set current for resistance measurement PXI 4071

[johanneschimpo]

Hi,

I would like to measure the resistance of some RTDs and strain guages using a 4-wire configuration.  These resistors will be at cryogenic temperatures (~4K), so we are concerned about self-heating.  As a result, we would like to pass only a small current (~1uA) to make some of the measurements.  I cannot figure out how to set the current for a resistance measurement with the PXI 4071, using NI-DMM.

In the resistance range that we need (we're measuring about 350 Ohms), the 4071 automatically pushes 1mA, and this is far too much.  I have found a subVI that sets the current, but it's labeled as working only for diode measurements, so I assume that if we set that property and then make a resistance measurement, the 4071 will ignore us and use the default current (as determined by the resistance range).

If it is not possible to set the current in this case, then we will have to use a separate current source, but it would be nice to have everything done by the 4071.  Any ideas?

k thnx

[ChrisR.]

Just use a larger resistance range. For example, the 10K range will put 100uA through the resistor. The measurement noise will be higher, but there's no way to avoid that if you insist on a very small excitation current.

1mA will cause only 350 microwatts of power dissipation in a 350 ohm resistor. I wouldn't expect that to cause even half a degree of temperature rise. How low do you need the dissipation to be?

Chris

[johanneschimpo]

We might end up having to set the range high and settling for less resolution.  In fact, that might actually be perfectly acceptable...I'll have to crunch through some more numbers to see for sure.

What I have found is that while operating near 5 K, we would like to have a 50 mK or better temperature resolution, so that means we need very low heat generation in both the strain guages and the RTDs.  To complicate the matter further, these resistors will be in an insulating vacuum, which will hinder heat dissipation, increasing the effect we want to avoid, and we will take several measurements a minute for days on end. 

That 50 mK temperature resolution would translate to about a 2 Ohm resolution on the RTD resistance, if that were our only source of error.  The most obvious other error factor is measuring the resistance of the strain guages themselves (which then have to be calibrated to the temperture--thus the RTDs), so I'm thinking a resolution on the order of 100 mOhm on the RTDs would be necessary?

...And the data sheet for PXI 4071 says that at a range of 100 kOhm, I should be able to get 10 mOhm resolution with 10uA current, so maybe that will work (lower digits resoultion--i.e. 6 1/2 instead of 7 1/2--are highly preferable for speed).  I'll have to check the RTDs and SGs to get more details on the self-heating to see if 10uA is OK. 

Thanks for that suggestion, although I would still like to have a way to control the current without compromising speed or resolution, if you, Chris, or anyone else has any more ideas?

k thnx

Message Edited by johanneschimpo on 06-22-2007 05:13 PM

[ChrisR.]

There's no way to change the current without affecting noise and resolution. Smaller current means less voltage to measure, so signal-to-noise ratio drops. However, I'm pretty confident the 4071 is quiet enough to do what you want. I just took some quick room temperature measurements of a 380-ohm resistor with my 4070 (the noise should be just about the same for the 4071), and got less than 0.1 ohm pk-pk variation at 1PLC (16.67msec aperture) on the 100K range. On the 1K range I could drop the aperture to 20usec and still stay under 0.1 ohm pk-pk variation.

You say you want to take "several measurements a minute," yet you indicate a concern for speed. How fast do you want to take readings? If you took really quick yet infrequent readings and made sure the current source was off between them, that would reduce the average power dissipation in your resistor, and presumably the self-heating as well, depending on how long it takes the resistor to heat up in its environment. So maybe you could keep the measurement current higher than you would think, if it turns out you need to; just make sure it's not on all the time.

I'm not clear on whether you need 100 milliohm accuracy, or just resolution. If it's accuracy you need, that will make it harder to use the 100K range. 0.1 ohms is only 1 ppm of 100Kohm; the 4071 can't guarantee that level of accuracy without manual offset nulling.

Since power dissipation is proportional to current squared, the 10K range will produce 100x lower dissipation than the 1K range. So going to the 100K range might be unnecessary.

Chris