I wasted a lot of time today trying to get a good current-vs-voltage plot for a diode-connected AOI514 nFET, which is the one I plan to have the students use this year. When I started on it this morning, I thought I was working on my book, but in the end I decided not to include the I-vs-V plot in the book, so most of the time ended up being wasted.
I breadboarded three different test jigs for measuring the current of the nFET:
The simplest test jig just uses the function generator to provide the power for the measurements and a Teensy board running PteroDAQ to make the measurements.
The second jig allows higher voltages on the function generator, hence somewhat higher currents (limited mainly by the 50Ω output impedance of the function generator, but also by the current limits of the function generator).
The most complicated jig uses an external power supply with the function generator controlling the current by changing the gate voltage of an extra nFET.
The 10µF capacitors for removing noise on A10-A11 are almost certainly too big, introducing bias into the measurements, but the 22nF capacitor works well.
The weird plots at high currents show the effect of temperature changes on the FET characteristics. At 3A the transistor got warm, but as the current dropped it cooled off a little, getting a bit warmer on each 5 second cycle.
At the other end, I had some difficulty measuring currents less than 1µA—current-sense resistors large enough to give sufficiently large voltages would be too high impedance to handle the noise injected by the sampling circuitry on the Teensy 3.1. I also went to 60Hz sampling, to alias out 60Hz interference from capacitive coupling to the breadboard. I still don’t trust the measurements below 500nA.
I decided that the I-vs-V curve here is too messy to put in the book, so instead of working on my book all day, I’ve wasted my time getting nothing more out of it than this blog post and a reminder that even “simple” concepts like I-vs-V plots are not so simple in the real world.