Dear all,
I have a problem with a circuit I’m building. A small n-channel mosfet gets destroyed - it becomes conductive between base and source (although no smoke is released…) and I can’t think of any reason for this. I have simulated the circuit in LTSpice, with no problems whatsoever, and it behaves as I intended.
I’m pretty beginner at electronics, and self-taught at that, so I may be missing something obvious here...
The circuit is the following (actually, part of a bigger thing, but the rest is not relevant). The mosfet that fails is Q205 (schema below).
And the PCB (the white arrow shows the problematic transistor).
This is a load switch, driven by a microcontroller, with a few “features”:
- The switched voltage (LDSW+) is 12V, same as SW_12V. But LDSW+ is present all the time, while SW_12V can be turned on and off (a microcontroller shuts down this voltage, and with that it shuts down itself too).
- An objective is to have very low current when SW_12V is turned off (i.e., only leakage current, less than 1uA).
- When SW_12V is off, I don’t want the LDSW+ voltage to leak back on the left part of the circuit through R206 and R205, so I added Q205. The idea being that if SW_12V is at ground, the mosfet will be off and no current can flow backwards. And when SW_12V is on, at +12V, the push-pull transistors Q203 and Q204 will be able to drive the gate of the switching p-mosfet, Q210, first via the body diode of Q205 then directly when Q205 turns on, as the voltage goes from 12V to 0V.
- The Q211 is not relevant to the problem, its role is to stop reverse current if load is powered directly, SW_12V is shut down and LDSW+ is connected to ground.
For my tests, the load is just the LED (but it’s expected to be a load of a few amps, driving a 25W motor under 12V (and a freewheel diode exists, but not shown here)).
So, when I power this circuit, I find that Q205 leaks current from gate to source -- the voltage at the source of Q205 never goes close to 0 when I turn the switch on, but remains at 3-4 volts. To make sure this is really the problem, I tested with a 2n7000 mosfet. When new, not in circuit, I can apply 15V in all directions: gate to source, source to gate, gate to drain, drain to gate, and there is no current passing. Then I put it in the circuit, remove it and measure again, raising the voltage slowly. Once I get over 2-3 volts gate to source, it starts conducting 1-2 milliamps (and goes quickly higher if I increase the voltage). So basically the mosfet dies. I tried this 2-3 times, always with a similar result.
I also tried with a 1k resistor at the gate of Q205 (which would have been a good idea anyways), with the same result.
If I remove Q205, the push-pull transistors work fine, the voltage goes from almost 0V to almost 12V.
So, I’m a bit out of hypotheses on why this happens. I haven’t seen anything strange with an oscilloscope. Here is a screenshot with the Q205 removed, the blue trace measuring the net at the source of (the removed) Q205, showing the two transitions: going low (switching on) and going back high (switching off). (The former being much slower than the latter, but that is normal.) The yellow trace is just the power, SW_12V. When Q205 is present, the trace does not go to 0V, but remains at 3-4V minimum.
I'd appreciate your opinions, and anything I can test. Thanks!
Home
Help
Search
Login
Register
