CN-0357 Design Details

Hi Willm_van. Sorry for the late reply! Here are a couple of my notes on this circuit:

1. U2-A is used in a potentiostat configuration. Electrochemical gas sensors usually have a bias requirement wherin the voltage difference between RE and WE must a certain value. So the idea here is to keep the voltage levels at the RE and WE pins approximately equal to VREF using the feedback action of the two op amps. In this way, all the voltage adjustments happens at the CE pin. This op amp configuration is pretty common for this application - here is a document by SGX Sensortech that might help you understand it better.
   1. The two RC stages (R5, R6, C3 and C4) are used to improve the stability of U2-A.
   2. R5 and R6, as mentioned above, have an effect on the stability. Using the same value for R3 and R4 however, is based on common recommended potentiostat designs by sensor manufacturers. For example, the application note for the CO-AX (the carbon monoxide sensor that was used to test the CN0357) has equal values for these resistors. Here is another application note where this is also the case.
2. The voltage output of the ADR3412 is VREF.
3. Virtual ground here is VREF (equal to 1.2 V). The block diagram is a bit confusing but '1.2V' is actually the same as 'VREF'.
4. U5 is indeed used as an external input buffer since the RC constant of the R8 + C9 low-pass filter can cause gain errors in an ADC. This is particularly relevant if you want to use an unbuffered ADC with the CN0357 instead of the on-board AD7790.

Hi Willm_van. Sorry for the late reply! Here are a couple of my notes on this circuit:

1. U2-A is used in a potentiostat configuration. Electrochemical gas sensors usually have a bias requirement wherin the voltage difference between RE and WE must a certain value. So the idea here is to keep the voltage levels at the RE and WE pins approximately equal to VREF using the feedback action of the two op amps. In this way, all the voltage adjustments happens at the CE pin. This op amp configuration is pretty common for this application - here is a document by SGX Sensortech that might help you understand it better.
   1. The two RC stages (R5, R6, C3 and C4) are used to improve the stability of U2-A.
   2. R5 and R6, as mentioned above, have an effect on the stability. Using the same value for R3 and R4 however, is based on common recommended potentiostat designs by sensor manufacturers. For example, the application note for the CO-AX (the carbon monoxide sensor that was used to test the CN0357) has equal values for these resistors. Here is another application note where this is also the case.
2. The voltage output of the ADR3412 is VREF.
3. Virtual ground here is VREF (equal to 1.2 V). The block diagram is a bit confusing but '1.2V' is actually the same as 'VREF'.
4. U5 is indeed used as an external input buffer since the RC constant of the R8 + C9 low-pass filter can cause gain errors in an ADC. This is particularly relevant if you want to use an unbuffered ADC with the CN0357 instead of the on-board AD7790.

Hello ALU6,

Thanks for the reply.

Could you please elaborate on 1.a? I am used to dealing with type 1,2, and 3 compensators, and, do not recognize or know about this style of compensating. References would be appreciated.

Also, if I wanted to simulate this circuit in LTspice, how do I model the 3 electrode oxygen sensor? 

Hi Willm_van,

ALU6 can elaborate on your 1.a, but on the question of modeling the sensor - we try to avoid speculating on how to model particular sensors, and wherever possible, copy the manufacturer's recommended circuit. A similar thread is here:

 RE: CN0396 AlphaSenseCOH-A2 LTSPICE model 

Not trying to dodge the question, but there may be aspects of the sensor that we don't fully understand.

-Mark

Hello, No problem, thought I would ask.

Although, I am still waiting for an answer RE my other question

Hi Wilm_van,

Are you referring to the topology of the U2-A feedback circuit? I believe this was just based off the recommended potentiostat circuit provided by Alphasense for the CO-AX (the sensor used to test the CN0357 - you can find this in the application I linked in my first response).

The component values were obviously adjusted for the ADA4528-2 (since the Alphasense circuit used a different op amp); probably determined empirically via testing. This is a pretty old CN though and I unfortunately don't have a record of that data.