SCN68681C chips have been discontinued for years.
It took a long time to find a source, and on the first try I managed to get NOS (new old stock) from warehouse stock.
Condition is excellent: clean markings and fully working parts.

Started testing Philips SCN68681C1A44 DUART chips.
DUART is the main interface for rosco_m68k, and even as an alternative chip in the BOM it managed to run reliably at 38400 baud — after modifying the firmware to reach that speed.
Pretty solid result for hardware that’s long out of production.

The chip contains undocumented instructions that still execute consistently on real hardware. Developers discovered them by experimentation, and some are faster or more powerful than official ones. They’re risky — but legendary.

Unlike many CPUs of its era, the 6502 executes instructions mostly through hardwired control logic. This makes its behavior very predictable and also why cycle-exact programming and demos are even possible.

The 6502 has a tiny transistor count compared to its competitors, yet delivers impressive performance per clock. Fewer registers, simple addressing modes, and tight timing made it fast where it mattered — real code on real hardware.

Apple I and II, Commodore PET and C64, Atari 2600, and the NES all ran on the 6502 or its variants. If you used a home computer or console in the late 70s–80s, chances are it had a 6502 at its core.

MOS 6502: 5 Facts You Should Know
When MOS released the 6502 in 1975 for $25, it undercut Motorola and Intel by almost an order of magnitude. The low price came from aggressive chip layout and cost-focused manufacturing. This single move opened the door for affordable personal computers.

Fixed the contact temporarily for testing — and the system booted immediately.
Lesson learned:
check IC socket pins, not just solder joints.
A single bad contact can make a computer look dead.

First thought: RESET button.
Desoldered it — no change.
While reflashing the PLD, I noticed the real issue:
a bad IC socket pin — no proper contact at all.

Assembled my rosco_m68k kit (~4 hours, careful build).
Powered it on — no boot.
Only START and RESET LEDs were on.
RESET pin on the CPU was sitting at ~2V. Definitely wrong.

PCBs finally arrived!
Boards for the rosco m68k, 6502, and SLC Free just landed. Half of the components are already sourced.
Also didn’t expect the rosco 6502 to be this compact.

Never seen a pile of 68010 CPUs like this.
Almost 30 years old, but still perfect for learning real low-level computing.

At launch, there will be two kits: - rosco_m68k — a single-board computer based on the Motorola 68010 - rosco_6502 — a single-board computer built around the W65C02S8P-10

I’ve been working on my project SolderDemon for about a month now. It’s a platform that helps you learn the basics of computer science, experiment with retro hardware, and write real low-level programs.

SolderDemon: solderdemon.com/en