Phase

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Drivers have whats known as the acoustic centre, this is where the waveform is created.

This acoustic centre can differ between two unlike drivers on a flat baffle. If you ask both drivers to produce a frequency at your desired Xover point, due to the difference in acoustic centre, one wave form could lead the other one by a significant amount.

This lead will automatically mean your two drivers are NOT perfectly in phase with one another at your desired Xover frequency.

Crossovers on the other hand alter the phase of the signal too. The job of the designer is to measure the relative difference between the acoustic centres of his or her drivers, then design a crossover that suits the drivers with regards to xover frequency, but at the same time make sure that the drivers are also phase coherent.
 
rinx said:
Take a look here, there is measured phase, so tell how the phase should be around XO point. Is that good on this picture?

http://www.geocities.com/woove99/Spkrbldg/DesigningXO_Appendix.htm

Yes, the phase is aligned well in that figure.

The acoustic phase used in my simulation is not measured one but a minimum phase predicted by the Hilbert-Bode transform (HBT) from a modeled frequency response. This technique is simple to use and known to be accurate enough for crossover simulation AS LONG AS the frequency response is modeled properly. Your FR to be used for the HBT should have rolloffs at both high and low extreme frequencies, which mimicks the driver's actual rolloffs. Read my xo design notes for further information.
 
Given that the absolute phase of drivers is difficult to obtain while HBT is very accurate, I have been thinking about doing XO design in this way. I think the trick is to find the accurate, relative acoustic centre offset of the drivers:

Measure individual drivers. Derive the individual phase of the drivers from HBT. Design and build an approximate XO. Measure the sum response of the XO network (whether it sums ruler flat or not is not important at this stage). Use SpeakerWorkshop and assume some driver offsets then simulate the response from the XO network. Try setting different driver offsets, until the simulated response matches that of the measured sum response with the XO. Now we have obtained the proper relative acoustic centre offset of the drivers. Looking in another way, we have obtained the accurate phase of all the drivers. We can now design the final XO network.

Any fault?
 
Your idea is plausible. But let me tell you one thing. If you can measure physical distance from a driver's frame surface to a point where the cone meets its spider, it will be a good estimate of the driver's acoustic center. If you design a XO with this estimate, you'll know that a small inaccuracy (e.g., .2 to .3 inch) of an acoustic offset won't much affect your xo simulation result. Keep in mind that phase alignment doesn't need to be perfect as long as you can obtain your desired FR at your listening position.
 
HiFiNut,

Since you can measure, do both methods in SW and check that the final responses with crossover derived appear the same.

Conversely, use the actual phase measured by SW to verify the acoustic offset you calculate when you HBT the measured FR curves to derive the same final FR response of the measured XO sim.

Hope that makes sense,
David.
 
Is there a way to align phase in an active or active-digital crossover?

Specifically, I'm wondering if FIR filters can do this.

It seems there's more to phase matching than delay correction between tweeter/woofer... is that right?

(someday I will understand the terms minimum, linear, and maximum phase in the context of FIR filters, but not today)
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.