[Amps] MOV - Varistor Reccomendations

[R.Measures]

On Jan 30, 2005, at 1:26 AM, Ian White, G3SEK wrote:

> David Pruett wrote:
>> At 10:44 AM 1/29/2005, aborg wrote:
>>> Can anyone suggest what MOV to use for my 1N5408
>>> rectifier board. I an using 32 diodes , 8 per leg. My
>>> transformer is a 4800 Vac pole pig. I want to use the
>>> MOV's instead of the resistor/cap combo on each diode.
>>>
>
>> Mike,
>>
>> While this is not the question you asked, I'd point out that in a 
>> bridge rectifier, each diode (or string) has to stand off the peak AC 
>> input voltage during the half-cycle that it's not conducting, and 
>> this is not referenced to zero, but to the DC output voltage when 
>> used with a capacitor filter.
>>
>> The 4800 volt AC input is around 6700 volts peak.  The filter 
>> capacitor is charged to this value, meanwhile on negative swings the 
>> rectifier has to stand off an equivalent negative peak.  You need 
>> about 13.4 KV worth of diode rating and you've got 8 KV.
>>
>> I've run the power supply for my 40M 4-1000A amp off a 4800V pole pig 
>> with four K2AW 14-1 (14 KV rated) diode blocks.  Indicated plate 
>> voltage no-load is about 6600V.  My understanding is that the K2AW 
>> blocks are fourteen 1N5408s in series.  The above analysis shows that 
>> I could probably use some design margin in this application, but it's 
>> run for more than a decade with no problems.
>>
>> 73,
>>
>> Dave/K8CC
>
> Agreed. In terms of reliability, the best value for money seems to 
> come from very long strings of well-matched diodes, totaling at least 
> 2x the normal calculated PIV requirement.
>
> In other words, spend your money on more diodes, not on parallel 
> "equalizing" components. When you look at the prices of Rs and Cs 
> rated for >>1kV with all the other ratings required, you will find 
> that adding some more series diodes works out *much* cheaper than any 
> other option.
>
> They don't have to be one of K2AWs potted modules, or one of the 
> modules sold for microwave oven spares. As Rich has pointed out, the 
> encapsulation increases the chip temperature, so bare diodes on 
> perf-board will be both cheaper and better cooled.
>
> The only remaining problem is to find well-matched diodes. The 
> military select their diodes by High Temperature Reverse Breakdown 
> testing at high temperature, and maybe K2AW does that too, though I'm 
> sure the microwave oven manufacturers don't.
>
> Unless you have your own HTRB test rig (and time on your hands) the 
> best practical solution is to buy named-brand 1N5408s from a reputable 
> supplier like Digi-Key or Mouser. You're taking them off the bandolier 
> in the same order as they came off the original production line, and 
> it will be hard to find diodes better-matched than that.

Ian -- RE:  The current amplifier project:  We are using 1N5408s taken 
from a reel of 1400 we obtained from Digikey.  After testing several 
dozen diodes from the reel, it was apparent that the PIVs were within 
100V of 1300v, so we stopped measuring PIV, and assembled the 
anode-supply PS FWB-FWD rectifier board.  It produces >9kV DC in FWD 
mode and >4500V in FWB mode.
http://www.somis.org/pb3_pic24.jpg
- cheers - ...

Richard L. Measures, AG6K, 805.386.3734.  www.somis.org