Analog Subwoofer Phase Control - How?

[ocinn]

Hi there,

I am not an electronics expert by any means, but something has been bugging me for a few years now. I have owned many powered subwoofers that have a fully variable 0-180deg phase knob (ex: Jamo E6SUB)

I’ve found schematics for RF phase shifters but from what I’ve read it’s exceedingly more difficult at lower (audible) frequencies.

I was curious (in intermediate-level-terms) as to how this is implemented in circuitry, and if it is truely a phase (frequency dependent) shift or a fixed (frequency independent) delay translated to “approx” phase at the crossover frequency, (bucket brigade device, digital chip?)

This also brings up the question of how phase inverting switches work on single ended input subwoofers, im assuming converted to differential and then flipped, and then sent to the amp?

 

[KSTR]

Phase control in analog subwoofers typically consists of a polarity switch and a phase-shifter circuit, comprising two simple 1st-order active allpass filters whose corner frequency is adjusted. The allpass filter allows for phase shift at the crossover frequency from (close to) zero degrees to (more than) 180 degrees, and the polarity flip then makes the whole 0..360 (or +-180) degree span available.

But, phase control is always a quirky compromise and often does not help for proper integration because of the additional delay. You can easily end up with seemingly matching phases but the subwoofer lagging behind by one full 360 degree cycle (or more when the sub is farther away from the listening position than the main speakers).

Timing properties of the classical 80Hz sub-to-mains crossover are already bad even without such an additional phase shift / delay and with that "compensating" phase shift you may end up with proper addition of signals around the XO frequency as long as they are stead-state continuous. But bass transients (kick drums, plucked upright bass, etc) will get dispersed, stretched in time, with the "thump" lagging behind.

IMHO, there is no way to proper subwoofer integration other than doing it like an accomplished speaker designer would do it if the sub were already integrated part of the speaker.

 

[ocinn]

Phase control in analog subwoofers typically consists of a polarity switch and a phase-shifter circuit, comprising two simple 1st-order active allpass filters whose corner frequency is adjusted. The allpass filter allows for phase shift at the crossover frequency from (close to) zero degrees to (more than) 180 degrees, and the polarity flip then makes the whole 0..360 (or +-180) degree span available.

Ok, I’m starting to understand the picture here now.

To clarify, this is a true phase shift and not a constant delay? And the polarity switch you mentioned must’ve required a differential conversion?

Apologies, I grew up in the digital era so analog signal processing circuitry is my Achilles heel.

But, phase control is always a quirky compromise and often does not help for proper integration because of the additional delay

Agreed in an idealist vacuum, however when I was a teenager, and didn’t own DSP gear etc, the knob was quite useful to ballpark (using band-passed pink noise and a measurement mic) general integration to the speakers at the crossover point. I look back on it now and realize how “technically” flawed it was, but it was better than having a sub (ex.) 90deg out of phase at the crossover.

 

[KSTR]

To clarify, this is a true phase shift and not a constant delay? And the polarity switch you mentioned must’ve required a differential conversion?

A first order allpass phase-shifter is also constant group delay -- but only to a certain frequency much lower than the RC-defined corner frequency. For a cascade of two such filters it gets even worse... but the point is that we only need to match phases at and slightly around the crossover frequency so it doesn't matter much.

Polarity switch requires only one inverter OpAmp (and yes, this makes up for a local differential signal) and a select switch.

Agreed in an idealist vacuum, however when I was a teenager, and didn’t own DSP gear etc, the knob was quite useful to ballpark (using band-passed pink noise and a measurement mic) general integration to the speakers at the crossover point. I look back on it now and realize how “technically” flawed it was, but it was better than having a sub (ex.) 90deg out of phase at the crossover.

Yep, the usual analog phase shifter is a non-ideal workaround but just as you say it does work sort of OK, especially to avoid deep nulls from cancelling when mains and sub happen to be ~180 degree apart at the XO frequency range.

If you have DSP and measurement tools and skills, a carefully aligned digital crossover with true time delay compensation and linear phase is king, and not by a small margin.

Some analog implementations are based on the sub "knowing" which mains it will run with and therefore a good sub-to-mains crossover which truly realizes the acoustic Linkwitz-Riley 4th order frequency split can be tailored by the designer. But it cannot cope with larger path-length differences which would require to really do a time-delay on either the sub(s) or the mains. Luckily, the 4th order Linkwitz-Riley is the most robust filter with regard to those issues.

 

[egellings]

Hi there,

I am not an electronics expert by any means, but something has been bugging me for a few years now. I have owned many powered subwoofers that have a fully variable 0-180deg phase knob (ex: Jamo E6SUB)

I’ve found schematics for RF phase shifters but from what I’ve read it’s exceedingly more difficult at lower (audible) frequencies.

I was curious (in intermediate-level-terms) as to how this is implemented in circuitry, and if it is truely a phase (frequency dependent) shift or a fixed (frequency independent) delay translated to “approx” phase at the crossover frequency, (bucket brigade device, digital chip?)

This also brings up the question of how phase inverting switches work on single ended input subwoofers, im assuming converted to differential and then flipped, and then sent to the amp?

Use all-pass filters. You can cascade two first order ones to get nearly 180 degrees of phase shift between the input and the output.