Connect with us

Alternate for SG3525

Discussion in 'Electronic Design' started by [email protected], Jun 21, 2007.

Scroll to continue with content
  1. Guest

    Is there a better alternative for PWMing a solenoid valve at about 20
    Hz than a SG3525? To get this low frequency with a SG3525, the timimg
    cap needs to be an order of magnatude greater than the max.
    recommended value. Is this a stress issue on the discharge circuit?
  2. I think the main concern with large timing capacitors is the
    energy dumped into the discharge transistors. You should be
    able to add extra resistance between the discharge pin 7 and
    the capacitor measuring pin 5 to limit the energy dumped
    into this transistor each cycle.

    But there may be a solenoid driver chip that includes the
    power switch transistor.
    or look through the rest of the L29X series.
  3. Winfield

    Winfield Guest

    I'd consider using a CMOS PWM chip, which would allow
    using higher resistor values. BTW, what's wrong with
    a higher-frequency PWM for the solenoid?
  4. MooseFET

    MooseFET Guest

    It likely has a very lossy core. 100Hz may work but not 100KHz.
  5. James Arthur

    James Arthur Guest

    But the a.c. flux will be low, so losses shouldn't be much.

    For strong pull-in force then hold-in at a reduced power, open-
    looping these things (initially on, followed by a fixed duty cycle
    PWM) is often good enough.

    James Arthur
  6. Guest

    The solenoid does not respond to PWM signals at higher frequencies.
    With increasing frequency, the response time becomes an increasing
    portion of the on - off cycle until it just turns on or off. The
    intellegence is in the duty cycle, current limiting or conservation
    is not a factor. In other words, the the solenoid has to bang on and
    off in relative proportion to the duty cycle.

    Is it true, the long term average power through the discharge circuit
    may not be much different at 20 Hz than from a proper size cap
    discharging at a rate magnatudes higher? If not, tell my old brain
    why? Adding a discharge resister causes an unwanted anomaly on the
    output pulse train, unless it is a few ohms.


  7. Have you estimated how long your valve will last, mechanically, if it
    continuously opens and closes 20 times a second?

  8. MooseFET

    MooseFET Guest

    It won't be too bad if the valve is designed for that service. The
    valves in an engine going at 3000 RPM will operate at a 3000 / (2 *
    60) = 25Hz rate. The valves in an engine used continuously last for
    years. The valves that control the water flow into an automatic
    washing machine wouldn't last very long doing this.
  9. James Arthur

    James Arthur Guest

    The solenoid's plunger won't move, but its coil will still integrate
    the effects of any applied voltage, creating a holding force.

    When the plunger is out of the coil, the poorer magnetic circuit means
    it takes a lot of juice to exert a given mechanical pull-in force on
    that plunger. Once the plunger has moved within the coil, more force
    is produced with less current. To save power (or heat in the coil),
    then you can start chopping.

    I've seen this done with relays, which also combine coils and a
    magnetic circuit with a gap that closes.
    I don't understand. It almost sounds like you're trying to get
    proportional motion from the solenoid, which it won't easily do--
    that's a linear motor.
    Discharge resistor? Please elaborate.

  10. Guest

    The end result is proportional pressure from the valve, which is made
    to operate in the 15 - 20 Hz range. The valve is in a closed loop so
    it does not need to be too accurate.

    The discharge resistor can be used in the SG3525 frequency oscillator
    timing circuit. To obtain 20 Hz the timing capacitor can be larger
    than the max recomended value. A discharge resistor is used to limit
    the current though the discharge circuit.
  11. James Arthur

    James Arthur Guest

    Ah, the timing resistor. Gotcha.

    Rather than fret about SG3525's guts, why not roll your own controller
    from an LM339 or such? Easy at these frequencies, & you'd have a
    section or two left over.

    E.g.: (oscillator to produce ramp)-->(threshold comparator) = pulse
    width modulator, with another section for the error amp.

    Then again, you could always use a PIC... <ducking for cover>

    James Arthur
  12. MooseFET

    MooseFET Guest

    [... PWM in a valve ...]
    You may be able to do all that is needed with just two LM339 or other
    comparitors. At these frequencies (20Hz) an LM324 would serve as a

    The error amplifier stage is as per normal. The oscillator stage
    needs to be a bit funny.

    R1 10*R1 1*R1
    Error sig ----/\/\----+-------/\/\-----+---/\/\/----Vcc
    ! !
    --!+\ !
    ! >--------+-+--- Out
    GND--!!------+--!-/ !
    C1 ! !
    R3(huge) R2

    C1 and R2 have to be larger than normal. At the two extremes the
    frequency decreases.

    The R3 ensures that the Error Sig can completely stop the output. It
    just has to overcome the biggest offset voltage you can see. An
    R4(huge) to Vcc could also be added to make the duty cycle go to 100%.
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