Curve Tracer Project

[soligen]

Hi,  I am working on a basic curve tracer circuit for use with my oscilloscope and I have some questions I hope to get help with.  I am trying to keep it basic, using jelly bean parts I have on hand and optionally running it from a battery.  Here is where I am so far.

I am generating a 6V peak to peak 1 Khz triangle wave with the bottom peak touching ground, then running it through a 1k resistor which both limits current and allows current to be measured.  This is working and I can see curves for diodes and the oval for some capacitors, however only small capacitors work.  Larger capacitors and all the inductors I have on hand to test show as a short circuit (which makes sense to me given the frequency and current limit).  A really rough schematic is attached.  It contains options for removing the DC offset and also going to 12V peak to peak, but I have not tried these options yet. It works from 7.5 up to 16, so if running on higher voltage, a 12V trangle wave should be easy to achieve. The op amps are speced for 30 ma max output.  The 1k output resistor keeps things well under this.

It seems to me that this would be much more versatile if I had selectable frequency and/or output resistances.  I am building this just because it seemed interesting - I don't yet have any specific use in mind for this device.  So, the question is: What other frequency and/or output resistances would be good to have?.  And, a follow up to this is how to size capacitor C6 for removing DC offset when these things can be changed.

I'd like to add a step generator at some point too, but I haven't started looking at this yet.

Any other traps for "rookie" players  that I should look for?

Thanks

[oldway]

I did not quite understand the purpose of what you are doing.

If it's a simple component tester, just use a little transformer and the 50 or 60 hz mains voltage, no need to generate a triangular wave.

The component testers are extremely simple and are sometimes even included in the oscilloscopes (Hameg, Hung Chang).
Search on google "curve tracer octopus"

The most sophisticated models are the HUNTRON, they even offer the ability to memorize curves and compare them.
http://www.huntron.com/

If it is a curve tracer for transistors like the TEK 586, 577, or other, it is a much more sophisticated circuit.
See for example the heathkit IT-1121.
http://tubularelectronics.com/Heath_Manual_Collection/Heath_Manuals_IT-NE/IT-1121/IT1121_man.pdf

For a curve tracer with multiple curve, it is best to use a high remanence or memory cathode ray tube oscilloscope.

[danadak]

The Tektronix 575 was a mainstay of the industry for a long time,
and its manual shows in the circuit chapter how it works. Even
though vacuum tube might be useful for ideas.

http://bama.edebris.com/manuals/tek/575-175/

I use PSOC UPs for a lot of stuff, and it has onboard DAC, A/D,
reference, wave generator (SAW / Step included) and is pretty easy
to program and use. I have often thought it would make the basis
of a good curve tracer. It can develop base drive, I or V as well, and
measure currents with its onboard delsig A/D.


Regards, Dana.

[Ian.M]

Some older CROs have rear panel outputs for the timebase ramp.   If so, a simple fast OPAMP buffer amplifier can be used to provide the drive signal to a curve tracer.   Sum in a pot across the rails for DC level adjust.   Then if you set the trigger to auto and ext, with minimum holdoff, it will free-run and you can control the rate with the normal timebase controls, so you can suit its speed to the reactannce of the component you are testing.

Also you can use a comparator to generate a signal off the falling (flyback) edge of the ramp to drive a counter to step another parameter.

The output stages required are highly dependent on the components you want to trace curves for - providing 0uA - 1mA into the base of a small signal transistor is very different from providing 0 to 20V to its collector with a peak current of several Amps.

However, any CRO based semiconductor curve tracer has severe deficiencies compared to a MCU based one - its very difficult to do pulsed tests to minimise the dissipation, and without a storage scope it is unsuitable for matching components, unless you screw around with an acetate overlay on the screen + an OHP pen.