Active crossovers - phase distortion?

rjon17469

Member
2004-08-01 8:03 pm
IL
GRollins said:
Assuming an analog circuit, there will always be phase shift at the crossover point. I don't know that I would go so far as to call it 'distortion' though.

Grey


Thanks for the info/correction.

What keeps getting me is the low pass crossover. The only thing in the signal path is resistors, and then the caps are going to ground. In this arrangement, would the caps still exhibit a phase shift on the signal?

Thanks!!

Reece
 
Yes.
Everything is in the signal path.;)

Green is gain, red is phase, black is group delay.
 

Attachments

  • low-pass.png
    low-pass.png
    29.6 KB · Views: 356
There's a tendency to think only of series connections as representing "the signal path" but, as Carlos notes, parallel ones count, too.
Think of the resistor/cap combination as a voltage divider. The resistor is frequency-neutral, but the cap passes high frequencies. In low pass configuration, the cap passes those frequencies to ground, where they are lost. What remains is sort of a photographic negative of the high frequency 'image' that went through the cap, i.e. the lower frequencies. Whatever the cap removed leaves an impression--in the negative sense--on what remains. Thus the cap does have an effect on the sound quality, by virtue of what it subtracts. This includes phase relationships, except that they too are reversed.

Grey
 

rjon17469

Member
2004-08-01 8:03 pm
IL
Thanks so much you guys for your help.

That makes sense. So, I guess, there's no real way to avoid the phase shift besides moving into the digital domain, and that has it's whole fair share of problems.

It seemed too good to be true! (my motto for audio)

Thanks all.

Reece
 
for me to get a grasp of it, i have to think of the characteristics of caps.

at DC caps are open
at AC caps are short

in the frequency (S) plane, DC is equivalent to 0Hz, so at low frequencies caps are open. So if a cap was in series with a driver it would block all low frequencies.

At high frequencies caps are shorts, so if one was placed in series with a driver only high frequencies get by.


But thats just for caps and the s plane, you are looking at how this affects the phase of the driver.

One thing that i would recommend you do is look at a 4th order Sallen-Key filters as there is no phase shift over all frequencies

http://www.linkwitzlab.com/filters.htm#13

check it out, if you aren't to adept with EE stuff the S plane and all the poles and zeros might mess you up a bit.
 
Any time that Nature gives you a non-flat frequency response, there's phase shift. Any time there's a time delay (e.g., between the time of recording and the time of playback), there's a phase shift. Getting rid of phase shift in a crossover involves doing some very unnatural manipulations to the signal.

I think that the real problem is that you don't really know what "phase shift" means but have read somewhere that it's a Bad Thing, probably written in an article by some audio writer who knows less than your cat about these issues. If you really want to understand them, it would be worthwhile to pick up a basic physics text like French's "Vibrations and Waves" and try to get the basic concepts down before deciding what to worry about and what not to.
 
Well, I understand that phase shifts correlate to the period of compression or rarefaction of the wave at a particular point in space as it moves.

The reason I'm trying to avoid phase shifts is to help in the transistion between drivers, and especially to avoid having two speakers 90 degrees out of phase with eachother at the crossover point.

Plus, and I guess it's just an issue I have, but at times it seems like to me that severe phase shifts detract from the realism of a speaker, although for me to actually make this claim would require a lot more hard data than what I have.

Reece
 
I understand that phase shifts correlate to the period of compression or rarefaction of the wave at a particular point in space as it moves.

Deep, deep misunderstanding. What's the level of your math background? Knowing that, I can send you to some appropriate books that will make things much clearer to you.



although for me to actually make this claim would require a lot more hard data than what I have.

That is seriously a very, very smart thing to say. And a much smarter thing than I hear most audiophiles say about ANYTHING.
 
As for my math background, it's iffy. This August I'll be entering my freshman year of college (so I'm a young DIYer). However, in regard to math, I've taken everything through calculus BC.

Can I ask, what is phase then? I have always understood it to be the period of the phase or rarefaction that the wave was in.

Thanks in advance!

Reece
 
They will cancel if they're of the same frequency and 180 degrees out of phase at the point you're sampling.

OK, here's a problem. You have two sine waves of the same frequency and amplitude, but 90 degrees apart in phase. How do they sum? What's the resulting wave's amplitude?

One more question, then I swear I'll stop torturing you. Do you have a copy of Vance Dickason's Loudspeaker Design Cookbook? There are some very enlightening figures and discussions in there pertinent to what you're really after here.
 
SY said:
Getting rid of phase shift in a crossover involves doing some very unnatural manipulations to the signal.

Yeah... I like a good op-amp, but too many of them in the way of the signal...:apathic:
I still think that in some cases, choosing the right drivers, a passive 1st order crossover (single cap in the tweeter) is much better than treating the signal that way.:bawling: