3 way - how to measure woof/mid xo point?

[Dave Bullet]

I am looking to build a 3 way speaker with a passive crossover.

My issue is to how to obtain correct relative phase between woofer and midrange for the woof/mid crossover point (and obviously at least an octave or 2 on either side to assess phase tracking through the crossover).

A gated farfield measurement in my largest room is no good as I can only get down to about 200Hz with very coarse resolution.

The only solutions I can come up with are:
a) Perform ground plane measurements outside
b) Perform near-field on both woofer and midrange, then hilbert transform the FR curves to get the phase data.

I haven't attempted a) yet (weather issues).

Regarding b) - my concerns here are:
1. How to specify acoustic offset with two nearfield measurements. I don't think this is going to work
2. Even if I could apply some sort of meaningful acoustic offset - how would I integrate the highpass part of the midrange curve (esp. phase) with the gated farfield response at the upper end of the midrange crossing over to the tweeter?

Thanks,
David.

 

[augerpro]

I use both nearfield and farfield, then merge and then use the HBT. As long as the HBT phase agrees with the farfield phase the phase constructed at lower (nearfield merged) frequencies should be correct I *believe*.

 

[Dave Bullet]

Thanks,

So just to clarify....
1. I take a gated farfield measurement of the midrange
2. I take a nearfield of the woofer
3. I run a HBT on the midrange

if phase from 1. compares to phase (at the lower end where the mid will crossover to the woofer) for 3. (ie. after doing an HBT) - then I can use the HBT generated phase for the mid/woof xo point... and should be able to use actual measured phase (gated farfield) for the mid/tweeter xo point?

Cheers,
David.

 

[Dave Bullet]

Sorru augerpro - re-read your post.

You mean
1. take a woofer nearfield measurement
2. take a woofer farfield measurement
3. merge the near and far field of the woofer
4. do an HBT on the merged response (3.)

If the phase generated in 4. compares with 2... then use 4 for xo modeling with the midrange.

Is that right?

Cheers,
David.

 

[augerpro]

Sorru augerpro - re-read your post.

You mean
1. take a woofer nearfield measurement
2. take a woofer farfield measurement
3. merge the near and far field of the woofer
4. do an HBT on the merged response (3.)

If the phase generated in 4. compares with 2... then use 4 for xo modeling with the midrange.

Is that right?

Cheers,
David.

yes. Just make sure that the HBT phase teh merged response is the same as the measured farfield portion of that merged response. Obviously the HBT constructed phase will not agree with the nearfield measurment. But as long as the HBT phase agrees with the farfield above the merge point then it will be correct below it also.

 

[Dave Bullet]

thanks - that makes sense.

One last question - could I do the same with the midrange?

I mean - if I can only measure down to say 200Hz and that is poor granularity, then could I do an HBT on the midrange and if the farfield midrange phase looks good above the planned xo point compared to the HBT generated one, then below should be ok.

Therefore, I would be using two HBT phase generated FR plots for xo modeling... factoring in acoustic offset of both drivers too of course.

I will try groundplane outside as a preference.

Thanks,
David.

 

[augerpro]

Yes I would HBT both the mid and woofer.

Groundplane and/or a long gate would be a good way to verify the truth, even if the long gate is a little jagged.

 

[Jay_WJ]

When you perform the HBT, it is always a good idea to extend your measured FR at both ends so that it can simulate the driver's roll-off to extreme frequencies.