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

Optocoupler not behaving as expected.

  1. Jun 12, 2009 #1
    Hi all,

    I am trying to use an optocoupler to isolate two circuits. Right now, I am trying to get a simple prototype going which uses a signal generator to drive the opto's input. On the output side, I have connected a resistor between the supply voltage (5V DC) and the output anode in the pull-down configuration. The signal generator is set up to generate a 4.5V square wave, which passes through a 330ohm resistor connected to the opto's input.

    So, as far as I understand optocouplers, one must bias the input side's anode with about 1.2V and 10mA (hence the presence of the 330ohm resistor). Whenever, the internal diode receives the right voltage and current, it should shine, making the optical switch connecting the output pins to start conducting.

    My problem, however, is that none of the optocouplers I have tried today have conducted during the on portion of the square wave from the signal generator. The voltage drop over the output pins remains at 5V. I have tried 10 optocouplers, all with the same result. I find it unlikely that they are all faulty. Its much more likely that I have missed something, I am not sure what though.

    The optocoupler I am using is an H11C6. I have attached the datasheet I am working from. Thanks very much for any assistance.
     

    Attached Files:

  2. jcsd
  3. Jun 12, 2009 #2

    cepheid

    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    Did you connect the gate of the output side SCR (pin 6) to the output side ground through a resistor? It looks like this is recommended in most of the application circuits shown. I don't know too much about SCR's but I think it's like a diode whose conductivity is controlled by a third terminal (the gate). I think the gate has to conduct if the anode and cathode are to be conductive.
     
    Last edited: Jun 12, 2009
  4. Jun 12, 2009 #3

    cepheid

    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    From Wikipedia article entitled Silicon-controlled rectifier:

    So this makes sense to me. I imagine that illuminating the gate creates electron-hole pairs, but in order for there to be a gate-cathode voltage, there has to be a voltage drop which is presumably across the external resistor connected between the gate and the cathode in the circuit. This external resistor gives those photoelectrons somewhere to go, resulting in a current across the resistor and a corresponding voltage drop across it.

    Hopefully somebody who knows more about these devices can step in and corroborate (or tear apart) my explanation.
     
  5. Jun 12, 2009 #4
    I have a 250kohm resistor between gate and ground.
     
  6. Jun 12, 2009 #5

    cepheid

    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    Can you attach a diagram showing exactly how it's wired up?

    What's the frequency of your square wave? Is it possible that the ON cycle has a minimum required duration for some reason?

    Can you just try the optocoupler with a constant HIGH input to see if the output responds (switches) as expected?
     
  7. Jun 12, 2009 #6
    I have tried a constant high input, but there is no change.

    Edit: Circuit diagram is on it's way.
     
  8. Jun 12, 2009 #7
    Here is my circuit. Please disregard the fact that it says I am using a 4N25 optocoupler. I took that block because it is the closest I have in the circuit application. V1 represents the signal generator, while V2 is my 5V DC power supply. I am measuring the voltage drop between collector and emitter. R2 represents my load, which should end up having close to a 5V drop over it, but doesn't because the optocoupler doesn't conduct.

    Thanks for the interest.
     

    Attached Files:

  9. Jun 12, 2009 #8
    It's difficult to say what the problem is without a schematic. The resistor values you've chosen make the SCR sensitive to turn on. The SCR once turned on will not turn off without removing the forward voltage.

    You could have a miswire.

    If the SCR is just triggered on, and not off, you could be blowing out the SCRs by exceeding the 400 mW power disipation or 300 mA current rating.
     
  10. Jun 12, 2009 #9

    cepheid

    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    I don't seen anything immediately wrong with it, assuming the mapping from this circuit to your circuit is:

    collector --> anode
    emitter --> cathode
    base --> gate

    I'm no expert on SCRs though. I did notice your pull-down resistor on your gate is much larger than those shown on the application circuits (by an order of magnitude). How did you arrive at the values for your output resistors? Edit: Why is the load resistor that big?

    By the way, the two grounds you've depicted (input side and output side) are isolated, right? Like you're using two different voltage supplies or something? Just checking!
     
  11. Jun 12, 2009 #10
    I am using two different power supplies - one is a signal generator, the other is a separate DC power supply. Technically, they aren't completely isolated since they both run of the same wall socket, but otherwise, there is not connection between the two supplies.

    As for the resistance, it is actually an arbitrary choice based on what limited supplies I have in my lab. I'm going to try connecting a few in parallel to get smaller values and see what happens.
     
  12. Jun 12, 2009 #11
    OK. nevermind all that. 900K seems awfully large.


    Try 100 ohms.
     
    Last edited: Jun 12, 2009
  13. Jun 12, 2009 #12
    There are a few things that could be keeping your circuit from operating properly. First, looking at the application examples on the data sheet it appears that this device is designed to operate from a fairly high anode voltage (typically 50 V or higher). For an anode current of 100 mA, the data sheet gives the on-state Vak as being fairly low (1 volt or so), which would lead you to believe that a 5 V supply (for the SCR) would work ok. However, generally it takes a much higher Vak to initiate turn-on of an SCR. You simply might not have a high enough Vak to initially turn on the SCR.

    Keep in mind that once an SCR turns on, it remains latched on. Turing off the LED will not turn off the SCR. In order to turn the SCR off, you must reduce the anode current to less than the holding current, Ih. Based on your gate resistor being 250 k, the holding current would be around 10 uA or less. Most often SCRs are used to control loads that are driven from an AC source (see the sample circuits on the data sheet), which automatically turns off the SCR on negative-going zero crossings of the line voltage (technically, it's called autocommutation).

    Hope this helps.
     
  14. Jun 12, 2009 #13
    I jsut looked at your schematic. I don't believe this circuit could possibly work with a 900 k load. The anode current would never even reach the holding current level.
     
  15. Jun 12, 2009 #14
    Time for some updates. I have decreased the resistor on the base to 110kohm and the load resistor to 2.7kohm. At first, that didn't change anything. Then I accidentally disconnected the probe ground and reconnected it. Now there is a nice 4V drop over the load. If I switch everything off and on again, the optocoupler doesn't conduct again, until I disconnect the probe ground and reconnect it. I observed on the oscilloscope that, when I disconnected the ground, the voltage becomes a sine wave with 60V amplitude and frequency of 50Hz. Weird.

    I think what I have realised is that this is the wrong chip for the job.
     
  16. Jun 12, 2009 #15

    Averagesupernova

    User Avatar
    Gold Member

    That chip is typically used for controlling line voltage. So if you are not controlling AC it will not likely work as you intend.
     
  17. Jun 12, 2009 #16
    What I actually want to do is to isolate a PIC microcontroller from a hobby servo fitted to a model helicopter. Having looked more closely at Texas Instruments website as well as a local electronics store's site, it seems what I actually want is some thing like a 3.3V/5V digital isolator or an optocoupler like the H11LX series from Fairchild. Does anyone have some comment about the revised choice of isolation - am I going in the right direction now?

    Thanks everyone for the help. Incidentally, the reason why I used these ones is I had a bunch lying around from a previous practical and thought that one optocoupler would work pretty much like another. Well, thanks for correcting me on that one.
     
  18. Jun 12, 2009 #17

    Averagesupernova

    User Avatar
    Gold Member

    I don't have any experience (that I know of specifically) with the numbers you gave. I think you want to stay in the small signal category of opto-isolators/opto-couplers. Stay away from anything with an SCR or TRIAC for the output. Have a look at the digi-key web-site. www.digikey.com. They have a pretty good search engine. Although I find myself picking up the catalog while browsing their site.
     
    Last edited: Jun 12, 2009
  19. Jun 12, 2009 #18
    Thanks very much Averagesupernova. I'll have a look.
     
  20. Jun 12, 2009 #19

    cepheid

    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    I used an optocoupler / solid state relay in my last project that you might want to check out. It's a Panasonic "GU PhotoMOS" solid state relay whose part number is AQV212 (or similar). As the name suggests, it has MOSFET outputs.
     
  21. Jun 12, 2009 #20
    Thanks cepheid, I will consider that one. First things first, I must see what is available in this country (not America unfortunately). At least now I know what type of equipment I am looking for.
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook




Similar Discussions: Optocoupler not behaving as expected.
  1. MOSFET vs Optocoupler (Replies: 10)

  2. Optocoupler turn ON (Replies: 1)

Loading...