Motor RTD Noise

[mls1]

We have several 450HP, 4-pole, 480V, induction motors on VFDs with embedded RTDs. We are experiencing significant noise on the RTD circuit and are in the process of adding filtration, etc. The RTDs are three wire and we have appropriate shielding, segregation, etc. but we noted that the RTD leads coming out of the motor auxiliary terminal box appear to not be shielded or even twisted. They look like standard motor wire (not sure of the exact type but similar to MTW). The motor is drive rated so it was expected that it would run on a drive. What is typical practice for wiring embedded RTD leads out to the terminal box? Shouldn't they at least be shielded? The motor windings are random wound and we have 6 RTD (3 DE and 3 NDE). Thanks!

 

[UKpete]

Hi, my limited experience with RTDs in electrical machines is that they aren't screened or twisted because they don't really need to be for they have high noise immunity. When fitting them inside a manufactured generator winding we were told by an RTD supplier that the screen wasn't necessary.

As the sensor is a relatively low resistance device (say 100ohms) fed from a constant current source which is inherently low impedance, this means that any induced noise is attenuated - when the sensor is actually connected up the measured noise voltage should be low. It may be misleading to measure the noise voltage whilst the sensor is out of circuit because the noise will not be attenuated by a voltmeter as these have a very high input impedance. Also in a three-wire set up (having a bridge circuit in the amplifier) there is common mode noise rejection.

You may already be doing this, in which case apologies for stating the obvious, but the ultimate test is to see if there is an instant change in temperature when the motor is started up. I have done this experiment on a machine on a variable frequence drive and noticed no step change in indicated temperature at starting, only a slow increase as the windings heated up. But every application is different I guess.

 

[Drivesrock]

I have one job where we had to change the digital temperature indicator and protection device as when the drive/motor was started we got a jump in the temperature reading. A better model and the problem was solved.
Then again on many jobs where we have RTD inputs to the PLCs and I have not seen this problem.

 

[mls1]

We're measuring them with Allen Bradley Flex I/O (1797-IRT8). The module has a week input filter but very good digital filter. Unfortunately, the noise is overranging the input so with the filter cranked all the way down to 0.2Hz we get a smoother measurement that is off by 20 degrees.

We spoke with the motor manufacturer this morning and they indicated that they have had applications where they shielded the RTD leads but only on request. Their standard practice is to use their normal high temp, high strand wire which is what we have.

We know some noise is coming from the motor and some from the drive (the Flex I/O module is in the drive cabinet). We're going to scope them tomorrow and see who is the worst offender and try a few filtering options. I'm thinking a larger capacitor applied at both ends will knock the noise down enough to let the digital filter take over. I'll post results.

Thanks for the feedback so far!