STRANGE LED PROBLEM-DISCO LIGHT

Discussion in 'The Projects Forum' started by dougalere, Apr 1, 2015.

  1. dougalere

    Thread Starter Member

    Mar 4, 2015
    128
    5
    hi,
    just got back to this repair after a spell in hospital, this is a circular display consisting of 7 seperate red, blue and green led arrays, the centre array was not lighting most of the blue and green leds,
    as the unit was oldish and the centre array does more work than the others.i put this down to age,
    so i labouriously replaced all the green and blue leds, and they all lit.and then watched in horror as they all failed one by one, or should i say three by three, as they are in chains of three.
    each chain is fed by a 13.8 volt supply through a 91 ohm resistor, now on all the other arrays this results in 9.5 volts at the leds, which of course is fine,
    but on the faulty centre array its reading 13.14volts to each chain of three, hence the demise of said leds.
    led spec 5mm


    3.2v – 3.8v


    465nm


    9000mcd

    for some reason, the new leds are not allowing the 91 ohm resistors to do their job?.
    please help
    doug
     
  2. Søren

    Senior Member

    Sep 2, 2006
    472
    28
    (13.8-13.14)/91 = 7.3mA, so the resistor works fine and the LEDs are only lightly fed.
    Check all tracks with a magnifying glass.
     
  3. dougalere

    Thread Starter Member

    Mar 4, 2015
    128
    5
    hi soeren
    thank you for your reply
    "(13.8-13.14)/91 = 7.3mA" ,can i be cheeky and ask what the voltage should be after the resistor.?-maths isnt my strong point.
    (all the leds on every array run very hot.i would guess this is normal

    found this. i will test for the forward voltage of my leds later, in the meantime i have assumed 3.3 volts fv and current of 20 ma for each led upload_2015-4-2_15-55-26.png
    regards
    doug
     
  4. dougalere

    Thread Starter Member

    Mar 4, 2015
    128
    5
    just a thought,the calculator suggests that the 91 r is way too low, but this value works fine on the other 6 displays, which total 30 chains of 3 leds.
    thanks again
    also the faulty leds comprise 8 seperate sets of 3 leds
    doug
     
    Last edited: Apr 2, 2015
  5. mcgyvr

    AAC Fanatic!

    Oct 15, 2009
    4,771
    971
    Garbage in = garbage out

    hot leds = short life
    wrong resistor = hot resistors
     
  6. dougalere

    Thread Starter Member

    Mar 4, 2015
    128
    5
    thanks mcgyvr
    "hot leds = short life
    wrong resistor = hot resistors"
    in general thats good advice, however the other 6 identical displays on the board are working fine, and have done for the last 10 years,with 91 ohm resistors.
    my comment on the hot leds might have been misleading, i failed to check wether the individual led bodies were getting hot, i just know that as i was working on the track side of the board, i was almost burning my hand on the board,
    i suppose this is to be expected with 300 mw per chain of 3,x20 chains of leds in a small area =over 6 watts.or 42 watts in total.
    it would be interesting to put this on a simulator,
    but as its so basic, i'll knock up an actual circuit.
    at the end of the day what matters is to get the thing working with the emphasis on safety,so i could put a 10 watt resistor in the 13.8 volt cable to the leds ?
    regards
    doug
     
  7. Søren

    Senior Member

    Sep 2, 2006
    472
    28
    You feed the chain with 13.8V
    You measure 13.14V after the resistor.
    Ergo the resistor sees a voltage drop of 0.66V
    0.66V over a resistance of 91 ohm equals 7.3mA (7.253mA to be exact).

    Can I be cheeky and say that with a 7.3mA drain, the voltage after the resistor should be... Tadaa... 13.14V ;)

    There's something wrong somewhere, that's clear as mud. 7.3mA will not hurt your LEDs!

    Measure if the 91 Ohm resistor is actually 91 Ohm.
    Only cause I can imagine is a deteriorated resistor which has somehow gotten on a diet, loosing Ohms, or that someone had mounted eg. a 9.1 Ohm (9R1) instead - with low light and/or tired and not-so-young eyes, this is not an impossible fault.

    So please measure the resistor with an Ohm-meter (without power going into the chain) and tell me who's wrong - the resistor or me :)
    And while you're at it - measure the current in the chain (with power on).
     
  8. dougalere

    Thread Starter Member

    Mar 4, 2015
    128
    5
    hi soeren
    thank you for getting back to me, me,i was getting worried,
    on the display thats causing problems, all the leds in the blue and the red circles are failing,thats a rotal of 7 chains of 3 and of course seven 91R sm resistors.
    the resistors all read fine,
    (the leds on the whole display, project 30 0r 40 feet, so they will be operating at high power)
    so to sum up, on the board there are 7 identical blue and green circular displays,so 49 chains of 3 leds, each with 7 identical resistors, so 49 resistors in all, 42 resistors on the good displays are dropping 4.4 volts, the other 7 resistors on the 'faulty' display are only dropping 2 volts -wierd!
    i will knock up a test chain using my leds and see what happens,in the meantime, any thoughts welcome
    thanks again
    doug
     
  9. dougalere

    Thread Starter Member

    Mar 4, 2015
    128
    5
    just another thought, bit of a long shot, could it be that the current through the new leds is lower (or higher)than the originals, giving a different voltage drop on the resistors.
    however that wouldnt explain why the original leds failed.
    and as i said the other 6 identical displays are working fine...
    it would make thing clearer if i could upload a pic of the led side of the board, but there might be copyright? issues, perhaps the moderator could advise.
     
    Last edited: Apr 3, 2015
  10. RamaD

    Active Member

    Dec 4, 2009
    254
    33
    On the good chains, the voltage on the leds is 9.5V.
    Drop on the 91 Ohms resistor is 13.8V - 9.5V = 4.3V.
    Current is 4.3V / 91 Ohms = 47mA.
    Not all 5mm LEDs are rated for this current. The datasheet I checked on digikey part from Cree has an abs. max rating of 25mA.
    Can you get a datasheet for your new LED? Check the value of the resistor in the working chains of blue or Green. Maybe, it is higher at around 220 Ohms, like your calculator, which would limit the current to 20mA. Also, power dissipiation on the resistor 91 Ohms with 4.3V across is around 200mW. The SM resistor looks too small.
    More likely, the resistor should be 220 Ohms or the working LEDs are higher rated.
    Also, Red LED forward voltage drop is around 2.2V only.
     
  11. dougalere

    Thread Starter Member

    Mar 4, 2015
    128
    5
    thanks RamaD
    yes, all the reds are working fine at 2.2 volts,

    i think you are spot on, i may need to get special leds ,rather than changing the ressistors.

    i will look into this and get back,

    thanks again
    doug
     
  12. dougalere

    Thread Starter Member

    Mar 4, 2015
    128
    5
    upload_2015-4-3_13-19-19.png
    just found these ,quite cheap,i suppose its whether the resistors will take them...
     
  13. mcgyvr

    AAC Fanatic!

    Oct 15, 2009
    4,771
    971
    "most" 5mm type LEDs will take 50mA.. Note though thats the absolute maximum rating and it WILL shorten the life of them to run them that high..

    Now the important question..
    Do you understand the math on how to size a resistor to provide the required current for a string of LEDs?

    For resistance needed
    Its
    (Vin - sum of LED VF)/current in amps = resistance

    For wattage of the resistors its
    (current in amps^2 x resistance) x 2 = wattage
    Note: the times 2 is a safety factor to ensure the resistor doesn't run too hot.

    Example (just an example)
    Vin = 12V
    sum of vf = 10V
    current = .02 Amps
    so resistance needed is
    (12-10)/.02 = 100 ohms
    wattage needed is
    (.02^2 x 100) x 2 = .08W (round up to next typical value so a 100 ohm 1/8W resistor)
     
  14. Søren

    Senior Member

    Sep 2, 2006
    472
    28
    OK

    That rules out the 7mA then.

    Why did the 0.66V turn into 2V???

    You didn't measure the current in the bad chain?
    Please do (we cannot troubleshoot via telepathy ;)) and in one of the good chains as well (break the connection between two LEDs and use the Ampere-meter to bridge the connection)

    Your advantage here is that you have both good and bad chains, so it's easy to compare and that should reveal the culprit in 10 minutes or less.
     
  15. dougalere

    Thread Starter Member

    Mar 4, 2015
    128
    5
    sorry soeren that is my next job,
    youve just reminded me of a gig years ago, with a full dancefloor, a female came up and asked me to put some good music on. ok i said, what would you like?
    oh i dont know, she replied. my reply was-not only do you want me to be the best dj in the world, you want me to be a mindreader as well,
    sorry about that
    regards
    dougal
     
    Søren likes this.
  16. dougalere

    Thread Starter Member

    Mar 4, 2015
    128
    5
    just checked the current on one of the good chains, at least 40ma, should have done that first, sorry...
    i cant be more specific at the moment,as the fuse on my milliamp scale has gone, (my next job,) so i have had to use the amp scale.
    quote "with low light and/or tired and not-so-young eyes," (i havent learned to do quotations yet- my next next job) i have an amazing aid,some time ago i bought some reading glasses, they were too powerful for reading, so i shoved them in a drawer and forgot about them.
    last year i lost my magnifier and stumbled on the reading glasses,i have never used a magnifier since,the glasses make a resistor look the size of a cucumber, with no distortion, i have both hand free, a true stereo image, ive no doubt there are dedicated similar things available, but these can be picked up at the local chemist for under a tenner ,
    sorry for rambling on.
    well it looks like this ones solved, my heartfelt thanks to all of you who gave up your time to help out.
    thank again
    doug/tibbles
     
  17. Søren

    Senior Member

    Sep 2, 2006
    472
    28
    Does that mean that you got it fixed?

    Beware of the glasses, I started out exactly as you, using a pair of +1.5 for soldering SMD's as a teen, but it's like drugs - someday you realize that you can't live without them (and going cold turkey doesn't help here) - I'm reding this screen with +3.5 (+2.5 would do in a pinch but gets me a headache after prolonged use) and with the naked eye I only see grey stripes on the screen. Back when I earned my living as a pro soldier, I had better than 20/20 and faboulous night vision as well - time's a bitch! ;)
     
  18. dougalere

    Thread Starter Member

    Mar 4, 2015
    128
    5
    EUREKA/ISH
    whats the worst that can happen?
    just put the new leds in( the 50ma 100ma peak ones)perfect,
    done all the green and most of the blue, miscalculated on the blue...been on for an hour,continously,
    bear in mind that in actual use they will go on and off.
    the new leds have a little heatsink on one of the legs.
    the snag is, the new leds look wonderful but they are taking 51.8 ma,and the drop accross the resistor is 5.4 volt
    the voltage accross the blues is 3.3 v
    nothing seems to be getting too hot.

    (as a matter of interest,confessing my ignorance now, am i right in thinking that each componant in a series drop voltage, but the current is the same for all?
    sounds a bit like a free lunch ,as in this case,there is 51 ma going into the chain,and yet 51 ma to each led.)
    i was going to p0st a similier question about batteries charged in series,am i right in thinking that if theres 200 ma at the beginning of the chain, each battery will still be getting 200 ma?

    ps just checked the current on the other arrays- 47 ma, i thought it was a lot lower, so
    it probably ok.
    sorry about that
    regards
    doug
     
    Last edited: Apr 10, 2015
  19. dougalere

    Thread Starter Member

    Mar 4, 2015
    128
    5
    thanks søren


    its a toss up what goes first, the eyes or the memory, and i'm only 21, i wish.
     
    Last edited: Apr 10, 2015
  20. RamaD

    Active Member

    Dec 4, 2009
    254
    33
    Right.
    Yes.
     
Loading...