use phase cancellation instead of a crossover

[joessportster]

As you may or may not know I have been playing with  OB  speakers..................................Steve at Decware had an interesting comment  where he said If you reverse the phase on  your bass drivers in an OB the drivers will NATURALLY form  a crossover point

 

"Yes, these are intended to be bass baffles to pair with the Alnico Betsy speakers. My hope was to find a lower efficiency woofer with a steep natural roll-off that I could either series or parallel with the Betsy and in either configuration not use a crossover. Since the Alnico Betsy driver is more sensitive and has over 3 octaves higher response, even though both drivers are hooked together, the high frequencies will naturally go to the Betsy driver, and the low frequencies will naturally go to the bass driver. Any overlap is seamlessly canceled by running the bass drivers out of phase relative to the Betsy drivers."

 

Thoughts

[rplace]

No direct comments on the above as I'm not familiar with the Alnico Betsy, I'll have to google it.

 

I do however find, more often than not, on my 2-way active set up I need to invert the polarity of one driver. I find I typically get better results inverting the HF not the LF, measuring and everything else the same. Not really sure why that is perhaps, @Chris A could explain. That said, you probably know I am running Open Baffle bass bins with H-frames and 15" drivers. Would be curious to know the connection between OB and phase. I, for some reason, was under the impression that every "order" of the crossover changed things 90 degrees. So by the time you have a 4th order XO you have rotated it 90 X 4 for a total of 360 but it would/could be now out of phase....I'm still not exactly clear as to how all that works but that is my basic understanding.

[Chris A]

5 hours ago, joessportster said:

Any overlap is seamlessly canceled by running the bass drivers out of phase relative to the Betsy drivers."

Perhaps this statement is true...perfectly on-axis.  However, if the drivers are mounted non-coaxially, i.e., a separate woofer and a coaxial full-range driver on a baffle, you'll get very interesting polar response, like the woofers are out of phase from the HF driver (which they would be if you did what he's saying), and the on-axis bass midrange response will be not very good.  It will sound very strange, I think.  I find a $205 miniDSP 2x4 HD and two stereo amplifiers will work wonders--far beyond what the guy in that thread is proposing. 

 

In this pastime, that isn't very much...and the sound quality will be outstanding.

 

Chris

[glens]

On 5/12/2020 at 1:47 PM, rplace said:

I, for some reason, was under the impression that every "order" of the crossover changed things 90 degrees. So by the time you have a 4th order XO you have rotated it 90 X 4 for a total of 360 but it would/could be now out of phase....

 

At the nominal crossover frequency it's 45 degrees per element.  Just an inductor on the woofer and a capacitor on the next one up, at that crossover frequency is +45 one way and -45 the other, for a combined total of 90 difference.  Two elements each driver nets 180 difference at that point, etc.

 

360 apart is as good as 0 but for the one cycle one driver does first and the one the other does last, at the nominal crossover frequency.

 

But if my imperfect understanding is correct, each element causes a 90 degree phase shift eventually (in its greater attenuation end of things), implying that it exhibits less phase angle change at the greatest part of its pass band.

 

So say you had a symmetrical second-order crossover (180 apart) on a two-way system.  The high-pass goes 90 one way and the low-pass goes 90 the other direction.  If you used a pair of monstrous coil and cap to form a high-pass at 30 Hz to feed your woofer and appropriate coil and cap to low-pass your tweeter at 20 kHz thinking it'll provide 0 degrees difference at the woofer/tweeter crossover (essentially undoing the phase angle shifts there, for a form of "minimum phase passive crossover network" ), I believe you'll have wasted good time and money (and wasted amplifier function) for no net gain in that respect.  I reserve the right to be wrong, though.  A cap and coil each the size of a suitcase driving the woofer might be just the ticket...