Connect with us

Comparator & Amplifier ICs with High Bandwidth

Discussion in 'Electronic Design' started by [email protected], Jan 26, 2005.

Scroll to continue with content
  1. Guest

    Dear All,

    I'm constructing a circuit which requires some comparators and
    amplifier with high bandwidth. What ICs should I use for bandwidth more
    than 10MHz? How about above 100MHz?? Is LM139 suitable for bandwidth
    above 10MHz? I looked at the datasheet but it did not show anything
    regarding bandwidth.
    Any comments would be very appreciated.

  2. There are comparators and line receivers being as fast
    as 200ps or so available as standard products.
    What input voltage range and logic levels were you thinking
    of ? The jitter can also be specified, though there the
    risetime of the signal comes in.
    The 100MHz can be done with a MC100EL16 & MC100ELT23,
    above ECL, LVDS and CML become standard.

  3. Andrew Holme

    Andrew Holme Guest

    They quote propogation delay for comparators, not frequency.

    The LM139 has a typical response time of 1.3us - not even fast enough
    for 1 MHz.

    I used the Analog Devices AD8561 "Ultrafast 7 ns Single Supply
    Comparator" for converting sine waves to TTL at 16 MHz. That's not
    it's maximum frequency - but it won't do 100 MHz.

  4. Being a comparator the LM139 datasheet doesn't have any "bandwidth"
    parameter specified, but it does provide a response time parameter. For a
    small amount of input overdrive (IE: 5mV) the LM139 will waste vaguely
    around 1.3us before the output transitions (slightly different for positive
    versus negative going output transitions, see the graphs). For large input
    overdrive voltages (IE: large voltage differences on the order of 100mV or
    more between the inverting and non-inverting inputs) the delay shrinks to
    something more like 300ns.

    Either way, the LM139 is way too slow for applications anywhere near 10MHz.
    Even at 1MHz it isn't very useful.

    Linear Technology makes some pretty fast comparators such as the LT1715
    (around 4ns response time).,C1,C1154,C1004,C1012,P1908,D2821

    Unfortunately fast comparators usually aren't very cheap. The LM319 offers
    fairly decent bang for your buck, but it has an 80ns response time.
  5. Fred Bloggs

    Fred Bloggs Guest

    Get lost!- and don't post to this newsgroup again.
  6. Tim Wescott

    Tim Wescott Guest

    A comparator data sheet will talk about "response time" or "delay". If
    it's a decent data sheet it'll have a chart of delay vs. differential
    voltage (they're generally slower with smaller inputs).

    Figuring out how the delay impacts the speed that your circuit can
    operate is left as an exercise to -- you.
  7. Guest

    Hi Fred,

    I'd like to apologize to you. I feel really sorry if I did anything
    wrong before.
    And thanks to those who helped me regarding my questions.

  8. Jim Thompson

    Jim Thompson Guest

    Will, Fred is our resident wacko. You would be best served by
    kill-filing him.

    ...Jim Thompson
  9. Rich Grise

    Rich Grise Guest

    Please don't pay any attention to Fred Bloggs. Stick around awhile, and
    you'll there are a few noisemakers here. I assure you, they're all quite
    harmless. :)

  10. Pig Bladder

    Pig Bladder Guest

    You might be well-served to take your own advice. I know
    everybody else would.
  11. Jamie

    Jamie Guest

    you didn't have your Java yet this morning Fred ?
  12. Guest

    As everybody has already said, the LM139 isn't suitable for bandwidths
    above 10MHz.

    Comparators are usually built with several stages of gain and no
    frequency compensation, unlike op amps - so their gain usually stays
    flat out to quite high frequencies, then falls off rapidly when each of
    the gain stages starts running out of gain at the same frequency.

    Comparator data sheets rarely specify this frequency, but the data
    sheet for Linear Technology "ultrafast" LT10106, which has a 10nsec
    propagation delay, includes an application circuit which uses it to
    make a crystal oscillator - specifically a 10MHz to 25MHz crystal
    oscillator, which suggests that it isn't much use over 25MHz.

    The application notes also includes the claim that the LT1016 has a
    gain bandwidth product of about 50GHz, which - for a part that has a
    typical gain of 3000, implies a bandwidth of 17MHz.

    There are faster parts around - the Honeywell, then SPT and now
    Fairchild SPT9689 has its 3dB point at 900MHz, but it costs $20 and
    Arrow Electronics list a two week delivery time.

    Maxim have parts that are cheaper, and tolerably fast

    which do seem to be able to cope with 100MHz signals - the data sheet
    examples are of 100MHz square waves.
    Newark have 175 MAX9691ESA in stock for $6.00 each.
  13. Fred Bloggs

    Fred Bloggs Guest

    Really? And who the hell do you think you're fooling? You've been doing
    NOTHING for the past several months but cluttering up the server with
    your worthless off-topic, pseudo-intellectual, ignorant crap- while
    producing NOTHING of any informational worth to electronics- because you
    know damned little about the topic. Why don't you just take a hike,
    disappear, die, and/or get the hell off the NG too. You won't be missed.
  14. Guest

    Hi Fred,

    Just want to say SORRY ...

Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day