Thermocouple / ground problems 1

[justin5x7]

Hello,

I'm having some problems with some thermocouples and hope someone can provide some assistance…

I have attached a few crude pictures for reference.

Type K Grounded thermocouples (SERP-K-1, Omega Engineering) are connected to a machine.

There are 20 thermocouples in the system.

The thermocouple leads are connected to a data acquisition system (OMB-DAQ-2416, Omega Engineering) with differential inputs.

Thermocouple A is not grounded to the machine.

Thermocouple B is grounded to the machine.

The machine is a food extruder, with a 28kW DC motor, which is grounded.

The data acquisition system is connected, through a usb port, to a computer which is also grounded.

At room temperature with the power off, both thermocouple A and B read correct values (~24C).

When the power is supplied and the motor is turned on, thermocouple B temperature increases 5 to 25 degrees.

Once the motor and power are turned off the thermocouple readings return to normal.

Thermocouple A reads correct values and is unaffected by power and motor being turned on and off.

Thermocouple B appears to be picking up voltage from the extruder drive motor when power is supplied.

Since the thermocouples output in mV, picking up any voltage from the ground is significantly effecting the temperature readings.

Do I need to establish a common ground between the extruder and the data acquisition system?

Is anyone familiar with "guarding" and can this technique be applied?
Here is a link:

http://www.sensorland.com/HowPage078.html

I don't think I can filter the thermocouple leads with an RC filter due to the step response, but I am not sure.

I realize that the thermocouples should probably be ungrounded type K thermocouples due to the presence of the DC motor, but this is what I have to use. Also they worked with an older data acquisition system. More thermocouples were added so a daq system with more channels was needed.

Does anyone have any suggestions? Any information would be greatly appreciated, thanks.

Justin

 

[Skogsgurra]

Your earlier system probably had differential (and perhaps also floating) inputs and that is what you need to have in your new system also.

Shielding, guarding, filtering, bonding does not work if there are millivolts DC between the two systems. Your best choice is to use a TC/mA or TC/V converter with floating TC input for TC B.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[hacksaw]

equipment grounding should not be intermingled with signal/computer grounding in any manufacturing environment. the signal grounds need to be brought to a common point (RTU,I/o system, etc.) then tied to the ground system at one point

equipment gounds are for electrical safety at the equipment not signal grounding.

you can use non-grounded sensors, or local signal isolation by means of a t/c transmitter.

 

[Comcokid]

I assume your computer is a standard desktop-type computer. The isolation in the power supply in these is not very good. Try a laptop running on batteries.

Although less likely you could also have a common-mode noise issue. Use a clamp-on ferrite EMI filter (radio shack or other) on any grounded thermocouple wire - wrap the thermocouple wire through 2 to 4 or more times through the ferrite to increase the attenuation.

 

[itsmoked]

Here's what I would do.

As Comcokid suggested bring in a laptop and while it runs on batteries run the data acquisition software on it and see what happens. If it's unaffected by the motor then the solution would be a USB isolator. You can find these on the web. They isolate the USB from the field from the computer.

If the laptop test doesn't work then replace the grounded probe with the correct un-grounded probe and don't take any asinine limitation like "this is what I have to use" for an answer.

And no, running pieces of wire here and there won't fix it.

Keith Cress
kcress -

http://www.flaminsystems.com