4 X 12 vdc signals into single pulse output, self re-setting.

Discussion in 'The Projects Forum' started by Keith Wells, Feb 26, 2016.

  1. Keith Wells

    Thread Starter New Member

    Dec 2, 2015
    5
    0
    Help..! I'm a pretty novice mech guy getting my hands into this dark art, so please excuse my fairly basic knowledge / ignorance......

    I have 4 individual 12 vdc signals coming from some NPN output proximity sensors. These 4 sensors form a "set" that I want to signal a single channel of my logger with. In order to record a sensor "event" I only need around a half second pulse to capture the event.
    My first challenge is to merge the 4 sensor signals into one output then the second is to convert each sensing event (which could be several minutes) into a single short pulse.

    The critical thing is that whilst one sensor may be active, I still need any of the other three to be able to signal an event (to the same logger input)

    The current required for the logger channel is very small (less than 100mA) so I am hoping an transistor solution is workable. I cannot use mechanical relays due to speed, reliability and noise issues.

    I have tried using a 555 timer, but it needs to be powered separately and only seems to trigger to ground, not by the application of the 12v line. They also need matched caps and pull up/down resistors and finish up having to build them onto a board.... I have 20 sets of these to complete.

    Does anyone have any ideas of a simple circuit to do do this? I need it to be as simple a solution as possible.
     
  2. #12

    Expert

    Nov 30, 2010
    16,298
    6,811
    This is called a summing junction. It can input as many inputs as you want, increase or decrease it's voltage gain, and the circuit on the output turns any sudden on or off event into a pulse. If the input is positive, the out-pulse will be negative. If this doesn't suit you, you can add an inverter stage to change it into a positive pulse. The time of the pulse is controlled by the output capacitor and resistor. Now that you have a signal which happens on either event at any number of inputs, you can process the pulses to the correct amplitude to measure them. Have I got you started or just went completely over your head?
     
  3. Keith Wells

    Thread Starter New Member

    Dec 2, 2015
    5
    0
    Hi #12,

    You have certainly pointed me in the right direction thanks..!! :) I didn't realise you could pair up op amp inputs in this way. The inverted signal is not an issue for me as my logger can record either + or - pulses.

    I'm going to have a play around with a couple of bits and pieces, it seems that the typical runnit resistor value to use is around 10K..?

    I'll give it a go..... :)
     
  4. #12

    Expert

    Nov 30, 2010
    16,298
    6,811
    I generally use 1 ma of current for my sensors because this pretty much makes them immune to noise. 12 volts...12K...that seems about right.
    The first problem with this circuit is that all four inputs going high makes the DC output level negative 4X the DC level of a single input. You have to reduce the feedback resistance to assure that the op-amp won't run out of voltage with 4 inputs all high at the same time.

    There is also the option to capacit the inputs so the output DC level stays at about zero DC and just flips pulses. That requires a capacitor for each input, so choose your poison. Decrease the DC gain or more parts required.
     
    Last edited: Feb 26, 2016
  5. Keith Wells

    Thread Starter New Member

    Dec 2, 2015
    5
    0
    Hi #12,

    Thanks for the advice.....

    I think it'll be almost impossible for all 4 to be outputting together but I can certainly see two or possibly three on the odd occasion being on together.

    I've just done a quick search and there seems to a bewildering array of op amps around..... any pointers as to preferences? I've found a couple of TI 6 to 18 V ones that should be OK they are actually 2A way more than I need but I can try them?
     
  6. #12

    Expert

    Nov 30, 2010
    16,298
    6,811
    This is such a simple circuit. Start with what you have or want for voltages. With a dual voltage supply, any op-amp ever made will center on zero volts and pulse either way. Even the 50 year old LM741 will do this job. So...any old POS will work as long as it can survive the voltages you want. Some of the new 5 volt chips won't work because they don't have the voltage range you need.
     
  7. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,542
    1,251
    To clarify something, are the four 12 V signals:

    1. Analog: the recorded voltage value is what is important, and can be anything from 0 V to 12 V)

    - or

    2. Digital: anything below 1 V is considered a logic 0, anything above 11 V is a logic 1, and what is important is the timing of the transitions between the two logic levels, not the actual DC voltages.

    Also, are the "NPN" outputs open collector, like the output of a tach sensor, and make a variable square wave?

    ak
     
  8. #12

    Expert

    Nov 30, 2010
    16,298
    6,811
    I assumed this was a digital event with very long time constants and the only relevant information is the moment of the event.
    Of course, it could probably be done in digital gates but I'm an analog person, so I worked with what I know.
     
  9. Keith Wells

    Thread Starter New Member

    Dec 2, 2015
    5
    0
    Yes it is a digital event. I just want to log a "present" or "not present" state across each of the 4 sensor beams. (I'm then manipulating the recorded data in Excel)

    You are correct about the time constants, I'm not concerned with how long each of the beam breaks (events) are, I just want to record the total number of state changes from each of the 4 sensors on one channel.
     
  10. Keith Wells

    Thread Starter New Member

    Dec 2, 2015
    5
    0
    Hi, Yes it is a digital event I'm trying to capture, the actual voltage is irrelevant to me, and you are correct in that like anything below 1 volt is "off"(I.e. my sensor beam area is clear) and anything above, say, 10 volts is "on". The issue I have is that when one beam detects presence,mother NPN open collector output signal 12 Volt correctly, but if another of the 4 sensors sees another break the same signal goes to the logger. I can make the logger sample really quickly, and scan a number of channels at many Hz, but I need to record every beams state of change each time it happens, and with one or more other sensors outputs giving an "on" 12 volt signal.

    The sensors are powered up and connected to my logger and they work beautifully one at a time, but if you close one beam the others obviously go "blind" with the same output line,menace wanting to split them and just recording a pulse.
     
Loading...