I was trying to sell this thing a while ago on Reddit. So I'm repurposing the timestamp pictures here to show what it looked like before I started tearing it down. You can also see the failed and failing pixels on the bottom left image which is why I was trying to get rid of it.
Neat, so let's take it apart. Several screws later, we arrive at the first layer. There's a switch on the top that prevents startup if disassembled. This switch needs to be held down for as long as the device is plugged in. I can bypass this later by shorting the switch. For now, I'm taping it.
After we've exposed the logic board, we can start to tear down the rest of the assembly, making sure that we keep things connected as they may be required for projector operation later. Removing the main circuitry board reveals the power supply underneath. We'll need this for sure, as well as the logic board itself.
As for the logic board, this board is a bit more sparse than I imagined. I suppose they packed a lot of the operation logic into the DLP chip on the right. If I had to guess, it probably has an ARM core on there and some video decoding logic to convert the video signals to raw data which is then transmitted to a driver chip (probably the one north west of it). We'll still need the whole board.
The next thing that was easy to find was the DLP chip.
The chip is directly driven by a many pin connector straight off of the logic board. With everything still plugged in, I tried to turn the projector to see if it would work. Ideally, I wanted to see an image on the DLP chip. And sure enough, you can see the warm up screen very clearly. On my camera, I'm so zoomed in and the DLP is changing so rapidly it causes a banding effect.
This is exactly what I was going for. But now the question is this: how much can I strip away before the projector refuses to work? Ideally, I would like to get rid of the fans, the color wheel, the lens, and the lamp. Everything else seems pretty necessary to the operation of the device. The projector, however, will not start without these components. So, how do we mock them so the logic board believes they're all in working order?
Now, the rest of this is only hypothesis because while I attempted to figure this out, I shorted and destroyed something. I'm not sure what, but there was a little bit of magic smoke somewhere, and now the thing doesn't turn on. So that's neat. So take the following with a grain of salt, it may have worked, it may not have.
I shouldn't be screwing with high voltage electronics. That's a given. The lamp ballast is the first (and primary) thing that needs to be mocked out. I tried doing what this site told me to do, but in a very, very careless way. I figured if I could short one of the communication pins from the logic board to the ballast control circuit, it would appear the lamp was always on and in good working order. This is, in theory, still a viable option. It's also possible that the communication is done over UART, but think about it... why would it be? I mean, yeah, sure it's possible. But still. It's so much simpler to have a few status wires that indicate the status of the lamp. Just don't be a big stupid and reverse engineer this the way I did.
So with that, we need to get rid of the color wheel. This is getting feedback from a magnetic strip on the side of the brushless motor. I think this technically makes it a hall effect sensor. As far as I can tell, the color wheel has no additional circuitry and goes directly back to the main board. To mock this out, I think we just need to add a PWM signal of a certain frequency to it. Is it worth this much work? Probably not, but again this is all in theory. I get to not try it because I broke my thing.
Seeing the mirrors move and change shape on the DLP chip is actually pretty amazing. I knew this was the way it had to work, but regardless, actually seeing it is very, very neat. I recommend doing this if you have a DLP projector you wish to dispose of. Just... just don't break it.
No comments:
Post a Comment