curve tracer

[ed]

I mentioned a curve tracer in another thread in regard to filling a spare hour. The idea has now taken root and I've got the urge to see this through. I'm sure there are countless examples of this on the interweb but I thought I'd treat it as an academic exercise and see what I can come up with without copying anything....

the very simple premise is:

provide a range of Vgs
provide a max Vd

for each Vgs
for every Vd from 2 to max, measure the Id

for every Id at Vd plot the curve.

here is a very simple(naive) circuit which I think will accomplish this. The left hand v.reg will be a 338, but I couldn't find a model for it so I can't hook the circuit up.:

and here is a plot of an integer sin wave to demo the user interface. The esp processor will control the digi pots and measure at the test points, reporting back a table of results for the user interface to plot.

If anybody is interested I'd welcome any comments, partcilularly if anybody thinks the method is over simplistic. If there is any interest I'll post the

[ed]

wierdest thing.........it cut me off before finishing, and wouldn't let me edit at the end......c'est la vie!!

I was going to say I'll post any results if there is any interest.....My excuse for this is probably that I'm going stir crazy with the lock down and there's only so much primary school teaching a chap can survive.

[Nick]

Only thing I can think of that may be worth considering is that a lot of devices will need the voltage applying in pulses to avoid exceeding dissipation limits.

[ed]

I originally drew a sketch of the off-board unit and it had a large heatsink with a screw and butterfly nut for the test device. My first Vd power supply was a FET and resistor which prompted the use of the heat sink....but I soon abandoned that idea in favour of the 338. It did alert me to the fact that to get realistic readings the device will need to be at least a little bit warm....

when you see characteristics on the datasheet do you know if they are arrived at by pulse driving???

[Nick]

when you see characteristics on the datasheet do you know if they are arrived at by pulse driving

When they are outside of the DC SOA I would assume so. I wonder if they maintain the heatsink at a constant temp for the tests?

[ed]

I've just had a rethink with this.

The v.reg method is proving tortuous because I can't get the supplied voltage down low enough with an LM device and the digi pots might be upset with the adjust current, so I've re-jigged it with bjt amplifiers instead. That way I can cope with the Vds current and also get the voltages down to 1 or 2. There will have to be a heat sink involved because the heavy duty power supply will need to go up to 30v and 3 amps, so the DUT can be strapped to this as well.

I'm still pondering on how to reverse the voltages though.

This is quite timely as I've had the urge to revisit the digital pre amps again...I've located some breakout boards and I've got some muses 72320 and also some MAS/DA6116 on order to do some comparisons.

forward

[ed]

a bit of a revision for the power supply:

V2 and V3 will be 12 bit DACs running at 3.3v or 5.0v and driven by Atmel or esp32.
it sims ok but If anybody can see any obvious error then please let me know. I keep thinking I've missed something silly.