Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

Thermocouple Interface Circuit

Status
Not open for further replies.
A

abdoalghareeb

Member level 5
Member level 5
Joined
Feb 5, 2010
Messages
84
Helped
1
Reputation
2
Reaction score
1
Trophy points
1,288
Activity points
1,981
Thi is my design for interfacing K-Type Thermocouple with AVR microcontroller.



I use Differential Amplifier.
for measuring temperature I use good digital temperature meter.
The Temperature range is (0 == 1000 °C).
this is the table of result.

T (°C) <===> V (mille volt)
---------------------------------
31 <===> -141
35 <===> -100
49 <===> -70
51 <===> -3
72 <===> +51
120 <===> +265
165 <===> +412
211 <===> +565
260 <===> +731
280 <===> +813
315 <===> +900
350 <===> +1063
380 <===> +1215
400 <===> +1300
----------------------------------------------

My questions are:
Is there any wrong in this design?
How can I make all the output values positive (shift up 0°C == 0 volt)
 

It appears that the ref temp is around 51C; you need to see that the ref temp is close to the room temp 25C (or 0C) as per your convenience. 50C looks like a odd choice but this is something you need to consider seriously.

Good digital thermometers can accept thermocouple inputs but I do not know the purpose of the table you give.

Suggestion: do 3-4 independent measurements with high quality thermometers and voltmeters (10-15 temp points with corresponding voltages) and fit these results to a quadratic, cubic or a quartic polynomial (depending your accuracy needed).

If you want to have all the voltages positive, you need to have the ref temp lower than the lowest temp in the table.
 
I use the digital thermometer to calibrate my project, digital thermometer has it's own sensor and my project has it's own sensor, Both sensors exposed to the same heating source .
The table is the results of my experiment , first column is temperature of digital thermometer and the second column is the output voltage of operational amplifier.
 

Let me see: I have plotted your values Temp vs Potential. See the graph below:



The solid straight line is a curve fit (cubic) least square fitted with your data. Clearly you will agree that your data has too much noise.

Also you will see that the relationship is almost linear; You can use a simple equation in place of a look up table.

Your project will benefit if you add lots of analysis that can show that you have done a very thorough analysis.

What is the mode of operation (how does it work?) of the digital thermometer and the accuracy? How much is the accuracy of the voltage measurements?
 
Hi,

your circuit simply is not a useful thermocouple amplifier. There are several issues with your circuit.
--> use a proven and dedicated thermocouple amplifier circuit. There are mayn in the internet.
And there are ready to buy thermocouple amplifier ICs.

Or read documents on how a thermocouple measurement circuit works.

Klaus
 

abdoalghareeb said:
I use the digital thermometer to calibrate my project, digital thermometer has it's own sensor and my project has it's own sensor, Both sensors exposed to the same heating source .
The table is the results of my experiment , first column is temperature of digital thermometer and the second column is the output voltage of operational amplifier.
Click to expand...

It appears that you have a input offset voltage of about 1.4 mv , with a gain of 100, you get 141 mv when you should get zero (room temperature)

And has far as noise in your measurement, getting two sensors to couple together is problematic. Best if they are both submerged in a oil bath (high temperature) or a water bath for lower temperatures.


Also add a capacitor across the 1 meg resistors to slow down the op-amp and reduce noise. Off hand I would say .001uf or 100 pf.
 

abdoalghareeb said:
I use the digital thermometer to calibrate my project, digital thermometer has it's own sensor and my project has it's own sensor, Both sensors exposed to the same heating source .
The table is the results of my experiment , first column is temperature of digital thermometer and the second column is the output voltage of operational amplifier.
Click to expand...

Your input voltage is small but the impedance is low and you should have no problem in measuring the voltage accurately. However, the one op-amp circuit is clearly not good enough- I strongly suggest you use an instrumentation amplifier for this project.
 

What if I use INA118 op-amp !
Is it necessary to add cold junction compensation?
 

abdoalghareeb said:
What if I use INA118 op-amp !
Is it necessary to add cold junction compensation?
Click to expand...

Only if you want to compensate for changes in ambient temperature.

TC voltage is based on a delta temperature, not absolute temperature.
 
abdoalghareeb said:
What if I use INA118 op-amp...
Click to expand...

I believe it is an excellent instrumentation amplifier and certainly suited for a thermocouple amplifier. For decent results, you should always add cold junction compensation feature because it takes only a little effort and makes your project look professional.
 
Hi,

As said before: Read about thermocouple operation, it's benefits, it's drawbacks and what is essential for an amplifier.
There are so many thermocouple amplifiers in the internet...why not simpky choose one of them?
And there are so many discussions in this forum.

INA:
There really is no need for an INA.
An INA if for
* high impedance input signals
* is for two differential signal
* introduces relatively much noise
* introduces relatively much offset (drift)
..all these are not important or not good features for thermocouples.

Any single simple low_offset, low_offset_drift Opamp will have better performance.

Cold temperature compensation.
We can not answer this, because it depends on your needs. And reading the documents already should have tod you:
* if you need high accuracy: usecold junction compensation
* if you need absolute temperature reading: use CJC
* if you have high ambient temperature fluctuations: use CJC

Klaus
 
Hi ...
Read AN28-20 Thermocouple Measurement by Jim Williams from Linear Technology Corporation will be help you to design t/c amp
 
Simple solution: use a good amplifier with offset and gain adjustments: with the reference junction at room temp (say 25C), set the gain so that we have 1mV/C slope and then adjust the offset so that 0mV reads as 25C. It is possible that the two adjustments will interfere and therefore you need to do the adjustment iteratively.If you use a millivoltmeter at the output, you can read the output directly as the temp in deg C.

Perhaps you will not get high accuracy but +/- 1 will be easy with care.
 

c_mitra said:
Also you will see that the relationship is almost linear; You can use a simple equation in place of a look up table.
Click to expand...

Just a side note: I always resist to the temptation of modeling any physical phenomena considering it behaving as linear, even if a straight line seemingly fits there, because in nature there are factors of all kinds that distorts the apparent ideal curve.

In addition, if we apply only the RMS method, this can induce to error by assigning the same weight to each sample, whereas the ones with the highest absolute value should have greater force in the regression. The following (non-linear) curve fits the measured table and evince how the smaller values are likely impacted by the environment noise.

Graph.png
Source: https://mycurvefit.com/
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top