Ok, the idea is by no means mine, but anyways...:
I'm currently working on a big mancave horn system that needs to be XO:ed at around 60-70Hz.
Right now I's using a DIY 70Hz 12dB/Oct Linkwitz-Riley crossover that seems to work quite ok but I'm not too excited about the build quality, especially since it's been modded to within an inch of it's life by now.
The whole rig is built in a retro 40's style (think Altec Voice of the theatre) so using a DSP would be almost blasphemous, even regular opamp based active crossovers would leave a nasty itch in my brain.
Would it be completely insane to try passive line level LCR filters as crossovers?
I have some Hammond 156C iron core chokes on the shelves (150H, Rdc 3,8kohm) and a bunch of precision 40,2nF polyprop caps. If my maths is correct, 150H and 40,2nF into a 30,5k load would form a 65Hz L-R filter.
My sources are CD and computer soundcard into Sowther TVCs, so the source impedance should be low enough to drive sucha crossover I believe.
Any thoughts on this?
I know Marchand sells something like this, I believe it's called XM46 if my memory serves.
I'm currently working on a big mancave horn system that needs to be XO:ed at around 60-70Hz.
Right now I's using a DIY 70Hz 12dB/Oct Linkwitz-Riley crossover that seems to work quite ok but I'm not too excited about the build quality, especially since it's been modded to within an inch of it's life by now.
The whole rig is built in a retro 40's style (think Altec Voice of the theatre) so using a DSP would be almost blasphemous, even regular opamp based active crossovers would leave a nasty itch in my brain.
Would it be completely insane to try passive line level LCR filters as crossovers?
I have some Hammond 156C iron core chokes on the shelves (150H, Rdc 3,8kohm) and a bunch of precision 40,2nF polyprop caps. If my maths is correct, 150H and 40,2nF into a 30,5k load would form a 65Hz L-R filter.
My sources are CD and computer soundcard into Sowther TVCs, so the source impedance should be low enough to drive sucha crossover I believe.
Any thoughts on this?
I know Marchand sells something like this, I believe it's called XM46 if my memory serves.
Inductors for higher impedance use tend to be specialised. This means if you go through your collection of high value chokes and small power transformer windings looking for suitable values, you need to check them to see that their self resonance isn't within the pass band. You may find that specific values are harder to find commercially and they will tend to be expensive.
Yes, my biggest concern here is the quality of the chokes. This specific model works surprisingly well as plate loads for small triodes but that is a completely different application. As far as I can understand the inductance in iron cored chokes will vary with signal strength and frequency, definitely not a good thing i a crossover although the airgap in the core should linearize things quite a bit.
Wire up the filter and measure the frequency response at say 2Vrms output.
The results should tell you the answer.
The results should tell you the answer.
"Wire up the filter and measure the frequency response at say 2Vrms output."
Yeah, that's what I intend to do next. I have this idea of wiring one choke and one cap in series and measure the resonance frequency and see if it changes at different levels.
Yeah, that's what I intend to do next. I have this idea of wiring one choke and one cap in series and measure the resonance frequency and see if it changes at different levels.
Keep in mind that if you only have a few values, you'll be relying heavily on the impedance value to help you get a frequency. Does the inductance work for you in that respect.
![IMG_20160331_093905[1].jpg](/community/data/attachments/1037/1037438-bc7d39aba699b5f862ee8c5cdf9a2b7a.jpg?hash=vH05q6aZtf)
Measure the sine wave frequency response from say 10Hz - 300Hz at a nominal 2Vrms,
and then also at say 200mVrms.
Running some initial tests didn't take long, and the results where not too exciting: At low levels (0,5V or so) the LC circuit resonates at 86Hz, at a few volts it decreases to 74Hz. Not impressively linear.
"Keep in mind that if you only have a few values, you'll be relying heavily on the impedance value to help you get a frequency. Does the inductance work for you in that respect."
The idea would require modding some amps to have 30,5k input resistors, which wouldn't be a problem as long as don't change amps too often 🙂
"Opamp don't have to be solid state"
True. Tube opamps would be a bit of an overkill, but tube or Jfet followers should get the job done quite nicely.
The idea would require modding some amps to have 30,5k input resistors, which wouldn't be a problem as long as don't change amps too often 🙂
"Opamp don't have to be solid state"
True. Tube opamps would be a bit of an overkill, but tube or Jfet followers should get the job done quite nicely.
I think maybe your problem is believing op-amp crossovers would leave a nasty itch in your brain. 🙂
There's absolutely no reason that should be the case.
Dave.
There's absolutely no reason that should be the case.
Dave.
Possible alternative: although you cannot make an exact second-order Linkwitz-Riley filter as a passive RC filter, you can get pretty close by cascading two first-order RC sections and giving the second a much higher impedance level than the first. I did some calculations on that years ago, especially about how to tweak the RC times for minimum error, I can look them up if you are interested.
In reality, there's of course nothing wrong with good opamp based filters. I just got this idea that potentially would grant me some "audiophile swag" but the imperfections of those chokes took the fun out of it... I'm actually collecting parts as we speak to assemble a semi-permanent opamp based filter to use until I can muster the energy to build something fancier.
If you can find (very hard), or make the required HUGE inductors with low emuff R it is a good idea. But that is a BIG if.
dave
dave
If you want passive, is there a problem with Harrison FMODs? They have standard values of 50 Hz and 70 Hz at 12 dB/octave.
I have some doubts about second order passive RC filters, not many positive things have been written about them online.
After 30+ years of struggling with primitive, cr*ppy soldering irons I finally bought myself a proper soldering station today.
Suddenly, soldering pcb's is almost fun... Here's three of mr Milletts Opamp Experimenter PCBs assembled as a 12dB/oct 72Hz L-R filter with mono sub output
After 30+ years of struggling with primitive, cr*ppy soldering irons I finally bought myself a proper soldering station today.
Suddenly, soldering pcb's is almost fun... Here's three of mr Milletts Opamp Experimenter PCBs assembled as a 12dB/oct 72Hz L-R filter with mono sub output
Last edited by a moderator:
I wouldn't say that a PLL crossover does things badly because if a design doesn't work well then it should be easy to measure that.
However it sometimes challenges your design skills. I have been in a situation where adding the complexity of different needed correction curves can make it fall into line, albeit with a commensurate circuit. On the other hand it may be easier to use simple single order rolloffs and just benefit from less drive to your amplifier... add the extra in another way.
Here's a generic attempt to do LR2. Quite reasonable for a simple circuit. In any case, it's rare that you need exactly that.
However it sometimes challenges your design skills. I have been in a situation where adding the complexity of different needed correction curves can make it fall into line, albeit with a commensurate circuit. On the other hand it may be easier to use simple single order rolloffs and just benefit from less drive to your amplifier... add the extra in another way.
Here's a generic attempt to do LR2. Quite reasonable for a simple circuit. In any case, it's rare that you need exactly that.
Yeah, that isn't going to work. 🙂
The essence of these PLLs is scaling the values such that they (pretty much) accomplish the filter slopes, but also can be driven by a typical line-level source.
A good (low output R) headphone amplifier is preferable in most cases.....vice a typical preamp which probably has an output of 50 ohms or greater.
The Harrison Labs filters approach the solution in that way. It's sort of successful, in some instances. 🙂
I don't consider them poorly designed, but rather designed as well as they can be considering the requirement for some sort of consistency with various source output resistances and load resistances that customers might have.
What I would consider poorly designed are the PLL's using 10 henry inductors (or similar) like the Marchand XM46. Ridiculous! Products like that are aimed at audiophiles who have a burr up their *** regards anything electronic in their signal path.
Dave.
The essence of these PLLs is scaling the values such that they (pretty much) accomplish the filter slopes, but also can be driven by a typical line-level source.
A good (low output R) headphone amplifier is preferable in most cases.....vice a typical preamp which probably has an output of 50 ohms or greater.
The Harrison Labs filters approach the solution in that way. It's sort of successful, in some instances. 🙂
I don't consider them poorly designed, but rather designed as well as they can be considering the requirement for some sort of consistency with various source output resistances and load resistances that customers might have.
What I would consider poorly designed are the PLL's using 10 henry inductors (or similar) like the Marchand XM46. Ridiculous! Products like that are aimed at audiophiles who have a burr up their *** regards anything electronic in their signal path.
Dave.
Last edited by a moderator:
I've made some pretty sophisticated PLL filters by sacrificing insertion loss... so if you're not afraid to throw away 10-20dB. As it happens the filter shown above is only down 1dB.
1k to 50k ohms gives you a couple of orders of magnitude to work with.
1k to 50k ohms gives you a couple of orders of magnitude to work with.