AD623 Thermocouple Amp

Hi Brett,

Sorry for the late post, but I just wanted to follow up on a few concerns.

You said you're not worried about cold-junction compensation. A thermocouple is a differential temperature sensor, and it reports a voltage proportional to the difference between its measurement junction (tip) and its reference junction (or cold junction, where it connects to the copper of your PCB). If your measurement junction temperature remains constant, but the ambient temperature goes up by a degree, then your thermocouple signal will go down by a degree. So if you don't use CJC, it's not just that you aren't measuring the absolute temperature, but you may not even be measuring changes in the temperature that you intend to measure, unless you can hold your reference junction temperature constant. Check this video and this article for more explanation.

The AD623 datasheet "Amplifying Signals With Low Common-Mode Voltage" section says that you should get an output voltage of 1.11V to 3.07V for -200°C to 200°C measurement temp (which, by the way, assumes that the reference junction is held at 0°C exactly). In reality, I don't think anyone uses a Type-J Thermocouple down to –200°C. NIST limits the accuracy range to 0°C minimum and temperatures below ambient are generally avoided for type J because condensation can cause the iron to rust. So a better question is: what is your temperature range? What is the temperature range that you tested when you observed that the AD623 output didn't change? Did you keep the AD623 and thermocouple reference-junction out of the oven, and place only the thermocouple tip in the oven so you could vary only the measurement junction temperature without changing the reference junction temperature as well? Alternately, many people use a thermocouple simulator to test their thermocouple circuits.

The AD623 can provide output voltages up to about ±1V from the REF pin for single supply when you have 0V common-mode voltage as described in the data sheet. In the years since the AD623 release, we have developed even better options, such as AD8227 and AD8237 which can amplify 0V CMV signals nearly to the rails. Here, I have linked a configuration for the AD8227 where it can measure the full 0°C to 750°C type-J temperature range (-1.277mV to 41.004mV at Trj = 25°C) with a single 5V supply.

I hope this helps.
Best regards,

Scott

Hi Brett,

Sorry for the late post, but I just wanted to follow up on a few concerns.

You said you're not worried about cold-junction compensation. A thermocouple is a differential temperature sensor, and it reports a voltage proportional to the difference between its measurement junction (tip) and its reference junction (or cold junction, where it connects to the copper of your PCB). If your measurement junction temperature remains constant, but the ambient temperature goes up by a degree, then your thermocouple signal will go down by a degree. So if you don't use CJC, it's not just that you aren't measuring the absolute temperature, but you may not even be measuring changes in the temperature that you intend to measure, unless you can hold your reference junction temperature constant. Check this video and this article for more explanation.

The AD623 datasheet "Amplifying Signals With Low Common-Mode Voltage" section says that you should get an output voltage of 1.11V to 3.07V for -200°C to 200°C measurement temp (which, by the way, assumes that the reference junction is held at 0°C exactly). In reality, I don't think anyone uses a Type-J Thermocouple down to –200°C. NIST limits the accuracy range to 0°C minimum and temperatures below ambient are generally avoided for type J because condensation can cause the iron to rust. So a better question is: what is your temperature range? What is the temperature range that you tested when you observed that the AD623 output didn't change? Did you keep the AD623 and thermocouple reference-junction out of the oven, and place only the thermocouple tip in the oven so you could vary only the measurement junction temperature without changing the reference junction temperature as well? Alternately, many people use a thermocouple simulator to test their thermocouple circuits.

The AD623 can provide output voltages up to about ±1V from the REF pin for single supply when you have 0V common-mode voltage as described in the data sheet. In the years since the AD623 release, we have developed even better options, such as AD8227 and AD8237 which can amplify 0V CMV signals nearly to the rails. Here, I have linked a configuration for the AD8227 where it can measure the full 0°C to 750°C type-J temperature range (-1.277mV to 41.004mV at Trj = 25°C) with a single 5V supply.

I hope this helps.
Best regards,

Scott