How can I measure capacitance using labview?

[Ashley H]

My research group would like to use accelerated testing ovens to test multiple packets, and the item of data most important to us is the change in capacitance.  The problem we keep finding is that no DAQ that we've been able to locate can read capacitance of a least 4 (optimally 😎 different pieces at the same time.  Our funds and available space are very limited, so stacking 8 LCR meters would not be a good option, but I do have access to a computer with LabView and I was wondering if there was some way of reading all 8 of our pieces using the program, probably with help of another piece of equipment.  If anyone has any ideas or knows about a DAQ that would fit the bill, I very much welcome all suggestions.

Thank You,

Ashley H.

[SnowMule]

What about one LCR meter capable of a four-wire measurement, then an array of relays to toggle between measurement points?

Use a digital I/O to drive the relays... switch in what you need, take a measurement, switch it out...

[Ashley H]

It would be a good solution except I need to measure the change in resistance constantly while the pieces are in the testing ovens. So all pieces need to be plugged in at all times taking constant readings. We know we have to stop the test when we see the capacitane change...our fail mode.  The wires coming off the pieces and out of the oven to the DAQ will be soldered in place on the pieces.

[CTF]

Take a look at http://www.pupman.com/listarchives/1998/April/msg00625.html It describes several methods of measuring capacitors.  Since you know the value of your capacitors, you should be able to tune one of the circuits so that you can monitor the voltage across one or more components that will change when your capacitor(s) fail.

[AnalogKid2DigitalMan]

What range of capacitance do you need to measure? I know there are off-the-shelf IC's to convert capacitance to voltage.

-AK2DM

[Ashley H]

technically the packets we're measuring aren't exactly capacitors, they will not be connected to each other, all independednt.  We are still in the process of making one test packet before we make the rest of the batch in case we messed something up in the procedure, so we don't even know what the C we'll be reading is yet.  We do know that the reading will most likely be in pF, and the change will be very very small.  The failure mode is actually going to be triggered by the growth of a tin whisker growing between 2 tin plated peices inside of the packet.  Our goal is to find the very beginning of the growth, and because the whiskers are small in nature the capacitance change will be very very very small.  If the item you were reffering to that converts the cap to volt can handle small decimals and can independently read multiple inputs that would be great.  for the single packet test we're going cheap and using the LCR meter we already own because we only need to read the one packet.  Our problem comes in when we try to test 8 of the packets at a time.

[AnalogKid2DigitalMan]

This may be a challenge, since (small values) + (high resolution) != (cheap).

You may have to have a controlled environment and fixturing to minimize effects of triboelectrics and use a controlled environment.

You will also want to control the test layout and connections to minimize parasitic capacitances.

Once use a chip to generate a voltage from a humidity sensor that varied its capacitance, but can't recall what it was.

Analog Devices makes the AD7745/46 series, 0-4pF with 24 bit resolution and communicated via I2C bus. Range can be extended to ~40pF with external circuitry.

They also feature the AD7151, 0-13pF.

Analog Microelectronics has the CAV424 and a couple others, 2.2nF-18pF range, analog voltage out.

Xemics, if still in business has the XE2004 that handles up to ~100pF.

You may want to read some white papers from some of the major test instrument companies regrading the caveats on doing precise, small value capacitance measurements and what controls to put in place to ensure accurate and stable measurements.

Hope this helps!

-AK2DM

[johnsold]

Perhaps you should look for the onset of tunneling current rather than capacitance.  AK2DM has pointed out the challenges you face trying to measure capacitance directly.

Lynn