incomplete in what way?

IIRC there are missing bits of the PC increment logic.

I found one ground symbol missing on bit 6 of PCH, but I'm not sure if that's the same one you found.

Edit: misread the message about begin incomplete. You meant Balazs' schematic, right? I ended up looking at the schematic I'm working on. Although I did find the mistake I mentioned

Oh, yes, indeed, I meant Balazs' schematic is incomplete! Sorry, I haven't carefully reviewed yours - when I first came across his, it was an extraordinary experience, and I printed it out over several sheets at large scale and taped together a big version to explore. Since then, visual6502 came along and I've spent many hours going over that. So, I'm very interested, but I'm afraid I've done it twice before. Indeed, the Breaking 6502 apart thread (possibly related Breaking NES website here, in Russian) was another one I had my eye on.

I hope it's not discouraging that there have been previous low-level investigations! See this post for a previous effort to link to previous efforts.

I mainly started this as a personal project for learning so it's fine if it has been done before.

One thing that to my knowledge has not been done yet is a logic gate version, although there are some mosfet setups that I'm not sure how to convert to a corresponding logic gate yet, so that might be an issue at this point.

One question I have is what is the difference between these two devices:



Both seem to produce the same output, but the other one has 2 additional buffers(?) between the NOT gate and output.

Logically, those are the same. When you see a pullup transistor driven by a logic signal, rather than the usual arrangement where it's driven by the output of the pulldown tree, that's called a superbuffer or a superinverter. The idea is that the usual NMOS structure doesn't pull up terribly fast, so when you need a lot of drive and when you need a fast turn-on, you use a superbuffer.

In the second case, the intermediate driving stage is also a superbuffer: I would expect we'd see that the final driving pair are very large transistors, which need a substantial drive just to turn them on and off quickly. If the final pullup and pulldown were both on there would be a lot of crowbar current straight from the power rail to the ground rail, which costs heat and causes electrical noise. So there's an incentive to minimise the time that they are both on, by driving them hard in both directions.