When a Project Isn’t Successful

For every project I post on here where I obtain the results I was looking for, chances are there are two other projects I’ve tried that didn’t quite work out. Such is the nature of tinkering and hacking for fun.

Just because it ended in failure (and occasionally, catastrophic failure), it doesn’t mean I’ve wasted my time – here are some lessons I’ve learnt from failure over time.

Most recently, I obtained an used battery pack from a friend’s laptop. It still held a charge, but only for 15-30 minutes. I figured, “Great, I can use this to finish another project!” I have an old ThinkPad R31 with a completely dead battery pack that I already disassembled, in hopes of replacing the cells. The project was shelved while I decided how I would proceed and weighing the risks, but getting a free battery pack was all I needed to get back to it.

After taking pictures of how it was all wired, I removed the old cells, and began to replace them with the new cells. This wasn’t particularly difficult, I just had to be careful with soldering to not let the batteries get too hot (Lithium Ion battery hate heat, short circuiting, punctures… and literally, air.) This perhaps falls into the category of a dangerous hack.

Once the pack was together, I installed it into the laptop. Rather than the 0% I always got from the other pack, this one reported 92%! I though I was in the clear, and let it charge (while carefully monitoring it). Although it said it was charging, the level didn’t climb. I did some measurements, and confirmed my suspicions: the pack protection circuitry was not allowing voltage in or out of the pack. At this point, I called it a night and a failed project.

Lessons

From this I learnt several things:

  • A half second short circuit will shoot cell temperature right up
  • These protection circuits seem to stop working when the cells discharge too low, to prevent recharging when they are too far gone
  • The cells are never directly connected to the laptop, instead passing through switching circuitry

Over time I’ve amassed a collection of working LCD monitors I’ve refurbished. I’ve also collected a graveyard of LCD monitors I’ve been unable to repair – these ones do not have tell tale capacitor bulging, thus leaving me to either replace every capacitor, or look for a different source of malfunction.

Most recently, I received a 19″ widescreen LG monitor, which would have made a very suitable replacement for my 17″ fullscreen second monitor. It suffered the exact same symptoms as almost every broken monitor that makes its way into my care – backlight on for 2 seconds, then darkness. The panel still works, there is just no lighting. This happens because the protection circuitry detects an error condition: usually over- or under-voltage. I believe bad caps would lead to under-voltage. This 19″ didn’t have any bad caps, so I looked online for more ideas.


Looks good to me?

There is an IC (OZ9938GN) that controls all backlight functions, such as on/off, dimming, and error detection. Pin 7 was connected to a voltage feedback circuit. I measured this pin, and the voltage didn’t match up with the datasheet-stated values, meaning that either:

  1. There was an under-voltage condition
  2. The feedback network was not working correctly

On the Badcaps forums, someone said that they once found the feedback circuit to be the problem, and to fix this, they simply bypassed it using a 1M Ohm resistor connected from VDD to pin 7. I tried the same, but to no success.

Lessons

I learnt a lot about the operation of LCD panels. From protection circuits to troubleshooting circuits based on schematics and the types of components present. Even though I didn’t manage to fix this LG (along with two other BenQs and a Cicero), I haven’t given up on them, and for every monitor I can’t fix, there has still been another one I have fixed and saved from going to the landfill.


There have been many more failures than this, however, due to their nature of being failures, no documentation exists for most of them. All that remains are the damaged circuit boards, broken cases, and blown components from a project with hours spent, ending in worse shape than when it started. Still, it takes these lessons to get better and continue expanding skills and knowledge.

Tags: , ,

Sunday, September 11th, 2011 electronics, life, projects

5 Comments to When a Project Isn’t Successful

  • Rex says:

    Hi, I also have a LG LCD (19” W1934S) with the same diagnostic, 2 seconds backlight then nothing… I’m looking for the problem. While measuring voltages in the OZ9938 I’ve managed to keep it longer on if i had the probes on pin 3 (TIMER). So I thought it might be the capacitor. I’ve added a 100uF in parallel and managed to get 2 minutes average of on time. But it’s still not a solution. Going to investigate more but could be a problem in the feedback network…

    • Dan says:

      HAHAH! That is EXACTLY what I noticed happen as well – the moment I probed it I found that it did indeed stay on longer. I too attributed it to capacitance.

      That being said, the problem could lie in the feedback network improperly generating the error condition, or the voltage going to the tubes could simply be beneath the “functional” threshold. I’m not all that familiar with the feedback network, I know which components are used, but I don’t yet understand what values I should expect at different points, or how to determine where the fault is. Being all surface mount, I do not quite have the kit necessary to do that work.

      It’s always rewarding to get an LCD monitor back to working condition – sadly as I’m stuck at the same issue with my monitor and haven’t dug into it more since, I have no ideas that would help you. If you get it going through your own brilliance, I would love to know how you did it!

  • Rex says:

    I’VE SOLVED IT!!!!!!! :DDDDD
    YES!!!! 😀
    Steps:
    1- removed the feedback diodes so that the Vsens wouldn’t be affected.
    2- found out that the “lower” lamp on the “upper” inverter would stay always on. With the “upper” lamp connected to the “lower” inverter it would also stay on but the inverter would heat up pretty bad (85º+).
    3- as the wires on the lamps are pretty short i’ve decided to disassemble the screen to change one for the other. I’ve also read from someone who had the same problem and that it was a bad lamp. At 35€ for 4 lamps on ebay it would probably be worth it…
    So, when i was removing the lamps from their frames I noticed that one of the wires had a “brownish” coloration near the lamp. As i was pulling the rubber that holds the lamps in place, the wire came off!!!
    The wire had broken due to overheating (and/or bad soldering).
    Cutted a bit off to remove the oxidated bit and soldered it back on to the lamp and…. TADA!!!! It works! And at “zero” cost 🙂

    You should disassemble yours too and check if it has a similar problem or some of the lamps are “dead”. I was about to leave only one of them on 🙂

    Hope this info helps you! 😉

    • Dan says:

      YES! Awesome! I’m very glad to hear you found it. I will be checking that out shortly on my display. Heck, if it was as simple as a bad wire – I’ll check them before taking out the feedback diodes, just in case.

      Thank you so much for getting back to me on this!

  • Rex says:

    😉 you’re welcome!

    Yes, check it out.

    Now i’m going for another battery recovery… Let’s see how it goes… the last one i accidentally shorted something on the board while opening it. Gave up on it and saved the cells for other recovery. 🙂

    I’ve found out that some controllers can be “reseted” by briefly connecting the pack Vcc to the output pin on the connector. Maybe you could try that too.

  • Leave a Reply to Dan