Active crossovers and notch filters - diyAudio
Go Back   Home > Forums > Loudspeakers > Multi-Way

Multi-Way Conventional loudspeakers with crossovers

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 24th February 2004, 11:11 PM   #1
diyAudio Member
 
Join Date: Sep 2002
Location: Oregon, USA
Default Active crossovers and notch filters

Hi,

When you build an active line-level crossover to bi-amp a speaker, do you put the notch filters in there as well? Or is it better to leave the notch filters as speaker-level circuits? I have a series notch filter in my speaker's crossover (I hope I got the name right, it's the one where the RLC components are in series and the whole thing is in parallel with the driver). If I understand things right, this is to handle the impedance peak of the driver, which would also cause a sharp response peak at that frequency if the notch filter weren't present. Correct so far?

Now if the XO moves out to be in front of the amps, what does one do with this notch filter? It seems to make sense to leave it with the driver, because that'll flatten the impedance seen by the amp. Moving it to the active XO would probably achieve the frequency response flattening though (I don't know how to make a notch filter with an opamp, but I'm sure I can find a schematic somewhere), and isn't one of the advantages of an active XO that the amp has direct control over the driver and you don't worry so much about impedance swings?

I know this is probably insanely ambitious, but I think I'm going to try an active XO and biamp my speakers. Use my 2A3 SET on top and build a gainclone for the woofer. The speakers are coaxes (heavily butcherd Adire HE12.1s), so I think I can ignore issues of time delay for now (based on the paperwork, the accoustic centers are aligned to 60us). S, I think I have some shot at making a somewhat decent XO if I start with duplicating the slopes of the passive filter. I'll probably go analog, because I don't want to give up my TVC and don't want to buy 4 more transformers to feed the DCX2496 the signal level it needs and put the volume control after the XO where it belongs.

Anyway... if this is really stupid and I shouldn't even begin to try this, you're welcome to let me know Any other advice is most welcome too.

Thanks,
Saurav
  Reply With Quote
Old 25th February 2004, 12:47 AM   #2
MarkMcK is offline MarkMcK  United States
diyAudio Member
 
Join Date: Aug 2003
Location: Ohio
Just a quick response to the notch filter question. A series notch filter wired in parallel with the driver will only change the frequency response when the source impedence is relatively high. If you use a speaker level crossover near the driver's Fs, then the notch filter is important to make the other crossover components act as they should. If you remove the passive speaker level crossover components and buffer the crossover with the amplifier, then the series notch filter may not be needed.

If, however, the notch filters are used to adjust frequency response, then the inductors between the amplifier and the notch filter are important. If this is the case, then to go with a line level crossover, the notch filter should be included with it.

When I design speaker level crossovers, I will use the interaction between reactive components in the crossover and the reactive structures in the drivers to produce a synergy greater than the mere sum of the parts. I have no idea if the person who designed your crossovers was as complex in their work. If they were, it might be difficult to duplicate the speaker level crossover in the line level active circuit. If not, then you can likely do just as well or better active.

Good luck and remember to have fun,

Mark
  Reply With Quote
Old 25th February 2004, 12:56 AM   #3
diyAudio Member
 
Join Date: May 2002
Location: Cambridge, MA, USA
Send a message via AIM to Kilentra
Hi Saurav,

The filter you describe is for impedance compensation and doesn't affect the frequency response since that impedance is all in parallel with the driver. If you had an R, L, and C in parallel with that whole thing in series with the driver, that would cause a dip in frequency response and a bump in the impedance at that frequency. So, I would leave it in if it's on the tweeter (to keep those amps happy), and get rid of it if it's for the woofer. While the Gainclones supposedly don't like tough loads, I dont think that is the kind of impedance variation that would cause a problem.

Michael P
  Reply With Quote
Old 25th February 2004, 01:30 AM   #4
diyAudio Member
 
Join Date: Sep 2002
Location: Oregon, USA
Thanks for the responses. The person who designed my speakers definitely knew what he was doing, but then I changed the tweeter and started messing with the crossover (hence my use of the word 'butchered') myself. I know I should buy the books and get the software/measurement tools to do this right, but for now I'm just having fun. So I kept the circuit on the woofer the same, and added an extra series resistor on the tweeter to bring its level down (I tried an L-pad too). You guys are probably pretty aghast at this by now

Anyway... when I ran a frequency sweep and took SPL measurements, the new tweeter had a nasty spike in its frequency response that is above the crossover frequency (I know, means I have the wrong tweeter for the job, but it was cheap and except for the spike I think it sounds better than the stock tweeter, and I'm mostly just fooling around anyway). The spike was fairly narrow - response at 3000Hz and 6000Hz was normal, and 4000Hz and 5000Hz were about 10dB higher. I assumed that this was an impedance peak (should have measured it) - it looked like it based on the reading material I found for series and parallel notch filters. I think I tried a parallel notch filter and it didn't do much help. A series notch filter did take care of this spike though. I got the L and C values from the online calculators (those set the frequency of the notch, right), then tried various values of R until the spike dropped to where I wanted it. The original schematic has another notch filter for the woofer, and it says that one is for impedance compensation.

So that's the full picture. I believe I have a crossover that's actually below my tweeter's resonance frequency (though not if you believe the tweeter's datasheets for 'recommended crossover frequency'). So what do you think - should I leave the notch filters in or not?

Thanks,
Saurav
  Reply With Quote
Old 25th February 2004, 02:19 AM   #5
diyAudio Member
 
Join Date: May 2002
Location: Cambridge, MA, USA
Send a message via AIM to Kilentra
Woah... what's that particular tweeter you're using?

Either that passive crossover would have to be revamped after doing some measurements on that driver mounted in the Beta-12, or you should just dump it entirely and go active.

I myself am personally using LspCAD active filters to simulate the known response of my drivers, and then add my own new active filters on top of that to see a 'general idea' of the acoustic sum response until I can get real measurements.
  Reply With Quote
Old 25th February 2004, 02:54 AM   #6
diyAudio Member
 
Join Date: Sep 2002
Location: Oregon, USA
I did try something like that, entering interpolated data for the frequency response plots I have for my drivers (measured and/or published). It got too tedious after a while, and I think I was limited in what I could do since I had the freeware version of the software.

The tweeter is a Selenium DT150. The paperwork says it's usable down to 1500Hz, though there's another line which says the minimum recommended crossover is 4000Hz if it's 12dB/octave. I just wanted to try a different tweeter, so I picked this one mostly because it was cheap and had a phenolic diaphragm (I'd heard bad things about cheap metal diaphragm compression drivers). I asked Dan if he'd considered other tweeters and he said that all the others he tried had a nasty resonance in their passband. And he was right

Anyway... other than that 4kHz spike, I like this tweeter better because it seems to go up higher before it falls off (pretty sharply, I think that might be a characteristic of phenolic diaphragms, or just the way the 12CXs horn works or something). This adds a bit of sparkle to the sound that seems better to my ears.
  Reply With Quote
Old 25th February 2004, 07:22 AM   #7
Svante is offline Svante  Sweden
diyAudio Member
 
Svante's Avatar
 
Join Date: Feb 2004
Location: Stockholm
Default Re: Active crossovers and notch filters

Quote:
Originally posted by Saurav
Hi,

When you build an active line-level crossover to bi-amp a speaker, do you put the notch filters in there as well? Or is it better to leave the notch filters as speaker-level circuits? I have a series notch filter in my speaker's crossover (I hope I got the name right, it's the one where the RLC components are in series and the whole thing is in parallel with the driver). If I understand things right, this is to handle the impedance peak of the driver, which would also cause a sharp response peak at that frequency if the notch filter weren't present. Correct so far?
At the risk of repeating what has been said:
When designong *passive* crossovers things get very much easier if the load impedance is purely resistive. Normal filters will behave differently if the load varies with frequency. So, in order to make the load seen by the filter purely resistive, the idea of a conjugate link, and in some cases this parallel RLC circuit that you describe pops up. But the purpose of this is soley to make the life easy for crossover filter. If you move to active filters, the reason to use the impedance compensation is no more. You can simply remove it. The amplifier will not mind a impedance that peaks at the resonance.

Now, there are other reasons to look close at the passive filter, though. There might be other compensations, like for the baffle step, or so, that you want to incorporate in your active crossover.


HTH
__________________
Simulate loudspeakers: Basta!
Simulate the baffle step: The Edge
  Reply With Quote
Old 25th February 2004, 02:15 PM   #8
diyAudio Member
 
Join Date: Sep 2002
Location: Oregon, USA
OK, that makes sense now, because the amplifier's output is a lot lower in impedance than the passive crossover components. Thanks. Though I'll be driving this with a SET which will probably have much higher output impedance than an SS amp, but I can deal with that when I get to it.

Yes, I know I'll have to look at the BSC that's built into the passive filter. It also handles a rise in the woofer's response right around the XO region, and I'll have to handle that as well.

Thanks everyone,

Saurav
  Reply With Quote
Old 26th February 2004, 01:42 AM   #9
diyAudio Member
 
Join Date: May 2002
Location: Cambridge, MA, USA
Send a message via AIM to Kilentra
Yeah, that tweeter looks somewhat interesting, it appears as if the Fs is low enough but there's still a significant (6db?) resonance which shows up as a broad hump across an octave or so (not just normal response variations; look at the impedance plot). A low Q notch filter will do the trick, that is somewhat more complicated with active filters since you probably have to use a gyrator circuit for the inductor. Bafflestep can be handled by placing the input of a buffer between an R from the input and an RC to ground. I will be starting with 33k, 47k, 6.8nf to get about 4.5db of step centered on 500Hz for my Kit281s, your circuit may need a larger value capacitor.
  Reply With Quote
Old 26th February 2004, 02:32 PM   #10
diyAudio Member
 
Join Date: Sep 2002
Location: Oregon, USA
I see what you're saying now. I had looked at those plots, but the tweeter behaved completely differently in my speaker. I'd also been confused about why that 5kHz spike was different when the full passive XO was in place, vs. when the tweeter was connected directly to the amp (but still mounted at the back of the 12CX). Now it makes sense.

So I'm not really using it below Fs (the XO is at 2.4kHz, 4th order electrical AFAIK).

Quote:
A low Q notch filter will do the trick, that is somewhat more complicated with active filters since you probably have to use a gyrator circuit for the inductor.
This would still be a series notch filter, right? I already have the L and C to build one for that frequency, all I'd have to do is adjust the R to get the correct Q. And I'll know that once I'm running it connected directly to the amp so I know what that peak looks like.

So by now I understand that I won't usually need the passive notch filter with an active XO. But if I still see a significant response variation due to the resonance, is there any harm in handling it with a speaker level filter? Would it sap a lot of power? I don't even know what a gyrator circuit is, I've seen a circuit that 'emulates' an inductor, maybe that's what you're talking about.

Thanks for the tip about BSC.
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Active notch filters neutron7 Full Range 2 19th November 2006 11:51 AM
Multiple notch filters? Mikael Abdellah Full Range 2 25th August 2006 05:09 PM
Help on series notch filters. bowdown Multi-Way 1 9th March 2005 05:51 AM


New To Site? Need Help?

All times are GMT. The time now is 01:12 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2