Virtual Impedance Analyzer

[SuperSnake428]

I'm interested in measuring the complex impedance of a sensor in order to characterize it.  This VI looks very promising, but I am not sure what the exact circuit setup is for it.  Can anyone explain?  Also, do you think that it would be possible to convert to an M series card or is the DSA card strictly required?

Thanks,

Drew

Message Edited by SuperSnake428 on 02-24-2008 10:13 PM

[jaced]

Hi Drew,

I believe that that circuit detail in the first post should work for your test; your sensor will be the unknown impedance. Perhaps I misunderstood where your confusion is, on this issue. The DSA card in question uses the DAQmx driver so in theory, you should be able to use this code with an M-series, with the phase considerations mentioned earlier in this post. One issue that could arise is a property node which is DSA specific is used. I have not gone through this code to determine if this is the case but it is a possibility to be aware of. I have also found this article which expands on other methods to accomplish this task. I hope this helps!

[SuperSnake428]

Does this VI make any assumptions about the input amplitude or the resistor?  What about the frequency, you have to actually sweep it externally?  It just seems like the oscillator could be the DAC right?

[EBL]

If you're looking for information on the test fixture, it's fairly well described in these:

http://forums.ni.com/attachments/ni/250/32912/1/Impedance%20Fixture%2002.png

http://forums.ni.com/attachments/ni/250/32912/2/P8251778a1.JPG

http://forums.ni.com/attachments/ni/250/32912/3/P8251779a1.JPG

As for using an M-series card in place of a DSA, the biggest concern I have is the lack of anti-imaging filters on the output and anti-aliasing filters on the input. This means that high-frequency components of the excitation signal (produced by the analog output) will create images at frequencies above one half of the sample rate. Similarly, the ADC will alias these images back down to their original frequencies below one half of the sample rate. And so impedance measurement results for frequencies in the band of interest may be influenced by impedances at frequencies outside the band of interest (i.e. above one half the sample rate).

This problem can be kept to a minimum if the sample rate is kept considerably higher than the frequency band of interest. For instance, if you could run at a sample rate of 250 kHz and limit your frequency band of interest (and hence your excitation signals) to below 12.5 kHz, the images will be attenuated by more than 26 dB. Even with 26 dB of attenuation you could still run into trouble, especially in cases where the impedance at the original frequency was close to the bridge impedance (1 kohm in the documented fixture) and the impedances at the images were very low or very high, producing a higher bridge indication signal than the un-imaged frequency impedance.

Of course, with such a large separation between the sample rate and the band of interest, it would be fairly easy to build analog low-pass filters to serve as anti-imaging / anti-aliasing filters. Then you would avoid the frequency ambiguity problem altogether.

Hope this helps,
Ed L.

[SuperSnake428]

Ed,

Thanks for the fixture diagram.  Looks like a pretty straightforward fixture, high pass filter on the output of the DAC and then just the resistive divider on the input.

I guess my biggest confusion is (and I haven't completely deciphered the VI, so correct me if I am wrong), this VI doesn't seem to sweep the frequency so how does it actually measure the impedance as a function of frequency???  Am I missing something here?

This is my interpretation of this VI:
- Measure V1 and V2 across the oscillator and the DUT respectively.
- From this calculate the S11
- From the S11 you can calculate L, C, R....

As for the imaging concerns, very understandable.  I would sample at 250ksps.  However, the M series do have an AAF on the input of the card so i don't actually think that this is a big problem.  Realistically, i'm really only concerned with the impedance up to about 50KHz anyway.

Drew

[EBL]

>>Does this VI make any assumptions about the input amplitude or the resistor?<<

The way the VIs are written, certain voltage levels are assumed, but only to the extent that certain voltage ranges are set as the defaults. Those, of course, can be changed. The VIs do assume the resistor is 1 kohm, and that -may- be hardcoded in, but of course that can be changed with a little fishing around inside the code. Incidentally, the resistor value should be chosen to be around the center of the dynamic range of impedances you expect. For example, the 1 k resistor covers complex impedances from 0.1 ohm to 10 Mohm nicely. However, if all the impedances you expect to measure were all around 10 ohms, for instance, you might use something like a 10 ohm resistor.

>>What about the frequency, you have to actually sweep it externally?  It just seems like the oscillator could be the DAC right?<<

The first VI in the set generates a multi-tone excitation signal out the DAC. No muss, no fuss.

Ed L.

[EBL]

>>I guess my biggest confusion is (and I haven't completely deciphered the VI, so correct me if I am wrong), this VI doesn't seem to sweep the frequency so how does it actually measure the impedance as a function of frequency???  Am I missing something here?<<

There's no "sweep" in the old-fashioned sense. The DAC generates a multi-tone that covers the band of interest, and the entire spectrum is acquired at once. An FFT sorts out the behavior at different frequencies after acquisition.

>>This is my interpretation of this VI:
- Measure V1 and V2 across the oscillator and the DUT respectively.
- From this calculate the S11
- From the S11 you can calculate L, C, R.... <<

Yep, that's pretty much it.

>>As for the imaging concerns, very understandable.  I would sample at 250ksps.  However, the M series do have an AAF on the input of the card so i don't actually think that this is a big problem.  Realistically, i'm really only concerned with the impedance up to about 50KHz anyway.<<

Very good. I wasn't sure which card you wanted to work with, but if you have an anti-aliasing filter already, you should be fine. An anti-imaging filter for the DAC would be nice, but it's not necessary, since it should be sufficient to have just one or the other.

Ed L.