Paula Internal Logic

[bloodline]

There have been a few fascinating threads recently about the Amiga's audio generation system, and we've had some great insights from people who have taken the time to probe the various outputs from the Paula chip, but there is some debate around how the chip actually functions.

I noticed that there is a decapped die image on Wikipedia, and I wondered if anyone here is able to provide some annotation to the structures which are visible?

It seems likely that the large regular structure to the right is the audio generation, as it is clearly divided into 4 identical units, and there each unit appears to have 16 subunits (two 8bit samples?). But beyond that, I can't really determine much else.

[roondar]

I'm always fascinated by die shots. Not that I really understand them, I've always been a software guy first and foremost

[KONEY]

!

[Rotareneg]

Or if you want a little more detail, how about an 22,836 × 17,625 version of the same image, which I strongly suggest you don't try to directly open in a browser:

https://upload.wikimedia.org/wikiped..._top_metal.jpg

Looks like the audio outputs are on the top right, above and left of the 1984 CBM text. Each pair of channels looks to be mixed right at the pads?

[bloodline]

Originally Posted by Rotareneg

Looks like the audio outputs are on the top right, above and left of the 1984 CBM text. Each pair of channels looks to be mixed right at the pads?

Hmmm, I’m having difficulty following the traces... I assume the analog section would be isolated from the digital section to reduce noise, so that would explain the apparent “gap” near the pads.

Really we need someone more familiar with the old MOS fab process to explain what we are looking at!

Trying to match the die shot with a pinout of the chip package is not helping me at all!