Schmitt Trigger troubles: multiple edges from a single analogue pulse

Discussion in 'General Electronics Chat' started by dan.col, Apr 3, 2016.

  1. dan.col

    Thread Starter New Member

    Mar 12, 2016
    4
    0
    Hi all,

    I'm attempting to condition an analogue pulse into a nice clean square pulse to trigger an edge-detecting microcontroller interrupt downstream. (see 20160403 cct.pdf)

    The analogue signal, from a phototransistor, (label A) looks like a rounded square pulse with Vmax=4.8V and Vmin=0.2V (phototransistor pulse.pdf). The pulses are variable duration in the range 1-10 milliseconds long. The signal ramp rate is approximately 7 Volts per millisecond (ie. taking around 0.5 milliseconds to transition).

    I noticed when hooking the collector signal directly to a 74LS02 input (labels A->B) that a significant number of additional edges were generated on the NORout output during rise and fall. These edges occasionally falsely trigger the downstream microcontroller interrupt, which is a problem.

    So I instead hooked up the collector signal directly into a 74HC244 Schmitt Trigger (labels A->C), and was surprised to note the Schmitt's output BuffOut also showed additional high frequency oscillation edges for around 100 microseconds after the first edge (see 'Input and Schmitt output.pdf' with input on top, output below at a different scale). I had hoped for a single square pulse output. I also note that somehow the input signal is oscillating now as well.

    My questions:
    1 Shouldn't the 244's hysteresis prevent the oscillation and give a single output pulse ?
    2 I found some online references to custom hysteresis calcs for using a handful of resistors around an LM393, but given I have a large number of phototransistor inputs to deal with, what solution would keep component count and cost low ?

    Many thanks.
     
  2. dannyf

    Well-Known Member

    Sep 13, 2015
    1,771
    358
    likedly coupled through some common source, like the supply or ground.

    I would try decoupling first. If that doesn't work, try to put a rc filter on the phototransistor's ouput. A small .1uf from the collector to ground should work.
     
  3. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,516
    1,246
    Each power pin on each device *must* be decoupled to ground. Keepthe leads as short as possible. 0.1 uF is the most common value. Ceramic or film caps are best.

    ak
     
  4. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    5,773
    1,103
    The 74HC244 is an octal bus transceiver IC, apparently without a Schmitt trigger function. Try a 74HC14.
     
    PeterCoxSmith likes this.
  5. dannyf

    Well-Known Member

    Sep 13, 2015
    1,771
    358
    it is likely that your mcu's input pin has a ST structure so you don't need a dedicated ST buffer.
     
  6. dan.col

    Thread Starter New Member

    Mar 12, 2016
    4
    0
    Thanks everyone !

    Danny and ak - I tried decoupling, but with no improvement.

    Alec_t - I think you're right. A 74LS14 (with datasheet showing 0.8V hysteresis) worked ok. I think what tricked me was that both the 74LS14 and 74HC14 have hysteresis, whereas the 74LS244 does but the 74HC244 doesn't. I'll purchase and try out the 74LS244.

    Appreciate everyone's help.
     
  7. PeterCoxSmith

    Member

    Feb 23, 2015
    148
    38
  8. dannyf

    Well-Known Member

    Sep 13, 2015
    1,771
    358
    Getting a new device doesn't fundamentally change the situation. There are some forms of coupling going on here. You are better off finding out what it is.
     
  9. crutschow

    Expert

    Mar 14, 2008
    12,991
    3,226
    Standard gates will often oscillate around their transition point with a slow rise or fall time input.
    As you found, a Schmitt trigger eliminates the problem.
    Why do you think the 74LS244 has a Schmitt trigger input? It does not state that in the data sheet.
     
  10. dan.col

    Thread Starter New Member

    Mar 12, 2016
    4
    0
    Hi Crutschow,

    http://www.ti.com/lit/ds/symlink/sn74ls244.pdf is an interesting datasheet, in that it mentions hysteresis, and shows a hysteresis symbol in the logic diagram, but doesn't actually say 'Schmitt trigger'. Isn't the presence of hysteresis in this kind of buffer equal, or at least functionally equivalent, to being a Schmitt trigger ? In any case, isn't hysteresis the thing I'm really after ?

    Cheers.
    Dan.
     
  11. crutschow

    Expert

    Mar 14, 2008
    12,991
    3,226
    Okay, well that may do what you want then, as it is hysteresis that defines a Schmitt trigger.
     
  12. dan.col

    Thread Starter New Member

    Mar 12, 2016
    4
    0
  13. RichardO

    Well-Known Member

    May 4, 2013
    1,230
    382
    I find it interesting that some '244's have hysteresis and some don't. :eek:
    This makes upgrading a circuit from LS TTL to CMOS problematic. Thanks for the heads up.
     
  14. hp1729

    Well-Known Member

    Nov 23, 2015
    1,943
    219
    Who's data sheets are you looking at? I never noticed it before either. Yes, 74HC244 is not Schmitt. 74LS244 is.
     
  15. ramancini8

    Member

    Jul 18, 2012
    442
    118
    Your data indicates that you need 0.7V hysteresis to be safe. Looks like the LS244 has this much hysteresis.
     
Loading...