Looking to put together a parallel notch filter for some Peerless 830869 drivers I have, to kill the peak and the associated sub-harmonic distortion.
Is this the right place to start? https://www.diyaudioandvideo.com/Calculator/ParallelNotchFilter/Help/
https://hificompass.com/en/speakers/measurements/peerless/peerless-830869
With f1 ~= 1800 Hz and f2 ~= 2700 Hz, I get values of roughly 66.7 uF, 1.67 mH, and 1.44 ohms. Does that make sense?
Is this the right place to start? https://www.diyaudioandvideo.com/Calculator/ParallelNotchFilter/Help/
https://hificompass.com/en/speakers/measurements/peerless/peerless-830869
With f1 ~= 1800 Hz and f2 ~= 2700 Hz, I get values of roughly 66.7 uF, 1.67 mH, and 1.44 ohms. Does that make sense?
Hmm... apparently there's a calculator that does the math for you, but the values don't change no matter what kind of target attenuation you're looking for. So much for that.
Also, does anyone know how the DCR of the inductor affects the result?
Also, does anyone know how the DCR of the inductor affects the result?
You can find it in a simulator by trial and error.
Keep in mind that a 0 degree peak is often not representative of power.
Keep in mind that a 0 degree peak is often not representative of power.
So, sure, we can take this on as an academic exercise, but as a practical matter, why? This driver appears to have a reasonably smooth response, and any rise in response at the top end is easily dealt with by the implied low pass filter, no? Depending on your HP filter position you may not even need a Zobel.
Exactly , Agree
OP, Depending on Tweeter used. It is a 8" So would have to cross around the usual 1000 to 1400 Hz
That peak / response/ cone breakup will disappear.
Fun part is finding the tweet, 3rd or 4th order helps with such a low crossover point.
For less distortion at higher levels. Tight range for many tweets, basically riding the Fs of a driver that can even dip that low.
Need Fs way down there like 600 to 700 Hz range
With same 3rd or 4th order on the 8" that peak goes Bye Bye
Last edited:
Yeah, the weird thing about this driver is that the excess HD (mainly HD5) between 300 Hz and 2 kHz does not seem to go down much with reduced signal level, and in a few areas is even higher with the lower signal level. Not many sample points, to be sure, but you can see from the HiFiCompass measurements that the HD5 is practically the same whether the signal is 2.83 V or 5.6 V. Extrapolating from that, I'm thinking it would be good to have the notch filter "just in case" the [active] filtered HF going to the driver still gives the same distortion products.
Passive notch component values would need to be entirely recalculated/simulated after your active low-pass filter has been applied; response ideally measured ofc.
Example here after applying a rando 2khz 2nd order filter
Example here after applying a rando 2khz 2nd order filter
Since the breakup peak is mostly at 0 degrees and many don't listen at that angle, the measured increase may not apply.
Correct, because it is H5 and not very audible.
The primary concern is H1 what we actually hear then H2
As with any speaker it is the reason for a crossover. We are using the bandwidth it is appropriate for.
Strange enough we usually see wideband distortion measurements of drivers. Be nice to see distortion with filter.
Would make more sense, basically removing what the speaker isnt designed to do.
All in All, agree. On having Virtuix Cad available
I wouldnt overthink H5 to much. I wouldnt spend more than 10 seconds worrying about it.
Since it would take about 10 to 30 minutes to design the filters and be done.
Adding the notch and whatever results it has with the overall phase. Is the actual concern, or what will be the most time involved
when tweeter phase is also in the picture, and of course the usual change with Baffle loading.
Im a big fan myself of 8" 2 ways. Having designed a few the story doesn't change much.
Most the challenge will be the phase response when integrating the tweeter. The woofer low pass will play a big role in that. Mainly vertical.
Horizontal is easy to get. Vertical is the tough one, with the spacing a large tweeter will present to cross so low.
All in good time, look forward to seeing the build once past the design phase. There is that magic feeling bolting nice cast frame drivers in place
Very good point. Is there a possibility that the peakiness might increase IMD as well? And those products might not be at high freqs?
Passive parallel notch doesn't seem to be doing the trick for me. Did you use a library block? When I try to either tune it or optimize it, it doesn't seem to do what I want to bring the peak down.
I think I see where you're headed with this. Modulation needs a nonlinearity in order to occur.
Ordinarily I consider IMD to be expected, being related to the underlying HD, and hence not audible in small quantities. On the other hand I don't think distortion is a problem for most modern drivers used within their design capabilities.