View previous topic :: View next topic |
Author |
Message |
Fabri
Joined: 22 Aug 2005 Posts: 275
|
cold junction compensation |
Posted: Sat Jun 30, 2007 3:02 am |
|
|
Hi to All,
I need to mesure temperature from thermocouple in stove.
I did software linearization and work fine. For cold junction compensation I use one NTC to read temperature near termocouple connector on PCB and I add it to thermocouple temperature. I have problem when stove temperature rise up and also cold junction compensation can mesure more then 35°C.
My question is:
Does add operation between temperature of thermocouple and cold junction is correct or must take account of something else ?
Thanks for answer,
Fabri |
|
|
kender
Joined: 09 Aug 2004 Posts: 768 Location: Silicon Valley
|
|
Posted: Sun Jul 01, 2007 11:56 pm |
|
|
You might run into a problem, when the oven temperature is below ice point. Consider an example. The temperature on the board (where the ice point and the NTC that measures the ice point reside) could be high (say, 35C), because some components on the board are dissipating heat. The oven is at room temperature. Voltage difference between the ice point and thermocouple becomes negative. If your analog front end can’t handle negative supply voltages, it will rail somewhere: either in the analog front end or in A/D.
You should also carefully watch for material that your thermocouple connectors are made of. I had a bad experience with that once. The device had an enclosure. There was a T-type (copper and constantan) thermocouple that had to measure outside the enclosure. There was a generic front panel connector that had to bring the thermocouple (and a bunch of other signals into the enclosure). Thermocouple measurement worked just fine without enclosure, but with the enclosure there was a considerable offset. It turned out that the constantan wire, the connector and a fair amount of heat that was generated inside the instrument were causing the problem. The connector pins were made of copper, and it have caused 2 additional parasitic thermocouples to form. Since the instrument was generating heat inside, there was a temperature gradient across the connector, that was causing the voltage offset.
If you describe your instrument, may be we can help you find the source of the problem.
Best,
Nick _________________ Read the label, before opening a can of worms. |
|
|
Fabri
Joined: 22 Aug 2005 Posts: 275
|
|
Posted: Mon Jul 02, 2007 12:49 am |
|
|
Hi Kender,
My thermocouple amplifier hand only positive signal, I can't handle negative supply voltage. The termocouple is J Type. The circuits ha it's enclosure but heating became from stove. I Have NTC to mesure outside temperature and I can use it for cold junction compensation, I've already thought about. I belive the problem is junction connector so I need to mesure it's voltage junction to compensate it in temperature.
Have you got any ideas ?
Regards,
Fabri |
|
|
kender
Joined: 09 Aug 2004 Posts: 768 Location: Silicon Valley
|
|
Posted: Mon Jul 02, 2007 12:58 am |
|
|
Fabri wrote: | I belive the problem is junction connector so I need to mesure it's voltage junction to compensate it in temperature.
Have you got any ideas ? |
There are 2 good ways to beat this problem:
1. Get a special thermocouple connector that has pins made of the same material as thermocouple wires.
2. Avoid the connector and place the cold junction compensation outside the enclosure. Effectively, the whole thermocouple with both of its junctions will reside outside the box. _________________ Read the label, before opening a can of worms. |
|
|
Fabri
Joined: 22 Aug 2005 Posts: 275
|
|
Posted: Mon Jul 02, 2007 3:47 am |
|
|
Hi,
In my project I have outside temperature and connector temperature (NTC on PCB) so I measure the real mV of cold junction and compensate it by table. Anyway the error is maximum 4-5° C in better case from outside know temperature.
Thanks for help,
Regards,
Fabri |
|
|
|