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.

INA116 question

Other Parts Discussed in Thread: INA116

We're using the INA116 in a differential coulombmeter probe.  On both sides of the pcb, guard tracks surround both electrode inputs.  The guarding is remarkably accurate as shown by the extremely low DC drift at the output of the 116.  Even with nothing connected to the inputs, the output has low DC offset and very low drift.  We are able to connect capacitive sensors to the inputs and maintain this type of operation for many hours.  However, we do observe small (50-100 microvolts pk) low frequency pulses (about 0.5 to 1 sec pw) appearing occasionally at the output (one every two or three minutes or longer).  The larger the input sensor capacitance,  the smaller and less frequent are the pulses.  We want to use small-capacitance probes and continue operating without input bias resistors, but need an explanation for the source of the pulses.  They appear regardless of the shielding or location of the device, so this does not seem to be an external E-M source. We also have voltage regulators (+and - 5V) near the IC, with a battery for pwr source.  Any ideas on what causes the pulses? 

  • Hello Ross,

    Can you please post scope shots of the pulses?

     

  • Ross;

    What you are seeing may be "popcorn noise"; this is something that is seen in semiconductors from time to time. It is process dependent and the internal circuit can make it more or less susceptible to this type of noise. Generally, it may be lot-dependent or even vary from unit to unit.

    I would not rule out an external influence such as ionizing radiation until I observe the effect of bringing a wristwatch with a radium dial near your circuit.

  •  

     

    Here are a couple of examples.  Time scale is 2 secs per division, amplitudes are about 350 uV on the top pulse and 100 uV on the 1st pulse in the bottom example.  BW is 0.005 - 100 Hz in top example, 0.5 - 100 Hz in bottom example.

  • We tried a packet of radioactive mud, radioactive enough to get a nice response from a cheap Geiger Counter at least, and did not see any increase in pulse counts.  Popcorn noise is an interesting conjecture.  We hadn't considered it because the rate is not very high.  We have seen this activity in at least four different IC's so far.

  • Ross;

    Your experiment shows that radioactivity can be eliminated as a possible cause of what you are seeing. One other thing to check (but unlikely due to appearance of the waveform) is the type of capacitors that you may have in the circuit. High-K ceramics are insidious in low-level circuits; they are both piezoelectric and pyroelectric, generating small charge displacements from shock and vibration and from changes in temperature. The best capacitors for ultra-sensitive amplifiers are NPO or COG ceramics, but they are not available in values above about 10nF. Above that, film capacitors are the next best choice.

    You might try to correlate the pops with the date codes on your INA116s.

  • Thanks Neil.  From what we've seen, the value of the capacitance across the inputs has an effect, but no real discernable difference between a NPO/COG ceramic and a mica capacitor.   Will check into the date codes.  Thanks!

  • Ross;

    Silver mica capacitors are excellent but they are getting hard to find.

  • We've been using the Cornell Dubilier CD15 series, available but pricey.  Regarding shot noise, is this hypothesis supported by the observation of less noise (amplitude and rate) with increasing capacitance across the inputs?  I could see a possible low pass filter...

  • Ross,

    My esteemed former colleague, Neil, has given you some good advice.

    You should also check out the power supply pins. The popping could be coming from a noisy (popcorn noise) or defective voltage regulator. A small voltage step on the positive or negative supply pins can create the type of pulse you are seeing.

    Regards, Bruce.

  • Thanks Bruce.  Will check into this.  We have seen the noise on all test units built so far, and have looked at our power supply pins, but not with same long-term monitoring as we have done with INA116 outputs. 

  • Ross;

    Do not confuse shot noise with popcorn noise-- they are two very different things with different characteristics and sources.