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

Struggle to fix a 36W dual channel LED driver

  1. Jul 11, 2017 #1
    Good morning and my compliment for their infinite knowledge contained in this forum.

    I'm struggling to keep being my girlfriend's hero.

    I have here a nail led 36w UV led lamp that failed last week.

    Bulbs are tested all ok and connected in two channels each channel in series of 9 LEDs of 2w

    Driver is a smart one, it includes a LCD display and a timer that can be setted for auto shutdown after lit up 5sec 20 sec or 30 sec (each time have a separate button to activate)

    To trigger the light there is a photocell like sensor which leads i measured during power on are around 5v.

    At triac i find 12v.

    The main power supply of the board is DC 12v.

    All powers up fine but when i blind the photocell it try to lit up LEDs with no luck (i just see a short message on LCD ). If i try second time neither the message come up i need to unplug and plug again power supply to replicate the behaviour.

    Change the two capacitors didn't help . Did today.

    I inspected the board again and i found a 1200 smd resistor slightly browned. Unsoldered it and measured firm on 120ohm (r14)
    The tracks beneath it looks well warmed up but i checked continuity and are ok , and no shorts are found.
    Picture attached
    I'm lost....

    Attached Files:

    Last edited by a moderator: Jul 11, 2017
  2. jcsd
  3. Jul 11, 2017 #2


    User Avatar

    Staff: Mentor

    Welcome to the PF. :smile:
    How did you test the LEDs? Did you use the diode check setting on your DVM to check the forward voltage under small current bias, and found around 2V? What color are these LEDs? Do you have access to them without any built-in resistors?
    What triac? Triacs usually are part of AC Mains circuits, not low-voltage 12V LED driver circuits.
    Did you use your DVM to measure the overall voltage driven to the LED strings when this happens?
    What 2 capacitors? What motivated you to change them?

    Can you post the model number of the driver and LED lights, and links to any other technical information you have found? Thanks. :smile:
  4. Jul 12, 2017 #3
    LEDs have been tested with another led driver which differs from the original one. One faulty led lamp have been found faulty in the serial chain of one bank and have been replaced.

    LEDs are 2w RMS? And in UV wavelength.

    When i said triac i should have referred to single n channel MOSFET. Attached data sheet.

    This one i didn't test yet....I don't know how to test one of these...


    I checked the 3 diodes and found one (d6) that shows 0v on DVM on both directions. I tested it in circuit so far. This diode links positive pads on the capacitor area to the driver output to led banks.
    I don't know how to replace this diode will any diode work the same? Attached data sheet

    The other 2 diodes are zener and if tested in circuit with DVM diode mode the show regular diode values. Should i test them differently?

    Attached Files:

    Last edited by a moderator: Jul 12, 2017
  5. Jul 12, 2017 #4

    I replaced the D6 with a generic 3 amp diode (voltages specs are far more higher compared to the original one data sheet but i don't think it's a problem) code is 1na5408 see attached datasheet. I'm not quite sure does the same job of the smd one ....

    Replaced the smd 120ohm smd with a generic one (not smd) however the original one is not faulty as measure correctly.

    Replace back the caps with the stock one as they were higher in temperature toleration.


    Timer works! It activate when photocells is triggered and count down.
    Led lit up but shows a very weak light flow .

    Can this be due to n channel MOSFET issue?
    Can it be an issue at the 2 zener?

    I'm still struggling but somehow seeing the light!

    Also a desoldering hot air station is on the way and i bought 2 MOSFETs spares.

    Attached Files:

    Last edited by a moderator: Jul 12, 2017
  6. Jul 13, 2017 #5
    Guys i still need your help.

    Im sure d6 is ok.

    I detached the n channel MOSFET and tested it off circuit. It behaves fine the the gate is triggered by diode tester i read high value from source to drain and when i ground the gate there is no value so expected behaviour.

    Now my issue is still weak power at LEDs.
    They power up but i barely see the light .

    LEDs are ok as they lit up properly with a different driver.

    What is the role of the inductor and of the 2 zener diodes in the circuit?

    Inductor seem ok not sure how to test it properly.
    Zener I'd need to remove them from circuit so not sure if they can be a culprit i will go on and remove them.

    Let me know if you have suggestions

    Thank you
  7. Jul 13, 2017 #6


    User Avatar
    Science Advisor

    At this point we are all flying blind; although D6 is a suspect. The 1N5408 replacement is intended for power supply applications at power line frequency, it would not work well at high switching frequencies.
    If you have access to an Oscilloscope, see what frequency, voltage, and waveform the diode is asked to handle. And also what frequency, voltage, and waveform is at the LED strings. Post 'scope photos if you can, noting both Vertical and Horizontal settings.

    Beyond that, it's time you drew a schematic of the circuit, everything that allows the chip to control the LED strings. Show all reference designators, part Nos. and pin Nos. and whatever values you can. Voltages at key points would also be a big help.
  8. Jul 14, 2017 #7
    Well, I've replaced the 1n5408 with another diode from a boost dc dc converter for led application to be sure it can deal with the specs.

    I'm just trying to find the root cause of the circuit output being fine (light turn on when photocell is triggered) but light emission is weak.

    Unfortunately I don't have the background to fill the root cause analysis :(

    what I can barely imagine is that it could be due to the switching capabilities being bad?

    Can the MOSFET despite the test i did having issue to reach is duty cycle?
    can the oscillator in charge of this duty cycle (i believe is the qartz on che circuit) be also involved in this issue?
  9. Jul 17, 2017 #8
    Guys i may have found other faulty components.
    The two resistance on the ground of led out leads have weird values.

    See attached picture for values in and off circuit The rest of the resistances in that area measure ok.

    Not sure about c5

    Attached Files:

  10. Jul 19, 2017 #9
    Guys !

    Replacing the two resistor that were going far above their stock values solved the issue.

    Lamp is now working.

    Thank you for your support
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook

Have something to add?
Draft saved Draft deleted