MOTION ALARM FOR LOW SPEED PUMP - suggestions?

Discussion in 'The Projects Forum' started by PILOT203, Sep 9, 2010.

  1. PILOT203

    Thread Starter New Member

    Sep 1, 2010
    9
    0
    I have similar project where I am trying to detect if a pump has stop or is in motion.
    I am running a pneumatic pump which operate from 0.5 to 250 HZ. The sensor feedback is NPN proximity switch.
    One of the problem I had is that when the pump stops, it can stop on the prox-switch and give a high (+Vd) or NOT on the prox-switch which gives a low signal (0V grounded).

    I seemed to solve this problem by using two 555 timers in parallel ( inverted one of the 555 input), then ANDing the outputs of the timers. This will give me High output if only if both timers outputs are high. This solves the problem if the pump is stopped on the with it on or off the prox-switch.
    This is basically two of the missing-pulse detector parallel then "Anding" the output of two 555's (one 556), the trouble starts at the pulse width threshold, this is the at low speed operation (~<1hz). When the input pulses are below the 1-2 hz, the output of the AND gate will oscillate when it drops below the threshold frequency.

    The output will be use to signal a computer that the pump is running. Their are some off the shelf frequency counters that have alarm outputs but are in the $300-500 range.

    I'll have to interface this with a 24VDC control circuit, so I'll have to use a regulator to get it down to 15VDC. It draws about 40-50ma. I would like to fit this in a 22mm x 22mm x 50mm space.

    Anyone have a suggestions for a simpler circuit.
     
    Last edited: Sep 9, 2010
  2. beenthere

    Retired Moderator

    Apr 20, 2004
    15,815
    282
    Welcome to the forums.

    It's lots easier to see what's going on if you start a new thread rather than hijack an old one - http://forum.allaboutcircuits.com/showthread.php?t=41215

    For clarification, would you verify that you measure the pump rotation in terms of frequency rather than RPM? -
     
    PILOT203 likes this.
  3. Kermit2

    AAC Fanatic!

    Feb 5, 2010
    3,761
    924
    Sounds pretty complicated for a circuit to tell if a motor stops or not.

    I presume this pump is running off of a power supply? Ac/Dc or something using both that goes to a motor?

    Why not sense the current used by the motor to tell when it stops running?

    Just asking, maybe it will give you an idea for a simpler direction to go in...;)
     
    PILOT203 likes this.
  4. hgmjr

    Moderator

    Jan 28, 2005
    9,030
    214
    I am interested in your reply to Kermit2's suggestion.

    hgmjr
     
  5. windoze killa

    AAC Fanatic!

    Feb 23, 2006
    605
    24
    How about a photo interupter (slotted wheel and sensor) on the shaft. When you stop getting pulses its stopped. You can also use it to calculate the speed of the shaft.
     
  6. PILOT203

    Thread Starter New Member

    Sep 1, 2010
    9
    0
    The pump is pneumatic (air-powered diaphragm). A remote computer system (PLC) energizes a solenoid which supplies compressed air to the pump actuator. The acuator causes the pump to recipicate via series of check valves. As the pump oscillates in a "back and forth manner" (recipicating), its speed is dependent on the amount fluid it is pumping, therefore the speed at which it operates can vary from .5 to 250 cycles/second. The speed or cycles/second is sensed by a prox-switch mounted on one side of the pump (no rotating shaft). If viewed on an osciliscope the feedback (switch) looks an asymetrical square-wave (0 to 15V). The +pulses are a minimum of 25ms up to 50ms.

    At this time, I don't have a choice on which type of sensor I can use. For now I'll have to stick with the prox-switch (three-wire, npn output, you could treat this like regular switch). It seems overkill for simple pump, but it is critical that we know if the pump is moving or stopped.

    p.s. sorry for hijacking that last thread, I'm a "new-be". I'll try to correct my forum "faux pas" in the future and thanks again for your replies and suggestions.
     
  7. marshallf3

    Well-Known Member

    Jul 26, 2010
    2,358
    201
    To eliminate the "dead spot" problem you'll simply need to differentiate between a steady DC output and that of one that's slowly changing state. Shouldn't be too hard to come up with a circuit for this, even a simple RC time constant could be employed.
     
  8. KMoffett

    AAC Fanatic!

    Dec 19, 2007
    2,574
    230
    Here's an AC coupled edge-detector/missing pulse detector 555 circuit. This should handle the fixed Hi or Lo levels from the proximity sensor at stopped positions, and the very low repetition rate at slow speed pump speeds.

    Ken
     
    Last edited: Sep 10, 2010
    PILOT203 likes this.
  9. PILOT203

    Thread Starter New Member

    Sep 1, 2010
    9
    0
    Here is the circuit I've been working on. See attached PDF (50KB). Note, I am currently using transistors for the inverters and the AND gate. some details missing.
     
  10. GetDeviceInfo

    Senior Member

    Jun 7, 2009
    1,571
    230
    If your PLC has an unused input, tie your prox into a watchdog timer. This gives you many more options in responding to a fault condition.
     
  11. PILOT203

    Thread Starter New Member

    Sep 1, 2010
    9
    0
    RE: GetDeviceInfo

    If I had control over what the PLC does, I would have easily programmed the PLC discrete input to look to see if it was stopped or moving.

    However, the programmers (off-shore engineers) have written their code (motor control templates) to look for a "OK" signals from the motor controller within a few seconds after it is commanded to run. However, the programmed templates are looking for a steady-state input if the pump is running. It the input cycles on and off, as it does, then the PLC will think that their is a pump malfunction

    The main object is to install an low cost hardware solution without altering the program templates.
     
  12. windoze killa

    AAC Fanatic!

    Feb 23, 2006
    605
    24
    You just need a 555 timer circuit that will stay in an output state (low or high, which ever you need) so long as it receives a pulse within a certain time. If the pump stops (no more pulses) for say 2 seconds the timers output state changes and will activate an alarm or what ever else you need to operate. I did something very similar for someone else on here a while ago.
     
  13. windoze killa

    AAC Fanatic!

    Feb 23, 2006
    605
    24
    Think it is called a missing pulse detector. Or a variation of that.
     
  14. GetDeviceInfo

    Senior Member

    Jun 7, 2009
    1,571
    230
    I thought you where just turning on an air valve solenoid
     
  15. PILOT203

    Thread Starter New Member

    Sep 1, 2010
    9
    0
    The programmers assume that all pumps are motor driven, whether they are rotory or recipicating pneumatic via a solenoid. I guess you could call the them a "pump driver template". These pre-programmed driver templates generate command signals (discrete or analog) and look for feedback to ensure a complete control loop.

    I hope this clear up any confusion.
     
  16. GetDeviceInfo

    Senior Member

    Jun 7, 2009
    1,571
    230
    Gotcha.

    Why look specifically at the pump's mechanics? Many industrial applications would look for power at the solenoid, or downstream pressure created by the pump.
     
    PILOT203 likes this.
  17. PILOT203

    Thread Starter New Member

    Sep 1, 2010
    9
    0
    This particular pump was shiped from the manufacture with a prox-switch mounted on the side of the pump to pick-up the diaphragm movement. When the pump is running (recipicating) the proximity switch picks up one half of the stroke cycle. The pump may be pumping a liquid or air. 1. This type of pump causes pressure surges, so a pressure switch is not practical in this application, but I have seen this used for other reasons. 2. Looking at the solenoid power would not be very helpful because its always on when the pump is running. Just because the solenoid is energized does not mean it is moving.

    here is the pump spec...notice the threaded hole on the side of the pump for the prox.
    http://www.yamadapump.com/specialty/specs_split_manifold.html

    UPDATE:
    I did find a solution to the proximity switch input. I used a simple switched capacitor input and comparator circuit for the output to the PLC. I'm testing it now. I hope to upload a schematic after I've tested it.

    That again for all the help.
     
  18. GetDeviceInfo

    Senior Member

    Jun 7, 2009
    1,571
    230
    The Grayhills version is relatively common in industry. I've found a couple of deviations in operation of these units, one being that the pump will stall against sufficient back pressure, and the second being that the check balls may not seat due to contaminents.

    it sounds like you have found a suitable solution.
     
Loading...