Safety & reliability: Cap as current-limiting device.

Discussion in 'General Electronics Chat' started by ebeowulf17, Dec 5, 2017.

  1. ebeowulf17

    Thread Starter Well-Known Member

    Aug 12, 2014
    1,335
    217
    So, I'm thinking of making some solid state opto isolator modules for switching a 5V signal based on 120-240VAC input. I know this can be done with a small relay, but unless this is a really horrible idea, let's stick with the opto idea for now. Here's the current incarnation of my plan:
    IMG_3771.PNG
    I simmed with two separate optos because LTspice didn't have an AC input (two emitter LED) model, but my plan would be to use a single opto suitable for AC.

    The question is about input current limiting. It seems I can get the LED current where I need it using just the impedance (react
     
  2. ebeowulf17

    Thread Starter Well-Known Member

    Aug 12, 2014
    1,335
    217
    Wow, my phone is losing its mind. Post got cut off and I can't keep the editor window open to finish it.

    Anyway, the impedance (reactance?) of the caps seems to get the job done, but I often see a resistor in series with current limiting caps like this. What are the pros and cons? Should I put a real resistor where my just-for-simulation resistor is? I only hesitate because if it contributes any meaningful amount of resistance it will need to be high wattage (large and expensive,) so I'd rather not add it if it's not helpful.

    As a side note, I haven't picked specific caps yet, but I'm aware of the need to choose for AC and high voltage. I seem to remember sourcing film caps rated 600VAC once before for a similar application.
     
  3. AlbertHall

    Distinguished Member

    Jun 4, 2014
    4,759
    1,111
    You do need a resistor. Consider what happens if the capacitors are discharged and you switch on at the peak of the input voltage. A very large surge current will flow through the LED and the capacitor may not be too happy either.

    The capacitors should be 'X' rated - suitable for continuous connection across the mains.
     
    Sensacell likes this.
  4. ebeowulf17

    Thread Starter Well-Known Member

    Aug 12, 2014
    1,335
    217
    Thanks!

    Ok, I think I understand the problem - I calculated based on 60Hz impedance, but switching could create high frequency voltage components which would pass much higher currents because the impedance (reactance?) drops as frequency increases.

    So then, any advice on choosing a suitable resistor value? I tried reading the datasheets for two optos I was considering, but neither gives any indication of an acceptable momentary or single-cycle current/wattage limit like some devices do. How do you know how much current limiting you need in this situation?
     
  5. AlbertHall

    Distinguished Member

    Jun 4, 2014
    4,759
    1,111
    I looked at the 4N35 and in absolute max ratings it says 1A for <10uS, max continuous 50mA, so that would give some sort of a starting point.
     
    ebeowulf17 likes this.
  6. ebeowulf17

    Thread Starter Well-Known Member

    Aug 12, 2014
    1,335
    217
    Guess I should've checked more datasheets before giving up. Sorry about that, and thanks once again!
     
  7. ebeowulf17

    Thread Starter Well-Known Member

    Aug 12, 2014
    1,335
    217
    Well, I'm still sort of guessing, but I figured the surge could easily last more than 10uS, so I chose a lower peak surge current limit of ~110mA, which is about double the continuous current rating for most of the optos I'm looking at. Based on that, I've come up with the following.

    I'm feeling relatively confident about this setup now. Thanks for your guidance! Let me know if anything in the schematic or notes below doesn't look appropriate.

    ***EDIT: Oops! I compiled the image below from two different screenshots taken at different times. The reference designators on the wattage and current traces are wrong. Magenta is R2 and Cyan is R1.
    Vulcanus_opto_07.png
     
    Last edited: Dec 7, 2017
  8. DickCappels

    Moderator

    Aug 21, 2008
    4,258
    1,230
    Make the 2.2k a fuse resistor or put a low current fuse in series with it and you won't have to worry as much about fire and smoke.

    For the capacitors, check the manufacturer's datasheet to see whether they say anything about whether the capacitors can be use in series with the AC Line (some X and Y caps carry a warning about this).
     
    ebeowulf17 likes this.
  9. Sensacell

    Senior Member

    Jun 19, 2012
    1,565
    420
    Looks like a good design.

    R1 can be bigger, why burn 1/2 watt?
     
  10. ebeowulf17

    Thread Starter Well-Known Member

    Aug 12, 2014
    1,335
    217
    Thanks for both of the warnings! I'll keep them in mind, both for now and later (I want to remember everything I learn, but especially the stuff that involves safety!)
     
  11. ebeowulf17

    Thread Starter Well-Known Member

    Aug 12, 2014
    1,335
    217
    Thanks!

    R1 discharges through the opto emitter LED when the switch is turned off (unless it happens to turn off at a zero crossing.) So, the larger I make R1, the longer it takes for the output to turn off.

    I'm not sure how time critical I need this to be - responding within 0.1s is definitely good enough for me, but I'm not sure if I'd be equally happy with 0.25 or 0.5s. I'm definitely sure that one second lag is more than I'll accept. So, it's a balancing act; I could maybe increase R1 a little, but not far enough to completely get away from heat dissipation concerns.

    *** EDIT:
    Come to think of it, could I put R1 across the capacitors (in parallel with them,) instead of across line to neutral? Then the discharge path wouldn't go through the opto and I could choose a much longer time constant.

    I wonder if I'm overlooking anything there. I'll have to give that some more thought.
     
    Last edited: Dec 7, 2017
    Sensacell likes this.
  12. DickCappels

    Moderator

    Aug 21, 2008
    4,258
    1,230
    I think putting R1 across the capacitors is the preferred position.
     
    ebeowulf17 likes this.
  13. ebeowulf17

    Thread Starter Well-Known Member

    Aug 12, 2014
    1,335
    217
    Ok, latest update. I think I'm pretty close now...

    I've moved R1 and changed its value to 1Meg so that it can comfortably be 1/8W. This of course also dramatically improves the response time of the optos when switching off. The caps may take a second (or a few) to fully discharge, but I have no problem with that - I just didn't want them lingering at high voltage for minutes or hours. One less thing to worry about when servicing the machine this is being built for.

    I've also updated some of the notes to reflect this change, and the capacitor safety notes so I don't forget them later when I'm shopping for parts.

    Vulcanus_opto_08.png
     
  14. Danko

    Member

    Nov 22, 2017
    98
    16
    Diodes D1...D4 any low voltage.
    opto-1.png
    opto-2.png
    Point 13 on schematic diagram means point 2.
    Edit: Ground on the mains is for simulation only.
    Because we need to see voltage on U1 LED.
     
    Last edited: Dec 7, 2017
  15. AlbertHall

    Distinguished Member

    Jun 4, 2014
    4,759
    1,111
  16. ian field

    AAC Fanatic!

    Oct 27, 2012
    5,711
    1,013
    Some telecoms opto isolators have dual (bidirectional) IRLED.
     
  17. Danko

    Member

    Nov 22, 2017
    98
    16
    Nothing bad! This ground on the mains is for simulation only.
    Because I need to see voltage on U1 LED.
     
  18. Bernard

    AAC Fanatic!

    Aug 7, 2008
    4,625
    497
    As for reliability, my AAC illegal line operated night ( & day ) LED light has been on for 10 years, 24-7. Like
    post # 7 but with a bridge rectifier & white LED.
     
  19. Danko

    Member

    Nov 22, 2017
    98
    16
    "The change of the CTR over operating time is mainly caused by a drop in the luminous efficiency of the LED. In general, the larger the LED input current (IF) and the higher the ambient temperature, the faster the CTR decreases."
    https://www.renesas.com/en-eu/products/optoelectronics/technology/ctr.html
     
  20. ebeowulf17

    Thread Starter Well-Known Member

    Aug 12, 2014
    1,335
    217
    Looks like a good setup to me.

    That said, I don't see any particular advantage in using a bridge rectifier feeding a standard opto vs. just using an AC opto instead.
     
Loading...