RTD fluctuations

Guys, This is killing me! I am commissioning some RTD's which are connected in some stainless steel tanks and pipes. Inside the tanks and pipes is milk. so far so good. When I connect my RTD up to my OMRON C1000 analog input card (with the screen connected to the screen terminal of the card and not in the feild) my temperature varies +- 4 degrees every display update on my SCADA screen, some times more, some times less but also appears to not be very accurate. If I remove my RTD from the pipe work the reading is dead steady and accurate. I connected my RTD calibration unit to the RTD wires while in the pipework and it shows the RTD resistance is steady so somethings upsetting the electronics in the pucken puck. Not a VSD on site. When I connected the card shield to earth it made the reading less accurate and not a lot steadier. Im guessing that noise is being induced through the Steelwork otherwise the "hanging in the air" test would not have worked as well as it did. Any suggestions on what it might be and how to fix it??

Regards
Steve D

perhaps there is a potential between your reference grounds. i've had similar problems in pH instrumentation. try grounding the card shield to the steel pipe.

Is you analong input single or double ended?

We always design our products to have differential analog inputs for noise rejection. Differential analog inputs cost a lot more but then so does not having them when you need them.

Be a little bit careful with kenc's suggestion. If there is a potential difference between the two points and the source of this difference can generate any appreciable current you could cook your input card. Just make sure that the way you tie your ground connections between the pipe and the card does not use the card as part of the path from the pipe to ground.

Keith

this probably isn’t going to help at all but I offer it up for free ... maybe it will point to something useful ...

I’ve seen the exact same effect while working with “grounded” thermocouple probes ... the piping was all non-conductive PVC plastic and I was using an individual “hockey puck” transmitter for each thermocouple ... when I unscrewed the probe and hung it in the air, the reading was rock steady ... as soon as the probe touched the water, the reading was all over the map ...

my solution was to go back and buy “ungrounded” thermocouple probes ... these were more expensive but at least they took care of the problem ...

as for your problem, I’ve never heard of a “grounded” RTD probe before so I’m probably just wasting your time ... but maybe this will be of some interest ...

I've seen the same thing as the other Ron... when using thermocouples with grounded sheaths. I have never figured out why they even make them. If it is a grounded sheath, just replace it with an ungrounded and the problem should go away.

Ground the milk!

Ground the milk!

Ummm...I think it has to be powdered milk before you can grind it..err...ground it...umm...

It sounds like grounding problems, but to my kowledg RTDs don't have gronded/ungrounded sheaths. Check your PLC I/O card, make sure your panel is well grounded, your power supplies are groundd, and you are using shielded twisted wire with the shield grounded at one end.

I experienced a similar problem many years ago. In my case I was using inexpensive RTD transmitters to convert to 4-20 mA. I switched to isolated transmitters and the problem went away. I never did find the root cause of the problem.

it is for sure a grounding problem.

the shield of the cable should be connected to earth only close to the plc. and be continued up to the end of the sensor.
however not connected to anything else. so maximum you do have 5 wires coming out of a cable. when using 3 wires just cut of the fourth wire. the sensor should never be grounded. as the current is 1 mA the voltage for 1 celsius is 100.4 millivolt. so the accuracy is in microvolts and that is very quiet.
or call me and i will fix the problem all over the world.