Looking to build some basic analog active crossovers to biamp large 4 way speakers.
I'm going from a fully balanced input to the floating individual single ended filters, then back to fully balanced, using input and output transformers.
My concerns are how the inductors will behave in terms of distortion depending on which core materials I decide to use. I want to use air cores wherever possible, but the lower frequency filters will almost certainly require cored inductors.
Anyone been down a similar road with their builds? Any input is appreciated.
I'm going from a fully balanced input to the floating individual single ended filters, then back to fully balanced, using input and output transformers.
My concerns are how the inductors will behave in terms of distortion depending on which core materials I decide to use. I want to use air cores wherever possible, but the lower frequency filters will almost certainly require cored inductors.
Anyone been down a similar road with their builds? Any input is appreciated.
I once tried Hammonds iroc cored chokes in a similar application but found them very non-linear, the crossover frequency point shifted changed with the signal levels.
Lately I've been looking at these for a passive (or buffered) line level LCR XO:
https://utmindustry.com/inductors/
Lately I've been looking at these for a passive (or buffered) line level LCR XO:
https://utmindustry.com/inductors/
Hi,
there was thread recently with discussion on the topic you might want to browse through. On a passive filter that is after amplifier, you use appropriate core you can do it without adverse effects. Adverse effects could come also from using aircore with too much added Re in a wrong place, so it's very likely there is some "best compromise" in between of air core and worst possible iron core just from system sound quality perspective, while also saving you some money.
Although, I'm not sure if I understood your application correctly. Usage of transformers sounds like the filters are before amplification in which case you can use air core inductor I think, high R on line level inductor doesn't matter as long as your transformers are happy. Your transformers also have a core, so little bit more in an inductor doesn't likely harm either. Are you making one xo before amplifiers to be able to biamp and more after the amplification? Bass gets it's own amp and the rest 3 ways another with associated passive filtering, in which case you need inductors both before and after the power amps? Sounds like very complicated system because it is expensive to fine tune, which you likely are going to have to do because typical home measurement data isn't accurate for low crossover points. I hope you get it tuned and working nicely!
there was thread recently with discussion on the topic you might want to browse through. On a passive filter that is after amplifier, you use appropriate core you can do it without adverse effects. Adverse effects could come also from using aircore with too much added Re in a wrong place, so it's very likely there is some "best compromise" in between of air core and worst possible iron core just from system sound quality perspective, while also saving you some money.
Although, I'm not sure if I understood your application correctly. Usage of transformers sounds like the filters are before amplification in which case you can use air core inductor I think, high R on line level inductor doesn't matter as long as your transformers are happy. Your transformers also have a core, so little bit more in an inductor doesn't likely harm either. Are you making one xo before amplifiers to be able to biamp and more after the amplification? Bass gets it's own amp and the rest 3 ways another with associated passive filtering, in which case you need inductors both before and after the power amps? Sounds like very complicated system because it is expensive to fine tune, which you likely are going to have to do because typical home measurement data isn't accurate for low crossover points. I hope you get it tuned and working nicely!
Last edited:
The topic here is I believe a line level active filter with transformers on its ins and outs...
//
Yeap in that case there is likely more information on GroupDIY.com than diyaudio.com, although some of the experts might be here as well. Unfortunately groupdiy is overrun by ads so very painful to browse around, but there is plenty of passive line level network schematics and discussion around them. Way out of my league so nothing more to contribute 🙂
If you haven't been there before, here some tips to get started browsing, I was actively following it some 15years ago. I remember best transformer performance demonstrated on groupdiy was by nick name Marik (Mark Fouxman). He founded a company later https://www.samaraudiodesign.com/index.html and seems to have many products available, perhaps even suitable inductors. I also remember Ian Thomson-Bell who is with nick name ruffrecords played a lot with passive filters, in EQ context though, but his posts likely lead to lot's of useful nuggets for this kind of passive xo work as well.
In general, passive line level filters are quite rare so not that much info around except from the oldies, which is why I think the information is found from the GroupDIY, from world of recording studio electronics. The IO transformers alone count 8 for biamped stereo, so cost gets into 1000$ territory before any filtering components.
If you haven't been there before, here some tips to get started browsing, I was actively following it some 15years ago. I remember best transformer performance demonstrated on groupdiy was by nick name Marik (Mark Fouxman). He founded a company later https://www.samaraudiodesign.com/index.html and seems to have many products available, perhaps even suitable inductors. I also remember Ian Thomson-Bell who is with nick name ruffrecords played a lot with passive filters, in EQ context though, but his posts likely lead to lot's of useful nuggets for this kind of passive xo work as well.
In general, passive line level filters are quite rare so not that much info around except from the oldies, which is why I think the information is found from the GroupDIY, from world of recording studio electronics. The IO transformers alone count 8 for biamped stereo, so cost gets into 1000$ territory before any filtering components.
Last edited:
But it's sometime strange with the concept as described - in an active filter and for the filtering sections, indictors are seldom used but rather only resistors and capacitors. And one never uses inductors on input and output... but here trafos are to be used...
Is the idea to build a crossover that uses the same schematic as a passive traditional crossover. Bonkers... why? 🙂
//
Is the idea to build a crossover that uses the same schematic as a passive traditional crossover. Bonkers... why? 🙂
//
@TNT I think you're misunderstanding my intentions. I'm not building a passive xover.
It will be using one filter stage for each 2nd order slope (LC, not RC), thus cutting the amount of filter stages in half for. That also reduces noise along with using coupling transformers. My entire signal chain is balanced, so I'll need to convert to floating single ended filter stages, then back to electrically balanced again.
@tmuikku Thanks for the links. Yes, 8 decent transformers are expensive. So far I spent about $750 just on those alone.
I'm using Jensen JT11P-1 input and JT11-DMCF output transformers. I also like the lower level 2nd order harmonics added by the iron in the signal path. That also includes the inductors. My pre- and power- amp are Parasound JC2 and JC5, which are balanced. I have two more modified Halo A21s to use for the lower power drivers.
It will be using one filter stage for each 2nd order slope (LC, not RC), thus cutting the amount of filter stages in half for. That also reduces noise along with using coupling transformers. My entire signal chain is balanced, so I'll need to convert to floating single ended filter stages, then back to electrically balanced again.
@tmuikku Thanks for the links. Yes, 8 decent transformers are expensive. So far I spent about $750 just on those alone.
I'm using Jensen JT11P-1 input and JT11-DMCF output transformers. I also like the lower level 2nd order harmonics added by the iron in the signal path. That also includes the inductors. My pre- and power- amp are Parasound JC2 and JC5, which are balanced. I have two more modified Halo A21s to use for the lower power drivers.
Why don't you make a balanced filter?
A disadvantage of air core inductors is that you can't give them a closed magnetic circuit as far as I know. You therefore have to be careful with magnetic coupling, for example with supply transformers. I once used two hand-wound air core inductors in a line level stereo crosstalk emulating/compensating circuit and could never get rid of its hum. When I needed inductors for a DAC reconstruction filter decades later, I went for TDK/Epcos N48 potcores with a big air gap. I never had any hum with those.
@MarcelvdG You do have a valid point there. Its challenging to shield these air core inductors in a practical application. I've had that issue before and did manage to shield them using small mu metal cans. The other issue is real estate on the PCB and having a layout which avoids crosstalk. The mu metal can take care of that as well but its not easy. They do make small laminated cores for coupling transformers which could be a good compromise. Ferrite is just to unpredictable when and how it affects the inductor's specific characteristics in HD.
In terms of active components in the filter stages, I'm opting for paralleled signetics NE5532ANs in mil spec ceramic packaging. I have a bunch of these date coded from the early 80s and they perform excellently. The voltage noise is under 2.5 uV/hz and offset is hand picked with some external trimming added. This is why I like transformer coupling, as it can eliminate most of the caps in the signal chain. As long as the output offset is servo trimmed, I won't need any series caps.
If someone has any preferences in circuit topology for LR2 stages, please share them. I have used staggered 1st order filters with better success than 2nd order elements using the op amps NFB loop. I find the input current offset balance is much better controlled having linear FB and only 22pf compensation. Open loop gain isn't the greatest on most NE5532 family chips, but it doesn't need to be that crazy as some higher performance off label op amps ie. designed for video circuits.
If someone has any preferences in circuit topology for LR2 stages, please share them. I have used staggered 1st order filters with better success than 2nd order elements using the op amps NFB loop. I find the input current offset balance is much better controlled having linear FB and only 22pf compensation. Open loop gain isn't the greatest on most NE5532 family chips, but it doesn't need to be that crazy as some higher performance off label op amps ie. designed for video circuits.