Thermocouple Specification 1

[peglor]

I've spent the last few weeks unsucessfully looking for a decent guide on the selection and the fitting of thermocouples to take measurements from a system.

I am looking for a means of collecting both surface temperatures for a block of Aluminium heated by power resistors and fluid temperatures for the water which I am using to cool it.

I currently have K Type thermocouples, which I think are suitable (Temperatures to be measured are from room temperature to 120degC Max), but I don't know how to mount them correctly for surface readings or whether they need to be protected if they are immersed in water to take the fluid temperature.

Any help would be greatly appreciated.

 

[abeltio]

The range of temperatures you are measuring is very low... RTD's could do the job as well, are cheaper and do not require special wiring (Chromel/Alumel) as the K-type TC's.
K-type are suitable for a very high range (into the 1000F range)
RTD's use just regular wiring (twisted shielded pair).
May be you'd like to investigate that road?
Are easy to check and when they fail (usually open) the control system will recognize the failure correctly (open RTD ==> infinite resistance (max temp) so the high temp alarm/trip should operate.
With TC's when they fail (usually open) they usually give a min (negative) signal ==> the control system will think the temp is low and try to heat more...

Also, consider using a 2 out of 3 voting for max reliability
(bring the 3 sensors into a median select block).

I did not solve your TC problem, in fact just added more unknowns to the problem...
HTH
saludos.
a.

 

[IRstuff]

The thermocouple approach should have no problems with the water temperature, although putting it inside a sleeve of some sort migh be prudent and the mounting to the block should simply use a reasonably good thermally-conductive epoxy.

The thermistors might be a better choice from the standpoint of ruggedness, since they're usually ceramic covered. Mounting is the same.

TTFN

 

[peglor]

So basically the reason there's no real documentation about mounting thermocouples is because it really is as simple as sticking the thermocouple where you want it?

Is a thermocouple rendered useless if it is put in a conductive solution? I'm wondering because I've heard about using water baths to calibrate thermocouples, and water is the liquid I need to take the temperature of. I'll be using deionised water, so that isn't conductive anyway.

Also, as a totally academic issue (Epoxies are electrically insulating?), if a thermocouple is placed onto a metal surface, which short circuits the thermocouple, will it still work. From what I understand of thermocouples it should provided both junctions are at the same temperature. Does the fact that the original thermocouple junction remains mean that two different emfs are generated? Can anyone answer this one?

 

[IRstuff]

A sufficiently conductive medium will render the thermocouple unreadable.

Most epoxies are non-conductive, thats why you have to be metal-filled or impregnated epoxy to get electrical conductivity. Likewise, standard epoxy is thermally non-conductive, so those are filled with ceramic or metal (you'll need to avoid those ;-)).

As to metal shorts, it may or may not work, if it works, it will probably work incorrectly, since the additional couples could interfere with the basic measurement.

As general rule, since the thermocouple is part of an electrical circuit, you need to take the basic precautions of insulating the wires, etc.

TTFN

 

[Astroboy]

Hi
Some reference on the use and mounting of Tcs are:
Omega,

http://www.omega.com/

and stuff at cooling zone,

http://www.coolingzone.com/Content/Library/index.html

check out articles by Tony Kordyban and for a real thermal hoot get his book "hot air rises and heat sinks or everything you know about cooling electronics is wrong.

Anyway, a Tc junction can be immersed in liquid mercury and work just fine. The junction between the two wires can be soldered together with fine results as long as both junctions are at the same temp which would be the case if they are really really close to each other.

For general mounting issues the one big big thing is to make sure that the heat flux from a surface or source into the junction/tip/bulb, is way greater than the flux carried away by the wires. A common problem is seen with a junction glued to a plastic component case with the wires flapping around up in a breeze such as from a fan in some electronic enclosure. The wires dissapate the heat from the junction faster than the plastic can heat it giving the engineer a happy joyous feeling that the component is much cooler than reality.

For selection of types you are covered well with K and T with the noted exception that T types are less sensitive to electrical noise.

Have fun