Active low pass filter phase

gupsta

Member
2008-04-30 9:39 pm
In my current setup I'm using an 8inch Seas full range and 15inch woofer in open baffle, with a 1st order passive crossover crossed at 200hz. Unfortunately the the roll off on the woofer is no where near steep enough. Rather than waste money expensive passive parts,I'm thinking of going active.


If I was to leave the full range driver running full range and build just the low frequency crossover part of a typical 24db active xo, would the woofer roll off and phase shift still be the same i.e 24db/octave and 360 degrees ?
 
As long as going active...

You will probably want at least a -6dB/octave HPF on full ranger to reduce distortion from cone movement for bass. That combined with natural roll off from the full range will probably give -12dB overall and that will electro acoustically form something closer to a Harsch XO, if you were to time delay your full range by half the period of the XO frequency. This would have not an360 deg phase shift but something closer to 60 deg bump and flat otherwise. The delay in te full range can be achieved with offsetting the tweeter back on the baffle or a waveguide. I would suggest a higher xo freq like 300Hz to 500Hz. You don't want 360 deg phase shift - music loses time coherency on percussions.

Read more in Harsch Xo thread.

http://www.diyaudio.com/forums/multi-way/277691-s-harsch-xo.html

Don't work XO's blind - use a mic and measurment software like REW. The model your XO in PCD. Then use minidsp.
 

gupsta

Member
2008-04-30 9:39 pm
Thank you for the advice, after which I started reading the S Harsch XO, and now I've just purchased a nano digi 2x8. I have two dacs (Audio Note and Teac) which are completely different. Hopefully that wont mess up the timming.

If the crossover is done in the digital domain will there still be any phase shift ?

I don't grasp what causes the phase shift. I presumed (no doubt incorrectly) that it is something directly related to the use of capacitors/inductors.

Dave
 
Regardless of analog or digital, phase shifts are present for all filters - that is just physics and math of a Butterworth, Linkwitz Riley, etc. You will need to implement FIR filters to not have phase shifts, but those require non real time DSP so there is latency. One method is to use IIR filters (BW, LR, etc) which have phase shifts, the use DSP convolution to reverse and flatten the phase shifts. There are entire threads on this.

A true first order symmetric Butterworth, is the only one with no phase shift. But achieving true 1st order BW is very difficult in practice. There are only a handful of speakers that do this.

http://www.diyaudio.com/forums/full...on-what-can-we-do-dsp-power-now-availabl.html

http://www.diyaudio.com/forums/mult...hase-linearization-eq-fir-filtering-tool.html
 
Last edited:
In my current setup I'm using an 8inch Seas full range and 15inch woofer in open baffle, with a 1st order passive crossover crossed at 200hz. Unfortunately the the roll off on the woofer is no where near steep enough. Rather than waste money expensive passive parts,I'm thinking of going active.


If I was to leave the full range driver running full range and build just the low frequency crossover part of a typical 24db active xo, would the woofer roll off and phase shift still be the same i.e 24db/octave and 360 degrees ?

The ultimate rolloff (slope) of the full range depends on what you are doing with it. Open baffle, Closed box, Vented, Horn give 18dB, 12dB, 24dB, 24dB per octave ultimate rolloff, respectively.

While it seems like you should match slopes, asymmetric slopes are not uncommon when crossing over to a subwoofer. If you are simply doing a FAST system and the crossover point is low (around 100Hz) then just use the 24dB/oct active crossover and play with the crossover point and try inverted or normal phase until you find the result acceptable. You should be able to get a 90 degree relative phase shift (between FR and sub) at the crossover point using this approach to your OB FR driver.

But I hope that you have already done other compensation for the loss of bass in an open baffle system (due to cancellation)... Anyone who just sticks a FR driver in a baffle and thinks they are done is sadly mistaken. I did this with a system back in the mid 1990's and having no idea what I was doing I could never get it to sound right. Have you made any measurements of the system to see what the frequency response looks like?
 
Last edited:

gupsta

Member
2008-04-30 9:39 pm
Thank you for the help.

I don't have a Microphone, so I'm going to try building the Mic described on Linkwitz website using the Panasonic WM-61a.

I have a sound card with a MIC input, and know how to generate test frequencies in foobar. Any suggestions on what measurement software to use on Windows 7 that will plot the frequency response. Preferably something simple thats freeware ? A search came up with HolmImpulse

Dave
 

krivium

Member
Paid Member
2009-10-13 2:43 am
For software use REW: free, great feature widely used if need advice.

For the mic don't DIY it. You'll need accurate measure and a diy one should be difficult to calibrate for measurement.

Behringer ECM8000 or copycats including calibrating files are easy to buy and relatively cheap given what they do. You won't regret to spend alittle more money on a mic like that.