Fail-safing Bleeder resistors for high voltage inputs/outputs

Thread Starter

SiCEngineer

Joined May 22, 2019
299
Hi All,

Pretty simple question here but can't find an answer on the board.

I am designing high voltage power supply and am currently choosing bleeder resistors. I had originally selected single resistors that are rated for the appropriate power and resistance and drain from 300V and 6kV to safe voltage of 50V in 5 minutes.

I have seen multiple reference designs from Texas Instruments which, rather than using single resistors across the input and output capacitors, use approximately 10 in series each rated lower than the power and voltage on a single resistor basis, but together meet the required criteria.

My question is, what is the point of having 10x200V 56k resistors, rather than having a single 500V resistor rated at 560kOhm? Is it for fail-safe proofing in case any one of the resistors fails? I would think that if this was the case, they would be connected in parallel. Because a series failure in a resistor, should mean the entire string fails, meaning the high voltage capacitors will not drain rendering the bleeder resistors unsafe and unable to drain the HV. In parallel, however, if one resistor failed, the other 9 or so would still be safe to discharge the capacitor.

Please let me know if there is any fundamental flaw in my thinking.

Best wishes,
J
 

wayneh

Joined Sep 9, 2010
16,769
I don't know anything about this but I speculate that it's simply cost and availability. A string of ten, 1/4W resistors is cheaper than a single, 2.5W resistor. Pick and place machines are more likely to handle the standard part easily.
 

ralphrides

Joined Jan 29, 2021
3
Hi All,

Pretty simple question here but can't find an answer on the board.

I am designing high voltage power supply and am currently choosing bleeder resistors. I had originally selected single resistors that are rated for the appropriate power and resistance and drain from 300V and 6kV to safe voltage of 50V in 5 minutes.

I have seen multiple reference designs from Texas Instruments which, rather than using single resistors across the input and output capacitors, use approximately 10 in series each rated lower than the power and voltage on a single resistor basis, but together meet the required criteria.

My question is, what is the point of having 10x200V 56k resistors, rather than having a single 500V resistor rated at 560kOhm? Is it for fail-safe proofing in case any one of the resistors fails? I would think that if this was the case, they would be connected in parallel. Because a series failure in a resistor, should mean the entire string fails, meaning the high voltage capacitors will not drain rendering the bleeder resistors unsafe and unable to drain the HV. In parallel, however, if one resistor failed, the other 9 or so would still be safe to discharge the capacitor.

Please let me know if there is any fundamental flaw in my thinking.

Best wishes,
J
Look at
https://www.digikey.com/en/resource...version-calculator-capacitor-safety-discharge
for calculating what is safe.

Carbon film resistors are typically rated for 300 volts with a 600 volt peak. So by putting them in series you can get a higher voltage rating. The other issue is wattage. The same applies, the wattage across each is reduced by the number or resistors in series. Regarding strings in parallel for a fail safe design the resistance will be 1/x + 1/x = 1/y so two 5 ohm in parallel are 2.5 ohms and double the wattage handled. Resistors are pretty reliable if you operate them below their rated wattage so its up to you to double up on them if one string fails but do the math and use the calculator for your caps and voltage, safety threshold voltage can be 40 or 50 volts. Higher current resistors are more expensive.
 

Marc Sugrue

Joined Jan 19, 2018
211
Hi All,

Pretty simple question here but can't find an answer on the board.

I am designing high voltage power supply and am currently choosing bleeder resistors. I had originally selected single resistors that are rated for the appropriate power and resistance and drain from 300V and 6kV to safe voltage of 50V in 5 minutes.

I have seen multiple reference designs from Texas Instruments which, rather than using single resistors across the input and output capacitors, use approximately 10 in series each rated lower than the power and voltage on a single resistor basis, but together meet the required criteria.

My question is, what is the point of having 10x200V 56k resistors, rather than having a single 500V resistor rated at 560kOhm? Is it for fail-safe proofing in case any one of the resistors fails? I would think that if this was the case, they would be connected in parallel. Because a series failure in a resistor, should mean the entire string fails, meaning the high voltage capacitors will not drain rendering the bleeder resistors unsafe and unable to drain the HV. In parallel, however, if one resistor failed, the other 9 or so would still be safe to discharge the capacitor.

Please let me know if there is any fundamental flaw in my thinking.

Best wishes,
J
Hi,
Thermal density could be a consideration, using multiple devices allow you to spread the load whereas a single device would have all the heat in one location - this reason alone in some applications may allow the use of SMT devices instead of a through hole or other form factor of device that that requires manual assembly stages. Multiple SMT devices may actually be cheaper to assemble than a single device using a manual process.

Derating factors may be applied to improve reliability of a product, Just because a device is rated at 500V doesn't mean you operate at that level continously. Surges and such like could temporarily exceed this rating and over time cause a device to fail. If you apply a 50% derating factor the chance of device failure due to stress is reduced even though the component quatity is increased.

Safety/Reliability - As you already point out parallel resistors reduce the probability of a single device failure whereas in a series configuration a single device you lose a chain. If that is a consideration then it may be a good reason for a more safety driven approach to used both parallel and series configuation. As the designer you should consider safety factors and failure modes and if theres a failure mode that can kill someone it probably needs adressing :)
 
Top