Damaging High Voltage Resistors with Arcs and Short Circuits - John Dunn, Consultant, Ambertec, P.E., P.C.

High voltage resistors that get subjected to an arc usually suffer damage near their high voltage end, less likely in the middle somewhere and almost never at the low end towards ground.

The reason why is evident in the following sketch:   

   

To emphasize the point, in the extreme case with just one resistance element and one capacitance element, we'd have this situation:    

     
The resistor sits harmlessly until the switch moves down and then that resistor gets hit hard.

Purchased high voltage dividers need to be very carefully protected from short circuits and arc-overs. Therefore, never allow an arc-over to happen when one of these is connected and in service.      
Someone once told me that my 50000 volt high voltage divider in a particular power supply at the time should not be allowed to see a dv/dt of anything more than 10¹¹ volts per second, so I put a 10K resistor at the top of my high voltage resistor and added a dv/dt limiting capacitor:   

 
It worked just fine. The divider never showed any discernable damage from short circuit tests of my power supply.

Comments

[Anthony Stewart EE'75]

We used 3KV testers to for HIPOT safety tests on all AC powered products in production. I also used special Human and Cart model circuits for ESD simulation of discharge up to 15kV for susceptibility testing of human interface products.

Often you want to test for voltage breakdown but you don't want to induce catastrophic failure in the test process. Also the capacitance of the tester may be high enough to cause amps of discharge when you are only concerned about uA of leakage. ... So my solution in this case was to use a low capacitance 10k resistor in series inside the voltage source probe or connector. Fusing a current limiting resistor is a lot better than fusing your DUT, but typical dielectric breakdowns are momentary. At the time we were testing $200 power supplies so when failures from supplier quality issues with insufficient gaps or insulation material were detected with less risk for damage by current limiting the source.

Much like the RC divider you added to your divider.. I simply added a low capacitance current limiting resistor closest to the source to prevent similar failures. The failure mode being the capacitance discharge from source to load under dielectric stress.

Axial leaded 1/4 w resistors are typically a few pF or less.

But your

[Kamau]

Interesting