[divie23]

After having completed a Regulated LM317 PS based on a torroidial tranny, I've discovered that the tranny is emitting noise that can be picked up on my amp. Although the amp and the PS are in seperate enclusures, I stack them one on top of the other to save space. The noise is certainly from the tranny.

How do I rid of this interference? Is there some kind of shielding I could try? Wrap the tranny in foil perhaps?!

 

[jeffreyj]

Quote:

Originally posted by divie23 ... How do I rid of this interference? Is there some kind of shielding I could try? Wrap the tranny in foil perhaps?!

The greatest amount of flux leakage from a toroidal transformer comes from where the wires exit. Try rotating the transformer such that the wires point at the chassis (away from any other circuitry, that is) and see if that doesn't clear it up.

There are a number of ways for hum to get into an audio amp but flux leakage from a toroidal power transformer is one of the least likely ones. Ground loops are always a suspicion but current pulses flowing from the rectifier to the reservoir caps is one of the more common culprits. This cause is minimized by using a short, fat connection between rectifier and cap(s), then taking the DC from a separate connection directly at the capacitor terminals, not at the rectifier's.

Power supplies are simple, aren't they?

 

[Whit]

Toroids allow more high frequency noise to be passed through from the mains to the power supply than standard laminated transformers. Try either a Corcom or Schurter input line filter or at least some ferrite beads.

 

[divie23]

Maybe I didnt make it obvious enough, my fault, let me explain again. Typically the PS and the amp, which are in seperate enclosures, sit side by side. The PS connects to the amp via a 2-core wire to supply the juice. Side by side I have absolutely no problems with noise of any kind. NONE.

When stacking the amp on top of the PS however, such that the amp circuit board lies directly above the torroidial tranny, thats when the noise creeps in. It seems to be like I simply need to shield the tranny somehow.

The simple solution would be not to stack the PS and amp, but I need the extra space.

Any further suggestions would be much appreciated.

TRhanks jeffert and whit

 

[jeffreyj]

Foil will not block magnetic flux leakage from a transformer; it is only useful against electrostatic noise. You a number of options available, but none will be pleasant, I imagine:

1. put a band of Mu metal (expensive!) around the toroid, taking special care to cover the wire exit point thoroughly.,
2. make sure there is at least a quarter inch, but preferably 3/8" or more, of mild steel plate between the toroid and the affected circuit board.
3. replace the toroid with a different style of transformer - toroid's are the best at minimizing overall flux leakage, but the winding exit point is their Achille's Heel.
4. reduce the loop area in the affected circuit (ie - make pc board traces shorted, use the shortest wire possible to connect jacks to the board or, better yet, use board-mounted jacks, etc.)
5. more physical distance

#4 is the real solution, by the way. Loop area must be minimized so that the traces on the pcb and the wires connecting everything together don't act like "induction coils" (or as antennas).

 

[divie23]

Interesting post there jeffrey. I'll get working through those suggestions - I have a feeling its going to be very hit and miss but oh well! Never realised it would be so hard to rid of magnetic flux leakage

Thanks again

 

[JohnFerrier]

Quote:

Originally posted by divie23 Interesting post there jeffrey. I'll get working through those suggestions - I have a feeling its going to be very hit and miss but oh well! Never realised it would be so hard to rid of magnetic flux leakage Thanks again

...I plan to use a steel enclosure around the torroid. Steel is better than copper or aluminum for shielding magnetics. And if that doesn't work I'm going to shell out $20 for some mu-metal foil. See: http://www.crutchfield.com/cgi-bin/S...uded&i=127COF2


JF

 

[antomas]

Hi Jeffrey,
Thanks for the very interesting posts! May I ask a couple of things about the "lost flux" problem?

First is about the toroid. Is the flux leakage related to the "exit point", not to the "exiting wires", isn't it? In both cases, would it be useful to twist the wires together?

Second question is about the rectifier bridge. I remember a post somewhere (sorry, I don't remember where...) saying that rectifiers generate strong magnetic fields. If this is true, is there a "best way" to lay the rectifiers out? For instance:
face to face: http://home5.swipnet.se/~w-50674/hif...0_overview.jpg
in a line: http://www.welbornelabs.com/images/ps1d.jpg
or again, like the four sides of a square (like in a integrated bridge, I guess?).

I do hope the latter question be pertinent to the original thread.

Thanks a lot for your very valuable answer,
Massimo

 

[jeffreyj]

Quote:

Originally posted by antomas ... Is the flux leakage related to the "exit point", not to the "exiting wires", isn't it? In both cases, would it be useful to twist the wires together?

That is correct, antomas. It is the wire exit, not the wires themselves, that produces most of the flux leakage (the wires do produce some as well, of course, and this will be reduced by twisting opposing ends of each winding together much as you stated). The reason the exit point is a source of magnetic flux is simply because the core is "bare" at that spot (usually - sometimes the windings are staggered around the toroid to prevent any of the core being left exposed).


Quote:

Second question is about the rectifier bridge. I remember a post somewhere (sorry, I don't remember where...) saying that rectifiers generate strong magnetic fields. If this is true, is there a "best way" to lay the rectifiers out?

The strength of a magnetic field is directly proportional to the current flowing through the wire. Given that basic fact of physics, there's no reason why the rectifiers should produce a stronger magnetic field than the traces (or wires) coming from the transformer, going to the caps, etc. The reverse recovery behavior of the rectifier *does* influence the amount of radiated noise (ie - radiated magnetic flux) with faster rectifiers generally making less total noise but of higher harmonic content until you get to Schottky's, which turn off so fast there is virtually no noise to speak of from their operation. The solution in switching power supplies is to either snub the rectifiers (with an RC network) or put a ferrite bead on each rectifer (to add enough inductive reactance to slow the rate of change in current flow). Keep in mind that switching power supplies typically operate at several 10's to several 100's of kHz, so applying these techniques down around 50/60Hz is a bit foolish (though not actually harmful).

 

[jeffreyj]

Quote:

Originally posted by JohnFerrier ...I plan to use a steel enclosure around the torroid. Steel is better than copper or aluminum for shielding magnetics. .. JF

The whole point of spending more money for a toroidal-core transformer is so that you don't have to resort to these tactics! If you find your design requires Mu-metal to block flux leakage from the transformer, you might as well save a few bucks and use one with a plain old EI-core. Rotating a toroidal transformer so that the winding exit faces the outside of the enclosure usually does the trick - if it doesn't, then perhaps a reassessment of the pc board layout is in order. In the case of pcb-mounted toroids, well, you've got no choice but to work on minimizing the loop area on the board!

 

[antomas]

Hi Jeffrey,

thanks very much for your kind answer.
Quote:

The strength of a magnetic field is directly proportional to the current flowing through the wire. Given that basic fact of physics, there's no reason why the rectifiers should produce a stronger magnetic field than the traces (or wires) coming from the transformer, going to the caps, etc.

What made me suspicious about the rectifiers (besides the post I mentioned) was the idea that current flowing into a bridge resembles more AC than DC and the fact that rectifiers of a bridge are often laid out in a semicircular (inductor-like) shape. So, in my mind, all this stuff could have been more "dangerous" than a "average" (straight) power supply trace.

Thanks once again and best regards,
Massimo

 

[JohnFerrier]

Quote:

Originally posted by jeffreyj The whole point of spending more money for a toroidal-core transformer is so that you don't have to resort to these tactics! If you find your design requires Mu-metal to block flux leakage from the transformer, you might as well save a few bucks and use one with a plain old EI-core. Rotating a toroidal transformer so that the winding exit faces the outside of the enclosure usually does the trick - if it doesn't, then perhaps a reassessment of the pc board layout is in order. In the case of pcb-mounted toroids, well, you've got no choice but to work on minimizing the loop area on the board!

Thanks for the idea of pointing the PC mount toroid away from the amplifier circuitry. If I'm going to take the time to contruct an amplifier, I'm interested in reference quality (0.01% bulk metal film resistors probably aren't necessary, but they won't hurt either). Therefore, I'll use a steel enclosure around the transformer/power supply (I'll try to save $20 on mu-metal foil).


JF