Electronics > Beginners
Oscilations in digital uCurrent source
(1/1)
Melon:
Hello, I've been lurking for a while but I have finally decided to make an account since I have a question myself. I'm quite a beginner so yeah, feel free to point out any dumb stuff. Anyway, to the point.
I'm designing a digitally controlled current source in the range of 0 to 100uA (see first attachment). Basically I use a digital potentiometer (100kohm total resistance) as a voltage divider for a reference voltage of 3.3V coming from a regulator. The OpAmp is powered by [5,-5]V voltage regulators. The jumpers are just for measuring the current when the target chip is not placed so I can connect the transistor to GND and have a path for the current. The circuit works, but when I measure the negative pin of the OpAmp with an oscilloscope I can see small oscillations at a frequency of 300kHz (or 600kHz, I can't remember exactly right now). On the other hand I don't see anything if I measure the amplifier output. Furthermore, when I use a better probe (not just the crappy coaxial) the oscillations aren't there either. I assume this might be due to the capacitance of the probe that filters them out but can't confirm.
The 3.3V reference is clean (at least at that frequency, I have some other small oscillations at 12MHz due to bad bypassing in an oscillator but thats another matter). I have also attached what I think is the relevant information from the datasheets regarding stability. So.. what's wrong here?
Gyro:
The first thing to try is a series resistor mounted close the to MOSFET gate. If that doesn't work then it may be necessary to put a capacitive feedback network around the opamp to improve its HF stability.
At the moment the opamp is directly driving a capacitive load.
Kleinstein:
The BSS84 is a rather small FET. So gate capacitance may not be so bad. Still a series resistor is a good idea.
If the oscillation is only seen with a poor probe - it could be in parts be due to the probe. A longer cable could add a resonance and make decoupling worse. If critical it could help to have additional series resistors (e.g. 10 ohms) where the supply to the OP comes in.
With some inductance on the drain side the FET stage can add some phase shift and the usual extra RC may be needed to slow down the OP.
Navigation
[0] Message Index
Go to full version