Regardless of the need or wisdom of driver phase reversal, the break before make issue is important and the solution is simply to wire the amp/input to the DPDT throws and the speaker to the poles. If one pole happens first, then the pole (speaker) shorts, but not the throws (amp).
Wrt the effect on response, real speakers and room acoustics are so far from flat that the transition glitch is barely noticeable. The OP suggested a first order filter where there should be no need for phase reversal. If you care about that level of accuracy, then you need some way of positioning the tweeter behind the plane of the mid/bass. This is like many other modifications that are done because they are easy and not because they make any/much difference, ie things a professional would never waste time doing.
Wrt the effect on response, real speakers and room acoustics are so far from flat that the transition glitch is barely noticeable. The OP suggested a first order filter where there should be no need for phase reversal. If you care about that level of accuracy, then you need some way of positioning the tweeter behind the plane of the mid/bass. This is like many other modifications that are done because they are easy and not because they make any/much difference, ie things a professional would never waste time doing.
I think we have a number of people contributing to this thread who need to read up on first-order crossovers. 🙂
The facility for easy flipping of the wiring polarity is an unrelated topic, really.
Dave.
That would definitely appear to be the case.
Next time consider the big picture of a query like this before you launch into knee-jerking ignorant comments.
Dave.
What I belive is that fixing some (non-existent) lobing issue by reversing the tweeter polarity will result in uneven fr response on axis, which is to me more important. Flat on axis, well behaved of axis is primary goal. Lots of compromises goes into speaker design. But flat on axis should not be compromised. Dunlavy used first order on the tweeter, and he worked on flat response and best impulse as primary goal.
Tweeter polarity switch was not the feature. As I said, you can connect it right way or wrong way. Cheers.
Tweeter polarity switch was not the feature. As I said, you can connect it right way or wrong way. Cheers.
Tweeter polarity reversal does not result in uneven amplitude response, nor will it change the summed frequency response in a first-order system. Assuming it was correct to begin with, it will be flat on axis with either polarity.
You might consider one way 'wrong' and one way 'right', but it's more complicated than that with a first-order system
As I mentioned, some people (you) need to do some brushing up on first-order crossovers.
Dave.
You might consider one way 'wrong' and one way 'right', but it's more complicated than that with a first-order system
As I mentioned, some people (you) need to do some brushing up on first-order crossovers.
Dave.
Many HT surround speakers switching tweeter polarity for Dipole or Bipole.
Well, no. If there is no attenuating resistor in the high-pass tweeter filter arm, audio content wil pass through the cap - that is esentially zero ohm load (short circuit) to the amplifier, although for a very brief time when switching. Is this a real danger to the amplifier? I think it is not wise to try this experiment - break before make switch is exactly the same price as make before brake switch.
Depends on the amplifier and how it copes with a brief capacitive load ie. If its prone to oscillation. It matters whether the switch is before or after the crossover (after isn't recommended), the size of cap, whether a signal is present when switching and how high the signal is to the tweeter. I personally would only switch polarity when the signal is very low or muted, but from my experience most dpdt switches will break before make anyways, so i wouldn't worry too much.
With a first-order cross-over, the low-pass output has a response 1/(s tau + 1) and the high-pass output s tau/(s tau + 1). Add them in phase and you get (1 + s tau)/(s tau + 1) = 1, a perfectly flat response with no phase shift. Do it in antiphase and you get (1 - s tau)/(s tau + 1). That's the transfer function of a first-order all-pass filter: perfectly flat magnitude response, but two times the phase shift of a first-order low-pass with the same corner frequency. So assuming a first-order filter and not too much extra phase shift due to the drivers, both would result in a flat response.
Do the same calculation for a second-order Linkwitz-Riley filter and you get the choice between a dip to 0 or an all-pass response.
The point is that the outputs of even-order filters are in phase or in antiphase, while those of odd-order filters are an odd multiple of 90 degrees out of phase (assuming normal filters with equal denominator polynomials and 1 in the low-pass numerator, (s tau)n in the high-pass numerator). You can therefore get the even-order outputs to cancel when you use the wrong polarity, but not the odd-order outputs: you get the same magnitude with a phase difference of +90 degrees as with -90 degrees.
This is still assuming driver phase response doesn't mess things up, which depends on the drivers as well as their placement (acoustic centres aligned or not).
The point is that the outputs of even-order filters are in phase or in antiphase, while those of odd-order filters are an odd multiple of 90 degrees out of phase (assuming normal filters with equal denominator polynomials and 1 in the low-pass numerator, (s tau)n in the high-pass numerator). You can therefore get the even-order outputs to cancel when you use the wrong polarity, but not the odd-order outputs: you get the same magnitude with a phase difference of +90 degrees as with -90 degrees.
This is still assuming driver phase response doesn't mess things up, which depends on the drivers as well as their placement (acoustic centres aligned or not).
Last edited:
With an odd-order filter, due to the fact that the filter outputs are never in phase, no matter what the tweeter polarity may be, you also get a tilted main lobe in the crossover frequency region (assuming just one woofer and one tweeter, one placed above the other, no D'Appolito configuration or coaxial placement). The signals from the woofer and tweeter are in phase (or at least closer to being in phase) when you measure in some weird angle where path length differences compensate (or partly compensate) for the 90 degrees phase difference. With the switch, you could choose in what direction the lobe tilts.
Thanks all for the replies and contribution.
The reason I'm asking is because, even I do not measure huge holes in the measured response (in room!), the listening feeling changes a lot in the two positons and I still do not understand which I prefer. I think that loudspeaker positioning, the listening room, the position of my head and the "real" tweeter offset are messing any my tempt of simulation and all these are influencing the mid-high response (3D soundstage, listening fatigue and brilliance) even in a perfect (?!) crossover design (first order for the tweeter). Hence the idea of the switch.
The point was to prevent the amplifier from seeing a direct short at DC but I see where you're coming from.
All good in theory. Besides the electric first order crossover, tweeters roll of acoustically (by design), so the resulting slope is higher order.
I measure in close proximity (ft or less) on tweeter axis when optimizing crossover. In the room response is totally different thing.
What matters is that the phase difference between woofer and tweeter is about 90 degrees in the frequency range where the sound of the woofer and the sound of the tweeter are roughly equally loud. If that's the case, then the tweeter polarity won't affect the magnitude response much, only the phase response and the lobing.
No, not if the electrical filters are designed to achieve acoustical first-order slopes. (That's what we've been talking about.) 🙂
Goodness, this whole topic seems to confound you.
Dave.
This sounds a little bit weird to me, but I will give it a try as well and let you know, thanks.
See photo 3 if it sounds weird to you. Unless you have anechoic chamber.
https://audioxpress.com/article/how...-speaker-response-without-an-anechoic-chamber
https://audioxpress.com/article/how...-speaker-response-without-an-anechoic-chamber