pulse generator with edge control

Discussion in 'The Projects Forum' started by james_foong, Jul 2, 2012.

  1. james_foong

    Thread Starter New Member

    Jul 2, 2012
    5
    0
    How to design a pulse generation circuit with falling edge control? Pls help shed some light. Thx
     
  2. james_foong

    Thread Starter New Member

    Jul 2, 2012
    5
    0
    To be more precise, I need a pulse generator with variable falling edge. THe pulse generator can include voltage regulator, etc.
     
  3. Sensacell

    Well-Known Member

    Jun 19, 2012
    1,130
    267
    Variable edge fall time/slew rate? or pulse width?

    Please clarify?
     
  4. WBahn

    Moderator

    Mar 31, 2012
    17,743
    4,795
    Also, please give some idea of what kind of range and resolution you are trying to achieve on whatever, precisely, it is that is being varied.

    Oh, and for what it's worth, congratulations. You are AAC's 30,000th member.

    (Sorry, no prize.)
     
  5. wayneh

    Expert

    Sep 9, 2010
    12,123
    3,047
    I'd do it in software, for instance Garage Band. Not sure of the max frequency. It's probably good up to 40kHz or so.
     
  6. james_foong

    Thread Starter New Member

    Jul 2, 2012
    5
    0
    yes, both variable pulse width and falling edge pulse generator circuit
     
  7. WBahn

    Moderator

    Mar 31, 2012
    17,743
    4,795
    So, aside from variable pulse width (i.e., when the falling edge occurs), what else about the falling edge needs to be variable (since you say 'both')?

    And, again, what range and resultion are you looking for. The discussion about how to choose from 100 different pulse widths ranging from 1s to 1hr is going to be quite different from one talking about 1,000,000 different pulse widths ranging from 1µs to 1s, which is going to be quite different from one talking about 10 different pulse widths ranging from 1ns to 2ns.

    Throw us a bone! You need to set a meaningful context for the discussion.
     
  8. james_foong

    Thread Starter New Member

    Jul 2, 2012
    5
    0
    sorry about the ambiguity. Let's assume a circuit that can generate a pulse with pulse width varies from 500ns to 1us and fall time that varies from 100ns to 500ns for 1kOhm resistance load.
     
  9. WBahn

    Moderator

    Mar 31, 2012
    17,743
    4,795
    Okay, we are getting somewhere.

    Now, how much resolution do you need?

    Can the rise time be the same as the fall time (i.e., can they vary together)?
     
  10. wayneh

    Expert

    Sep 9, 2010
    12,123
    3,047
    Well, so much for an audio-based source.
     
  11. james_foong

    Thread Starter New Member

    Jul 2, 2012
    5
    0
    pulse width and fall time vary in 4 steps (2 bit resolution). rise time is fixed.
     
  12. WBahn

    Moderator

    Mar 31, 2012
    17,743
    4,795
    Jeez. I feel like I'm pulling teeth, here.

    What is the voltage of the pulse need to be?

    What does the fixed rise time need to be?

    What is the load capacitance (you've already given the resistance as 1kΩ)

    So far we have:

    ================

    Pulse width: 500ns to 1µs with 4 steps.

    So is something close to 500ns, 625ns, 825ns, and 1000ns good enough?

    Notice that these are all in increments of 125ns. Thus, if you have a clock running at 8MHz, you can simply count 1, 2, 3, of 4 periods to determine the pulse width. If you run your main pulse through a 4-stage shift register, then you can use simple logic to generate the four different pulse widths.

    ================

    Fall time: 100ns to 500ns with 4 steps.

    So is something close to 100ns, 233ns, 367ns, and 500ns good enough?

    Changing the fall-time without changing the rise time complicates things, but is doable.

    One way to do it is to have multiple pull-down transistors that are enabled based on the fall-time selection. You would need at least three pull-down transistors, each having a different on-resistance. This gives you seven different pulldown strengths. Order them from highest resistance to lowest resistance and pick four of them. Now put a capacitor across the load and size it so that the highest of the four chosen gives you a 500ns falling edge. You then size your pull-up transistor(s) so that the rising edge is what you want. You now simply make a decoder to map from your two control bits to the resistance combinations you have chosen.
     
Loading...