[sdiy] Problem with 723 PSU

[Magnus Danielson]

From: Jim Patchell <patchell at silcom.com>
Subject: Re: [sdiy] Problem with 723 PSU
Date: Wed, 18 Apr 2001 06:52:49 -0700

>     Without actually seeing the thing, I just making wild guesses, but here is something to check.
> 
>     The good old capacitor input power supply filter causes a very high current spike while charging the capacitor.  You can estimate what this spike is by looking at the ripple on the filter cap.  The diode is only conducting while
> the voltage is going up, which means the diode could be conducting for say only about 1/10th of the time, for instance.  If you are drawing 1 amp from the power supply, this means the current flowing though the diode will be 10 amps
> durring this short period.  By having long wires going back to the transformer, you might be creating a bad situation. (Short 10 amp current spikes).
> 
>     To keep this spike down in some of my applications, I have used either a small resistor, or better yet, a choke.  In either case, you might have to adjust the power supply voltage, because a capacitor input filter is basically a
> "peak" detector.  A choke input filter will give you something closer to the average voltage.
> 
>     I don't know if any of this applies to your problem, but it is something to consider.

I'd tend to agree with Jim. The diodes are only effective over a short
period of the waveform and clips rather ugly and quick. One way to
handle these quick changes of current is to bypass the diodes with
plastic caps while trying to keep inductances to a minimum. When the
diode clips will the spike pass through the capacitor and the clipping
becomes less distinct. Sharp spikes and inductance is never good
friends. The classic way of handling EMC problems is:

1) Make the source less efficient noisemaker
2) Make the transmissionpath less efficient to transport noise
   (unwanted signal)
3) Make the receiver less efficient receiver of the noise

When you look at strong currents and sharp edges you start at the
source if possible. Since the source most likely is the clipping
diodes those are certainly to be checked up.

Hmm... Jörgen, why don't you come over?

Cheers,
Magnus - living 2 km from Jörgen.