This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

protection on ADC of MCU or additional instrumentation amp



I need to get some additional analog inputs to the microcontroller. These signals are very slow (1 sample per second is enough), 12-bit res is just ok. They are also nicely in range, and required gain would be 1. However, protection (esd/ overvoltage) needs to be there.

Now, which approach is better?

1. A series 10k resistor, then tvs, then RLC filter, then clamping diodes, then directly to ADC?

2. Doing it half externally. A 10k resistor and tvs for each channel, then multiplexer (which selects these channels instead of mcu), then RC filter, then instrumentation amplifier (Where V- and Vref is ground anyways), and then going to ADC.

Is 2 simply overkill, and does not offer any extra protection?

Would 2 provide more isolation maybe? If so, what would be a recommended design?

Having channels directly on MCU is certainly more convenient software-wise, but I wonder if the mcu may get damaged because of not doing the right thing.

If using INA, this is how I am depicting it.

  • Why the external amplifier? Would its input be more robust than the microcontroller's one?
  • Main matter is that it is better than some amplifier or multiplexer dies than the MCU. The analog inputs are secondary matters, and not having them is not necessarily awful, but if MCU dies, it is awful.

    Now, these inputs can be very slow. Literally, they need to be sampled say once in a minute. So, input series resistance on it could be large. MCU has 1M inputb resistance on ADC. Some amplifier can have 10M, and make it even easier to do.

    Now, say by pcb size/price compromise, which one is better:

    - 10k ohm resistor + 400W TVS + RLC filter (stronger) + 1kW clamping diode x2 + MCU

    or

    - 10-50kohm resistor + 100W TVS + RC filter (tiny) + multiplexer + INA + MCU

  • Hi Morteza,

    I am going to more this over to the ADC forum as the team their would be better suited to answer your questions. We know a lot about protection on the interface team but not a lot about the ADCs so I will let those guys handle this.
  • Michael,

    The input protection that you choose might depend on the application and what problems might come up with the protection.

    In the products that I support, we often make measurements of voltages coming from resistive elements, and comparing that to a voltage from a reference resistor. Since both the resistive element and reference resistor are driven by the same current, we can make precision measurements. However, if there are any leakage currents, they become errors in the measurement.

    Because of this, TVS diodes are generally not used. With TVS diodes, the leakage currents are rather large and adds significant error.

    I've seen a combination of series resistance and Schottky diodes that shunt current to either to positive supply or ground to protect against overvoltage events. I think BAV99 diodes have been popular because of the low leakage. As an example, I've shown how it's connected below.

    I'd note that the series resistance for these oversampled ADC generally shouldn't be larger than 10k. Since the inputs are sampled much faster than the data rate, you might get other errors from the measurement with large series resistances.

    This might not give the level of protection that you need, but it does suit many applications. Your application might have different requirements, so you might want to post more about what your application is and what it requires for protection.

    Joseph Wu