[stef_d]

hi folks,
I am trying to work out the correct sizing for a bleed resistor, which is supposed to empty the capacitors on my home made PSU, as soon as the psu is switched off. Here is what i've got, can someone tell me if i am miles out ?

So,
Voltage at t=0 is 80V
Capacitor value is C=50mF (milli Farrads)

I take the time constant to be T=RC, which gives T=5sec for R=100 Ohms.. i can live with 5 seconds. now, for the power rating.

a 100Ohms R would give me a DC current of 80V/100=0.8A. and there for a power rating requirement of 64W for the R. However, this is not DC, it's exponential decay, going at exp(-t/RC).
If i integrate this decay over 1sec to get the max power rating, i end up with 58W... not much of a saving !

I plugged all this into excel, and couldnt find any better way to do this. If i increase R, I reduce the Power rating, but increase the time constant. i worked out 8W for 40sec for a 800Ohms R.
am i handling this the right way ?

[Pandinus]

The Time constant is "Tau" and IIRC it takes about 5 Tau (logarithmic function) to get to about 98% discharge
so your discharge would be 25 seconds...

How about adding a relay so when the supply is powered on the bleed resistor is bypassed and not wasting power/overheating the resistor.

for a short term, high power load on a test circuit I have used a kettle filled with "watt"er...

a 240V 120watt light bulb would be about 50 Ohms... but be careful

add a LED and resistor to indicate if the circuit is charged up.

[Pandinus]

Just a thought, are you going to add a soft start for charging the capacitor without damaging the supply???

[stef_d]

Just a thought, are you going to add a soft start for charging the capacitor without damaging the supply???

why would the supply get damaged at switch on ? it's a basic linear supply with a 640VA transformator, a rectifier, and 50mF of cap. as far as i understand, the voltage is just going to "slowly" raise to 80V whilst charging the caps... no ?

yes, the resistor is on a relay bypass, which comes across the caps when the power is switched off. this is the resistor i am sizing up.
I think i can live with a 470Ohm/25W.
This costs about 3euros, and will have a tau of 23 sec..i just need to add a viual signal, as you mentioned to remember not to mess around the caps before a minute or so.

[RomanLini]

Assuming the PSU is actually attached to something (like stepper motor drivers!) it is probably sourcing 2A or 3A into the drivers, so that 50mF cap will be discharged very quickly after power down.

Normally a bleed resistor like that is for safety situations where the PSU may be on test etc with no load and you can use any reasonable value resistor. Normally manufacturers will pick a physical size resistor they wish to use (say 1W) and spec the bleed at a safe continuous heat dissipation in a warm environment (say 0.2W) which in your case would be; 80v @ 0.2W = 2.5mA which is about 33k 1W resistor.

Of course if you have specific reasons why you need to quickly discharge that PSU when it has no load then there are options for relays, SCR crowbar circuits etc.

[stef_d]

Assuming the PSU is actually attached to something (like stepper motor drivers!) it is probably sourcing 2A or 3A into the drivers, so that 50mF cap will be discharged very quickly after power down.

Normally a bleed resistor like that is for safety situations where the PSU may be on test etc with no load and you can use any reasonable value resistor. Normally manufacturers will pick a physical size resistor they wish to use (say 1W) and spec the bleed at a safe continuous heat dissipation in a warm environment (say 0.2W) which in your case would be; 80v @ 0.2W = 2.5mA which is about 33k 1W resistor.

Of course if you have specific reasons why you need to quickly discharge that PSU when it has no load then there are options for relays, SCR crowbar circuits etc.

Of course, i didnt think about that. The caps will remain connected to Geckos G202. They will take care of any charge left over !
oh well.. anyway, a 33k resistor would not work for me.. well it would, but it would take about 1650secs to get to tau. over an hour to be safe.

I purchased and hooked up a 25W, 470ohms today. no other load on the supply (it is still under test). it's hooked up on a NC contact of the main relay, so only comes in when the supply is switched off.
emptying takes roughly what i calculated. around 1 to 2 mins to go down significantly.