Discovery 2 Impedance Measurement and Compensation Issues

[Morgan MONROE]

Hello All,

I recently bought a Discovery 2 with the Impedance Analyzer attachment in order to characterize the dielectric properties of some capacitor samples I'm making for my PhD. I recently discovered an issue with the system, though, and I was wondering if someone can help me resolve it.

 

My measurements take place in a vacuum chamber, so I've had to assemble cabling from the Discovery2 through a vacuum feedthrough. When I take measurements of reference capacitors, they measure correctly when directly attached to the Discovery unit, but as soon as they are connected to the feedthrough, the values of my capacitors are consistently offset. When I measured the capacitance at 10kHz, for example, all of my reference capacitors were offset to a value 63% of the nominal capacitance. This multiplicative value varies depending on my cabling situation, but never goes away.

All of my cabling is shielded, and I have correctly implemented the Open+Short compensation feature in Waveforms, but this has not resolved the issue.

I suspect strongly that the issue has to do with a combination of:

• (1) The method that the Discovery 2 completes its compensation function
  • I measured the system with an Agilent 4294A Impedance analyzer (which is powerful, but old and very slow, hence the switch to Digilent tech!), and that system was able to correctly measure the capacitance of my reference capas under the same conditions,
  • Since the Disco2 measures correctly when the vacuum chamber isn't involved, this implies that neither the cabling nor the samples are causing the issue
• (2) some form of grounding loop
  • My vacuum chamber must (of necessity), be grounded to the mains of the lab workbench. When I unplug the grounding line of the chamber, the error varies, implying some correlation of the chamber grounding to the error
  • I noted that the impedance analyzer attachment does not connect the sample Low directly to ground, so that could be one source of the error. When I try to ground the sample to the chamber, the error I see gets much worse. 

I tried to look into how the compensation has been conducted, but was unable to find documentation. The change in phase I see when the full system is assembled is not compensated out. I would like to know how the compensation works to see if there is a way I can hard-code this strange behavior. Capacitors should have an ideal phase of -90deg, which is true for a directly-connected sample, but not the case for samples going through the chamber feedthrough, so there is some extra electrical elements causing the shift, but the Disco2 cannot compensate for them, and I am unclear on precisely what is causing it.

Does anyone have any experience with this weird situation, or could maybe propose some solutions? I am a materials scientist, not an Electrical engineer, so this is not my area of expertise. 

I've attached a few figures for clarity. Any assistance is greatly appreciated!

Cheers,

Morgan

Edit: Moved images to bottom of post

Edited June 4, 2021 by Morgan MONROE

[attila]

Hi @Morgan MONROE

On the IA adapter the contacts are Scope1&2 like in the following setup.
You can find the reference manual and schematic at: https://reference.digilentinc.com/add-ons/impedance-analyzer/start

The compensation is Zdut = Zopen*(Zshort-Zmeas)/(Zmeas-Zopen)

The open/short compensation may not remove the addition capacitance added by the chamber/wiring.
You may try to adjust the ProbeCap under Option dropdown to have a flat response. This, with Auto option for AD2 w Adapter is 43pF.