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.

LM2904: Comparator accuracy

Part Number: LM2904

Hi Team,

How do we calculate the accuracy of a comparator circuit(without hystesis) like below?

https://www.ti.com/lit/an/sboa219a/sboa219a.pdf?ts=1646117361920&ref_url=https%253A%252F%252Fwww.google.com%252F

Is a higher comparison threshold voltage effective to alleviate the error caused by Vos?

Regards,

Charles     

  • Hi Charles,

    the question sounds a bit academic, because the comparison threshold is usually given by the application itslef. But you are right: If your input signal is a sine wave with 1Vp and the threshold is 0.5V, then an input offset voltage of comparator of 10mV looks worse than when the amplitude of input signal is 10Vp and the threshold is 5V.

    For an error analysis the manufacturing tolerances of resistors have to be considered. In many cases they play a big role. Assume you want to generate a threshold of 5V from a reference voltage of 10V, then 1% toleranced resistors will give you an error of up to +/-50mV. This can be quite more than the input offset voltage of comparator. Using 0.1% toleranced resistors can do the trick then.

    Then, the tolerance of reference voltage has to be considered. If the supply voltage is used as reference voltage, then the output voltage tolerance of voltage regulator has to be considered, which is in the range of about +/-5%, usually.

    But not only the input offset voltage of OPAmp has to be considered, but also the unwanted voltage drop which is generated by the input bias currents flowing across any resistance being connected to the input. As an example: If one input of comparator sees a resistance of 100k to signal ground, an input bias current of 500nA will cause a voltage drop of additional 50mV. This voltage has to be considered as being in series to the input offset voltage of comparator. Or by other words, the input offset voltage of comparator is worsened by 50mV.

    Happily, you can minimize this error voltage by keeping the feedback resistance sufficiently small. If the input sees a resistance of only 10k, this error voltage decreases to 5mV.

    Another factor which can totally sabotage all your fine calculations is noise. If any noise is superimposed to the threshold voltage or the input voltage, the comparator can toggle earlier or later than calculated. If your threshold voltage is 5V and the signal comes with a noise level of +/-100mV, then the comparator can toggle at 5V - 0.1V = 4.9V or 5V + 0.1V = 5.1V. And if your input signal is rising or falling only slowly, this noise can totally ruin your timing. Because of this, I always use hystersis.

    Kai