problem with K-type thermocouple

[mutthunaveen]

i have designed a small project which monitors the temperature
i went for K-type thermocouple due to its wide range of temp measuring
i used AD595 as the conditioning ckt for my project

K-type thermocouple to AD595 and form AD595 to PIC 16F877A ADC. this is how my project build...

i initially when i power-up my ckt . the temperature read by the PIC is good with -3°C error... thats tolerable.. but after 10 min when i see the temp read there is a difference of nearly 14°C to 18°C

i cont figureout why this is happening. any suggestions please...

ADC read value with 10bit resolution and converted to mV and displayed in 20x4. i am checking the temperature by co-relating the mv to the K-type thermocouple table

 

[FvM]

Did you understand, that the thermocouple's cold junction must be placed at the AD595, respectively hold at the same temperature to make the cold junction compensation work correctly?

 

[mutthunaveen]

Hi FvM

can u explain me plz in detail how can i compenciate this
1. should i hold the K-type fused point in AD595 for some time
2. or should i have to add some electronic circuit in order to get good results

please reply me

 

[mrcube_ns]

First, physics:
Thermocouple (TC further) have cold and "hot" point. When you put Hot point on temperature T1, and cold on temperature T2, on cold point pins of TC you will get voltage Vtc = K*(T1-T2).

Problem is that cold point temperature usually are changing regards to athmosphere conditions and because of that there are many ways how to compensate cold junction temperature.

AD595 has cold junction temperature compensation inside (this means that this IC add or substract some voltage from measured value to get accurate measurement).
This means that cold junction and AD595 MUST BE as close at can, and must be on same temperature.


Mr.Cube

 

[mutthunaveen]

Thanks a lot mrcube_ns. i understood now, this is a great help for me to at least attain the better accuracy....
but 1 quest i can't place the AD595 near the furnace around 900'C because the soldering and wires get melt (i am not sure) and AD595 operating temp is also less than what it measures. any idea how can i get better accuracy still

 

[FvM]

i can't place the AD595 near the furnace around 900'C

The suggestion is to keep the cold junction at the same temperature as the AD595, not the hot junction. There's no length restriction for the thermoelement wire between both junctions. In a limited temperature range, you can also use extension wire. But it's necessary to have the cold junction either at a known temerature, or to measure and compensate it, as the AD595 does.

 

[mrcube_ns]

Lets go back to physics again:

When you have 2 different metal connected (hot spot), and apply different temperature (between hot spot and cold ends) you get some voltage => main principal of TC working.

Example:
You use Ni - Cr+Ni thermocouple. If you use copper wire to connect TC to some circuit, on connection Ni - Cu and Cr+Ni - CU you will theoretically made new TC's.

So, you must use extension cable,this cable is made from Ni and Cr+Ni wires. So when you connect this cable to TC, Ni to Ni wire, and Cr+Ni to Cr+Ni, no new TC is made (same material is connected to same material), and you can safely get TC signal to normal temperature area.

Mr.Cube


PS. See this
Thermocouple - Wikipedia, the free encyclopedia