Use PXI Analog Output for RTD Simulation

[John1986]

I'm trying to test a system that has a 4-wire RTD input, but I want to be able to just simulate this input.

If it was two-wires, it's straightforward enough, because I can just drive a suitable voltage into the wires.

But I don't know what to do with the other leads that I guess are used to compensate for the lengths of the leads?  (I've read about the 4-wire measurement system).

I'm worried that if I don't use these, I won't get a very accurate input into the system under test

[AJost]

Hi John,

In principle this should be possible, but it depends a bit on how smart your measurement device is to react on fault conditions:

However, there are a few things to consider:

- most RTD devices deliver a current to the RTD (check for this in the manual! - it might also drive a voltage instead to power the RTD and then everything is turned around). This current needs to go somewhere, so put a resistor on the resistance wires. This should also prevent the open circuit detection mechanism from triggering

- then you should be able to drive a voltage through the voltage detection leads to simulate an RTD using the other wires.

I hope you realize that this would not be possible in a 2-write set-up, where the same wires are used to drive the current and measure the voltage drop, because the current could destroy the analog output of your PXI card.

Andreas Jost

Applications Engineer

National Instruments

[John1986]

Thanks Andreas.  This makes sense to me.  I think you're saying to simply drop the two current-supplying wires across a resistor, and then connect the other two wires to the AO for a voltage drop.  Basically making two circuits (as with a real RTD the current-supplying wires would end up connected to the voltage detecting cables as well)

I think I was just worried about losing accuracy of control for the simulation, since I guess I lose the "benefits" of the lead loss compensation?

[AJost]

The wire resistance only comes into play, when it is comparable to the resistance you would like to measure. So for a 100Ohm RTD it might play a small role. The idea of the 4 wire scheme, is that the current flowing on the voltage leads is very low because you have the very large (usually >10MOhm to > 1GOhm) internal resistance of your voltmeter. Thus the resistance of the wire does not play a role anymore.

This will be the same for the simulation, so the biggest error I expect is the accuracy of the output module, which is typically lower than the accuracy of the RTD module. However, without knowing which ones you are using this is difficult to say.

You should probably have another look on the ideas behind 4 wire sensing and it will become much clearer to you.

Andreas Jost

[John1986]

Thanks again Andreas - I think I just got myself muddle, worrying about measuring voltage drops across the leads from the voltmeter. 

[Henrik_Volkers]

What do you know about the DUT?

Assuming a PT100  and constant DC current (no all devices do so, can be pulsed or AC ..)

Usual current are in the 1mA to 100µA range  -> 100mV to 10mV (offset)  1°C is coarsly 0.4Ohm change resulting in 400µV to 40µV change...   (and 0.4Ohm is quickly done with a copper wire ... that's why you measure with 4 wires)

Well, how precise can you control the voltage?

What temperature (or Ohms) range do you want to simulate?

and maybe you need to provide that voltage as a floating source :0

Hook up DMM to measure the sensing current and hook up a scope to see if it's DC current.

(Use a scope that do not ground one input 😉 )

[Minions]

Using a 100ohm Platinum RTD, you can use the following for a close approximation of desired temp to resistance.

R = (T * 0.385) + 100

[Henrik_Volkers]

Here is a schematic for RTD simulator I just happend to find 🙂

I think was AN313 from analog devices ... hint: the TIA resistor needs a low tempco