Connect with us

Sloppy flanks from a 74HC14, any fix?

Discussion in 'Electronic Basics' started by [email protected], Oct 14, 2005.

Scroll to continue with content
  1. Guest

    I have a S/P-dif signal which I send over cat.5 like this:

    S/P-dif source - External USB audiocard. Where TOTX is tapped for TTL signal.
    RS-422 TX (DS 26LS31)
    RS-422 RX (DS 90LV032)
    74HC14/Ph Cmos schmitt trigger
    650 nm, 4°, 1000 mcd

    S/P-dif TTL -> RS-422 TX -> cat.5 stp -> RS-422 RX -> Inverter -> LED

    The TOTX module of the soundcard is tapped for it's TTL signal (verified with
    scope). It then feeds the RS-422 transmitter chip which has 10 uF over
    GND-Vcc @ 5V. Delivered to cat.5 sheilded with a 333 series resistance to
    get from 5V swing downto 1V (approx).
    At the RS-422 receiver the differential is terminated with a 100 ohm resistor.
    And feeds the RS-422 receiver chip. Which controls the inverter input.
    On the inverter one input is used to invert the original signal. Which is then
    feeded into the rest 5x parallell wired outputs. That feed the LED with a
    140 ohm series resistance towards ground.

    The S/P dif signal duration is approx 300 ns. Where _rise_ flank is something
    like 100 ns. Which is way to slow. Everything works rise/fall wise if LED is
    disconnected (observed on scope).

    Any ideas on how to solve this?
  2. w2aew

    w2aew Guest

    Driving an LED directly with CMOS logic doesn't work well if you need
    good speed. You might try adding a simple transistor driver. This
    could be done in many ways. The simplest two are:

    1) Switch: Put the LED and current limiting resistor in series, between
    VCC and collector of NPN. Emitter to ground, base to logic output
    through base current limiting resistor. This approach though, you'll
    have to deal with saturation characteristics of the transistor.

    2) Shunt: Connect current limiting resistor from VCC to LED anode, LED
    cathode to ground. Connect NPN collector to LED anode, emitter to
    ground, base to logic via resistor. Now the transistor acts as a
    shunt. The approach draws a bit more power, but gives pretty fast
    optical rise/fall times.

    There are other driver schemes, but this should be enough to get you
  3. Guest

    Tried this approach with BC337/16 (NPN) and a 100 ohm series resistor.
    Base connected to 3.3V logic through a 10k ohm resistor.
    What's the fastest approach?
    I'm suspecting the LED behaves like a significant capacitor was connected in
    parallell over it.

    Right now it I'm considering useing a readymade TOTX module (rated 12 Mbps).
    So the inverter/transistor approach was hoped to be cheaper & better. But.. =)

    Anyway it can be useful to know how to transmitt optically at high frequencies.
    Esp when going _real_ fast without ruining the wallet.
  4. Jasen Betts

    Jasen Betts Guest

    You have 333 ohm (+ RS422chip impedance) input to a 100 ohm impedance
    transmission line that's not right. (probably not your problem though)
    pulse width or burst duration?
    bigger/better power supply decoupling caps on the inverter,
    use a eparate inverter to power the LED, (or use a BJT etc...)

    what's the capacitance of your LED?

  5. Guest

    Well the datasheet of the receiver says max input is Vcc of that chip. Which
    in turn is 3.3V. And the transmitter sends 5V (or maybe 4.. 4.5). So I had to
    limit it to ensure nothing blows up =). However it strikes me as odd to have
    Vcc as max on a line _receiver_
    However I checked pulses. And they seem alright.
    pulse width.
    I use a 3.3V XC6201P332PR 150 mA capable. With 10 uF electrolytic caps on
    both input and output. And separare ceramic caps directly over the chips.
    No idea. It was just a vague thought when I tested the circuit with the scope.
  6. Apparently your HC14 is not able to provide the current required by the LED.
    In any case not fast enough. The HC14 also seems to be overloaded as it
    disturbs its input as well. There are some "classic" ways to drive a LED.
    For one of them, see schematic below.

    | |
    |\ ___ | |R1
    +--| >O---|___|-+ '-'
    | |/ | |
    | | |
    |\ | |\ ___ | |
    ---| >O-+--| >O---|___|-+ |
    |/ | |/ | ___ |
    | +--|___|-+
    | |\ ___ | R2 |
    +--| >O---|___|-+ |
    | |/ | || |
    | +---||---+
    | |\ ___ | || |
    +--| >O---|___|-+ |
    |/ V
    - LED
    created by Andy´s ASCII-Circuit v1.24.140803 Beta
    (Draw only five of the six inverters available.)

    Component values highly depend on the properties of the LED. R1 provides
    some bias current. Enough to reach its knee voltage so it draws some current
    already without emitting too much light. The voltage is pretty temperature
    dependent so stay on the safe side. Other resistors mainly depend on the
    available current of wich you wish to send all through the LED without
    overloading the HC14.

    petrus bitbyter
  7. Robin

    Robin Guest

    -------- Vcc
    \ emitter
    _\ /_ LED
    \ resistor
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