1
\$\begingroup\$

I'm trying to get IV curves of the BFP720FESD NPN bipolar junction transistor using a PXIe-4162 SMU.

I do a sweep of 0-3V on the collector for six base currents between 5uA and 85uA and measure the collector current.

DUT schematic (please ignore the erroneous JFET symbol representing a BJT):

enter image description here

enter image description here

Here is a picture of the setup:

enter image description here

In the picture of the setup, the SMPM3 has a loop back I attached between the signal and shell of the SMP connector so that the emitter is connected to ground.

Currently, my IV curves appear correct up until roughly the same percentage of their maximum Ic (or to the same slope depending on how you want to look at it). Please ignore the random disappearance of the curve in the graphs (I believe that's a separate issue).

enter image description here

For reference, here is an IV curve I took of the same transistor on a different board with no capacitors and with its emitter connected directly to ground on the PCB (no external loop back to ground).

enter image description here

enter image description here

What is causing my IV curves to dramatically dip as if partially turned off and the other visible odd behavior?

\$\endgroup\$
12
  • \$\begingroup\$ Please show board layout or a photo of the wiring (preferably both). Why was a JFET symbol used in the schematic? I suppose it doesn't matter for the question but it is confusing to see. Also, do you have any RF probing equipment available? \$\endgroup\$ Commented Jan 4 at 1:07
  • \$\begingroup\$ yeah I'll get that updated. The JFET symbol is what the schematic designer used... I suspect he was just being sloppy. I do have rf probing equipment like VNAs, Spectrum analyzers etc. It's not clear to me why the behavior in the RF domain would matter here. Could you elaborate? Hold tight for the updated info \$\endgroup\$
    – Andrew
    Commented Jan 4 at 1:14
  • 3
    \$\begingroup\$ Apart from a possible junction voltage breakdown, could your test fixture be breaking in to oscillation at those bias points? You only have to look at them sideways before they go squirrelly. \$\endgroup\$
    – glen_geek
    Commented Jan 4 at 1:20
  • 1
    \$\begingroup\$ @FabioBarone ok awesome. Thanks \$\endgroup\$
    – Andrew
    Commented Jan 4 at 1:58
  • 1
    \$\begingroup\$ Sorry, PXIe-1095 is the chassis. What I meant is PXIe-4162 which the card installed which is a Source Measurement Unit. Here's a link ni.com/en-us/shop/model/pxie-4162.html \$\endgroup\$
    – Andrew
    Commented Jan 4 at 2:08

1 Answer 1

4
\$\begingroup\$

I've seen this frequently when using excessively inductive (i.e. excessively long) cables to connect an SMU to the device under test. What's happening is that your SMU is oscillating, thus failing to reach its set point, so the measurement ends up being a fraction of what the true value would be.

Note that this is not your transistor turning off at higher collector voltages; this is either the collector voltage or the base current not actually being what the SMU thinks it is.

The simplest solution for this is to just shorten up the cables between the DUT and the SMU. If that's not an option, experiment with ferrite beads or resistors in series with the SMU's force (not sense) lines.

\$\endgroup\$
5
  • \$\begingroup\$ Shouldn’t this be avoided if you wait long enough before taking a measurement? I’m waiting about 1ms between steps \$\endgroup\$
    – Andrew
    Commented Jan 4 at 2:49
  • 1
    \$\begingroup\$ @Andrew This isn't ringing that dies down over time; this is oscillation that is actively driven. \$\endgroup\$
    – Hearth
    Commented Jan 4 at 4:14
  • \$\begingroup\$ @Andrew Most SMUs can't deal with active RF oscillators very well. That's exactly what you got here. \$\endgroup\$ Commented Jan 4 at 18:08
  • \$\begingroup\$ @Kubahasn'tforgottenMonica Ah, right, it could be an oscillating DUT and not an oscillating SMU--the solution for both should be about the same, though. \$\endgroup\$
    – Hearth
    Commented Jan 5 at 1:26
  • \$\begingroup\$ So I looked at it with an 8ghz scope as a quick way to find oscillations and I definitely see a growing oscillation on the transistor that doesn’t work and no oscillation on the transistor that does. So the problem is definitely what @Hearth suggested. Why this one is oscillating and not the others on the same test setup is still a mystery to me. Thank you guys \$\endgroup\$
    – Andrew
    Commented Jan 5 at 2:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.