Issues with time alignment tool

[Ofer]

Hi,
I have a miniDSP 4x10HD controlling my 3way active xover system. Lately I have added a sub. The sub is between the speakers aprox in the middle and 38cm behind them. The distance from the sub to the MLP is 382cm and from each speaker is 380. I wanted to use the alignment tool mach phase. Here is what I did:
1. measured each speaker and sub
2. did trace arithmetic on left and right speaker
3. used the result to align the sub at cursor, 80hz my xover point.
For some reason I get an error that the delay is above 70ms. Considering the distances described this is strange. I tried aligning speakers to sub and sub to speakers always an error. Attached are my measurements. Would love to understand what am I doing wrong.

Thanks.

 

[jtalden]

Move the cursor to about 50 Hz, i.e., more in the middle of XO range. I timed the sub manually and got -50 ms using the current sub polarity and -41.5 ms using the sub polarity inverted. The exact position of cursor can make a big difference when the room effects are large. Using the alignment tool 'align phase slopes at the cursor' I got results close to these values.

It appears the alignment tool limits itself to a 50 ms range so a manual adjustment to the sub impulse must be made if the sub starting position is beyond that. It's a good idea to recheck the timing after any significant PEQ is applied.

I don't know why the needed delay is so negative when the distances suggest the delay should be closer to 0 ms. Is it a wireless sub?

There is a big difference in the Left and Right channel woofer levels, maybe 5 or 6 dB.

I am not sure why you truncated the woofer at 100 Hz. It's best to use full range sweeps on all drivers when all XO's are active.
Attached is my mdat.

 

[Ofer]

Thanks @jtalden for taking the time to go over the measurements. The sub is and 18" PA sub. I have truncated the sub at 100hz because I didn't want to run it to areas well outside the xover range. Anyway the huge delay doesn't make sense. In the 4x10HD the maximum delay you can apply is 9ms. When I get longer cables I will try different positions maybe that will help.

 

[Ofer]

OK. The sub is pretty close to the back wall, like 10-10cm close. So I tried a few things.
1. inverted polarity
2. turned the sub 180deg to face the wall
3. turned the sub 180 deg and invert polarity.
After step 3 I could get the alignment tool to suggest 6.14ms delay to the A+B trace arithmetic. A I am using a 4x10HD each driver has its own channel. I dialed the delay to each channel. For some reason it didn't make much different to the measurement after adding the delay. I can say that the base sounds better.

 

[jtalden]

How are you limiting the sub range? Are you not applying a low pass filter in the DSP or sub amp to do it? If you are, we want to see the impact of that setting.

• All measurements should be full range.
• All measurements (L, R, sub) should be measured using the same reference channel (left?).
• Always measure the sub by using measurement channel (right?) and muting the right main speaker.

The measurements posted don't reflect the 6.14 ms delay change. The overlay impulse chart shows almost no change in sub timing. Did you remeasure both subs and mains after the change? That is necessary if the mains delay was changed.
Increasing the mains delay is the same as reducing the sub delay. We wanted to reduce the sub delay by ~40 ms, but the largest reduction in sub delay is with the mains at the 9 ms maximum delay and the sub at 0 ms. You could measure both mains and sub using those settings to see how far we still are from a favorable timing.

Rotating the sub is not a major change. A different place in the room is a good idea. Closer sub distance to the LP will also help our timing issue. You may want to try different positions of the mains and/or the LP. These also can have a big impact with a significant change.

In the end finding the most favorable timing is not the largest factor to consider. We are mostly looking for resonable SPL response through the bass range. We can hopefully find an initial response that allows application of a resonable amount to EQ to achieve that.

Correction

 

[Ofer]

Changing the sub position is planned, I am waiting for a longer cable to arrive. Anyway the sub is huge (and not very pretty) so the number of options is limited. The mains can't be moved. They are 90kg and are carefully placed. After rotating the sub and delaying the mains by the maximum possible I managed to reduce the difference between the combined response (A+B max delay) and the sub to 0.58ms. I will leave it at that until I can play with placement.
Thanks a lot for your time and help.