Testing phase at the crossover point.

Helpful hints on how to build 'em, and where to get the stuff you need.
Post Reply
Message
Author
Grant Bunter
Posts: 6915
Joined: Mon Nov 28, 2011 11:12 am
Location: Ilfracombe Queensland Australia
Contact:

Testing phase at the crossover point.

#1 Post by Grant Bunter »

During my time here I have learnt a lot.
And the more you learn, the less you think you know.

Even in the last couple of weeks, there has been discussion in a thread about polarity vs phase.

Polarity is relatively easy, plus or minus, as is, in an electrical connection, the result being a speaker cone will move in the opposite direction if swapped.
If one of two cabs has polarity swapped, frequency cancellation will occur.

Phase is different, it's frequency dependant. And, if you view phase as frequency vs time, it could be described as delay, given different frequencies have different wavelengths.
So some frequencies take longer to get to the listening position, hence they are delayed.

One of Bill's sayings , in essence, with our systems, is to make sure "the subs are in phase with the mains at the crossover point.
If the crossover point is say 100Hz, the goal will be to get the highest possible SPL at 100Hz.

Yes, we could use a polarity swapped lead to induce 180 degrees out of phase, or hit a phase swap button in DSP, and use RTA at 100Hz while playing a 100Hz wave through a tone generator to see which is louder or loudest.
Is that the simplest solution?

What do you do to test phase response at 100Hz without FFT?

Let's throw in a curve ball, namely delaying mains to backline. I totally get why that's a good thing as a musician up the back more often than not.
But what does that do to phase alignment at the crossover frequency?

Love to hear your thoughts on this complex subject...
Built:
DR 250: x 2 melded array, 2x CD horn, March 2012 plans.
T39's: 4 x 20" KL3010LF , 2 x 28" 3012LF.
WH8: x 6 with melded array wired series/parallel.
Bunter's Audio and Lighting "like"s would be most appreciated...

User avatar
Bill Fitzmaurice
Site Admin
Posts: 28620
Joined: Tue May 02, 2006 5:59 pm

Re: Testing phase at the crossover point.

#2 Post by Bill Fitzmaurice »

This brings up the fact that you can't hear phase unless you have two different sources playing the same frequency and the phase is at or close to 180 degrees apart between the two sources. The easiest way to test it is to play a tone at the crossover frequency, do a polarity flip on either the subs or mains, and note if the level goes up or down. If it goes up leave it flipped. If it goes down flip it back again. If you're going to delay either the subs or mains do this after setting the delay, as delay changes phase. Don't use delay to account for a polarity issue. Use delay for time alignment, use polarity flip to fix polarity. If you can't hear any difference with a polarity flip don't worry about it. What you can't hear doesn't matter.
delaying mains to backline. I totally get why that's a good thing as a musician up the back more often than not. But what does that do to phase alignment at the crossover frequency?
It might not do anything. But it gets complicated when you have folded horn subs that also add delay via the path length. You may need to delay the mains to time align to the backline, but not the subs, or by a lesser amount, because of the path length. Whatever you need to do with time align do that first, then do the polarity check.

Grant Bunter
Posts: 6915
Joined: Mon Nov 28, 2011 11:12 am
Location: Ilfracombe Queensland Australia
Contact:

Re: Testing phase at the crossover point.

#3 Post by Grant Bunter »

Thanks as always Bill.
Great answers.

If one decides to move on to FFT, then it is quite easy to produce 2 signal sources out of phase, as one can play a tone at the crossover point through both sub, and mains, with polarity flip if required to make them out of phase compared to each other.
Is this correct?
Built:
DR 250: x 2 melded array, 2x CD horn, March 2012 plans.
T39's: 4 x 20" KL3010LF , 2 x 28" 3012LF.
WH8: x 6 with melded array wired series/parallel.
Bunter's Audio and Lighting "like"s would be most appreciated...

User avatar
Bill Fitzmaurice
Site Admin
Posts: 28620
Joined: Tue May 02, 2006 5:59 pm

Re: Testing phase at the crossover point.

#4 Post by Bill Fitzmaurice »

There's no need to take a phase reading with FFT. You may, but this is one case where your ears as tools are sufficient.

Grant Bunter
Posts: 6915
Joined: Mon Nov 28, 2011 11:12 am
Location: Ilfracombe Queensland Australia
Contact:

Re: Testing phase at the crossover point.

#5 Post by Grant Bunter »

Bill Fitzmaurice wrote:There's no need to take a phase reading with FFT. You may, but this is one case where your ears as tools are sufficient.
Thanks again!

For those who may not/don't understand why testing phase at the crossover point is important, I'm going to throw some stuff in to explain:

It's often said here in the forum, Don't mix cabinets of different designs in the same band pass.
Why? Because every design has a different phase response. It's pretty much impossible to address all the phase response issues when using more than one design.

There should be only one situation where you do have different designs at the same frequency, and that is subs with tops/mains, at the crossover point. Hence the need to test.

How often do you need to do this?
Once, but only if the system is in an install that is never moved ever again.

That's not how most of us use our systems though.
Multiple venues mean we set up differently every time. Even returning to a venue again might see us setting up differently. Subs further forward or backward, mains/tops further forward, or backward, or wider apart or closer together, means phase response is altered. Sometimes this can also be due to relative placement, mains/tops not perfectly in line as an example.
So we need to test phase every time.

Some may say "my venue doesn't allow me time to do that".
Sometimes the answer to that is, get there early.
If you aren't allowed that either, I'm unsure about the driverack series, but the DEQ2496 will allow a certain frequency to be allocated. RTA the crossover frequency alone, with music playing, hit the invert button on subs or tops/mains and see if the RTA changes up or down (or your ears tell you one setting was better than the other). It takes 10 seconds. That shouldn't be a problem.

Tips:
As Bill said above, enter delay before testing.

Also, be aware that delay causes phase shifting.

Only EQ in the chain up to the input side of your DSP. EQ alters phase response. So if you EQ on the output side of your DSP (which is post crossover) after testing, you alter phase, meaning you're back to where you started...
Built:
DR 250: x 2 melded array, 2x CD horn, March 2012 plans.
T39's: 4 x 20" KL3010LF , 2 x 28" 3012LF.
WH8: x 6 with melded array wired series/parallel.
Bunter's Audio and Lighting "like"s would be most appreciated...

Post Reply