Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Measuring delay for solenoids

  1. Dec 27, 2011 #1
    I have a solenoid that is controlled by a PLC. I want to know what the delay time is from the time the PLC sends the signal to engage the solenoid to the time the solenoid is fully engaged. The time will be very short (milliseconds), but I need to know what the delay time is to within 5ms.

    Is there a device that I could use to measure the delay? The solenoid draws 37A when is it moving, and the current drops to about 2A when fully engaged. I could monitor this with an oscilloscope, but those are pricey. The PLC isn't able to record time or counts either.
     
  2. jcsd
  3. Dec 27, 2011 #2

    jim hardy

    User Avatar
    Science Advisor
    Gold Member
    2016 Award

    it'll depend somewhat on where in the line cycle you close.

    but you'll have to measure it.
    put a microswitch under the moving part and use an oscilloscope.

    i've seen relays that'll close in ten milliseconds, but your device sounds like it's substantially bigger.
     
  4. Dec 27, 2011 #3

    vk6kro

    User Avatar
    Science Advisor

    You could make up a small circuit that switches the relay on for a second and off for a second. Then take it to a repair shop, a technical college or a Radio Amateur and ask them to measure the delay time.

    If you have made an effort to make the measurement easy, then you will probably get good cooperation.

    The price of oscilloscopes is probably not as bad as you imagine.
    There are some good ones coming from China that don't cost any more than a good multimeter.
    Like this one for $72 delivered:
    http://www.ebay.com.au/itm/Mini-SD-...159?pt=LH_DefaultDomain_0&hash=item3cc02f06c7
    The bandwidth is only 1 MHz, but for relay testing or audio work it would be fine.
     
  5. Dec 27, 2011 #4

    AlephZero

    User Avatar
    Science Advisor
    Homework Helper

    If you want a really cheap do-it-yourself solution, use your PC's audio input to digitize and record electrical signals. Just make sure the signals are the right amplitude to match the "line" or "aux" inputs (say +/- 0.5V maximum). Don't use the mic input!

    That will give you a timing resolution of about 1/44100 sec. You can record two different signals as the left and right stereo channels.

    Use free software like http://audacity.sourceforge.net/ to do the recording and look at the data.
     
  6. Dec 27, 2011 #5

    es1

    User Avatar

    If you have access to the CPLD HDL and can reprogram it you could create a counter in the CPLD itself. Have it start when the CPLD's IO changes and stop when the solenoid opens. Wire the solenoid output to another IO on the CPLD. You might need to wire the solenoid output with a resistor divider level shifter depending on the voltage levels.

    You'll then know the time to fire the solenoid in units of CPLD clocks.
     
  7. Dec 27, 2011 #6

    jim hardy

    User Avatar
    Science Advisor
    Gold Member
    2016 Award

    sorry - i missed your desire to avoid an o'scope.

    alephzero has a good idea there. use your PC's audio input for an o'scope

    there'll be a magnetic field around the airgap of your solenoid, if that's accessible.
    a small coil there used as a "flux detector" might show enough change between open and closed to be of use.

    experiment !!!!
     
  8. Dec 28, 2011 #7

    vk6kro

    User Avatar
    Science Advisor

    The line inputs of a computer sound card are capacity coupled, so you will get a small spike in the output rather than an actual change in DC levels.

    Also, you will need to use both stereo channels. One for the driving circuit voltage and the other produced from the relay contacts connecting a resistor to ground and letting it conduct current.

    Note that great care is needed with the line inputs of one of these cards as they are easily damaged. To be sure it is safe, you really need to examine it with an oscilloscope, especially when there is a relay coil with 2 amps flowing in it nearby.
     
  9. Dec 28, 2011 #8

    jim hardy

    User Avatar
    Science Advisor
    Gold Member
    2016 Award

    ""The line inputs of a computer sound card are capacity coupled, so you will get a small spike in the output rather than an actual change in DC levels.""

    the solenoid is presumably AC driven (37 amps inrush?)

    so a small coil nearby should pick up leakage flux
    which will be large when solenoid has airgap, ie open
    and small when solenoid is closed

    if the computer sound card will record 60hz

    one might try connecting a small coil to PC's mic input , for use as a flux detector. perhaps an earbud or tiny speaker would work?
    wave it around the solenoid when it's energizeded, at 2 amps and find a spot where there's measurable 60hz signal. It should be larger while solenoid is closing.
    there should be a "sweet spot" near the solenoid's armature airgap where one finds a large change in signal as solenoid closes. that waveshape will give you a clue as to time required for solenoid travel.
     
  10. Dec 28, 2011 #9

    vk6kro

    User Avatar
    Science Advisor

    All digital cameras now do movie clips.

    I wonder if you could take a close-up movie of the contacts closing?

    60 frames per second gives a resolution of 16 mS. 30 gives 33 mS.
    Might be close enough.
     
  11. Dec 28, 2011 #10
    Is this a relay or contactor, or only a solenoid?
    Is this 60 hertz or DC operated?
    If relay, are you interested in when the contacts close or when they quit bouncing.
    If this is a relay and 60 hertz, it should be easy to make the measurement. If you are in the USA, there are probably viewers with home laboratories that would make the measurement if you are willing to pay the shipping there and back.
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook




Similar Discussions: Measuring delay for solenoids
  1. Time delay (Replies: 6)

  2. Delayed AGC (Replies: 5)

  3. Delay in encoders (Replies: 1)

Loading...