I 100% agree, But the schematics that has the video RW circuitry came from a YouTube video where it was shown the designed worked. But you still have a point!

"The Design" is more than just the connections on the schematic. It is also the stackup and layout of the PCB, which affects Signal Integrity, Cross-Talk, Power Integrity and Timing Integrity. Unless you're building a 100% carbon copy of the existing design right down to the layout and stackup, you are not building the same design as the one you saw on YouTube: you are building one inspired by it, yes, but not the same one.

On top of that you've already given a long list of your inspirational sources that you've taken bits and pieces from to create your design; each of these designs may work individually but there is no guarantee that the parts you've taken from each will work in isolation, or with the parts you've attached them too.



Sorry for being blunt, but it is obvious you don't fully understand the rest of the circuitry, because you've made (that I know of) at least four fairly basic errors in your design: The one I mentioned earlier about the IRQ pins, the one Garth mentioned about how you have the two 32K SRAMs connected in parallel, and on a further review of your schematic an additional two I've spotted: I don't see any decoupling capacitors on any of your ICs, and you're using PHI2O as part of your address decoding (which, strictly speaking, isn't an error per se but is not a good idea and strongly discouraged by the W65C02's datasheet).

But that's okay! We all have to start somewhere, we all make mistakes or have gaps in our knowledge. You've come here asking for help, and we want to help, but our collective biggest concern is that you're charging head-first into a design more complex than your current knowledge level really permits, you're going to spend a lot of time and money manufacturing your board, it is not going to work and it will be difficult if not impossible to salvage. Believe me I understand your enthusiasm; when I first got into this hobby and joined this forum however many years ago it was, I had much the same mindset.

At the very, very least I would strongly recommend partitioning your design into two PCBs, one for the main computer board, the other as your "video card". Plan for the video card and provide the necessary signals on an easy-to-use connector on your computer board, but build the computer board first and get that working, then worry about video output. Just having a functioning 6502-based computer you can talk to over serial is already a significant achievement in its own right.

_________________
Want to design a PCB for your project? I strongly recommend KiCad. Its free, its multiplatform, and its easy to learn!
Also, I maintain KiCad libraries of Retro Computing and Arduino components you might find useful.

PC Building Yoshi, I've been thinking about Alarm Siren's point (above), and also the advice from others saying your design is too complex and too ambitious.

I agree that you're almost guaranteed to have some pretty serious difficulty. But to me the design doesn't seem that complex. As an alternative to splitting the design between two boards, would you be willing to do multiple iterations of the single-board version? Really it comes down to effort and cost. And PCBs have become quite affordable.

Let's imagine you do a first attempt at a single-board version. When that first PCB arrives, I wouldn't even bother populating the sites for the video components -- just leave those locations empty. Instead, just see if you can get the CPU to wake up and communicate via the serial port. Even that minimal goal isn't as simple as it sounds, and may require a re-spin of the PCB. But, assuming you've used sockets for the ICs, most of the parts can be reused.

Once you've got a version where the CPU will wake up and behave, then you can install the video parts and move on to the next level (and very possibly some further re-spins of the PCB). Are you daunted by this prospect of iteration upon iteration, or are you eager to get started? (I'm NOT trying to coax you!)

No matter how you proceed, it's true you have quite a lot to learn... about memory maps, to name just one example. I'm sympathetic to your enthusiasm, but unfortunately you're going to be forced to slow down -- a lot.

A few other points before I sign off. I'm afraid the schematic you've posted is little more than a net list. That may be alright for someone who already has a mental picture of how things connect, but the rest of us are forced to go hunting around and matching up signal names... and there are just too darn many! Since we're volunteering our time in order to help you, please provide a schematic where all or at least most of the connections are actually drawn. (Creating and studying that schematic will probably prove illuminating for you as well!)

Finally, you might consider replacing the 4116 DRAM chips with 4164. The latter are slightly less obsolete, and have quadruple the capacity. 3/4 of that will go unused, so no real difference there. But the benefit is, the 4164s require only a single +5V supply, not +5, -5 and +12. (And IIRC the pinout is almost identical -- you'd have almost nothing to change on the PCB.)

ETA: if there are to be multiple iterations then you'll do well to minimize PCB fabrication expense by shrinking the design. One way to do that (as suggested upthread) is to use a single, 128KB RAM (50% utilized) rather than two 32KB RAMs. Smaller oscillators are available, BTW. And overall there's a fair bit of empty space on the board; you might wanna tighten that up.

-- Jeff

_________________
In 1988 my 65C02 got six new registers and 44 new full-speed instructions!
https://laughtonelectronics.com/Arcana/ ... mmary.html

Good idea to consider a partially populated board! It's pretty much certain that there will be fixes needed, and probably that means revisions. (It's helpful to have a well-reviewed design but it doesn't guarantee anything.)

I would also recommend drawing up a block diagram. It would be indicative of intent: which things connect to which other things, over which busses, and what controls what.

Yes, I agree a partially populated board is also a good way of doing it, either way works.

_________________
Want to design a PCB for your project? I strongly recommend KiCad. Its free, its multiplatform, and its easy to learn!
Also, I maintain KiCad libraries of Retro Computing and Arduino components you might find useful.

A possible advantage of using two boards is that if you can keep each board to 100mm square you can get them made very cheaply. JLCPCB are currently offering 5 x 4 layer boards for $7 - unfortunately the shipping will cost more than the boards in that case. But if you go over 100mm square the cost jumps up considerably. That could be something to watch out for if you may potentially need several board revisions.

To simplify things quite a bit you might want to consider the TMS9118. It's programmatically identical to the TMS9918A, more modern, uses less power, uses only 2 RAM chips (4416s) and only needs a single 5V supply.

There is also a way to use static RAM with the TMS9918A. This involves adding some circuitry to re-direct I/O, but it manages to eliminate a few chips and the need for -5V and +12V.

_________________
Bill