In a LCLC filter, my chokes each have a yellow lead and a black lead indicating the winding direction, flux direction, phase, etc. Is there any (maybe humbucking?) advantage of hooking up the first series choke as yellow followed by black and the second choke as black followed by yellow? Or hook them up each the same?
The resulting sound difference is up to your taste.
From the POV of technical description, using one lead in some circumstances might result in higher breakdown voltage than the other. It will also result in different electrostatic field to the surroundings than the other. This is valid especially for choke input, but for secondary chokes where AC swing is slow, you don't have to worry.
I label my chokes for the highest breakdown voltage (considering core is grounded), but customers are free to wire them for the best sound they achieve.
First rule for highest breakdown voltage is, you want to keep the highest voltage layers from the ground potential (core).
In the end if you don't know how the choke is built, it's a bit of a gamble.
From the POV of technical description, using one lead in some circumstances might result in higher breakdown voltage than the other. It will also result in different electrostatic field to the surroundings than the other. This is valid especially for choke input, but for secondary chokes where AC swing is slow, you don't have to worry.
I label my chokes for the highest breakdown voltage (considering core is grounded), but customers are free to wire them for the best sound they achieve.
First rule for highest breakdown voltage is, you want to keep the highest voltage layers from the ground potential (core).
In the end if you don't know how the choke is built, it's a bit of a gamble.
For LCLC, do not put the chokes right next to each other if the laminations are oriented the same way . . .
if you do, the choke with the most voltage (hum) across it will couple magnetically to the other choke.
If they must be close to each other you might be able to rotate them 90 degrees from each other.
And, do not use a magnetic steel chassis, or in varying amounts, all the magnetic devices will couple to each other . . .
power transformer, choke(s), interstage transformer, output transformer.
There was a time when the engineer breadboarded an RF CLCLCLC filter with 60dB of ultimate out-of-band rejection.
Unfortunately, the PCB layout person lined up all 3 L in the same orientation; the ultimate out-of-band rejection was 20dB.
The next PCB had the middle L rotated 90 degrees, problem solved!
if you do, the choke with the most voltage (hum) across it will couple magnetically to the other choke.
If they must be close to each other you might be able to rotate them 90 degrees from each other.
And, do not use a magnetic steel chassis, or in varying amounts, all the magnetic devices will couple to each other . . .
power transformer, choke(s), interstage transformer, output transformer.
There was a time when the engineer breadboarded an RF CLCLCLC filter with 60dB of ultimate out-of-band rejection.
Unfortunately, the PCB layout person lined up all 3 L in the same orientation; the ultimate out-of-band rejection was 20dB.
The next PCB had the middle L rotated 90 degrees, problem solved!
Last edited:
Yes, even small coupling factors do matter, but in a supply filter it is difficult to predict what is going to be the best solution: it depends on the actual circuit's specifics, cap values, etc.
The best way to eliminate those effects, to make the real circuit work like a sim, is to minimize the coupling. Distance is your best friend, orientation works well too, but if you want to play with the phase of the windings, you need to make real world tests. You may achieve humbucking for certain configurations, current ranges.
The best way to eliminate those effects, to make the real circuit work like a sim, is to minimize the coupling. Distance is your best friend, orientation works well too, but if you want to play with the phase of the windings, you need to make real world tests. You may achieve humbucking for certain configurations, current ranges.
