[Amps] Bleeder Resistor Inquiry

David Kirkby david.kirkby at onetel.net
Sun Apr 24 10:28:26 EDT 2005


aborg wrote:
> I am currenty building a HV Power Supply using a
> 4800vac pole pig, 40uF 5kv oil filled cap , 10kv PIV
> rectifier board, 20 amp variac and a 100k 225watt
> bleeder.
> 
> The initial power up started well without the bleeder
> connected to the circuit after 30 min burn-in no
> problems. Discharged the circuit and added the 100k
> bleeder and 5 mins the bleeder starts to smell and the
> exterior paint starts to bubble. Shut it down and
> examined the circuit, all seem accurate. 
> 
> Question : Is my bleeder to small ?
> 
> P.S. : The variac was set at 4KV at all times.
> 
> Mike-KK4MS
> a_borg at yahoo.com

I guess it is not sufficiently heatsinked, but you seem to have very 
little safety margin on what is an important component. I would tend to 
allow a much greater safety margin than you are.

That 225W is going to be on a decent sized heatsink.

Also, is your 40uF cap rated at the ripple current you need for this PSU 
? My guess is that the capacitor is being overrun on ripple current too.

Is the resistor rated for 5 kV?

The time constant of your bleeder/cap combination is
T = C * R = 4 seconds.

When you switch off, the voltage is going to fall to V volts after a period

V = Vmax Exp[ -t / 4 ]
where Vmax is the offload voltage.

So in 30 s, it would have fallen to 0.06% of its off-load voltage. 
Assuming it is 5 kV off load, that means it is down to 2.7V in 30 
seconds. That is being a bit too paranoid I think.

My suggestions would be

a) Using a high value bleed resistor

b) Putting two in parallel, in case one fails. (if the resistors reallly 
are rated for this high voltage use).

c) Not running any bleed resistors close to its maximum.

Note, high power resistors often have a maximum working voltage below 
that which the power dissipation would suggest.

-- 
Dr. David Kirkby,
G8WRB

Please check out http://www.g8wrb.org/
of if you live in Essex http://www.southminster-branch-line.org.uk/




More information about the Amps mailing list