1. We will be in Read Only mode (no new threads, replies, registration) starting at 9:00 EDT for a number of hours as we migrate the forums to upgraded software.

I Need a GOOD Design for a PWM-Specific Parameters.

Discussion in 'The Projects Forum' started by donho, Sep 16, 2012.

  1. donho

    Thread Starter New Member

    May 16, 2010
    I want to create a PWM for my workbench, there are many ways to go about it, however I thought to task others expertise. First of all, the PWM is to be done with op-amps and/or a simple 555. I will list the parameters and you make the call on which setup to use.
    I intend to make a PWM power supply for my workbench, and since ALL of my test equipment is precision, I want the PWM generator to be precision as well. Here is what i require:

    The PWM must operate from 0Hz to maybe 100mhz (range may be limited by the selection of the semiconductors used).
    The PWM must be variable from 0 to 100% duty cycle.
    The PWM must have available 4 ranges of frequency which do not overlap more than a few Khz
    The PWM Duty Cycle and Frequency controls must NOT interact with each other.
    The Amplitude of the PWM output must be variable. (0~12v)
    I will design the power supply section with a 24v source and will have both a +12v and -12v rails (fully regulated) for the op-amps. I may also choose to utilize 10-turn precision wire-wound pots for the frequency & duty cycle. I may also add an internal frequency counter and % Duty cycle meter, that can come later. I have thought of other methods as well using a simple function generator chip or microprocessor. In my opinion, both the op-amps and the 555 have been around forever and are also very forgiving to any overloads as well as being extremely versatile. As the "core" of the power supply, it will in the future be subject to many different stress factors. I have weighed the options and I choose the "simple" approach.

    I will use the signal output from the "core" to drive various setups of a PWM Power stage using its own power source and drive transistor(s).
    If you have any other questions, please contact me.
  2. Wendy


    Mar 24, 2008
    High freq is a bit unrealistic. PWM generally uses a triangle wave, which requires a lot of harmonics. Your circuit will go past 1Ghz in same places in freq response.

    Here is a simple design, it will go up to 2Mhz.

    The base frequency is rock stable, and can be easily changed.

    This is from LEDs, 555s, Flashers, and Light Chasers, Chapter 5.


    While not what you asked for, it illustrates some basics. It would help if you told use what you actually needed.
  3. bountyhunter

    Well-Known Member

    Sep 7, 2009
    This makes absolutely no sense. Aside from the fact that trying to switch ANYTHING at 100 MHz is ridiculous, the way any PWM power supply works is that the ENTIRE design of the power supply is dictated by the frequency. The transformer, L-C filter, filter caps, inductors, rectifiers, snubbers etc all depend on what switching frequency you choose. Put another way: once you figure out what to build, you select the most advantageous switching frequency and design the thing.

    The way this works is you figure out what kind of a power supply you need: input voltage, output voltage/current and then decide how to approach. Design topology and switching frequency are two of the basic decisions that flow from this.
  4. wayneh


    Sep 9, 2010
    Most of those design parameters can be met by a LabJack, which is a computer-controlled, USB data acquisition device.

    Oops, nevermind. The clock is 48MHz but not the PWM output options.
    Last edited: Sep 17, 2012
  5. donho

    Thread Starter New Member

    May 16, 2010
    Bounty Hunter....
    Thanks for your response.
    There is a typo, should be maybe 10Mhz, not 100. I just want to be able to reach the highest frequency attainable with the op-amps and/or the 555. (my regular job is working in the Ghz range, VCO's & Synthesizers) Of course it makes sense. I do a fair amount of PWM designs, all from the ground up. This power supply will enable me to select the optimized frequency and duty cycle for the "project." I power the "project" temporarily with the Bench PWM and when optimized, take frequency and duty cycle measurements. Then the "project" would be constructed with fixed frequency, duty cycle and amplitude components. This will save a huge amount of time, just as having a good variable power supply on the workbench (or several) does. Make more sense now?

    Last edited: Sep 17, 2012