Hardware Misc Strange issue with a Japanese Pokémon Pearl cartridge

[grub]

Basically, it freezes during the in-game event below, when initiating dialogue with the first Gym Leader. I found mention of it in this Reddit thread.

I dumped the game and save data with godmode9 and loaded them with MelonDS, and the same issue occurs.

The most likely issue is that it's just counterfeit and an untested ROM was loaded onto the memory, right? I've attached images of the taken-apart shell and board if you're knowledgeable about spotting fake copies.

The labels on each IC seemed correct to me, and the cartridge itself seems fine too. The serial number on the back looks like it was worn away through usage, it's only visible if you look really closely.

There were 6 cartridge revisions of Pearl in Japan, but the later revisions trimmed down the ROM size from ~128mb to ~64mb. Comparing the dumped ROM with the only early revision ROM i have, 7 hexadecimal bytes are different between them.

When the save file is loaded with this other ROM, the dialogue in-game completes successfully and the battle starts.

please help me, gamers

Post automatically merged: Mar 17, 2025

A small update: I tried reflowing the solder on the ROM chip, still nothing. So I looked into the ROM itself with Tinke to see where the mismatched data actually was in the game files by comparing it to the working reference ROM.

The first six mismatched bytes belong to the field_encounteffect file, which I think means the game is failing to pull up the Gym Leader challenge animation. It would explain the freeze at this specific point.

There's also an additional byte in the font file that doesn't match.

The one Reddit thread I mentioned above posted an image of their cartridge, and the numbers above the contacts seem to match mine. So, it's either an obscure batch of bad Diamond/Pearl cartridges in Japan, or a (very convincing) counterfeit cartridge...?

I guess I'll try to find where I bought it in the first place. I think it was eBay, but it has been a few years.

 

[Deleted member 702243]

it's more likely that it's hardware damage that damaged the data on the chip somehow, than a counterfeit that happens to have just a few bytes wrong. no one makes bootlegs of japanese versions anyway

 

[grub]

it's more likely that it's hardware damage that damaged the data on the chip somehow, than a counterfeit that happens to have just a few bytes wrong. no one makes bootlegs of japanese versions anyway

Oh! That makes more sense! Maybe this batch of cartridges was more prone to whatever the damage might be caused by. Thank you for your response

 

[Wokann]

Hey guys, seeing the filename "field_encounteffect" in your post really excited me.
Because I've been investigating why the checksums of my two Japanese Pearl rev0 cartridges don't match the data from nointro.
After unpacking and comparing them, I found the difference lies in the "field_encounteffect" file, especially the file within its subfile id=23 (field_encounteffect_23.RGCN).

This is an image of a smiley face (the color palette was randomly selected and does not represent the actual colors). The top left corner shows from the normal file, while the bottom left and bottom right corners show from my two jpn cartridges, both with damaged pixels, especially the bottom right one, which has more damaged pixels than left.
I wondered if the cartridges had deteriorated over 20 years, or if this batch of cartridges was already damaged when it left the factory. So I started searching for similar posts, until I saw "field_encounteffect" again in your post. This makes the latter possibility more likely!

In addition, while reviewing the decompiled Pret project, I investigated where this image is invoked. I discovered that it seems to be triggered only during battles against Gym Leaders. I searched gameplay videos but didn't find this image during battles. I initially thought I had misunderstood the code, but since your triggering issue is stuck during battles against Gym Leaders, it actually proves that this image is very likely involved during those battles.

Considering that the damaged areas on my two cartridges are different, and they functioned normally without getting stuck in gym battles, it's very likely that your random damage data just happened to fall into the address where affecting the game's operation.

You can use Tinke to select the "field_encounteffect" and continue unpacking it, then compare the results to see which file is causing the problem. Is it the same as me, that field_encounteffect_23.RGCN is corrupted?