ADuCM355 probe configuration

I need the ADuCM355/LTC6078 to support both a pH and ORP probe and also 2 temperature sensors. Is this the proper way to connect the 2 probes and temperature sensors?

  • 0
    •  Analog Employees 
    on May 1, 2020 2:55 PM

    Hello,

    This looks fairly similar to what was done on CN0428. It could cause extra noise and error to have separate DACs connected to the references of the two sensors. If they're submersed in the same solution, any difference between the two DAC voltages could cause current flow between the two reference electrodes, which would disturb the measurement. It's probably better to use VZERO0 for both reference electrodes to prevent this.

    On the temperature side, the names on the temp sensor nets are a bit confusing here (net SE0 on the schematic connects to pin DE0 and same for ch1). Was that intentional? CN0428 used RE1/SE1 as the pair for temperature measurement. In either case, I would suggest a DNP 0 ohm resistor from CE0 to RE0 and another from CE1 to RE1 to give you some flexibility for how to set up the switching matrix internally. What types of temperature sensors are you supporting? PT100? PT1000? 30k NTC Thermistor? Have you checked the temperature sensor connector? They're not as standardized as the BNC for pH and ORP electrodes. What are your calibration resistors and how are they connected?

    Thanks,
    Scott

  • Thx for the feedback.

    I have include the entire schematic for your perusal. As you mentioned this implementation is based on CN0428, but I don’t need to measure impedance and I want to support both a pH and ORB probe connected to only one ADuCM355.

     At this point I will likely use NTC so I changed the RCA jacks to 2 wire terminals. But I need to know the best way to connect an NTC to ADuCM355. The way I show it, I don’t believe is right.

     I also connected the calibration resistors like they are connected in the CN0428, but I’m not sure why so many are needed.

     Regards,

    PDF

  • 0
    •  Analog Employees 
    on May 7, 2020 9:05 PM in reply to JTek

    Hello,

    Apologies for the long post, but there's quite a bit to say here.

    Regarding the connection of the temperature sensors, ac excitation for temperature measurement is typically used in this application, so the important thing is to be able to hook the temperature sensor into the impedance measurement loop. (Refer to page 118 of UG-1262 for the detailed impedance measurement loop connections). So the way you connected the sensors will work. Though I would still recommend adding a provision for the CE0 and CE1 connection I mentioned before (if you trace the signals in the UG-1262 switch matrix diagram, you can see where not having a CE connection forces you to use the PL switch, which puts the P5 or P6 switch in series with the impedance you're trying to measure and reduce accuracy).

    Most of the calibration resistors on CN0428 are for the main sensor input, which needs to take in a very wide range of impedances (from single ohms to 10+ Meg-ohms). If all you are measuring in this case is an NTC thermistor, the resistance range can be much more limited. You may even be able to use a fixed gain resistor for the internal TIA (whereas CN0428 had to do auto-ranging, which adds complexity). The main issue that can happen if the calibration resistor is too small is that the ADC can saturate.

    For example, imagine the circuit is measuring 30k from the thermistor, so the firmware chooses a 40k TIA resistance. 0.6V across the thermistor causes 0.8V at the TIA output, which is within the ADC max range. One of the next steps (see the example in page 137 of UG-1262 for reference) is to measure the calibration resistor in the same range so we can take a ratio of the two and get an accurate result. The problem would come if the system tries to calibrate the 40k TIA, but the only calibration resistor available is a 200 ohm Rcal (for example). 0.6V across 200 ohms is 3mA, which would make the output of the TIA want to go to 120V peak, which it can't do, so the TIA output and ADC are saturated and there's no valid calibration measurement.

    So in short, the most important thing is to make sure there's a calibration resistor large enough to fit the maximum TIA gain that you plan to use (while also considering pga gain and excitation amplitude) to meet the conditions outlined above. It's also important that there's a way to connect the calibration resistor pins to the impedance loop. Smaller calibration resistors can be added to improve accuracy for lower TIA gains if you're using multiple ranges. This is a smaller problem, but it's on the opposite end of the spectrum, meaning if your cal resistor is too large for the range you're trying to measure in, the signal can be relatively small at that ADC input, which doesn't invalidate the measurement as long as you can still read a signal, but it's just not the best for noise or error.

    I hope it helps.
    Best regards,
    Scott