Analog Discovery 2 Impedance Analyzer questions

[Jared]

So I see that that the impedance analyzer add on board utilizes 0.1% resistors based on the response by Attila on Oct 17, 2018. 

Could someone tell me if these resistors are non-inductive as well?

Also, after performing the calibration/compensation for the impedance analyzer, the impedance of a short circuit is nice and flat all the way up to 25MHz. Can I reasonably trust the values in this region of 1 MHz to 25 MHz when testing actual components? I didn't know if compensation would be enough to ensure the readings are accurate up this high.

Finally, after performing the compensation, are the values stored permanently or does this need to be performed each time the instrument is powered up? If they are stored, how can I access the values to back them up if I wanted to?

Thanks,

Jared

[attila]

Hi @Jared

The compensations are stored in project/workspace. You can save such for each device and/or circuit setup.

The IA module uses generic resistors 0.1% 0603.

This is not a metrology device. The measurement accuracy depends on the scope input voltage, but the math gives good results. 
Here is a ceramic cap of 39pF nominal, trace is compensated and ref 1 is not.

 

[Jared]

Thank you. A few more things:

1. Say I perform a compensation using using average or averaging and some particular number of samples, do I need to use the same settings to perform an actual measurement afterwards? For example, if I compensate the device using 2001 samples and a 100ms averaging, but then I choose to do a check of a device at 501 samples and no averaging, will the device compensation still apply the compensated values to my measurement?
2. I think the average and averaging explanation in the instructions needs to be clarified. I realize now that the average is a fixed value and the time becomes variable whereas, averaging is fixed in time and the number of averages accumulated is variable. As simple as this might be it took me along time to realize this. I couldnt figure out what the difference was between the two for a long time.
3. Is it possible for users to contribute to the documentation? I really feel it needs to be expanded/clarified on a lot.
4. Why are the number of samples always some value plus an additional one? for example, why 2001 samples or 501 samples rather than just 2000 samples or 500 samples. I've seen this on other instruments which is why I'm curious about it.

Thanks,

Jared

[attila]

Hi @Jared

1. It is recommended to use the same start/stop frequency and number of samples for compensation and measurement.
Otherwise the software may interpolate the closest compensation from frequency steps, which may not be ideal compensation value.
The 2001 compensation and 501 samples are integral multiples, if you use the same start/stop you will have compensation for each step.

2. The "averaging" expressed in time and the "average" of measurement counts has AND logic, in each step it will satisfy both conditions.

3. Feel free to post if you think something is missing. Just like the application the documentation is also under constant development.

4. The X * /Decade +1 is to have nice intermediate frequency steps.

Like here between 1kHz and 1MHz with 5/Decade you get 3*5+1 = 16 steps to have values for 10kHz and 100kHz

 

With 15 steps you wouldn't have 10k or 100k: