[Amps] original thread: Peter Dahl transformers, remark about capacitors for switching supplies

[Carl]

----- Original Message ----- 
From: "Bill Turner" <dezrat1242 at yahoo.com>
To: "Amps" <amps at contesting.com>
Sent: Wednesday, April 03, 2013 2:35 PM
Subject: Re: [Amps] original thread: Peter Dahl transformers,remark about 
capacitors for switching supplies


> ORIGINAL MESSAGE:
> On Wed, 03 Apr 2013 11:16:25 -0700, you wrote:
>
>>In a conventional 3.2 kV HV supply that I've built for my new amplifier, I 
>>bypassed each
>>of the electrolytics with an 0.01 uf disc ceramic capacitor. My thought 
>>was that there
>>will undoubtedly be RF leakage back to the power supply, and especially on 
>>160 meters the
>>regular bypassing of the HV line might not be adequate to keep RF out of 
>>the capacitors.
>>RF would be another source of heat that could shorten their life.
>>
>>It didn't occur to me that it might be possible to 'resonate' the 
>>capacitors at some RF
>>frequency and make the problem worse! I'm interested in the answer to 
>>this, too.
>
> REPLY:
> Just a suggestion:  I think it would be better to focus on keeping RF out 
> of
> the power supply in the first place. Check with an oscilloscope to see if
> there is a problem to begin with. You might be trying to solve a problem
> that doesn't really exist.
>
> Even if the RF is there, the HV line should be easy to filter.
>
> My 2 cents.
>
> 73, Bill W6WRT


Start with the 160M lowest reactance disc cap you can scare up at the base 
of the plate choke. A .0047 is adequate for a sane power level and a good RF 
choke.

Then add a healthy glitch resistor and bypass the other end also.The 
resistor has inductance and is self resonant somewhere in the 50-150 MHz 
range for the 20-25W variety. For the 3-500 and similar I also use a 10 Ohm 
20W in series and .001 bypassed as that is good for the parasitic frequency. 
Some amps have an unstable layout and need all the help they can get; in a 
SB-220 RFC-2 is replaced by that resistor.

Any further potentional damage to diodes and filter caps comes from 
switching spikes and each complete string should be bypased with an adequate 
.005/.0047 disc and another to ground at the HV output point.

 Some add a 1N5408 reverse connected across each cap as a crowbar and a very 
fast fault trip.

Carl
KM1H