Short circuiting power supplies - Premature failures?

[Flaw]

On the production line I'm working on, we do a high current 4-wire measurement to get the resistance of a battery switch.

The way we've done it up until now is run 50A over the switch and use a multimeter to measure the volt drop over the switch.
The limit we set for the switches is 10mV (0.2mΩ), which is a difficult resistance to measure quickly.

We currently use BK1900B power supplies to achieve 50A, but we've had a power supply fail in about a year and I'm getting this feeling that the other 3 we have on the lines are going to have a similar failure.
I'm now looking at getting the AIM TTi QPX1200S, as that's a much larger supply which isn't targeted to be a low cost, high output.

Do you have any recommendations of a power supply which will be able to shrug off thousands of short circuits a week at 50A?
Or is there an instrument which could achieve the 4-wire resistance measurement by itself?

Thanks in advance for any recommendations!

[Slh]

What's the resistance of the circuit that you're putting across the power supply terminals? Why not just increase the resistance to around 0.2 ohms and set the power supply to around 12V, 50A.  That prevents the output from being properly shorted suddenly so might be less likely to fail. I guess you could have been having issues with overshoots of some sort. This should reduce them a bit and be a bit kinder to the supply.

[Gyro]

Agreed. The PSU could be seeing very significant over-current transients as its output capacitors discharge into the short. I would be tempted to drop the resistor value though (maybe 0.1R) to drop the dissipation from 500W to 250W. Either way, it's going to need fan cooling.

It would be worth doing a voltage drop test on your current wiring to see how much resistance is already present.

How are you switching the current? By closing the switch under test, or cycling the PSU.

[Flaw]

The current setup has a battery switch directly over the power supply. So a positive and negative 6awg wire with two large alligator clips. We effectively have 0.1V, 50A coming out of the supply, so a tiny wattage but high current. The ripples on the output could be causing some damage.

On the topic of switching, we're using a pair of alligator clips to connect to the battery terminals. I didn't think much of it, but now I'm typing it, 50A connected via alligator clips sounds like a potential issue with the arcing and all.

@Gyro @Slh, cheers for the suggestions!
I'll try a couple of things
- I have a few large 0.25\$\Omega\$ resistors, so I'll put a couple in and scope the output to see if it looks cleaner.
- I'll investigate a more appropriate switch, so the alligator clips can be attached without any current flowing
    + Our battery switch isn't designed to switch more than 10A (although it can run 200A while on)
    + The BK doesn't have an on/off button, but the remote port does have an off pin, so I might be able to wire that in to turn the output off

An appropriate switch and a higher voltage may be the key to a more reliable setup!

[vu2nan]

A starter solenoid switch, rated 12 V 150 A or 24 V 150 A, would be appropriate.

Nandu.

[Gyro]

You could maybe consider a shorting switch which you close before the clips are disconnected and then open for the measurement. That would keep the PSU in constant current mode rather than repeatedly shorting it from open circuit. This would reduce thermal shock to the internal components (at the cost of constant dissipation). It would avoid sparks at least.

[m k]

We currently use BK1900B power supplies to achieve 50A

Why not a battery as a power supply?

[Gyro]

It needs to be a constant current source for accurate 4 wire resistance measurement.

[m k]

Constant current source must measure and maintain its constant current.

Selecting a battery and adding some stuff can be pretty decent.

[Dave]

Definitely disable the output of the power supply before connecting/disconnecting the DUT. The output capacitors are probably not going to enjoy being subjected to large transients over and over again.

Did you perhaps have the chance to take the ruined power supply apart and check what gave out?

[Marco]

You can just buy high current micro-ohm meters, it's a common type of test equipment.