Sorry, but had to share. This has been a side project of mine for most of covid, working evenings and odd weekends. If you recall, Ben Eater built a 6502 computer and made it say “Hello, world!”. Once I had mine version working, I thought it would be cool to build my own 6502 and replace the 65C02.

Today’s the day it came alive and said hello!




Hopefully the images will show up. The CPU itself is fully wired on the breadboard, and my next step is to convert it to a pcb. The wires you see bottom right are to connect the 65c22 from Ben’s build to the Led display, as well as to the nano on the board that’s acting as both the debugger, memory, and microcode eproms. This allowed me to load test programs during development and program the microcode. Now that the microcode is complete I’ll program 3 16-bit roms to provide the 48 lines

So happy it works!!!

the images didn't come through for me -I think that's due to the way your image host (imagebb) does things - it probably doesn't like them being embedded that way.

I've uploaded them here for you though:




Looks good!

Cheers,

-Gordon

_________________
--
Gordon Henderson.
See my Ruby 6502 and 65816 SBC projects here: https://projects.drogon.net/ruby/

Very breadboard! Much wires! Well done!

Thank you, Gordon.

Congratulations on your success, Noggin! An exciting moment, I'm sure!

Are you planning to run Klaus's test suite? It has a long history of revealing unsuspected bugs. I think it can be found somewhere near here: ttps://github.com/Klaus2m5/6502_65C02_f ... /bin_files

-- Jeff

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

Oh, yes. That's a great idea! I have one more test to write for interrupts and RTI, so once that's done i'll give Klaus's test suite a run. Should be fun

Very impressive work Noggin! It must take a great deal of perseverance to get to this point. Have you hit any manor roadblocks along the way or did everything work nicely together?

_________________
Jonathan Foucher

Take a look at the Planck 6502 computer.

Definitely lots of hurdles along the way. Most of the hardware was built without much understanding of the timing or nuances of the design (especially where they weren’t covered in Donald’s block diagram). So as I programmed the control lines for each opcode I had to make quite a number of ‘hacks’. These usually meant pulling control lines half way across the board to reach a free logic gate in order to hard wire a conditional fix, such as the final one which was to skip the program counter increment whenever an interrupt fired.

Next, I plan to make a pcb from the design. This will likely take longer than the building itself as I have to reverse engineer the whole circuit and all of the hacks.

I tried to run the Klaus test but not sure I built it correctly. Having first installed a windows VM so I could compile the test, I loaded it into my debugger and each starting vector (nmi, reset, irq) just jumped to a loop? Am I supposed to change this to $1000?

Edit: Actually, the next step is to remove the circuitry on the bottom right two breadboards and replace this with the 3 programmed proms that will hold the microcode. If I also add some dedicated memory then it'll be completely standalone and I see how fast (or not) it'll run

You have to load $400 in the PC and then set it loose : https://github.com/Klaus2m5/6502_65C02_ ... st.a65#L37

I wonder if / how you can achieve that on your build though...

_________________
Jonathan Foucher

Take a look at the Planck 6502 computer.

Thanks for that... I guess that's what I get for not paying attention to the read me . I'll see if I can run the test over the weekend

I had some time today to run klaus’ test and I’m beginning to think it’s failing. Due to the debugger it only runs at around 4 cycles per second and has been looping over and over around what appears to be the same set of 256 dex instructions - give or take a few branch / jump instructions. So far it’s completed 14500 cycles and I’m not sure if this is expected or if it’s stuck in a loop? It’s definitely not a tight loop so I’ve left it running

The full test takes a little under 100 million cycles - it's best to run it as fast as you can, with the only overhead being to detect the failure case. By default, I think, it fails into a tight loop, which can be detected in various ways.
See this thread
viewtopic.php?f=8&t=6202