I bought a DMG on eBay from a guy who had tried to backlight it but apparently gave up. He listed that the screen worked, no sound, left on d-pad broken, possibly a broken bivert chip. He also included a second CPU board which he stated was "untested," and he had already removed the reflective "sticker" from the back of the LCD panel.

I get the package and indeed left on the d-pad does not work, every other button works, up, down, right, a, b, select start. I fix the bivert chip problem, he had messed up soldering it on and it looks like he managed to mangle one of the traces so that the right of the two wasn't actually making any contact, and the wire leading to ground had broken loose. To repair the sound he had attempted to replace the pot with a monaural pot, with some effort I was able to remove this and replace it with a stereo pot, and whether through his action or mine some of the traces were "messed up," but the audio is working again.

As for the dpad, I tried everything I could think of to get it working, cleaning the contacts etc, and I tried to trace all the connections from it to other points on the LCD board; I eventually swapped over to the "untested" CPU board he had included and now all of the buttons work. I've read that if the input on the CPU is fried then both left and the B button won't work, I guess they use the same register, so I'm not sure if this is a CPU problem or not.  I don't know if he ever attempted to separate the two boards while the unit was on.

I have a soldering iron and a multimeter, no flash carts or anything like that. I don't know if there are any connections I can test for continuity to see if there is a broken trace someplace, or if there is any way to test the CPU for damage. I'm hoping that there is something simple that I missed that one of you might have run across, and I'm hoping the fact that all of the other buttons being functional means that the CPU isn't damaged.

Sorry for being long winded about this but I wanted to give a complete picture of what's going on. I appreciate any help that I can get. It may seem strange that I want to fix this board when I have a second, functioning one that fixes my issue, but I just hate to throw away something that I might be able to salvage; especially when I already put work into fixing what he messed up. Thanks for any help you can give me.

You've done your research. Well done. Indeed, left and B share the same physical input on the CPU. If you insert or remove the ribbon cable while the DMG is on, and misalign the cable ever so slightly so this connection is short circuited to the next one, which happens to be a -19 V supply for the LCD, the left/B input on the CPU is permanently damaged. Moral of the story, always turn off the DMG before inserting/removing that connector.

Are you absolutely sure the B button still works? There's really no way the left button could be broken without the b button unless there was some kind of problem, like a broken trace, around the area before the the B and left traces meet up on the LCD board. But then it wouldn't make sense that the button would start working, unless you accidentally rejiggered something in the process. If indeed it's an intermittent problem with the LCD board button connections, the good news is that you should be able to fix it with a little bit of soldering.

If it turns out to be a CPU problem, don't throw it away. You could give this 'boy some new life as a LSDj slave mode player, or MIDI synth.

As for probing, these are the relevant signals on the ribbon cable:

Pin 04 Left & B buttons
Pin 05 Button Diodes 1 & 2 (directionals)
Pin 06 Down & Start buttons
Pin 07 Up & Select buttons
Pin 08 Right & A Buttons
Pin 09 Button Diodes 3 & 4 (buttons)

The way it works is you have to put the negative (black) multimeter probe on the "diode" connection and the positive (red) probe on the other corresponding connection. As you press a button those two connections will short. Since there is a diode in series, you need to use the diode mode rather than the resistance mode, and beeping may not work so you might have to look at the multimeter screen for a readout instead.

Schematic for reference:

OK, I'll take a look when I can. Right now I don't have a backlight so visualizing the screen on this thing isn't easy, but when I tap B I can hear the Tetris piece rotate (Tetris has been my test cart). I didn't think to see if the piece rotates in the opposite direction. I can see the piece move right, I can just see the piece disappear when I press select, start pauses and gets me into the game, I can hear the piece thud when I press down, and up changes tracks in the music selection part of the menu when I press up. I could try plugging in the "damaged board" again to see if I just, as you say, regiggered something, however the board came to be with the problem described so I'm not hopeful that something so simple will fix it. I should have the backlight tomorrow and I can install that so that this isn't so much of a guessing game in terms of what the buttons are doing.

Thanks for getting back to me.

Ok, electronics are voodoo and black magic, and I only know the base level incantations; that rejiggering argument wins. I swear that left on the do as was broken, and the only thing broken; I pushed so hard on that button listening for the sound of a Tetris piece moving that I thought I might break something. Well I got the back light, and the new bivert chip, installed them (bloody easy) and got everything working. I played with it for a bit, went to dinner and came back.

So for shits and giggles, I swapped boards, because now I was actually able to see what was on screen. And left on the dpad works. So I've got two good CPU boards now. I thank you again for your advice, and the schematic. I'm hoping to see if there are any Gameboy on eBay with a dead CPU board, maybe I can piece together another Gameboy.