Noisy RTDs in Measurement

I have an electrical noise problem that I need help with. Please bear with the rambling:


Our lab is measuring the heat transfer performance of automotive diesel fuel heat exchangers. The equipment consists of a diesel fuel supply cart and a wind tunnel in the explosion safe test room, while the data acquisition equipment and supply cart/wind tunnel control box is located in a separate control room. For the data acquisition / control hardware we are using National Instruments Field Point with an RS232 link to the computer.

On the DAQ side, we are reading inlet fuel and outlet fuel Temperatures at our heat exchanger with Laboratory grade, shielded RTD's, as well as the diesel mass flow rate with a Coriolis mass flow meter. We are also measuring the wind speed in our tunnel with a 4-20 mA differential pressure transducer. There are other various measurements being made with thermocouples and other 4-20 mA devices. The sensor data is fed through the wall, away from any and all power lines, into the Field Point DAQ.

On the control side, we are using the Field Point to output signals from the DAQ box to the control box. Control level signals are passed from the DAQ box to the Control box to control the Diesel Flow (100 Vdc motor controlled by a Saftronic DC Motor controller), the Diesel Temperature (600 Vac heaters controller by a 0-10 V controlled 3phase SSR), and the air speed in the wind tunnel (Leeson VFD driver 600 Vac,

3 HP motor). These power lines are fed from the control box, through the wall, into the appropriate equipment.

Note the sensor lines are fed through a separate hole in the wall than the power lines, a good 10 feet apart.

Our facility is fed from the 27kV outside into a 480V delta output transformer. This 480 V delta is then enters our facility to the main switch gear / distribution panels. Power for the Diesel test equipment is derived from this 480 V delta supply from a distribution bus bar, into a 480 Wye input , 600 delta output transformer. Note that this transformer is hooked up in reverse, i.e. the label stated that the 600 V Delta is the primary, and the 480 V Wye is the secondary.

The Problem:

When ever the VFD to control the air speed is turned on, the temperature reading from the RTDs drop by about 5 °C, and become very noisy. The other 4-20 mA sensors do not drop in value, but they also become noisy. To my surprise, the temperatures being measured by thermocouple are stable like a rock. We have tried every thing to solve this problem, but I am at my wits end. All the sensors are shielded well, and as far as I can tell, properly. We have tried to power the DAQ by battery, to remove any influence of power line noise, but this didn't help. We have tried different shielding points, and this either has no effect, or a worsening effect. I scoped the ground with a battery powered (thus isolated) oscilloscope, with the +ve and

-ve leads connected to measure the common mode noise. It is noisy at the control box, but clean at the ground entering out building. This tells me that the noise may not be sinking to earth, but when traced, the integrity of the ground seems to be intact.

Again, sorry if I've been rambling, but I am very frustrated by this, and I would love some input from any of you people out there. There is a lot of knowledge among this group, and if anyone can help, this group can.

Thanks for the time, Dan Wright, Dana Corp., Long Mfg.Div.

Reply to
Dan Wright
Loading thread data ...

I once had a chance to locate some noise in a wastewater treatment plant. Turning on VFD A would turn on VFD B. It was a mess. I asked where all the controls and power was grounded. Control voltages ranged from 24v DC to 277 v AC. A lot of them were in the same conduits. I asked where the controls we re grounded. Some here some there. I suggested that they all be grounded in one place preferably at the PLC main cabinet. That was where most of the signals originated. The EC contractor called me nuts, the electrical engineer scratched his head a bit and the controls contractor said it would not matter. We started grounding everything once in the main plc cabinet and it took less than an hour for most of the problems to go away. There were 5 of us checking everything and it was mostly in two adjacent rooms. It really went pretty fast.

This is a suggestion only. I have found that shielded control wire need to be grounded every 1000 feet or so, but only once. Pick a spot as close to the source as you can and see if my suggestion helps. I always use the ground bus in the gear not a screw in the sheet metal. I hope my suggestion helps I am sure others will have ideas as well

Reply to

This advice is spot on!

At least 90% of the noise issues in control systems I have been engaged to solve have been related to improper bonding, & grounding.

One of the most common problems involving noisy control loops I find are, control cables with the shielding grounded at each end, shield not connected to ground at either end, and corroded connections.

It's been my experience that grounding the shield of a control cable at the source end only, and at a dedicated point common to all shields, usually in the control cabinet with a ground conductor back to the main service ground point of the facility, will in most cases result in a relatively noise free control system.

The adequacy, and condition, of the facility electrical grounding system is also critical. Bad connections, ineffective grounding electrode systems, and connections between neutrals and ground other than the service, can cause issues throughout the facility that frequently defy any reason.

In the OP the problem being most evident when VFDs are switched on might suggest a starting point to look. Reactors on the input, and the output of the VFD, good grounding, and bonding procedures, for everything connected to the VFD, metallic conduit & sealtight (no PVC), better yet shielded VFD type cable inside metallic conduit, on all VFD to motor runs. In addition keeping the distance between VFD, and motor as short as possible is advisable.

As you sort the noise issues out keep us informed. I for one am very interested in what you find, and how you successfully resolve it.


********************************************* Remove the two fish in address to respond
Reply to
Louis Bybee

Maybe you should put some RTD transmitters in there, if you're using direct wiring.

Best regards, Spehro Pefhany

Reply to
Spehro Pefhany

I completely agree with all these guys have said. But I would also like to commend you on your description of your setup and problem.

In spite of you thinking it is "Rambling" I thouhgt it was a very good definition of the problem at hand, and I sincerely wish that my customers could follow suite when they call up with technical questions.

with proper shielding practices We have made RTD's work reliably in some really "HORRIBLE" places, much like you describe, So I am quite confident that, following the advice of SQUit, Spehro and Louis you will find a suspicious connection somewhere that will make your worries go away.

Tom Grayson

Reply to
Tom Grqyson

Reply to

Have you tried output reactors on the VFD? Your RTD cabling may well have become a long low frequency antenna reacting to the waveforms being generated by the vfd.

Keep in mind that at higher frequencies such as the VFD uses, a normal safety earth ground is pretty ineffective. have you tried running ground braids along side the existing ground tie-ins?

Reply to
Bob Peterson

PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.