Large 200w CFL Ballast keeps failing

Discussion in 'The Projects Forum' started by highlumen, May 10, 2013.

  1. highlumen

    Thread Starter New Member

    May 10, 2013
    6
    0
    Hello all,

    This is my first posting here and I'm hoping somebody can help with re-occurring ballast failure problem. The ballast operates normally for 3 or 4 days and then fails, always in the same place. Only about 15% of our inventory seem to be effected which tends to make me think it's either a manufacturing or component issue.

    I have absolutely no experience with electronics and my attempt at trying to draw a circuit diagram from the tracks on the PCB have failed (and given me a bad headache) and so I'm hoping if I post images of the PCB itself somebody may have an idea.

    These ballasts are part of a super sized (200w) CFL which is purchased as a complete lamp from a manufacturer in China, unless I can prove to the manufacturer there is an issue I am likely to be disastrously out of pocket.

    There are 3 images, one showing the top of the board and components, the second showing the tracks on the back of the board (this image has been horizontally flipped so that it matches the first image), the 3rd is a close up of the failed area showing a short which is hot enough to vaporise the solder track.

    The point the failure happens is between what I think is an Inductor and a Bridge Rectifier.

    Any helps you can offer would be very much appreciated.

    Thanks
    Mike

    Image One:

    [​IMG]

    Image Two:






    [​IMG]

    Image Three:

    [​IMG]

    Thanks for Looking.......
     
  2. crutschow

    Expert

    Mar 14, 2008
    13,016
    3,235
    It sounds like a design deficiency. Can you check the temperature of all the parts on the PCB when they are operating normally (finger test)? I would expect there is one (or more) that is getting too hot, which leads to an eventual failure.

    If all temperatures appear normal then I would expect some component is experiencing voltage beyond its design limits. That can be more difficult to locate without a schematic and an oscilloscope.
     
    highlumen likes this.
  3. GopherT

    AAC Fanatic!

    Nov 23, 2012
    6,058
    3,820
    If you have a newer version of PowerPoint, you can paste your two top photos in an then adjust the "transparency" of the components overlaying the PCB traces. That way, you can see the components behind the connections all at once.

    However, even with the info about component values, we will not know the load from the LEDs. We also don't know if it happens as the current surges during power on (or off) or if heating in the fixture is the issue (imagine 200 watts that should not get over 100 degrees C). You should have some major sized heat sinks on the rest of the lamp and good (excellent) airflow around the lamp.

    I just don't see this ending happily after you invest time and effort to have the autopsy of the circuitry. I think your best approach will be to contact the seller directly and immediately. The less you invest in this problem, the less heartbreak you will have.

    Also, be creative in pricing the balance of your inventory (4 for the price of 6) - customer gets a deal if only 15% failure. You also recover some cash. Just inform your customer why they are on sale and see who bites.
     
    highlumen and #12 like this.
  4. #12

    Expert

    Nov 30, 2010
    16,298
    6,810
    I hate to be a wet blanket, but I'm with Gopher. This isn't going to end well.
     
    highlumen likes this.
  5. highlumen

    Thread Starter New Member

    May 10, 2013
    6
    0
    Hi, thanks very much for your post.

    I will attempt a 'finger test' to see if I can identify any particular component that's over heating. I think a friend has a thermal imaging camera so I might try to borrow that for the task.

    This isn't the only super large CFL we sell, there is another which is not as quite as powerful and this which is no problem at all, I had noticed that the type I originally posted about did in general get much hotter than the others.

    If this were, as you suggest, suffering from a Design Deficiency wouldn't all the lamps have the same problem? Is it possible that the components (or just the one responsible) are not of the highest quality and 15% failing for this reason?

    Thanks again for your help with this, I find it so frustrating not having the knowledge to figure it out myself. Beginning to wish I hadn't dropped Physics when I was at school, but that was a long long time ago..

    Mike
     
  6. highlumen

    Thread Starter New Member

    May 10, 2013
    6
    0
    Hi, and thanks for your suggestion about using PowerPoint, I had done exactly the same thing in Photoshop placing the PCB tracks in a layer over the components, that's how I came to 'flip' that image so it matched the components underneath it.

    Cheers,
    Mike
     
  7. highlumen

    Thread Starter New Member

    May 10, 2013
    6
    0
    Hi #12,

    Thanks for your post, I'm trying to stay positive as we have a decent investment in these. Currently sitting here with a bag over my head waiting for the fairy godmother to arrive and set a spell on my troublesome lamps.

    Cheers
    Mike
     
  8. kubeek

    AAC Fanatic!

    Sep 20, 2005
    4,670
    804
    Mind you that if that thing is powered from the outlet you need to be very careful with that finger test. Better turn the thing off and wait a few seconds before touching anything.
     
    highlumen likes this.
  9. highlumen

    Thread Starter New Member

    May 10, 2013
    6
    0
    Hi again,

    One thing I did forget to mention in my original post, not that it probably makes any difference, is that there are two separate lighting ballast circuits on this board. The lamp has an array of 8 x 10" U Tubes and each ballast drives 4 of those.

    When they fail its pretty much 50/50 as to whether its one or the other that blows - could this be another strike for the dodgy component theory? Perhaps some small Chinese component manufacturer doesn't operate as close a quality control as they should??

    Thanks again,
    Mike
     
  10. highlumen

    Thread Starter New Member

    May 10, 2013
    6
    0
    Yes, I agree, there's some hefty looking capacitors on the board that could probably wipe the smile off my face. All the more thinking that I should borrow that thermal imaging camera!

    Cheers
    Mike
     
  11. #12

    Expert

    Nov 30, 2010
    16,298
    6,810
    While designing commercial quality power supplies, my company expected, and had, 3% "infant mortality", knowing full well that the design was good (conservative). We knew that because the broken power supplies that we were repairing averaged more than 20 years old..if they didn't die in the first 90 days. If the design was skimpy, they wouldn't have an average survival time of more than 20 years.

    Any batch of components has variation. When they get sorted, anything that passes, passes. Some just barely pass and some are significantly better than the minimum requirements. Some are just sturdier in ways that aren't tested.

    Get the drift? I'm suggesting that this one design has parts running at the edge of survivability and the infant mortality rate is higher because of that. You wish you could find one part, or maybe a quad of rectifiers, that could be replaced with a better part, and thus fix the problem for 30 cents worth of parts. It's probably impossible to do that over the internet with no schematic and no part numbers. If I had half a dozen of these failures, all the documentation, and a few weeks to test them, I could probably fix the problem. Unfortunately, I am not the manufacturer, and it's THEIR problem. I'm sorry they saddled you with their problem.

    I can suggest one method. Run each light for a week before you sell it. That isn't the best solution, but it could avoid ******* off your customers and salvage most of the product. You can either refuse to sell "used" light fixtures or you can brag that they have all been tested to assure reliability.

    Edit: Oh look! Auto correct for naughty words! I didn't know this site had that:p. (Probably because I've never used a naughty word before.)
     
  12. kubeek

    AAC Fanatic!

    Sep 20, 2005
    4,670
    804
    Also there is another question, if you replace that burnt trace with ta piece of wire, does the light work again?
     
Loading...