A question about phase

[activexp]

I've made a 3-way crossover where:

Low pass = 2nd order
Band pass = 1st order
High pass = 2nd order

My question is which driver should have the polarity reversed?

I know that with a 2-way 2nd order the tweeter should be reversed and with a 3-way 2nd order the mid reversed.

My XO is a both 1st and 2nd so I'm a bit confused. Logic suggests that it should be the tweeter.

Any advice greatly appreciated.

Steve

 

[jayb77]

Well honestly, and know this is probably what you already knew, but really you should just take some time and play with the design. Build a proto box or whatever needs to be done to not damage the finished cabinet assuming youve made it already, and just listen to some test tracks and see which configuration works best. Its hard not know anything more than your network and nothing more to make a accurate decicion and I dont want to be responsible for it sounding bad but a 1st order will only change you phase by 90 degrees regardless of the notion that it doesnt change phase concept. 1st order does change phase. So 1st 90 degrees and a 2nd order changes 180 degrees so electrically it shouldnt matter but! acoustically is where most of us DIYers lose it because some designs regardless of phase shift of the cross over require wiring something out of electrical phase. Things like the drivers not being aligned and whether the the drivers are flush mounted vs receesed and there way too many to discuss, but I recomend just starting with the tweeter or the woofers because those are the main drivers that usaully benefit from a phase shift , but also I would change the 1st order back and forth to just see if it makes a "night and day" difference. There is relly no stright answer I can give as definative do this not that answer. You will just have to do some R & D on this one. Sorry I couldnt help more!

 

[Lynn Olson]

Unfortunately the crossover that matters (the one you can measure and hear) is the acoustic crossover, which is the sum of the electrical network and the driver responses. If the drivers were precisely flat, had bandwidths far wider than the intended frequency ranges, and had purely resistive impedances, the electrical networks you have would work perfectly.

No such driver exists - not even in principle. All drivers deviate from flat response, and commonly behave badly at the edge of the working range (which is why a crossover is necessary in the first place). Nearly all direct-radiators have significant reactances (variations in impedance) that strongly interact with the crossover. This means a "textbook" crossover cannot work as designed; the system has to be measured and the crossover adjusted to the behavior of the real drivers. This also includes the polarity of the drivers themselves; the correct polarity has to be determined by measurement.

 

[activexp]

Well honestly, and know this is probably what you already knew, but really you should just take some time and play with the design. Build a proto box or whatever needs to be done to not damage the finished cabinet assuming youve made it already, and just listen to some test tracks and see which configuration works best. Its hard not know anything more than your network and nothing more to make a accurate decicion and I dont want to be responsible for it sounding bad but a 1st order will only change you phase by 90 degrees regardless of the notion that it doesnt change phase concept. 1st order does change phase. So 1st 90 degrees and a 2nd order changes 180 degrees so electrically it shouldnt matter but! acoustically is where most of us DIYers lose it because some designs regardless of phase shift of the cross over require wiring something out of electrical phase. Things like the drivers not being aligned and whether the the drivers are flush mounted vs receesed and there way too many to discuss, but I recomend just starting with the tweeter or the woofers because those are the main drivers that usaully benefit from a phase shift , but also I would change the 1st order back and forth to just see if it makes a "night and day" difference. There is relly no stright answer I can give as definative do this not that answer. You will just have to do some R & D on this one. Sorry I couldnt help more!

Many thanks for your input jayb77. Judging by a thread on here regarding the significance of phase it appears that the jury is out on this one.

At the moment all 3 drivers are wired normally and the sound is OK but I feel that it could be a little more 'out of the box'. For a start I'll reverse the tweeter and see if I can hear any difference, probably not!

 

[activexp]

Unfortunately the crossover that matters (the one you can measure and hear) is the acoustic crossover, which is the sum of the electrical network and the driver responses. If the drivers were precisely flat, had bandwidths far wider than the intended frequency ranges, and had purely resistive impedances, the electrical networks you have would work perfectly.

No such driver exists - not even in principle. All drivers deviate from flat response, and commonly behave badly at the edge of the working range (which is why a crossover is necessary in the first place). Nearly all direct-radiators have significant reactances (variations in impedance) that strongly interact with the crossover. This means a "textbook" crossover cannot work as designed; the system has to be measured and the crossover adjusted to the behavior of the real drivers. This also includes the polarity of the drivers themselves; the correct polarity has to be determined by measurement.

Thanks Lynn for your explanation regarding textbook crossovers and realise that without taking measurements they are a compromise. At the moment I lack the necessary gear and just use trial and error until I get the sound I'm looking for. Not ideal I know. 🙄