EEVblog Electronics Community Forum
Products => Test Equipment => Topic started by: Kronkulus on September 01, 2022, 06:57:17 pm
-
At work I was testing a constant current sink circuit. It uses a zener and a resistor in combo with a p-fet to set the sink current. It is designed to work up to some high DC voltages, around 150VDC.
I had a current clamp on the supply input, and a differential probe across the current set resistor at the same time. The current clamp sees some really high amplitude 125kHz oscillations, while the differential probe sees a steady voltage across the current set resistor. The Vgs of the fet is steady as well. There's a logic feedback signal as well thats only active when the current is flowing through the current sink. This all points to this just being the current clamp going nuts.
Any ideas as to what could be causing this oscillation on the output of the current clamp?
-
Assuming p-fet means P-Channel MOSFET not P-Channel JFET. I've seen oscillations in a related circuit before. Zener noise could be coupling onto the MOSFET gate. Can you post a diagram of both the circuit with part numbers and the test setup?
In my case, putting a 100Ω resistor between the Zener and the MOSFET gate helped remove the oscillations I was seeing.
If you can glean the transconductance of your MOSFET at the circuit's operating point from the datasheet or from a SPICE model, you can work backwards from the channel current oscillation amplitude to a ballpark Vgs oscillation amplitude. It might be small. I should also note that, unless you're using an Exicon EXC10P20, I am not aware of any modern P-Channel MOSFETs that are designed for linear applications like this. You will be risking thermal instability. Older, higher voltage planar parts with wide SOA ratings have the best chance of being ok, but they're still not designed to operate in this mode. IXYS makes power N-Channel MOSFETs designed for linear applications; those are worth a look. You might also want to consider a depletion mode N-Channel MOSFET, as those are designed specifically for this kind of application. A circuit based on BJTs would also be ok from a thermal stability standpoint provided you avoid second breakdown.