Mystery regarding UA6528 lot numbers

Discuss hardware-related topics, such as development cartridges, CopyNES, PowerPak, EPROMs, or whatever.

Moderator: Moderators

Post Reply
loglow
Posts: 48
Joined: Mon Sep 19, 2022 1:47 pm

Mystery regarding UA6528 lot numbers

Post by loglow »

I have a small mystery on my hands involving UA6528 (clone PPU) chips. Here goes:

Having purchased and tested a decent number of UA6528 chips, I've noticed that some of them appear to work fine, while others produce wildly incorrect graphics, such as wrong colors, garbled sprites/tiles, unstable video, no video, and sometimes even strange scrolling artifacts. For the purposes of this post, all the chips I'm talking about appear to have their original markings, aka they probably haven't been re-marked, since their markings are printed in white ink, the package details are consistent among identical lot numbers, bottom markings match, etc.

For a while I just assumed that these chips were old and kind of flakey, so some of them had simply failed over time, presumably with malfunctioning or stuck register bits, and it had something to do with past storage, exposure, or damage. However, I finally decided to go through and catalog all of the working and non-working ones, and to my surprise, a very consistent pattern emerged. Here are the results:

Screen Shot 2022-09-25 at 12.24.50 AM.png

To summarize the obvious: CS and CB variants don't work, while CM and CA variants work fine. This was remarkably consistent across all of the 189 chips that I tested, even though the non-working chips all seem to fail in seemingly unique (and sometimes visually interesting) ways. Note that each individual non-working chip always exhibits the same issues whenever that particular chip is used (aka the specific glitches aren't random).

Can anyone explain to me what exactly is going on here?
calima
Posts: 1745
Joined: Tue Oct 06, 2015 10:16 am

Re: Mystery regarding UA6528 lot numbers

Post by calima »

Could be the failing series is typical Chinese "fourth shift", where the employees used the production line for their own benefit, and used inferior parts (which then failed quickly). Of course an official screwup is also possible.
loglow
Posts: 48
Joined: Mon Sep 19, 2022 1:47 pm

Re: Mystery regarding UA6528 lot numbers

Post by loglow »

Possible, but...

The results indicate basically that UA6528 chips from 1992 are fine, but chips from 1991 and 1993 aren't. So, it seems to me like something else (not factory shift related) is probably going on.

Also, I just thought I'd note that the tested chips were acquired from a variety of different sources over time and were not previously stored sorted by date or lot number on my end.
lidnariq
Posts: 11432
Joined: Sun Apr 13, 2008 11:12 am

Re: Mystery regarding UA6528 lot numbers

Post by lidnariq »

Is there any consistency at all across a group of broken PPUs?
loglow
Posts: 48
Joined: Mon Sep 19, 2022 1:47 pm

Re: Mystery regarding UA6528 lot numbers

Post by loglow »

lidnariq wrote: Tue Sep 27, 2022 10:36 am Is there any consistency at all across a group of broken PPUs?
Nope. All the failures are basically different, including background color faults, sprite color faults, sprite tile faults, overall tile faults, sprite position faults, scrolling faults, severe video instability, or no video at all. But I’d categorize the failures as “consistently inconsistent” across all of the non-working chips.
loglow
Posts: 48
Joined: Mon Sep 19, 2022 1:47 pm

Re: Mystery regarding UA6528 lot numbers

Post by loglow »

Just adding a link to the raw (copyable) data pictured above. Make sure to click the UA6528 tab at the bottom.

https://docs.google.com/spreadsheets/d/ ... C0BEb4VNkw
User avatar
zoinknoise
Posts: 50
Joined: Fri Mar 04, 2016 9:43 am
Location: Ice Cream Island

Re: Mystery regarding UA6528 lot numbers

Post by zoinknoise »

just thought i'd chime in... i bought some UA6528 chips a couple years ago to replace a totally fried PPU, and i did end up with chips made in 1991, with a CS suffix. the one i installed into my NES *does* work perfectly fine although i admit the colors are noticeably oversaturated. but aside from that, no problems that i've noticed.

you can see a photo in this thread.

but... after reading this thread, i wouldn't mind getting a UA6528 from 1992 to replace it!
User avatar
Individualised
Posts: 310
Joined: Mon Sep 05, 2022 6:46 am

Re: Mystery regarding UA6528 lot numbers

Post by Individualised »

zoinknoise wrote: Fri Nov 04, 2022 9:19 pm although i admit the colors are noticeably oversaturated. but aside from that, no problems that i've noticed.
UMC-manufactured PPUs have a more saturated video output than the Ricoh originals, so the chip should be working as intended.
Post Reply