[Amps] Big cap in HV PS

[Jim Thomson]

Date: Mon, 22 Mar 2010 02:55:04 -0400
From: Roger <sub1 at rogerhalstead.com>
Subject: Re: [Amps] big capacitor in HV PS


John Lyles wrote:
> The stored energy in that capacitance at 3.6 KVDC is 8.1 kilojoules. 
> This is a tremendous amount to deal with for a power supply. Roughly 
> equivalent to the mechanical force of a part of a stick of dynamite when 
> it lets loose. I would not want to have to deal with the safety aspects 
> of this design. For instance, if there was an internal arc in the power 
> supply, before the series resistor (mandatory) to the load, it would do 
> a lot of damage, such as rupture capacitors, blow holes in metal.

##  well, I have seen > 5 kilo joules short to chassis, by error, and sure
it made one helluva bang, but it's not that bad.  [ b4 glitch] 



 This 
> is just from the stored energy. An active crowbar might be needed if a series resistor 
> isn't enough to prevent more than a few Joules of energy to be dissipated 
> in a thin wire shorted across the power supply output, to simulate a tubes grid wires.

##  The secret appears to be using the CORRECT sized HV fuse.. just prior to any
'glitch R.   I use the  Buss HVU series  hv fuses, filled with silica, designed to interupt
super high faults currents /energy levels.  They also make HVJ series  for up to 20
kw dc/30 kw AC.  





> Then there is the inrush current. Rectifiers would need to be designed for high peak currents, 500 amps or more.
> Step start would need to be carefully designed to prevent taking out the diodes.

##  standard  6A10 /10A1  diodes  work just fine.  When t=0... the entire 240 vac  gets dropped across the
step start R...leaving nothing across the plate xfmr primary.  The real problem is the mag current in the
plate xfmr primary.  The mag current flows through the step start R... and there is always a drop across it. 
Caps  will never get fully charged up, and when step start R shorted, you get a big sec surge.  The trick is to
use a lower value step start R,  when using high C filters.  Plan B is to use a small variac to ramp up, then it's
relay switched out of the cicuit.... THEN u can charge the caps up to 102% of their normal value. 


>
>   
Having that much stored energy in a small area just isn't safe if 
something goes wrong.

##  If it's all wired correctly, it's really not a problem..or an issue. 






> Seems like a lot of trouble to get stiffer high voltage that isn't really needed for am amateur
> amplifier for SSB, CW, AM. I would not recommend going above 1/10 of that capacitance, unless prepared
> to deal with the effects mentioned above. I have a 250 uF capacitor at 28,000 volts at work, and 
> it took months to design the inrush and crowbar protection, to prevent blowing a 30 AWG wire when shorted across the output. 
> This bank of capacitance is in a steel vault with chicken-wire reinforced glass in the windows. Of course, this
> is far more stored energy than 8.1 k, about 10X this amount, needed to prevent more than a kilovolt of droop 
> during the pulses of an RF amplifier. 

### Up to 10 kvdc...  a 50 ohm glitch R does the job as a current limiter..just fuse it..



>   
We used to use 50,000 (50 Kilojoules) MOVs across our power supplies. 

##  what voltage ?  Used for what ?   Why weren't the MOV's fused ? 

##  I use MOV's acoss the sec of the plate xfmr... fused.  I also use em on the 240 vac side.,. also fused. 




When they went it was spectacular from both the light flash and almost 
unbelievably loud noise. We used #6 wire to them and you'd find the two 
wires sticking straight out and I do mean straight! They were even 
stretched a bit.  If you were working in the cage next to the one 
containing a running PS when that MOV went it was certainly an adrenalin 
rush.<:-))

##  all that shoukd have happened, is a blown mov fuse. 

Jim


73

Roger (K8RI)