Will an isolation power supply prevent ground loops? 1

No, the isolation power supply permits isolating the ground connection for the power source. The isolation transformer does not eliminate signal ground loop problems.

Assure that the signal shields are not grounded at both ends. This is the typical ground loop that is problematic. Typically, if you use shielded cable, cut and tape the shield at the field end. Assure that no drain wire or foil shield touches the field junction boxes. The shields should only be grounded at one point.

Some engineers and IEEE may suggest grounding at the signal source. Thus, a thermocouple shield would be grounded in the field and a 4-20 mA transmitter grounded in the control room. In practice, ground everything in the control room and nothing in the field to avoid installation errors.


John

It could help. Depends on where the ground loops are; before or after the power supply transformer. You have to consider, carefully, the actual wiring diagram of your system. Most instrumentation has isolated outputs/inputs already so it will not help with these devices. Also, instrument cases are typically isolated from any I/O as well.
It would be best to draw out exactly how things are wired before trying to use an isolation transformer or similar. It is likely that the problem can be fixed without using any isolation of the power.
Any physical connections to earth ground should only be in one spot as far as your controls go (assuming all run of the same power supply).

Buzzp,

Our situation here is that we have a beaded thermocouple connected to a National Instruments instrumentation chassis. The beaded end (i.e. the signal source end) is touching an internal surface of a pump we are testing. The pump is directly mounted to a 3 phase motor, in which the motor case is connected to ground.

When we connected the signal wires coming from the thermocouple to a hand held themocouple reader the reading we get is correct. However, when we connect the signal wires of the thermocouple to the computer instrumentation the signal biases up or down by a magnitute of 10 or more.
In addition, when we prevent the bead of the thermocouple from touching the metal pump case then both the hand held thermocouple reader and the instrumentation will read the correct temperature

The instrumentation is powered via a 115VAC outlet in the lab, which in turn happens to share the same netural and ground as the 3 phase power to the motor (this is done via a step down transformer).

Given that the hand held thermocouple reader is powered via a battery, thus it is inherently isolated, I thought that powering the instrumentation from an isolation transformer (or isolated power supply) would help removed the apparent ground loop that has formed. There is no shielding on the the thermocouple connection wire.

In your case, yes it would help. However, your not dealing with large voltages so this might not be practical. Your forcing the junction at that point to be at zero potential. Normally, a small voltage is generated across the two dissimilar metals and a temp change is proportional to the voltage measured.
The instrumentation should already provide isolation between the thermocouple and the instrument. If this is the case then the only real solution is to change the sensor to something electrically isolated (see below).
Some thermos have compensation for the lead length as well so this could also be a factor (on more than two wire devices).
Can you mount the thermocouple to something other than the pump casing? Or maybe look at a different type of thermo like the kind that "tapes" onto the device and is insulated, electrically. This might be the best solution.

I am sure others will contribute who know a bit more about thermocouples than I.