OPA2388: Signal Conditioner and Amplifier for Thermopile

Hi Shyamsundar,

For starters, take a look at a thermopile amplifier design that was developed by Neil Albaugh, who used by an Applications Engineer in this department for many years. The circuit was designed at least a decade ago, and used the OPA335 auto-zero op amp. It provided very high performance in the circuit. The OPA2388 is an newer amplifier with several enhanced performances that should make it a good choice as well. Some component value changes may be needed in the circuit to optimize the gain for your sensor, and to assure stability.

The circuit can be found on a TINA Spice blog page. You can simulate the circuit with TINA-TI which is downloadable from the TI web site. You can find the blog here:

www.tina.com/.../

Alternately, you could set the OPA2388 up as a high-gain dc amplifier and use a second stage for an active filter to limit the noise bandwidth. I do suggest that you take a look at the information that is on the web regarding the subject of thermopile amplifiers. There is useful information on the subject available.

Regards, Thomas
Precision Amplifiers Applications Engineering

Hi,

There are a few approaches to this. As a starting point, you may want to take a look at TIPD209, which is a TI Design for a thermocouple amplifier with cold-junction compensation (schematic below).

Additionally, I found the following article from Amphenol Sensors that includes an example configuration using an op-amp: 

Depending on your accuracy requirements, bandwidth, and the modulator frequency of your ADC, you may run into issues driving a 24-bit delta sigma converter with an INA333 or OPA388. The INA333 does not have the bandwidth typically required to drive an ADC, and the output impedance characteristics of the OPA388 make it difficult for the device to drive converters with high modulator frequencies (this is true for most zero-drift amplifiers). For better resolution into the ADC you could include a buffer amplifier (such as the OPA320) between your input circuit and the converter. 

Hi Thomas,

Thanks for the inputs.

I have Downloaded TINA and OPA335 Thermopile Signal Conditioner and simulated it by modifying it. Also i have used OPA388 instead of OPA335.

it is giving gain of 1000 V/V.
As i am a digital designer, i am not sure about this circuit. Is this circuit ok ?

Hi Shyamsundar,

Increasing the closed-loop gain from the original 101x, to 1001x, does reduce the loop gain with which the circuit operates. That can affect the gain accuracy of the circuit and reduce the bandwidth; however, this circuit is essentially a dc circuit so the latter shouldn't be an issue. Keeping the closed-loop gain at the original 101x will increase the loop-gain by 10x and the gain accuracy benefit gained by using the lower gain is maintained.

You might want to consider using the OPA2388 and simply add a second gain stage having a gain of +10 V/V. That will cause the operating current to approximately double and that needs to be considered. If you intend to drive an ADC you may need to use a different op amp for the second op amp, one best suited to drive the particular ADC. This is another reason to consider a two op amp approach.

I noticed that you removed the -5 V bias applied to the output resistor (R4). That is okay, but note that the OPA388 output will only be able to swing to its negative supply limit, which is millivolts or tens-of-millivolts above 0 V depending on the output current.

Regards, Thomas
Precision Amplifiers Applicaitons Engineering

Hi Thomas,

Thanks a lot.
Noted your suggestions.

i will use OPA388 with gain of 101 as an inline cable pre-amplifier, as the thermopile sensor is mounted on 3 meters high pole with 3 meter cable.

And a second stage of OPA320 with gain 10V/V on the Data acquisition module. The output will be fed to the ADC ADS1258-EP with Unipolar (+5V) operation. is that OK?
OR would you suggest to use one more stage of OPA388 DC amplifier with Filter before OPA320 to compensate cable loss and to nullify the cable noise ?

Any other suggestions please?

Regards,

Shyam.

Thanks.
I have decided to use www.tina.com/.../ by replacing OPAMP with OPA388.
And also i have decided to use OPA320 as an input buffer to ADS1258-EP.

The only issue is my thermopile sensor signal is coming from cable with 3 meter long. i will use the above infrared-thermopile-detector-amplifier circuit as a cable inline pre-amp. what is the second stage to be used..
My main concern is accuracy and the noise (as my thermopile sensitivity is 5 micro volts).

Thanks again.

Hello Shyamsundar,

The OPA320 is a capable ADC driver because of its wide bandwidth, good output drive current capability, and flat open-loop output Impedance vs frequency characteristics, all important for that particular application. I think it is a more capable choice for the ADC driver than the OPA388, which is better suited for the very precision sensor interface application.

Designing an analog signal chain that results in maximized 24-bit performance is very challenging. Since the ADS1258-EP is supported by TI's Precision Data Converter products group, I suggest you make an iquiry on their E2E forum about the best practices when driving the ADS1258-EP. They may be able to provide some suggestions that will help you design the driver circuit to achieve maximum performance.

Regards, Thomas

Precision Amplifiers Applications Engineering

Hi Shyamsundar,

I support the catalog ADS1258. I don't see any issues using the OPA320. As Thomas said, that op-amp has good driving capabilities and would likely work well in this application.

Just keep in mind that the ADS1258 can only measure voltages slightly larger than the reference voltage. Therefore, if the OPA320's output reaches 3.5V, you'd likely want to use a 4.096V reference, such as one of the REF6x40 devices.

Best regards,
Chris

Hi Chris,

Thanks.

The output of the OPA320 swings from 0.1V to 4.9V, i would like to use this entire voltage band.
If i use 4.096V reference, then the ADC will read full scale at input of 4.34V. am i right ?

So, i am planning to use the 4.5V or the 5V reference. Is that ok ?

i will tune my OPA388 and OPA320 stages such that ADC reads the full swing of the thermopile sensor output.