Thermocouple in inductive surrounding

[roro36]

We use induction furnaces to melt aluminium. I want to be able to read the temperature of the water in the pipe flowing in and out of the furnace. I am looking at a spot about 3m away but the induction from the furnace registers about 20mV on a scope which the thermocouples equate to 400V roughly.

If I move the thermocouples away, the water temp may have cooled too much and I still won't get an accurate reading.

Can anyone suggest a possible solution in either minimising the interference or using another sensor?

 

[0xDEADBEEF]

I suppose you meant something like 400C and not volts. It depends on how much electronics you are willing to get into. There are lots of options. One would be to use two thermocouples of opposing orientation and twist the lead wires together, then use a difference amplifier. With a bit of luck the interference cancels. Another way would be a simple PT100 thermometer that should give you much more signal and using an oscillating current to read the temperature would reduce interference even further. shielding might help. Really there are too many options to name and I am sure that there are ready made sensors that will work.

 

[Topher925]

Depending on the EMI that's being generated by the furnace a thermocouple or an RTD may not work at all no matter how hard you try. What type of thermocouple reader are you using? Do you have the option of adjusting the integration and/or acquisition time? What kind of accuracy are you looking for?

One option is just to ditch thermo-electric sensors all together and use something like fiber optic sensors. I've built optical temperature sensors from scratch and have had great success but they are can be difficult to build and very expensive to buy.

 

[roro36]

I'm using delts's dvp series thermocouple card, which was a failure. I can use a pt card or a normal analogu input an use somehing like an lm35? Are there any other options under this high inductance field that surrounds the probes?