Beware PSU current limits...
Posted: Tue Nov 17, 2026 5:39 pm
I have become used to the idea that if I want a constant current from a bench supply, I just dial the current limit to the required current then twiddle voltage to be adequate. This works on my Agilent E3648A and E3649A, and on most TTi supplies. It's a very nice way of working. A Keithley 2400 is different. Actually, everything about a Keithley 2400 is different, inconsistent, or just plain awkward. But if you can get over that, it's very capable. With a 2400, you have to explicitly tell it to be a voltage or current source, then you set compliance. And that turns out to be better...
The picture shows three stacked tag boards carrying a total of ninety nine Zeners. I have traditionally tested that I've wired such things correctly by setting the 2400 to 3mA and compliance to 210V. I can then check that I get the expected voltages at various points. But a 2400 only goes to 210V and I needed 300V, so I stacked a pair of TTi PLH250 and set the current limit to 3mA. This has worked fine before. But today, I didn't gently wind the voltage up, I dabbed it whilst watching the voltmeter. It turns out that the output capacitor (charged to umpty-ump volts) is capable of converting an entire series string of thirty eight Zeners into a short-circuit. That was an entire tag board that had to be thrown away and remade. Naturally, it was the bottom board (most difficult to get at) that was totally immolated, plus five more on the middle board. I was not pleased.
So, be warned. If you want to use a bench supply in constant current mode to test Zeners, set the current limit and gently wind the voltage up from zero. Do not dab it or switch the output on.
The remade (and very carefully checked) stack has gone into a die-cast aluminium box, been filled with white oil, had silicone sealant intended for engines applied to mating surfaces, and lid screwed down. I'll give it a day or so for the sealant to go off before disturbance and measurements.
Oh, and the feed through connectors are at Stage 2 of the dreaded "Won't solder" disease. When scraped back to bare brass, they just about soldered, but they didn't wet properly. Naturally, these are the feed throughs that I have plenty of. I have seen valve socket, solder tags, and connectors that have caught this disease. It's always tinned brass that is afflicted. Has anyone else seen this?
The picture shows three stacked tag boards carrying a total of ninety nine Zeners. I have traditionally tested that I've wired such things correctly by setting the 2400 to 3mA and compliance to 210V. I can then check that I get the expected voltages at various points. But a 2400 only goes to 210V and I needed 300V, so I stacked a pair of TTi PLH250 and set the current limit to 3mA. This has worked fine before. But today, I didn't gently wind the voltage up, I dabbed it whilst watching the voltmeter. It turns out that the output capacitor (charged to umpty-ump volts) is capable of converting an entire series string of thirty eight Zeners into a short-circuit. That was an entire tag board that had to be thrown away and remade. Naturally, it was the bottom board (most difficult to get at) that was totally immolated, plus five more on the middle board. I was not pleased.
So, be warned. If you want to use a bench supply in constant current mode to test Zeners, set the current limit and gently wind the voltage up from zero. Do not dab it or switch the output on.
The remade (and very carefully checked) stack has gone into a die-cast aluminium box, been filled with white oil, had silicone sealant intended for engines applied to mating surfaces, and lid screwed down. I'll give it a day or so for the sealant to go off before disturbance and measurements.
Oh, and the feed through connectors are at Stage 2 of the dreaded "Won't solder" disease. When scraped back to bare brass, they just about soldered, but they didn't wet properly. Naturally, these are the feed throughs that I have plenty of. I have seen valve socket, solder tags, and connectors that have caught this disease. It's always tinned brass that is afflicted. Has anyone else seen this?