About the comparator output ripples

Discussion in 'General Electronics Chat' started by JiaZheng, Mar 22, 2013.

  1. JiaZheng

    Thread Starter New Member

    Jan 28, 2013
    20
    0
    I'm using an IC (MAX 999 EUS) and it is a comparator. I found when the input signal frequency is under 1 MHz the output signal is the square wave. But When it reaches 4 MHz, the output becomes to have very large ripples. I have no idea why comparator's output could have such ripples. Should not it always be either 0 or 1 ? Also when the input voltage is small, the output will have oscillations (looks like sine wave but with much noise).

    In datasheet, it presents a figure (figure) that it is capable to respond to a near 50 MHz sine wave. But my circuit even failed to exceed 4 MHz. Quite confusing about comparator's ripple output.
     
  2. GopherT

    AAC Fanatic!

    Nov 23, 2012
    6,061
    3,830
    What load is on your output (resistance)?
     
  3. shortbus

    AAC Fanatic!

    Sep 30, 2009
    4,015
    1,531
    Have you follow the rules of circuit lay out for the chip shown on page 10 of the data sheet? These are important for good use of the chip. #6 is very important.

    Also you don't say what voltage range your comparing. Page 8 under timing has rules to be followed also. Your ripple could be internal oscillation of the comparator.
     
  4. Papabravo

    Expert

    Feb 24, 2006
    10,152
    1,793
    Read the datasheet and look for a parameter called "slew rate". The units are going to be "Volts per microsecond". It will give you an idea that if the input goes too high in frequency the output will NOT be able to keep up. Not only that but the output may not even have time to ago from rail-to-rail and thus may appear to be "like a sine wave" centered around Vcc/2.

    Ideal components exist only in a ideal world. Last time I checked we're not in one.
     
    shortbus likes this.
  5. JiaZheng

    Thread Starter New Member

    Jan 28, 2013
    20
    0
    The output is floating. I just use oscilloscope to check its waveform and I don't know whether the oscilloscope probe has some influence to it.
     
  6. JiaZheng

    Thread Starter New Member

    Jan 28, 2013
    20
    0

    I'm using MAX999 that datasheet doesn't show any reference circuit:(.
    Do you think the comparator's input traces must be the same length and as short as possible? This picture is part of my circuit. U7 is the comparator and the upper two pins are the inputs. I found the left pin is stable but the right one always oscillates. It may be too long? The comparator input voltage are around 100 mv, this two input are from an differential amplifier output. I think I need to redesign my circuit layout to make this two input traces having the same length.
     
  7. JiaZheng

    Thread Starter New Member

    Jan 28, 2013
    20
    0
    Do you think there is any bad layouts and any recommendations to it ?
     
  8. JiaZheng

    Thread Starter New Member

    Jan 28, 2013
    20
    0
    Hi man, I dont understand why the load resistance is so important. I notice that this comparator output seemly is driven by two current source. Will it cause the oscillation?
     
  9. JiaZheng

    Thread Starter New Member

    Jan 28, 2013
    20
    0
    I find the slew rate is very important. If I give a low frequency signal and with small amplitude, then the slew rate is small. The comparator may do not work well. But with that low frequency, if I increase the amplitude, later on it will work fine. (probably because of higher amplitude can give larger rising speed, slew rate). But now although it can give me the same frequency output, its waveform is not a square wave, but looks like triangle. I suspect the probe may cause this problem. What do you think?
     
  10. patricktoday

    Member

    Feb 12, 2013
    157
    42
    The oscilloscope may be affecting the circuit. It creates a new path to ground including the capacitance of the probe and also the resistance. Do you have switchable X1/X10 probes? If so try switching to X10 and see if the rise time improves.
     
  11. GopherT

    AAC Fanatic!

    Nov 23, 2012
    6,061
    3,830
    Add a load transistor to ground (1k to 10k if you are using 5v rals). The output should not float on most comparitors.
     
  12. JiaZheng

    Thread Starter New Member

    Jan 28, 2013
    20
    0
    Thanks. I will replace a better one to test it.
     
  13. JiaZheng

    Thread Starter New Member

    Jan 28, 2013
    20
    0
    right. I will try it out
     
Loading...