I am designing a crossover, I took frd and impedance measurements and modeled the crossover in Passive Crossover Designer from the FRD consortium to a pretty flat theoretical response. (2nd order L-R @400hz target)
I checked the woofer response with the LPF alone, and I checked the mid response with the HPF filter alone, and they both match the predicted response. When I hook them together, I get awful frequency response with -20db troughs at about 40hz and 90hz. Switching polarity on the mid doesn't make much difference. Using a cap alone (1st order HPF) on the mid causes the same problem. Removing the HPF altogether from the mid eliminates the frequency troughs, but the mid is trying to output the lows and would not be able to handle much power without any filter. So it seems there is some resonance between the cap on the mid filter and the woofer.
Any ideas how to eliminate this problem?
I checked the woofer response with the LPF alone, and I checked the mid response with the HPF filter alone, and they both match the predicted response. When I hook them together, I get awful frequency response with -20db troughs at about 40hz and 90hz. Switching polarity on the mid doesn't make much difference. Using a cap alone (1st order HPF) on the mid causes the same problem. Removing the HPF altogether from the mid eliminates the frequency troughs, but the mid is trying to output the lows and would not be able to handle much power without any filter. So it seems there is some resonance between the cap on the mid filter and the woofer.
Any ideas how to eliminate this problem?
Are you talking about the simulated response when you say -20db troughs? Or is it the measured response?
There should be good phase tracking between the two drivers through the crossover region and until individual driver outputs have fallen to -25 to -30db. The summed response is a vector sum, so the phase of the individual drivers also matter if you want flat summing. Therefore, even if you get the slopes to match the targets, the sum may not be flat if the phase angles between the two drivers are far apart at or near the crossover.
A second order HPF at 400 Hz will give -12db at 200 Hz and -24db at 100 Hz. So, troughs of -20db at 90 Hz don't really make sense, unless they are though some weird interaction like you suspect.
There should be good phase tracking between the two drivers through the crossover region and until individual driver outputs have fallen to -25 to -30db. The summed response is a vector sum, so the phase of the individual drivers also matter if you want flat summing. Therefore, even if you get the slopes to match the targets, the sum may not be flat if the phase angles between the two drivers are far apart at or near the crossover.
A second order HPF at 400 Hz will give -12db at 200 Hz and -24db at 100 Hz. So, troughs of -20db at 90 Hz don't really make sense, unless they are though some weird interaction like you suspect.
Last edited:
Is your cross-over network parallel or series? Most likely it is parallel, but it doesn't hurt for me to ask. If parallel, then it would seem highly unlikely that there is some kind of electrical interaction between the hp filter (cap) and the woofer.
Is your crossover network on a PCB? If so, is it possible that the PCB has a short or open connection on it? Are you perhaps making some kind of wiring error when you combine HP and LP filters?
Could the problem perhaps be too large of a distance between the mid and woofer?
Hope these questions might help you to solve the problem,
Pete
Is your crossover network on a PCB? If so, is it possible that the PCB has a short or open connection on it? Are you perhaps making some kind of wiring error when you combine HP and LP filters?
Could the problem perhaps be too large of a distance between the mid and woofer?
Hope these questions might help you to solve the problem,
Pete
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.