Fail of the Week: Projector LED Retrofit

fotw-projector-repair

That’s a deal for a project, how hard could it be to fix it up?

If you’re a real hacker we’d wager you’ve fallen for this type of thought process before. [Luft] bought this used Sharp XR-10X-L projector about a year back and planned to retrofit it with an LED bulb. He gathered all the parts and got to work, successfully testing and installing the modifications. But as luck would have it, the project is stuck in some type of boot loop.

This fail is certainly not for lack of preparation. The first post documenting his adventure shows that the hack has been done before, he acquired the service manual for this particular hardware, and he did his homework when ordering the parts. Success requires circumventing some sensors which ensure the case and internals are in place, and making sure the electronic status of the ballast is reported correctly event though it’s not needed for the LED source. Power-on gets as far as illuminating all the indicator lights in green as it should, but is then followed closely by a reboot sequence.

He tried watching the serial port to see if he can get any status info there but no dice. In keeping with the nature of this column, let’s see if we can provide any constructive troubleshooting advice in the comments.


2013-09-05-Hackaday-Fail-tips-tileFail of the Week is a Hackaday column which runs every Wednesday. Help keep the fun rolling by writing about your past failures and sending us a link to the story — or sending in links to fail write ups you find in your Internet travels.

Comments

  1. mike says:

    I know my projector has a time out wherein if the bulb s not lit It reboots. You may have shorted the “LAMP ON” signal but the micorcontroller might be expecting that to be asserted sometime after the lamp is struck and/or a warm-up period.

    • andrewjhull says:

      Furthermore, it may expect it to be a certain logic level, or indeed be an analog signal from a light sensor…. or… the projector may have a damaged PSU as a result of the lamp blowing… or the temperature of the lamp might be used to indicate that it is on… there are quite a number of possible rabbit holes here…

      • Nonya-Biz says:

        some balasts send a pulsed response to tell the status of the light. that way it can send a more detailed error.

        • RBMK says:

          Or some ballasts comunicate on full serial link on those optocouplers. I encountered that on ACER projector, so i sold it to more advanced and motivated hacker and got another projector (toshiba) and that had classic “start the lamp-its ON” control.

  2. atomsoft says:

    why not get a original bulb test out the booting and signals and mimic them with LED source.

    • Bill says:

      Projector bulbs are expensive. He probably wants to avoid paying more than the projector is worth.

      • fartface says:

        Not really, china knockoffs are $60 a lot cheaper than spending 10 hours scratching your head trying to guess at what it wants. Plus he doesnt know of the projector ever worked, he may be stuck with a borked projector that was sold as “only needs bulb”.

        Used projectors are not worth more than 10% of new retail price for a projector that is 1-2 years old. no more than $50 for anything older than 3 years. too much goes wrong in the things.

        • Jan Lukeš says:

          It that case I’m an idiot. I gave 200 bucks for 6 years old HD projector with 450 hours of usage and I had a chance to inspect it before I bought it (with service menu inspection). At least I kept it 10% of original price :)

        • Elmars says:

          Please note that he is not in the USA, where labor costs are high and parts costs are low and easily available. He most likely has to pay list price plus 25% for a new bulb. It is well worth his time to find a workaround.

        • DainBramage1991 says:

          Not everyone has $60 to throw at a project that may or may not work. Everyone, however, can come up with 10 hours without spending any money.

  3. dcroy says:

    ive perfomed a similar conversion on an lcd projector of mine, the problem is lamp on is hardwired so when the projector starts up expecting the bulb off it is likely seen as a fault and shuts down

  4. oneuglymonkey says:

    You talk about there being a feedback signal on the lamp. Do you know if that signal is just a simple on off deal or is there more to it. For example I know there are rpm feedback signals on certain computer server fans. If the signal is not detected the system will shout at you. If you can replicate the bulbs feedback signal then maybe that will fix your issue.

  5. Jim says:

    Are all the fans originally fitted still connected ?
    A projector i fixed had a dodgy connection to a fan in the cover and wouldn’t startup because it couldn’t see it spinning.
    Did the same thing as your, continuous restarts.

  6. hyperjer says:

    Some Projectors use bulbs that have a built in counter or serial identifier. Protects you from using the bulb too long and having it explode destroying the projector. or some may say its to make you buy more bulbs…

    • qwerty says:

      must be the first option. it is widely known how printers manufacturers put chips into ink cartridges only because they care about our precious documents and would never let us print them badly using dubious quality or almost empty cartridges:^)

      • Darik says:

        I’m honestly tired of people that bring this up. Yes, it absolutely is to protect profits, but not entirely in the way that you, and others, usually state. If someone purchased a printer, and later one of the ink cartridges happened to be counterfeit/after-market and wouldn’t print well, or worse, ruined the printer (and or furniture/carpet due to a leak, which happens on the very cheap refill/remans), who is the *average consumer* going to call and complain to? Further, good *quality* ink that maintains the same color and other properties from batch to batch, and year to year is *not* cheap. It is overpriced on printers sold at/below cost because that is the business model. Continuously complaining about it and how “ink is so expensive per ml” is as useful as saying “look at how expensive gold is when you buy it from Intel!”

      • You forgot to mention that printer manufacturers don’t really sell printers, but ink :D

  7. andrewjhull says:

    … or it could simply be that the “photosensor” doesn’t see any photons…

    http://newcar-audio.blogspot.co.uk/2013/12/dmd-replacement-sharp-xr-10x-l-xr-10s-l.html

  8. Chris T says:

    Projectors typically monitor the current flowing through the arc lamp and the ballast’s voltage output to determine it’s condition. If the current goes outside a set threshold, the projector assumes the bulb is bad (e.g. pulling too much current, or not enough). This is sometimes handled as a “go / no go” by the ballast and relayed to the main board via a single pin (e.g. +5 if everything is okay, grounded if things are no bueno), but sometimes it’s an analog value passed to the main board. It depends on the projector make/model.

  9. Nonya-Biz says:

    my wd57733 has a status in the manual for status LED red “balast UART communication problem”. going to try and simulate the signal and replace the lamp with a 50w or 100w led in mine.

  10. Andrew C says:

    I had this problem on my projector— when the colour wheel would start spinning, the 5v rail (IIRC) voltage would drop and the projector would essentially power-cycle. I fixed it by using bigger, higher-current transistors in my power supply (the original had died, prompting the hack) and adjusting the voltage slightly higher.

    If you have a scope, it’d be interesting to watch the power rails while it boots.

    • Andrew C says:

      My projector was not a Sharp, however. If you’re consistently getting 30 seconds out of it, the problem is almost certainly related to some failsafe signal not being read.

  11. SK says:

    The screenshot shows that the projector resets because of a “Data Abort Exception”, i.e. the processor is trying to access data it shouldn’t (NULL pointer?). I do not think this should happen in any scenario expected by the firmware programmer (like lamp signal detected too early). In these cases I would suspect a more specific error that that showing.
    I think the projector probably has this problem since purchase. Maybe the firmware image in Flash is corrupted or a supply voltage for the processor is flaky.

    • Greenaum says:

      It *might* be that some subtle combination of errors is sending the firmware program somewhere it shouldn’t go, and that this was never discovered in manufacturer testing. So then getting it back up to spec would mean this problem would stop occurring and he could carry on despite faulty firmware.

      It’s probably more likely what you said, though. It’s brain has gone. Whether that’s replacable determines whether our man was ripped off or not.

      I can see a small, evil, business opportunity in collecting broken projectors, taking the bulb out, and selling them on Ebay as “needs bulb”…

  12. Marvin says:

    He has the Service Manual, everything is there, he just has to check.. There even is a Diagram for “Check when the lamp turns
    off soon after turning on”

    There mentioned a Photosensor and a 180Hz Signal and so on… easy to check :)

  13. BotherSaidPooh says:

    555 timer, trimpot and a very high voltage pair of series capacitors :-)
    (Please PLEASE use verified non conductive trimpot tool, thanks!)
    The idea is that it simulates the PWM signal from the lamp driver, all you need to do is approximate the signal and it will phase lock to the original.

    A somewhat safer method is to use a reverse biased blue LED and two HV series capacitor to operate as a high voltage signal coupler.
    This can be driven from software on the Tx side, just use an IR LED as the sensor and red emitter then pot them in clear heatshrink to make sure they don’t degrade.

  14. Haku says:

    Yet another started-but-not-finshed project of mine is to do this, I bought a 2nd hand Sanyo PLC-XU41 projector which works fine except for the dead bulb, I discovered a full schematic PDF online and that I only needed to cut a single track and bridge the line to GND to fool it into thinking the bulb isn’t dead, meaning it can be undone if I ever felt the need to spend lots of money on a proper bulb…

    But attempts at using my 50w LED didn’t work so well, I think I need a lens or something.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Follow

Get every new post delivered to your Inbox.

Join 93,784 other followers