"1st electrical filter" means nothing. It's the combined response of driver and filter that matters and that should be adjusted to the target functions, whatever that may be.
A true 1st order acoustic crossover is almost impossible to realize (which requires the drivers also to be time-alinged mechanically, typically by recessing the baffle for the tweeter or having a waveguide) mainly because of the phase shifts that the driver rolloffs produce which kick in way before the frequency response starts to fall.
Therefore, with 1st order electrial filters often the result will be a 2nd order filter overall, and that one requires the tweeter to be inverted. And with time-of-flight not compensated you may or may not have to flip the tweeter again and "de-tune" the filters so that they still give a good summed frequency and phase response.
That's why you see either tweeter polarity in actual speakers with claimed "1st order crossovers" which most often they are not, acoustically.
EDIT: This can be tested by flipping tweeter polarity: in a true 1st order response the polarity doesn't matter, only phase is affected (where only one polarity gives a linear-phase "transient-perfect" crossover). If you get a notch or a peak at the XO freq then it's not 1st order (no consistent 90 degree phase offset between drivers around the XO freq).
A true 1st order acoustic crossover is almost impossible to realize (which requires the drivers also to be time-alinged mechanically, typically by recessing the baffle for the tweeter or having a waveguide) mainly because of the phase shifts that the driver rolloffs produce which kick in way before the frequency response starts to fall.
Therefore, with 1st order electrial filters often the result will be a 2nd order filter overall, and that one requires the tweeter to be inverted. And with time-of-flight not compensated you may or may not have to flip the tweeter again and "de-tune" the filters so that they still give a good summed frequency and phase response.
That's why you see either tweeter polarity in actual speakers with claimed "1st order crossovers" which most often they are not, acoustically.
EDIT: This can be tested by flipping tweeter polarity: in a true 1st order response the polarity doesn't matter, only phase is affected (where only one polarity gives a linear-phase "transient-perfect" crossover). If you get a notch or a peak at the XO freq then it's not 1st order (no consistent 90 degree phase offset between drivers around the XO freq).
Last edited:
So the answer is really: if you have two drivers with straight FR say 4 octaves in their (to become) passband and the acoustic centers of the two drivers is in the same vertical plane, then you shall connect the two drivers "in-phase" i.e. the terminal for each of the drivers that push the cone outwards when applying a positive voltage, shall both be connected to the same polarity. Then a first orderfilter is phase linear. It believe it dosen't sum correctly (straight) however. The perfect filter re: phase and level (given ideal drivers) is a 2:nd order with reversed polarity of one of the drivers, supplemented with a "filler driver" ala' B&O.
//
//
Sorry I miss to say that the drivers centers are aligned.
But I didn't realize the effect of roll-off on the relative phase.
Someone today in another post reminded me of the 2 Way Series crossover.
i.e.:
8 Ohm Input Terminal connects to the tweeter and inductor;
The other ends of the tweeter and inductor, connects to the woofer and capacitor;
The other ends of the woofer and capacitor connect to the Common Input Terminal.
For clarification:
The tweeter and inductor are in parallel; network 1
The woofer and capacitor are in parallel; network 2
These two networks are connected in series.
Most commonly, the tweeter has a level compensating resistor in series with it,
and the inductor connects across the series combination of the tweeter and resistor.
I tried that, but the tweeter and woofer were not time aligned. So I can not tell you which way of tweeter and woofer polarity will work best for you.
However, there may be more phase shift in a one pole Series crossover than in the more Traditional one pole cross over.
Try it, you may like it.
i.e.:
8 Ohm Input Terminal connects to the tweeter and inductor;
The other ends of the tweeter and inductor, connects to the woofer and capacitor;
The other ends of the woofer and capacitor connect to the Common Input Terminal.
For clarification:
The tweeter and inductor are in parallel; network 1
The woofer and capacitor are in parallel; network 2
These two networks are connected in series.
Most commonly, the tweeter has a level compensating resistor in series with it,
and the inductor connects across the series combination of the tweeter and resistor.
I tried that, but the tweeter and woofer were not time aligned. So I can not tell you which way of tweeter and woofer polarity will work best for you.
However, there may be more phase shift in a one pole Series crossover than in the more Traditional one pole cross over.
Try it, you may like it.
As the phase shift sum is around 90° (af fXO), it should not an issue reverse or not the polarity. (-90°=90°) If it is less, as you said for certain frequencies closer to fXO, then positive connection should be slightly better.
Far from fXO we have 2nd order behaviour, thus far from fXO the reverse polarity is preferable.
Did I summarize correctly?
So at the end if the speakers are physically time aligned and the drivers in the box are sufficiently linear away from fXO, positive connection should win.
Far from fXO we have 2nd order behaviour, thus far from fXO the reverse polarity is preferable.
Did I summarize correctly?
So at the end if the speakers are physically time aligned and the drivers in the box are sufficiently linear away from fXO, positive connection should win.
I went through this on series xo in the 2-way reference thread. Different versions of the xo had different polarity and you have to model or measure it to check.
diyAudio reference speaker project - Page 5 - diyAudio
diyAudio reference speaker project - Page 5 - diyAudio
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.