Typical bench supply CC response time?

[RossS]

I recently attached an LED to a bench supply set to a reasonable current limit. The LED was destroyed by a power spike before current regulation kicked in.

I tried shorting the supply (HMC 8043) at 3V/500mA while attached to a current clamp probe, and it took nearly 1ms to settle. An older (fully linear) TTi supply settled in just 50uS.

What's reasonable to expect from a supply here? Is my Rhode & Schwarz supply broken? Or just designed with a high output capacitance? Current limiting has protected me against dumb mistakes in the past while prototyping, but now I'm questioning what I can actually rely on.

Thanks!

[Algoma]

The output smoothing capacitor(s) are often fully charged up to the output voltage if the supply is on before you connect your load.

Those caps will happily discharge their full voltage and capacity into your poor LED long before the current limits from the rest of the supply will register and react. As those Capacitors are directly at the output, they're after the current regulating system, so they have the potential to damage sensitive items if they're already holding a voltage above the safe limits of circuit that you suddenly load them with.

[Kleinstein]

It is quite common that the current limit is not really fast. So it may not fully protect sensitive circuits.  The problem is that some circuits take quite some current spikes and really fast engaging current limit can than also be confusing, as it could trip way before the average current reachse the limit.

The shown example with some 500 µs delivering some 10 A shows quite a bit of charge before the CC mode kicks in. This is some 5 mC, which would be some 1500 µF charged to 3 V.

Besides the actual input capacitance there is usually also some delay in the electronics that can also add to the real capacitor.
One could check this when using an even smaller voltage like 1 or 1.5 V - if the current spike gets considerably smaller it is likely a real capacitor. The electronic part often does not change much. The shape of the current spike suggests this is more like an electronics delay.

It may be an unplesant surprise, but likely not a defect, more like a not so great design. Of cause this depends on the power rating of the supply - for a 20 A supply this may be OK.

[dougg]

Since I'm looking at documenting RD60xx power supplies, perhaps now would be a good time to measure this on my RD6018. V-SET=10V, I-SET=10A with a 5 Ohm load (two 10 Watt, 10 Ohm resistors in parallel) followed by a short circuit. After starting at 2 Amps the current goes up to ~ 27 Amps (perhaps higher) for about 200 microseconds then settles back to around 10 Amps after about 600 microseconds. That is shown in the first attachment.

Put the 80i-110s in high current mode for the second attachment. It warns not to exceed 100 A(peak), but it was only for 50 microseconds :-( The power at the first voltage and current intercept was around 500 Watts. Ouch.

[nctnico]

I recently attached an LED to a bench supply set to a reasonable current limit. The LED was destroyed by a power spike before current regulation kicked in.

I tried shorting the supply (HMC 8043) at 3V/500mA while attached to a current clamp probe, and it took nearly 1ms to settle. An older (fully linear) TTi supply settled in just 50uS.

What's reasonable to expect from a supply here? Is my Rhode & Schwarz supply broken? Or just designed with a high output capacitance? Current limiting has protected me against dumb mistakes in the past while prototyping, but now I'm questioning what I can actually rely on.

The problem is between the seat and knobs. You need to use the output enable switch to turn the output on AFTER connecting the load. That way the PSU will ramp the output voltage up until the voltage or current limit setting is reached (whichever comes first).

If you connect a load with the output enabled, then the capacitors will discharge before the current limit can kick in. I have modified quite a few of my bench supplies with an output enable switch; I consider it a mandatory feature on a bench supply.