Actuator and digital counter issues! Fix what 20 engineers failed at.

Discussion in 'The Projects Forum' started by Matt-Halliburton, Aug 5, 2012.

  1. Matt-Halliburton

    Thread Starter New Member

    Aug 5, 2012
    11
    2
    Well if someone figures this seemingly-retarded issue, ill try and scroung some presents to ship.

    No joke a dozen engineers have been working on this.


    Counter not in sync, randomly generates a number???!


    [​IMG]

    link http://i680.photobucket.com/albums/vv167/mattr11992/Work/Schematic.jpg


    It takes a year for this company to design and make and ship this item so i need to have this fixed, out of the box its not working.

    What has been done so far...

    isolated and individually tested all componontent for functionality
    (theres not much components as you can tell)

    still stumped figured it was wired wrong, tested and traced all wires. Nothing came up faulty, Then i got wise and decided to draw up a schem and just rewire it,,, was the same schematic essentially.

    WHAT IT DOES

    once u push the button it activates the actuator, which pushes a ball which drops on the lever/switch which counts and displays on the screen.

    Its a very essential tool in Fracing process in the oil fields... pumps a ball down hole which is pushed into a crack in the shale formation that holds it open.

    *the switch is not electrically connected to the counter.

    THE ISSUE

    when you press the button, it randomly counts and displays a number. If you push it, it will randomly count a number from 1-7 and add to the display?????

    figured there was backfeed from something but cant figure it out. Someone suggested a loose ground or some ground issues?


    Here are some more pics to show what we're working with, all ripped apart for testing and what not.

    This is the main unit
    [​IMG]

    This is the lever that trips the counter
    [​IMG]

    This is the Display You can see the activation switch on the left half of the bottom.

    [​IMG]

    The display has the digital flip flop actually built into it on the back of the unit. This is what we figured was bad and have another on order or spair.

    [​IMG]




    photobucket.com/Matt_Halliburton has all the pics

    Any ideas how to get the unit to accurately count would be great, always on the comp will try to answer everything soon!
     
  2. JohnInTX

    Moderator

    Jun 26, 2012
    2,348
    1,029
    Since the switch/actuator circuit wires lie alongside the counter wires for some distance, you are likely coupling the current spikes created by the bouncing actuator switch (presumably driving some decent current based on the 5A fuze and with a healthy di/dt because of the mechanical bouncing of the switch). These may be picked up easily by the digital counter.

    Try twisting the two switch / actuator wires together in one bundle and the two counter input wires together in another. Twisting helps cancel the magnetic field generated by the currents in the wire.

    Some low pass filtering on the counter inputs may be in order as well or at least loading the input with a resistor to make it less sensitive.

    If the actuator is DC, be sure it has a suppression diode or diode/zener network across its leads. A snubber (arc-suppressor) across the switch contacts might also be in order (you want to make the current in the actuator circuit slow-rise and smooth to minimize coupling).

    Don't forget that the ball counter switch should have some hysteresis to avoid multiple clicks from a bouncing ball..
     
    Last edited: Aug 5, 2012
  3. strantor

    AAC Fanatic!

    Oct 3, 2010
    4,302
    1,988
    That's what I was going to say; "what's being done to debounce the contacts"? but I'm more concerned about the contacts themselves bouncing and not the ball.
     
  4. JohnInTX

    Moderator

    Jun 26, 2012
    2,348
    1,029
    Usually not a problem compared to the necessity of debouncing but it can be an issue in some cases. I've worked in the vending/coin-op industry and one case involved a coin acceptor that would double-vend if you snapped the quarter hard into the slot just right. It would sail through the gates, make the light-force lever switch, bounce off the back wall of the ramp, unmake the switch then roll forward slowly and make the switch again as it fell into the coinbox. A switch with more make/break hysteresis fixed it.
     
  5. BMorse

    Senior Member

    Sep 26, 2009
    2,675
    234
    I would agree with Strantor, by description of the OP, the ball seems to just fall by the switch, so a switch debouncing circuit will most likely fix the issue, especially if the user can manually press the switch and get multiple counts from it (as the OP stated in his original post), so if the issue is persistent with out the ball, then switch is definitely causing issue as mechanical switches always seem to cause this type of issue with digital circuits, adding a capacitor/resistor pair to the counter to stabilize input may alleviate this problem......
     
  6. JohnInTX

    Moderator

    Jun 26, 2012
    2,348
    1,029
    .. but I thought I'd throw it out there anyway.

    Back to the OP. On the premise that coupling between the actuator circuit and the sensitive counter input is to blame, if the switch was debounced, I'm thinking you would still get one count when it was closed and/or released. Its still coupled.

    But looking at the skiz again, what's at the dot at the 14 between the cables on the wiring detail? Hard to read but is that a battery common connection between the actuator and counter circuits? That would provide a handy reference for any induced signals to the counter.
     
  7. Matt-Halliburton

    Thread Starter New Member

    Aug 5, 2012
    11
    2
    im gunna start tackling the ideas discussed



    To clearify something, the unit is not installed its being tested on the bench. no balls are being dropped onto the lever.

    the actuator pushes/loads a ball into a tube and hits the lever as it falls through the tube. The counter records the data so you know how much have been pumped down hole on each zone you frac.

    I hooked up a fluke oscilliscope and do notice some spiking i will try twisting and seperating them, i also noticed it was picking up **** from some communications equipment in the room... ???
     
  8. BMorse

    Senior Member

    Sep 26, 2009
    2,675
    234
    Spiking when the switch is pressed?? or with the unit just sitting idle??

    The **** you are picking up may be solved by twisting your wires as JohnInTX mentioned....
     
  9. Matt-Halliburton

    Thread Starter New Member

    Aug 5, 2012
    11
    2
    Also to note the counter has a debouncing circuit built into it.
     
  10. bertus

    Administrator

    Apr 5, 2008
    15,652
    2,348
    Hello,

    Are there RF signals in the surroundings?
    If so a shielded cable would help.
    Also ferrite beads around the cables will help in that case.

    bertus
     
  11. Matt-Halliburton

    Thread Starter New Member

    Aug 5, 2012
    11
    2
    I take it these counters are very sensitive?
     
  12. JohnInTX

    Moderator

    Jun 26, 2012
    2,348
    1,029
    From the photo its rated at 28V/1.5ma which is about 18.6K. That's not too much but any induced currents will flow through the count input. A cheap and dirty thing to try is to load the count input to ground with a resistor to make it less sensitive. Put 1K across it and see what happens. You could parallel the R with a 10uf capacitor as well and add a little series resistance to boot. (Just picked the values out of the air BTW, scope it to see what's going on) but you get the idea. See how long the real ball-count signal is, give yourself some cush then roll the heck off of the signal.

    If you can't keep energy from getting into the circuit, the idea is to dissipate it outside the counter, leaving enough at the end to run the count input. The components should be very close to the count and GND inputs of the counter itself using very short leads.
     
  13. Matt-Halliburton

    Thread Starter New Member

    Aug 5, 2012
    11
    2
    Well i rewired the entire thing with heavy industrial 4 wire cable, 2 twisted pairs heavily shielded with metal. still same prob.


    its got to be some sort of counter issue?

    If you forget about the lever, and actuator. and if u just hit the switch it will count 2-5 or whatever and add to the current number. which you shouldnt even be counting since the levers not activated..
     
  14. cork_ie

    Member

    Oct 8, 2011
    348
    58
    Surely the obvious thing to do is to connect an oscilloscope at various points in the circuitand try and see exactly what might be causing the false triggering. As previous posters have already indicated a schmidt trigger to debounce the switch would be a help
     
  15. BMorse

    Senior Member

    Sep 26, 2009
    2,675
    234
    Even though the counter has a debounce circuit, it may not be "long" enough to counter react the "noise" from the mechanical switch. . . Do you have a model/part number for that counter??
     
  16. JohnInTX

    Moderator

    Jun 26, 2012
    2,348
    1,029
    Which switch, the one operated by the ball counter or the one that runs the acutator? If its the former, its a debouncing problem. If the latter, its a decoupling problem.

    Agreed. Time to see what's happening right at the count input under various switch actions. Compare to the debounce time of the counter (or max count speed) as Bmorse suggests.

    I'm not sure a schmidt trigger will do it. They sharpen up slow-transitions. That doesn't seem to be the problem here.
     
    Last edited: Aug 5, 2012
  17. Matt-Halliburton

    Thread Starter New Member

    Aug 5, 2012
    11
    2
    John, the main actuator switch is triggering the count, which its not supposed to.

    I do have a scope hooked up i thought it was so obvious not to meantion.


    I put a capacitor to ground on the counter wire, it now counts only one at a time, yet it still counts one when the switch is pushed.. this is where im currently at.


    a resister to ground on this wire counts the actuator switch not the lever.

    now im not sure if im making progress with that.
    Im assuming that the counter is picking up the spike from the switch activating the actuator.
     
  18. Matt-Halliburton

    Thread Starter New Member

    Aug 5, 2012
    11
    2
    So i dont think the counter has speed issues since it is counting in correct intervals its just not supposed to be counting the actuator switch which isnt connected to the counter.

    Maybe i can try having seperate power sources for the actuator and the counter?????
     
  19. JohnInTX

    Moderator

    Jun 26, 2012
    2,348
    1,029
    As I indicated in #6, debouncing (which you have achieved with the cap) will solve the problem of many contact bounces coupling into the counter input but will still leave one. This implies coupling between the two circuits. Does it count when the switch is pushed or released?

    Me too. Do you have a diode across the (assumed DC) actuator terminals at the actuator?

    Presumably, flipping the ball lever generates the correct count? Ok, I see that you said that it did..

    Which wire is that?

    How hard would that be to try out? It would answer some questions. EDIT ** Probably can fix this without doing that, though **

    Since you are driving the counter with slow, ugly, low impedence, sparky switches and things, maybe a slow, ugly, low impedence, sparky electromechanical counter?
     
    Last edited: Aug 5, 2012
  20. BMorse

    Senior Member

    Sep 26, 2009
    2,675
    234
    Try a decoupling capacitor across the counters power connection, this should reduce the "noise" caused by the actuators switch (Maybe. Its hard to say without actually being there to see whats what.)
     
Loading...