Have you done any impedance sweeps of both raw free air woofer and in box? This can help identify cabinet resonances that are disturbing cone movement to be damped. A raw sweep is to isolate inherent driver behaviour from box induced colouration.
Last edited:
I don’t have the dats system so i have made a cable jig a few years ago but god knows where it is… But you are right, i should absolutely find that thing and take some measurements
In a pinch it can be done with 2 resistors. 1k to act as the current source and 10 ohms to calibrate to 10 of something, then substitute the speaker. You can reduce the 1k if you have a high noise floor.
It looks as though you are suffering a little from the impedance curves. Ignore the Woofer break up mode for a little bit, and try these suggestions.
Consider using a plain old Zobel in the woofer. You want to reduce the woofer's impedance rise so it's flat, but does not go below the nautural minimum impedance of the woofer (Re). See the first chart in my blog about those here. Do the Zobel design in isolation of the other components. That is, just the woofer and Zobel network attached, maybe in a separate simulation file. Keep the Zobel closest to the woofer when you copy it to your main design file. Put nothing in between the Zobel and woofer.
A quick simulation with the Dayton supplied ZMA suggests 14uF with 8.2 Ohms or there about to be an excellent starting point.
You'll then need to rethink the woofer filter due to the new impedance curve. When you do, try adding 0.5 Ohm resistors to the even order filter components (they go to ground) to improve your impedance. Note that these can improve the impedance of the OTHER section. That is, adding a small value resistor to the L1 (in the HP filter) can improve the impedance of the woofer and vice-versa. 0.5 Ohms may be too much but they'll give you a good indicator if they are needed because the minimum impedance will jump up (a good thing) by large amounts (1-2 Ohms).
At some point near the end of this process consider a notch filter if it still seems necessary.
PS - Include DCR and ESR in your simulation of caps and coils, especially in these even order components they are critical. In these cases having 0 DCR is often not a good thing.
Consider using a plain old Zobel in the woofer. You want to reduce the woofer's impedance rise so it's flat, but does not go below the nautural minimum impedance of the woofer (Re). See the first chart in my blog about those here. Do the Zobel design in isolation of the other components. That is, just the woofer and Zobel network attached, maybe in a separate simulation file. Keep the Zobel closest to the woofer when you copy it to your main design file. Put nothing in between the Zobel and woofer.
A quick simulation with the Dayton supplied ZMA suggests 14uF with 8.2 Ohms or there about to be an excellent starting point.
You'll then need to rethink the woofer filter due to the new impedance curve. When you do, try adding 0.5 Ohm resistors to the even order filter components (they go to ground) to improve your impedance. Note that these can improve the impedance of the OTHER section. That is, adding a small value resistor to the L1 (in the HP filter) can improve the impedance of the woofer and vice-versa. 0.5 Ohms may be too much but they'll give you a good indicator if they are needed because the minimum impedance will jump up (a good thing) by large amounts (1-2 Ohms).
At some point near the end of this process consider a notch filter if it still seems necessary.
PS - Include DCR and ESR in your simulation of caps and coils, especially in these even order components they are critical. In these cases having 0 DCR is often not a good thing.
Last edited:
You are probably well past these problems, but for fun I went ahead and implemented the Zobel I suggested and your original crossover values the woofer response and impedance now seems pretty smooth:
Last comment, honest. Given the dimensions of the mid-woofer (5") I'd attempt a crossover closer to 3 kHz if the driver response allows. You'll still have excellent off-axis response and improved power handling.