If you follow the link both nodes(1329 and 1394) are at the bottom right corner:

http://visual6502.org/JSSim/expert.html?nosim=t&find=1329,1394&panx=503.0&pany=117.0&zoom=12.0

They both have an orange powered diffusion layer to their left but this is apparently separated by the polysilicon. But as I figured out, I guess in this case the nodes are actually connected and thus powered because in both cases they are depletion mode transistors.

Some questions remain though:

1. If these are transistors, why can't I find the corresponding transistor name? For example what is the name of the transistor separating node 1394 from the left powered layer in the above simulator window(see link)?
2. How can I be sure that a node is powered by default? Besides visual inspection is the correct way to look up the node nr in the segdefs.js file(warning big file) and see if there is a '+' after the node number?

Nodes 1329 and 1394 are ir5 and notir5 - they are always complementary.

They are driven by the inverters to the right, which form the Instruction Register.

To the left of their right ends is indeed a diffusion connection to the positive rail - I think of that as red but I suppose it's a deep orange!

Two things are going on: to the south, the poly stops over diffusion. Here we'd find a connection: it's a buried contact directly from poly to diffusion. There will be a mask for this but it's not part of the visual6502 polygon collection. To the north of that, the vertical segment of poly crosses diffusion, and here we get the classic self-aligned transistor: there's a thin oxide layer between the poly and diffusion. As you say, we infer that this transistor is depletion mode - again, there will be a mask for this but that mask has never been recovered because it's very difficult to make it visible to microscopy.

As usual, we see this combination of a weak transistor connected to the positive rail in combination with a connection from the gate to the non-rail side as a pullup. We can guess that it's a depletion-mode pullup but for a logic simulation that's not too important. Most likely the only non-depletion pullups are those on the pad output drivers.

Indeed, the defining annotation for visual6502 simulation purposes is the '+' in the segdefs, which annotates those nodes with a depletion-mode pullup: the simulator uses that information to pull up the otherwise undriven node, if it isn't pulled down.

Hope this helps
Ed

On my screen this appears as a deep orange. It is a powered diffusion, it will disappear if I uncheck the corresponding checkbox. What is the positive rail? It doesn't appear on the simulation but I suppose it exists in the real chip and all "powered diffusion" are connected to it, right?

Right. If you look only at metal you'll see there are gaps all over the chip. These are the locations of the power rails, which visual6502 has removed for clarity.

If you take a look at the SVG version of 6502, either at
http://biged.github.com/visual6502/
or by downloading the SVG from https://github.com/BigEd/visual6502/blo ... /6502N.svg (perhaps by saving the 'raw' link) and then viewing in your browser or a program such as InkScape, you should see all the metal and the contacts: