 
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
|
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
67.169.141.45
In Reply to: RE: Low RFI power supply design posted by Ed Sommers on March 30, 2009 at 15:59:51
I've done a lot of experimentation and listening with this. There are two (at least) differing issues here and different attempts to alleviate them.
One is the RF noise generated by the reverse charge of the diode. The other is the resonance of the transformer secondary. In many cases this resonance falls right in the range of noise generated by the diode. The diode noise excites the resonance which can amplify it to a large degree. Not only is this an issue for the box the PS is in, but the signal from the resonance passes right through the transformer and onto the power cord where it can go to other boxes and be radiated.
This resonance can not only be excited by the diode noise but also from the high current pulses you get from cap input filters.
Part of the confusion is that there are two different circuits called by the same name (snubber) that do very different things.
One type is what you have, a resistor and capacitor in series sitting in parallel with the transformer secondary. This attempts to damp the transformer resonance so it will not resonate no matter how it is "excited". Part of the problem with this solution is that different transformers have very different parameters which theoretically need different values to damp them properly. I did a bunch of tests on every transformer I had and found that a good compromise was 330 ohms and .022uf. With those values I couldn't get any of my transformers to resonate. One big advantage to this "snubber" is that it works for both diode noise and high current pulses.
The other type of snubber is an RC network across with each diode. This attempts to lower the noise at the source. Again you can get different behaviors with different diodes. I found this to be less effective than the secondary damping, it can help but it rarely gets rid of all the noise, and whats left still is amplified by the transformer resonance.
My solution is to use Schottky diodes which don't have reverse charge at all in addition to the secondary damping network. And then just to be safe I carefully design all my supplies to minimize current pulses in general. The overall result is a drastic reduction in PS noise, both fed to the electronics and fed to the outside world.
John S.
Follow Ups:
Well said John! I can see that you have been doing good work in lowering the amount of RFI in your equipment. Glad to see that you like the RC across the secondary as much as I do. Placing one across each diode gives RFI a path around the diode and into the rest of the power supply,
I do not like this way of using a RC.
Keep up the good work!
BTW, what Schottky part numbers do you like?
Hi.
Simulation shows great result with C//C+R snuuber across power iron secondary winding soldered immediately upstream of the SS diode. That's why I installed them to cover 1N4007 type diodes & they bring me noticeable sonic improvement.
Depending on what voltage & current of the PS you need, check up CREE 'Zero Discovery' diodes: www.cree.com. I don't think you need any RFI snubbers with Cree diodes.
c-J
The current pulses from conventional diodes switching from forward to reverse bias have very steep falling edges and a lot of high frequency content. Snubbers made of individual capacitors and resistors with leads are likely to have self-resonance problems that would keep them from absorbing much of this energy.
It seems likely that zero-recovery diodes would be cheaper than snubbers with enough bandwidth to work with conventional diodes.
Hi.
Unless we are on MHZ RF amp designs, resonance of the leads of the parts in audio amps should not be an audio issue.
A simple yet effective C//R+C snubbber only cost a buck or so max! It costs me nothing as I gots tons of those common parts sleeping in my parts bin.
c-J
Hi.
But they cost much much more than a 1N4007 or the like which only costs 25 cents a pop.
A simple cheapie C//C+R snubber is enough to tame it.
As I already posted below, RFI can get into our HIFi gears via the powerlines as well as airborne. A RFI snubber built inside the amp power supply is only one solution to kill RFI generated INSIDE our components. We need to do more to tackle the whole situation.
This is O.T. from the original post.
c-J
John, if I am using a trafo with two secondaries wired in series, and wiring it as if it were a c.t., do i put an rc network across each winding, or just one across the series connection?
thanks!
Mark
Post a Followup:
| FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
||||||||||||||
This post is made possible by the generous support of people like you and our sponsors:
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2025, Audio Asylum, all rights reserved ]
