Repairing a CFL that stopped working after a few hours of use

Discussion in 'Technical Repair' started by seanspotatobusiness, May 25, 2017.

  1. seanspotatobusiness

    Thread Starter Member

    Sep 17, 2016
    141
    2
    A long time ago I bought a multipack of 11 watt CFL bulbs and I no longer have the receipt. I put one of them into service and after a few hours of use (and about three ignitions) it has stopped working. I've opened it up and I see nothing obviously wrong. What parts can I test? How many hundreds of volts does the energised ballast circuit produce?
     
  2. SLK001

    Senior Member

    Nov 29, 2011
    1,169
    431
    Look for bad diodes, then transistors. These are the most likely culprits
     
    seanspotatobusiness likes this.
  3. seanspotatobusiness

    Thread Starter Member

    Sep 17, 2016
    141
    2
    Thanks! There are two diodes that allow reverse current. What kind of specification do you suppose they would be?

    [​IMG]
     
  4. DickCappels

    Moderator

    Aug 21, 2008
    4,958
    1,508
    One of the diodes --the black onw which is a rectifier should have less than about 5 microamps of reverse leakage at room temperature. The blue one is (probably) a diac and with less than its firing voltage (usually 28 to 40 volts) should have less than about 10 microamps reverse leakage at room temperature.

    Is that little TO-92 transistor the only one?
     
  5. SLK001

    Senior Member

    Nov 29, 2011
    1,169
    431
    I don't see the switching transistors (there should be at least one). Maybe they are surface mounted on the back?
     
  6. DickCappels

    Moderator

    Aug 21, 2008
    4,958
    1,508
    Years ago Philips developed some TO-92 transistors for these low power lamps but I have only seen them used in pairs. For higher power lamps it is common to see a pair of husky versions of the MJE13003, which is much larger.

    The TO-92 looks like a transistor, maybe you can use your DVM to tell whether it is ok or not.
     
  7. seanspotatobusiness

    Thread Starter Member

    Sep 17, 2016
    141
    2
    I don't know whether this video is clear enough but there are two TO-92 transistors.



    In this video I test the transistors and the diodes (not the diac) and I think the transistors are okay but the diodes do not resist reverse current.

     
  8. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    8,744
    2,019
    You should lift one leg of the diode to test it. Any low impedance in parallel with the diode would confuse the test otherwise.
     
  9. seanspotatobusiness

    Thread Starter Member

    Sep 17, 2016
    141
    2
    Sorry if this is a stupid thing to ask but wouldn't a low impedance in parallel with the diode make the diode pointless?
     
  10. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    8,744
    2,019
    If the low impedance were intentional and very low, then yes. But if the low impedance were due to some other, failed component then a healthy diode could be mis-interpreted as a dead diode.
     
    seanspotatobusiness likes this.
  11. DickCappels

    Moderator

    Aug 21, 2008
    4,958
    1,508
    It can only be a stupid question if you don't ask it.

    There is a possibility of the timing transformer appearing across the diode and looking like a short to the ohmmeter but looking very different to high frequency signals.
     
  12. seanspotatobusiness

    Thread Starter Member

    Sep 17, 2016
    141
    2
    The diodes are actually fine so I still need to find what is causing the short circuit.

    Incidentally I have a capacitor which I may have overlooked before but looks damaged to me and has a measured capacity of 1.2 nF:

    [​IMG]
     
  13. R!f@@

    AAC Fanatic!

    Apr 2, 2009
    9,664
    1,110
    47nf 250V cap is the culprit as far as I can see.
     
  14. seanspotatobusiness

    Thread Starter Member

    Sep 17, 2016
    141
    2
    Is it called a "film capacitor"? Probably a stupid question but can it be replaced with a ceramic disk capacitor?
     
  15. R!f@@

    AAC Fanatic!

    Apr 2, 2009
    9,664
    1,110
    I am not sure.
    Getting ceramics of that voltage is not easy I think.
    Better to replace with the same type
     
  16. seanspotatobusiness

    Thread Starter Member

    Sep 17, 2016
    141
    2
    Okay, I will; thanks.

    Unfortunately, there is still a short circuit across where the diodes are (and it's not the diodes at fault) so I think there must be another flaw somewhere. Based on my reading, it looks like it will not be trivial to test the DIAC since it needs to be used at high voltage.
     
  17. DickCappels

    Moderator

    Aug 21, 2008
    4,958
    1,508
    Take note if the markings on the capacitor. I have a strong suspicion that it has a safety rating of "X1" or something similar. In that case it should be replaced with a capacitor with the same rating.

    The capacitor almost certainly failed because it was defective to begin with. There should be a fuse resistor in series with the capacitor at the AC input to the circuit that is normally a few to a hunderd ohms. The failure if the capacitor may have caused that resistor to fail open.
     
  18. seanspotatobusiness

    Thread Starter Member

    Sep 17, 2016
    141
    2
    Thanks DickCappels, I'm still trying to interpret the circuit but I think it might roughly match this schematic with some slightly different values. The 47 nF capacitor is the one next to the fluorescent lamp. Edit: it can't match the schematic because in my circuit I think the diodes and transistors are in parallel across collector and emitter (assuming the middle leg is the base) and also the diodes are not in series.

    [​IMG]

    In the below pictures I've already removed the capacitor, one transistor and disconnected on leg of the two through-hole diodes. Also, I did flip one of the images horizontally in order to line them up for the overlay. I'm wondering whether the short circuit across the diodes doesn't matter if the circuit is a high frequency (I mean if it's part of the design)?

    [​IMG]

    [​IMG]
    [​IMG]
     
    Last edited: May 31, 2017
  19. R!f@@

    AAC Fanatic!

    Apr 2, 2009
    9,664
    1,110
    The short might exist if the diode has a low value resistor across it or if it is across a coil.
    Remove it and check
     
  20. seanspotatobusiness

    Thread Starter Member

    Sep 17, 2016
    141
    2
    I've removed one leg of both diodes and they are not at fault. I'm thinking that the 220 resistors are maybe 22 ohms rather than 220? Then the path is as below; it goes through some coil thing which presumably resists AC and therefore isn't really a short circuit over the diode?

    [​IMG]
     
Loading...