A very precise comparator

Discussion in 'General Electronics Chat' started by sharmilla, Nov 4, 2008.

  1. sharmilla

    Thread Starter Member

    Nov 3, 2008
    17
    0
    hi all,

    can anyone advise on a very precise voltage comparator, i would like to compare two voltages in a very precise manner. Any suggestion of precise voltage comparator or circuit layout would much be appreciated.
     
  2. thingmaker3

    Retired Moderator

    May 16, 2005
    5,072
    6
    Have you considered driving the voltage comparator with an instrumentation amplifier?
     
  3. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Last edited: Nov 4, 2008
  4. sharmilla

    Thread Starter Member

    Nov 3, 2008
    17
    0

    hi what is an instrumentation amplifier....can you please explain
     
  5. beenthere

    Retired Moderator

    Apr 20, 2004
    15,815
    282
    Right here in our Ebook.

    If it's gotta be really tight, convert the voltage to digital with a 16 bit ADC and compare it to the precise value.

    Could you expand on "precise"?
     
    Last edited: Nov 4, 2008
  6. dileepchacko

    Active Member

    May 13, 2008
    102
    1
    Hi
    From the question it is not clear that what you mean by precise. whether it is in the sense of voltage, speed and accuracy. Any way some devices are available for your application.
    1. AD790 from analog device
    2. LM139 from national
    You can go through the two data sheet which I have uploaded.
     
  7. sharmilla

    Thread Starter Member

    Nov 3, 2008
    17
    0
    hi all,

    Actually what i meant by precision here is, i would want to compare a voltage that is very noisy to a reference voltage....but the output of the comparator should not be noisy and it must be precise.

    many thanks
     
  8. Wendy

    Moderator

    Mar 24, 2008
    20,766
    2,536
    The 1st step is to quieten the incoming noise without distorting the signal. This can usually be done with a simple RC filter.
     
  9. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Just to show what Bill's talking about, see the attached schematic.

    The signal input is a simple 1kHz sinewave that measures 2v peak to peak, and has a 2v offset. When averaged over a period of time, you should just get the offset voltage.

    The output of the signal generator is displayed in the simulation as the yellow trace.

    The output of the circuit is displayed as a green trace. Note that it takes around 120mS for the voltage on the capacitor to stabilize. This is due to the RC time constant.
     
  10. steveb

    Senior Member

    Jul 3, 2008
    2,433
    469
    You need to specify the signal bandwidth. In other words, how fast will your signal be changing. The previous suggestions to filter, are good ideas, but you can only filter out the noise that is above your required bandwidth.

    If you are just looking at DC levels, then you can filter the heck out of it and be in good shape, otherwise, your signal noise is going to cause your comparitor output to oscillate once the signal is near the comparison level.

    Once method to prevent the oscillations is to use hysteresis on the comparison level. Here you have different trip points with rising and falling voltage levels. However, now the precision comes into question. Your trippoint is not a fixed level, but is a votage band that is larger than the noise level.

    So basically, you need to tell us more information as far as bandwidth and precision needed.
     
Loading...