Discrete Instrumentation Amplifier for J-Type Themocouple Sensing

Hi Guys,

I am designing a discrete instrumentation amplifier from op-amp MCP601 for temperature sensing upto 200 degree C using J-Type Thermocouple. I tried simulating the circuit in pspice but result is not satisfactory. Attached is the circuit and simulation results.

For J-type thermocouple, output voltage at 200 degree C is approx. 10.779mV. Ideally gain of INA is set to 100 times. So at 10.779mV of input, output of INA is approx. 1.0779V and also provided a reference voltage of 1.25V to INA. Hence at 10mV of input, output should be 2.3277V. But simulation results doesnot shows same. Also output voltage doesnot changes when either changing Gain of INA by varying R5 value or also by varying input voltage.

A single supply instrumentation amplifier can't work with common mode voltage at negative rail. Look at the individual circuit node voltages, particularly output of U1 and U2 to understand why.

I wonder why you are using a differential amplifier. It only makes sense for a non-isolated thermo couple.

Could you please explain in detail why "A single supply instrumentation amplifier can't work with common mode voltage at negative rail. "?

From what i understand, for amplification of low input voltage signals INA is best suitable. That's why i opted for INA. In this case which circuit is suitable since i want to amplify the input signal approx. 100-150 times?

I will really appreciate your help here. If you do also have some link to share where i can build more understanding on, it will really help.

Hi,

My recommendation.. when to use an INA:
* small signal level
* large common mode range
* common mode range well within the specified range of the OPAMPs
**
OR: when the signal voltage reference is not the same as the output voltage reference.

***
There are a lot ready to buy thermocouple amplifiers available.
There are a lot of good thermocouple OPAMP circuits available.
There are a lot of thermocouple circuit discussions in this forum.
So the problem is not new.... It seems you want to reinvent the wheel. But why?
I recommend to look for reliable sources (Application notes at OPAMP manufacturers, for example) and follow their design rules.

***
I see no need for an INA with a thermocouple circuit.
Specify:
* thermocouple temperature measurement range
* cold junction temperature measurement range
* what accuracy and precision do you expect (values, please)
* what supply voltages are available
* your other specifications.

***
--> you need low offset voltage drift. I doubt a J-FET is a good choice.
J-FET is good for high input impedance ... which you don´t need here.

Klaus

As said, looking at U1 and U2 output voltage would clarify what the problem is.
U1 and U2 need tp swing symmetrically below and above input common mode level, +/- 1V for the G=100, Vin=20 mV example. But if the input is already at 0, one output can hardly swing to -1V. You can e.g. bias the input voltage to Vcc/2 to solve the problem.

You would preferably use low offset voltage OP(s) for a thermocouple amplifier, MCP601 isn't optimal, but can work for 200 °C range if the initial offset is removed by software calibration.