Right so, I picked up a couple used TI-83+ Silver Edition calculators from eBay for super cheap, hoping I could repair them.

What I didn't expect was the odd nature in the way they were broken.

Both power up, and the screens are corrupt - but the calculator's seem to be working fine otherwise. This would lead me to believe that it's an issue with the flex-cable going from the main board to the display board.

One calculator is like 90% usable - only the top row of characters really corrupts. The other is really bad and very difficult to see anything, but it does work if you fight through the corruption.

Okay, but here's the WEIRD PART.

Both TI-83+ Silvers have the Detached Solutions "Puzzle Pack" installed in the Applications menu. (One is older version that had copyrighted names, and the other one is newer).

Anyway, once I boot into Puzzle Pack, the screen works 100% perfectly.

Not kinda.

Not sorta.

100% Perfectly!

This is really odd to me, because it means that the flex cable is working perfectly fine. I was able to play all four games in the puzzle pack, on both calcs, and there was ZERO corruption or issues playing. It worked flawlessly.

The title screens also looked fine.

It's weird that ASM games would work perfectly but the OS does not.

One of the two calcs has MirageOS on it. Mirage loads up, but unlike PuzzlePack MirageOS does have corrupt graphics, as well as UncleWorm, but it doesn't corrupt as bad as the OS. Strangely though, despite MirageOS and UW being corrupted, PuzzlePack still works flawlessly on that same calc.

Anyway, I don't wanna goto the effort of soldering my own flex cable if the issue is with the OS.

On the other hand, maybe timing of the PuzzlePack writes to the screen quicker than the OS or something, and it compensates for a weak signal on the flex cable?

Not sure. Any ideas? Anyone seen anything like this before? I tried googling and couldn't really find anything.

I installed PuzzPack & MirageOS (among other things) on my TI-83+ & then did not use it for a while (at least a few years). The last time I tried to use it, the screen had the left column of pixels of some, but not all, characters corrupted in the OS & MirageOS, but in PuzzPack there was no corruption, similar to your situation. Some people suggested coarse ribbon cable failure, but I did not want to have to fix that if that was not the problem, & since it did not happen in PuzzPack, I suspected something could be wrong with the OS itself. I also wanted to install some other app & found that I could not despite having plenty of free space, so I reinstalled the OS in order to fix that (nothing else having worked to fix the app installation problem). This fixed the corruption (& also let me install the apps together)...so apparently the OS can somehow get partially corrupted? Or else the ribbon cable coincidentally decided to work properly after reflashing the OS.

Per chat in IRC, you can try reinstalling/upgrading the OS with version 1.19. If it's really coarse ribbon cable failure (which happens to a significant percentage of TI-83 Plus Silver Editions, far more than other calculator models), well, at least the repair process is well documented. ¯\_(ツ)_/¯
Taking screenshots with your preferred linking program will also leave artifacts if there's a coarse ribbon cable issue.

Oohh, that's fascinating. I'm guessing the video memory for the display is on the display daughter-board?

Welp, thanks guys - this weekend I'll test out the screen shots and also do an OS reinstall and I'll report back here with my findings!

Yep, the video memory is stored on the LCD controller itself.

Quick update:

I tried using the screen shot tool in TI-Connect and the results were completely unrecognizable from what the display was actually showing. Whatever, weird.

I updated both calcs, and it fixed neither. Though on one, the "receiving os" screen was flawless without corruption.

Anyway, I eventually decided to replace the ribbon cable with some crappy soldering, and... it worked!

So, as usual, it was the flex/ribbon cable thing.

How the PuzzlePack mysteriously bypassed this issue and worked flawlessly, may never be known.

Hi Greg
Now that was an interesting behaviour.

Hey Lionel! We cross paths again

I've seen calcs with damaged ribbon cables mis-interpret commands as contrast setting among other things. What I suspect is the games in PUZZPACK refreshed the screen contents often enough that any corruption didn't last long where as TI-OS and other apps which update just part of the LCD contents caused issues to be long lived enough to be noticeable.

It could just be that the wrong value was written to one or more port(s).
More specifically ports $29, $2A, $2E, and $2F.
That would explain a lot.
I'd say send the calcsys app, and using the port monitor, let us know which values you see.

I would like to point out that when about 3 years after the TI-83+SE came out, some TI community members started experiencing problems with the LCD similar to this, but I don't remember if ASM games like PuzzPack worked fine. In the years that followed more TI community 83+SE calcs failed the same way, which led me to think that this model, or at least the first hardware revision (if there were more than one) had a slightly higher failure rate than most other models. Early 84+ seemed to have issues early on as well.

Interesting, i wish i had one to investigate.
If someones who reads this does, please let me know, i may be able to help fixing the OS display.

Indeed the 83+SE seems more susceptible to LCD failure, the ones that I've owned all ended up with LCD issues, despite calcs 10 years older than them being fine.

This is a classic symptom of ribbon cable failure. Because of the high-impedance characteristic of some of the ribbon traces, the read commands that the calculator uses to fetch the contents of the LCD to send to TI Connect are not interpreted correctly (and/or the data that the LCD returns is corrupted).