xo of each driver and put them on a simulation software. Think to measure offsets dz.
The work begins...
The crossover mid-tweeter could be better 12dB acoustic for the tweeter and 24dB for the woofer. The phase tracking MT is not bad but far from a big hole and not very symetric.
Mid woofer. The phase tracking is not good, you don't see any change if mid is reverse or not. And at 1kHz the woofer has a bump, could have negative effect on the high mid, I think you have interferences with the tweeter.
You should see the big picture, the entire response on the same graph with all responses.
Measure raw response without xo of each driver and put them on a simulation software. Think to measure offsets dz.
The work begins...
Measure drivers separately, check first break point of woofer, first of mid, double both, go in middle
I cannot see images you attached th relation is weak
I cannot see images you attached th relation is weak
OK some one asked for frd files. I will redo the measurements. but will be away for 2 days. Will post upon my return. Thanks all.
You need the ZMA and the FRD, as installed in the final box. Measured from the same point.
It takes years to design proper XO. It takes wisdom, patience and experience.
These are only good to design the XO, then it must be built and adjusted a lot, XSIM helps a bit looking at what happens when values change, but you have to test everything in real life to confirm everything in the simulator.
It takes years to design proper XO. It takes wisdom, patience and experience.
These are only good to design the XO, then it must be built and adjusted a lot, XSIM helps a bit looking at what happens when values change, but you have to test everything in real life to confirm everything in the simulator.
OK I remembered from 3 years ago I have measurements I took using DATS3. However when I try to attach them they are an unsupported file (.dats) Looks like they are text files and zma only. see examples
I just finished a xo for my speakers and the sound is so bad , however it has a more relaxed frequency response than reference loudspeakers that I have, and it has slightly less noise and less distortion and a similar construction, it shouts at you and it seems volumes are not right and misses a lot of information, maybe it is just bad placement, I kinda kept in check the electrical impedance...
The phase is also quite good, way better than the reference, but I is so bad, on my vinyl setup and on budget dac,
I'll see if it helps using my tube dac. It was designed with it, but seriously I am putting everything I know in question, maybe it is the placement is too low, idk what to think....
The phase is also quite good, way better than the reference, but I is so bad, on my vinyl setup and on budget dac,
I'll see if it helps using my tube dac. It was designed with it, but seriously I am putting everything I know in question, maybe it is the placement is too low, idk what to think....
