Connect with us

50 ohm output ideas

Discussion in 'Electronic Design' started by Dave Boland, Oct 5, 2005.

Scroll to continue with content
  1. Dave Boland

    Dave Boland Guest

    I designed a circuit that produces some low frequency (.1 Hz
    to 100 KHz) square wave outputs in the TTL range (Voh =>
    2.4, Vol <= 0.8) using a 74LVC244. A customer called and
    asked if it will drive a 50 ohm load like a signal
    generator. The '244 won't do this job well, so my question
    is the best way to proceed? The options are another driver
    chip, op-amp, or discrete transistors. Anyone been through
    this and have some sage advice?

    I looked through questions similar to this, and a popular
    concept was a chip like the '244 with outputs parallel so
    the max. current is not exceeded. My concern is for
    reliability and waveform glitches.

  2. Joerg

    Joerg Guest

    Hello Dave,
    That seems to be a popular method. However, never parallel unless it's
    all on the same chip.

    Another option is the usual push-pull follower. At 100kHz almost any old
    npn/pnp combo should do. Considering that CMOS chips drive rail-to-rail
    that would still leave enough swing, provided you are supplying the
    driver ahead of it with 4V or more. 3.3V won't cut it with this setup.

    Regards, Joerg
  3. John Fields

    John Fields Guest

    Micrel makes some MOSFET drivers which would be perfect for what you
    need, it seems.

    seems like a likely candidate, but if not, check the rest of their
  4. Jon

    Jon Guest

    I suggest a video amplifier configured for again of 2. Insert a 50 Ohm
    back termination resistor in series with the output. The gain of 2
    compensates for the 2:1 voltage divider formed by the 50 Ohm back
    termination resistor and the 50 Ohm load.
  5. Jeff

    Jeff Guest

    I eventually used an AD817 Op Amp in a similar requirement for a 50 ohm
    output buffer for a VCO.
    As below, the AD817 needs a 50 ohm output resistor, and set to a gain of 2.
    Works good on single supply voltage as per the spec sheet example - but
    needs therefore a DC blocking cap at the output.
  6. Robert Baer

    Robert Baer Guest

    I would use a CMOS hex inverter, all 6 in parallel.
  7. Guest

    Thinking out loud: how about something like an MC1488 or MAX232, followed
    by a Schottky diode (to keep the 'low' voltage reasonably low) to throw
    away the negative swings you don't care about?

    Matt Roberds
  8. If adding another IC, there are several 50ohm drivers readily available.
    For logic signals like this, I had a similar need a while ago, and a look
    in my 'parts bin', found the MC4427 MOSFET driver, which worked very well
    indeed, and had the advantage of being really small.

    Best Wishes
  9. Roger Hamlett wrote...
    Did you mean TC4427? Sounds like you have a pretty good parts bin!
    Those parts do moderately well when powered from 5 volts: 15-ohms
    Ron typical, 50-ohms max, according to the datasheet. It's amazing
    how high the ON resistance of these serious 1.5A MOSFET drivers can
    get at low supply voltages. 50 ohms is pretty wimpy.
  10. Yes TC4427.
    However my data sheet only gives the Ron Max at 25R, not 50R at 5v. I was
    driving 10v though, so this aspect was less of a problem for me.

    Best Wishes
  11. Fred Bloggs

    Fred Bloggs Guest

    A high speed non-inverting line driver operating off the 3.3V with TTL
    compatible output and short-circuit tolerant would be something like this:
    View in a fixed-width font such as Courier.
  12. Chris Carlen

    Chris Carlen Guest

    There is also the preferrable although more expensive TC4427A and
    related parts (TC4426A) that offer better matching of rise/fall times
    and delays.

    Good day!

    Christopher R. Carlen
    Principal Laser&Electronics Technologist
    Sandia National Laboratories CA USA

    NOTE, delete texts: "RemoveThis" and
    "BOGUS" from email address to reply.
  13. The trouble with these things is the propagation delay, like,
    40ns is forever. Last time I had to do this for just a few
    channels I rolled my own, using two transistors and a few
    resistors. Sort of like this:

    No good if you need many, though. A simple emitter follower
    works in fact very well, especially if you drive it with high
    speed CMOS.

    Jeroen Belleman
  14. The original question was low speed (only 100KHz max). Things get
    different if you need to go faster.

    Best Wishes
  15. Fred Bloggs wrote...
    Sheesh, is all that stuff really needed, and is it all pulling
    its full weight? At the other end of the spectrum, Dave could
    consider the mind-blowing shortcut solution offered in The Art
    of Electronics (page 612), which works amazingly well:

    .. +Vcc --/\/\--, 10 ohms, 2W
    .. |
    .. logic output |/ 2n4401
    .. ---|>o-------|
    .. not critical |v output destination
    .. | jack logic - 74LVCxxx, etc,
    .. '----o -------- o--+---|>o-------
    .. cable | 51 ohms term
    .. '--/\/\-- gnd

    The destination sees a logic HI of Vcc - 0.7V and the 50-ohm
    termination acts as a prompt pulldown for the 2n4401 transistor.
    Any pulse traveling down a 50-ohm coax is properly terminated
    at the end, so there's little reflection energy, allowing one
    to get away with an absence of source termination. The jack on
    the destination should be labeled "50 ohms" or "50-ohm term."

    Alternately, one can employ an open-circuit at the destination.
    This reflects back to the source, at which point the cable input
    also appears as an open circuit. But to insure the destination
    gets the intended source amplitude (and doesn't spike to more!),
    the source must present at least 50-ohms of impedance (for 50-ohm
    coax). For example, with a roughly 25-ohm CMOS gate Rout, one
    could add at least 35 ohms, to be on the safe side, or one could
    go for it and add 25 ohms. This simple scheme works surprisingly
    well, and only requires one resistor. If you use a scope to look
    at the signal at the source, it starts out reduced, and only pops
    up to Vcc after the reflection arrives back. But if you look with
    your scope at the end, you'll just see an ordinary full-size pulse,
    as intended. Just one resistor... Cool, huh?
  16. John Larkin

    John Larkin Guest

    As time goes on, ICs keep killing discrete circuits. If you parallel
    all three sections of an NL37WZ16 and run it from 6 volts, it will put
    5 volts into 50 ohms in about 600 ps, for 13 cents.

  17. John Larkin wrote...
    A very nice factoid, thanks John! That implies three parallel
    NL37WZ16 stages have Zout = 20 ohms, and therefore needs about
    30 ohms in series to transparently drive a coax?
  18. John Larkin

    John Larkin Guest

    Something close to that, maybe 10+40 or something in-between. I'll let
    you know if the opportunity arises, and you can tell us ditto.

  19. John Larkin wrote...
    Sorry, right, I meant to say 10-ohms Ron, 40-ohms series out.
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