Connect with us

Multi Channel F-V converters?

Discussion in 'Electronic Design' started by [email protected], Mar 2, 2007.

Scroll to continue with content
  1. Guest

    Hi All,

    I'm looking to monitor a number(~10) of slowly varying low frequency
    outputs(100Hz-5KHz), I want to be able to sample & report the
    frequency on each channel once every 100mS. The micro processor I'm
    using has a number of ADC channels so I was hoping to use these to
    monitor the output of a Frequency to Voltage converter for each
    channel. The required range is 0-5V on the output of the converter.

    I don't think I can use the micro to time the inputs directly because
    I may not have the overhead avalable to do this every 100mS for every

    Does anybody have any ideas on the easiest way to achieve this?

    I had a look at the usual Semiconductor manufacturers and can't seem
    to find a multi channel IC to do this, does anyone know of any?

    Also what is the best method for pretty precise measurement of these
    frequencies? If you think I'm approaching this from the worst possible
    way let me know this also,


  2. Joerg

    Joerg Guest

    Don't know any multi-channel ones, I usually do such things with the
    LM331 which unfortunately doesn't come in small footprints.

    If the uC has lots of MIPS you might be able to squeeze in a fast
    interrupt scheme. Zero crossings would trip comparators where each goes
    to a port pin. When any of the port pins changes state an interrupt is
    generated. That's going to be lots of interrupts and you'd have to check
    states again after it comes out of the ISR to catch any that might have
    changed state while processing the ISR. Each time a state change is
    detected a time stamp versus a long timer is generated and you'd have to
    also handle the timer rollover. Again, this must be a powerful uC
    because the interrupts will come faster than they can pour Guinness in
    one of your pubs ;-)
  3. John Larkin

    John Larkin Guest

    Why not counters? A $2 CPLD could implement a bunch of frequency

  4. Didi

    Didi Guest

    Also what is the best method for pretty precise measurement of these
    At these low frequencies, and wanting to see 100 Hz every 100 mS,
    F/V will not really work.
    If your MCU has input capture, use it to measure the period,
    then calculate the frequency on a per signal period basis.
    If your input capture inputs are not as many as you need,
    you may still be able to multiplex them (digitally, externally
    to the mcu) to the various signals, even 1 input at 100 Hz
    (10 mS period) can be switched to measure all every 200 mS worst

  5. linnix

    linnix Guest

    Why not?
    I am sampling at 1 ms (with hardware timer) using a $5 micro?
    A decent micro should easily meet your requirements.
  6. linnix

    linnix Guest

    I am sampling 4 channels at 1KHz (1ms) because
    I can't pull data out over internet fast enough.
    By the way, my target is also flashed and debugged
    over the internet. My target (Linux/USB) powers up halted,
    until being jtag scanned (and flashed if necessary)
    by a remote machine (Window).

    If you can digest the data first, you can sample
    at much higher rate. For example, you can sample
    at 10KHz (100us) without any problem.
    The sampling can be timer based for minimum overheads.

    The LM3S828 has 8 channels of 10 bits A2D.
    The chip also has an internal sequencer for batch
    conversion. Namely, doing all 8 channels at once
    per interrupt, and at 125KHz rate.
    Even if you have to couple an external mux,
    you can still sample at 60KHz, more than enough
    for your 5KHz signals.
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