Stuffing and phase alignment - is there a relationship?

[fjhuerta]

I'm still revisiting some old designs of mine, hoping to learn a bit more. This is something that has me baffled, so to speak.

A couple of years ago, I built a 3 way speaker which I voiced too bright. Foolishly (or not, I don't know) instead of taming the treble, I removed all the stuffing, thinking I'd get more bass.

Fast forward to today. I took the original measurements, re-designed the crossover and listened. The speaker was still bright, even though the measurements seemed to indicate it had a flat on-axis FR and nicely controlled off axis curves.

So, I took an MLS measurement, and there was some extra treble response (or so I thought). After days of wondering what was wrong, I noticed that phase tracking between the woofer and midrange was off (this is a 2nd order L-R xover, so even though it was at 350 Hz, at 1 KHz the tweeter and woofer are supposed to contribute some output to the midrange). If I delayed the woofer by 200us, they tracked perfectly (as in my original measurements) and the speaker was flat on axis again.

Now, the only thing that I changed in those years was the stuffing. But I don't know if stuffing can alter phase so much at around 1 KHz!

I actually had to change the x-over topology to make them whole thing sum properly again.

[billfitzmaurice]

But I don't know if stuffing can alter phase so much at around 1 KHz!

It can, by as much as 180 degrees. You can model it in HornResp.

[Jeff B.]

I'm still revisiting some old designs of mine, hoping to learn a bit more. This is something that has me baffled, so to speak.

A couple of years ago, I built a 3 way speaker which I voiced too bright. Foolishly (or not, I don't know) instead of taming the treble, I removed all the stuffing, thinking I'd get more bass.

Fast forward to today. I took the original measurements, re-designed the crossover and listened. The speaker was still bright, even though the measurements seemed to indicate it had a flat on-axis FR and nicely controlled off axis curves.

So, I took an MLS measurement, and there was some extra treble response (or so I thought). After days of wondering what was wrong, I noticed that phase tracking between the woofer and midrange was off (this is a 2nd order L-R xover, so even though it was at 350 Hz, at 1 KHz the tweeter and woofer are supposed to contribute some output to the midrange). If I delayed the woofer by 200us, they tracked perfectly (as in my original measurements) and the speaker was flat on axis again.

Now, the only thing that I changed in those years was the stuffing. But I don't know if stuffing can alter phase so much at around 1 KHz!

I actually had to change the x-over topology to make them whole thing sum properly again.

Let's look at this objectively. We know that driver responses are minimum phase, and so are crossover roll-offs. Therefore, the phase of the driver in its stopband is tied directly to it's frequency response and roll-off. Without something like DSP, delay circuits, or offsets changes involved, you will not change the phase without changing the frequency response and vice-versa.

So, if the phase alignment changed with the change in stuffing, then the frequency response, and more precisely the roll-off likely changed too. Now, if it did change, since we are talking about stuffing we are likely talking about the contribution of the sound from inside the box contributing to the output coming directly from the cone. If the stuffing absorbs this, then there will be less contributution, so a steeper roll-off, and the associated change in phase. If the stuffing is removed, then this affect is removed too, there will be increased stopband output and a change in the phase alignment to go with it.

Make sense?

[fjhuerta]

Thank you Jeff. It did make sense.

I got tons of stuffing and re-stuffed the speakers. Guess what? I suppose I took my measurements wrong the first time around, because phase was exactly the same.

I added even more stuffing. No luck. I got less bass response after a while.

So... I think that stuffing a box may alter phase a bit. But not as much as I originally thought. Well, now I know.

Thanks again!

[Jeff B.]

Thank you Jeff. It did make sense.

I got tons of stuffing and re-stuffed the speakers. Guess what? I suppose I took my measurements wrong the first time around, because phase was exactly the same.

I added even more stuffing. No luck. I got less bass response after a while.

So... I think that stuffing a box may alter phase a bit. But not as much as I originally thought. Well, now I know.

Thanks again!

It can only alter the phase if it first alters the frequency response.

[fjhuerta]

It can only alter the phase if it first alters the frequency response.

It did alter FR, but not enough for it to show up as a huge change regarding phase. It was minimal....

[Jeff B.]

It did alter FR, but not enough for it to show up as a huge change regarding phase. It was minimal....

Exactly, it's a minimum phase relationship.

[duanebro]

Exactly, it's a minimum phase relationship.

This gets my vote for geek joke of the day!

[fjhuerta]

Exactly, it's a minimum phase relationship.

Ha!