In a second order filter, I know that it should be reverse the polarity of one driver to prevent a deep hole in crossover region. But what if the two drivers are designed to already have a hole at crossover point. Says the drivers are not meet at crossing point, to create some eq. For instant, woofer is crossed at 2kHz and tweeter is crossed at 3kHz — with the same Q — the crossover point should meet at 2.5 kHz. As can be seen, there’s an absolute hole on a crossing region. So, does it need to reverse the polarity of one driver?
Reversing polarity will not fill the hole in this case.
If you have the driver phase arranged all the same, then you will get the hole expected, rather than an even larger one. In this case global EQ will make corrections to both drivers at once.
If you have the driver phase arranged all the same, then you will get the hole expected, rather than an even larger one. In this case global EQ will make corrections to both drivers at once.
In a XSim simulation, I found a little difference between reversed and non-reversed polarity for a dip response there. But in reality, I had tried to reverse the polarity of one driver on my DIY speaker and I found a larger difference. The crossover point is at about 650 Hz for woofer and midrange — with non-meet crossover point. Frequency around 400-500 Hz and 800-900 Hz seem to be doubled when not reversing. On the other hand, The 650 Hz seems to be boosted when reversed polarity. Since these are DIY speakers and without measurement, I couldn’t sure for the accuracy and precision for what I’ve heard and the theory. So could anyone please help to ensure me? And please share your opinion. Thanks in advance.
Last edited:
This is because your phase responses do not follow the same slope. One is steep and one is shallow.
This is when the driver response is second order. The filter is just part of the story.
In the original statement, I intended to say about the result at frequencies nearby crossover point. Values of Hz are only approximate.
Do you mean if the slopes of both driver response are +/- 12 dB/octave, there will be 180-degree phase shift and require reversing polarity on one driver? If not at +/- 12 dB/octave, says woofers at -10 dB/octave and tweeter at +11 dB/octave, there won’t be 180 degrees phase shift and may not require reversing the polarity?
Yes..
Sometimes the response will look right but will be different. It can be difficult to see, but easier to measure or predict in a sim.
1. Xo around 500Hz is pretty difficult to measure indoors, because of reflections
2. Spl summation depends on slopes and pathlength to mic - basically it is about phases
3. it is a 3D phenomenom, design axis and mic distance at simulation or measurement give just one thruth. When phases don't match off-axis spl response may be better or worse than desing axis/spot.
When phases don't match well, spl response shows untypical shape and undulations
More about this Woofer crossover & offset
2. Spl summation depends on slopes and pathlength to mic - basically it is about phases
3. it is a 3D phenomenom, design axis and mic distance at simulation or measurement give just one thruth. When phases don't match off-axis spl response may be better or worse than desing axis/spot.
When phases don't match well, spl response shows untypical shape and undulations
More about this Woofer crossover & offset