When you think about vintage computers from the 1970s, the first thing that should spring to mind are front panels loaded up with switches, LEDs, and if you’re really lucky, a lock with a key. Across all families of CPUs from the ’70s, you’ll find front panel setups for Z80s and 8080s, but strangely not the 6502. That’s not to say blinkenlights and panel switches for 6502-based computers didn’t exist, but they were astonishingly rare.
If something hasn’t been done, that means someone has to do it. [Alexander Pierson] built The Cactus, a 6502-based computer that can be controlled entirely through toggle switches and LEDs.
If you’re wondering why something like this hasn’t been built before, you only have to look at the circuitry of the 6502 CPU. The first versions of this chip were built with an NMOS process, and these first chips included bugs, undefined behavior, and could not be run with a stopped clock signal. These problems were fixed with the next chip spin using a CMOS process (which introduced new bugs), but the CMOS version of the 6502 would retain the contents of its registers with a stopped clock signal.
The specs for the Cactus computer are what you would expect from a homebrew 6502 system. The chip is a WDC 65C02S running at 1MHz, there’s 32k of RAM and a 16k EPROM, dual 6551s give serial access at various baud rates, and there are 16 bits of parallel I/O from a 65C22 VIA. The ROM is loaded up with OSI Basic. The real trick here is the front panel, though. Sixteen toggle switches allow the front panel operator to toggle through the entire address space, and eight flip switches can set any bit in the computer. Other controls include Run, Halt, Step, Examine, and Deposit, as you would expect with any front panel computer.
It’s a fantastic piece of work which I missed seeing at VCF East so I’m really glad [Alexander] made the trip between coasts. Cactus is truly something that hasn’t been done before. Not because it’s impossible, but simply because the state of the art technology from when the 6502 was new didn’t allow it. Now we have the chips, and the only limitation is finding someone willing to put in the work.
I don’t understand. What technology didn’t exist in 1979 or so, to do a front panel for a 6502? Was the original 6502 not fully static? It seems to me that the only chip available now and not then would be a CMOS version of the 6502 itself.
I thought we saw something like this a while back, and I pointed out there was a 6502 with front panel described in Byte.
I can’t remember how the 6502 differs from the other 8bit CPUs.
One way to get a panel is to use software. The KIM-1 was successful because it had a great monitor, you could single step and stop programs, but it was done with software. So one could modify the monitor, and have a string of LEDs instead of 7 segment readouts, and use toggle switches rather than a calculator type keypad. Have some more LEDs to show the registers. It all depends on how fancy one wants to get.
Michael
That’s exactly the difference.
The cmos variant can deal with the clock being stretched or stopped and still work.
The nmos variant can’t, a clock that falls under a certain frequency (including zero) will lose the data stored internally.
For example your program counter register contains the address of the instruction being run so that the chip knows where the next instruction will be, which is incremented by 1-3 (4?) depending on the particular instruction, and outright changed drastically on a branch instruction being true.
Stop the clock, and some if not all of the bits in the program counter register fall to zeros. Now this registers value is corrupted and your program can’t “resume” where it left off once the clock is restarted, because it no longer knows where “left off” actually was.
Stopping the clock is pretty important with a front panel like this, as most of us humans can’t flip all the switches needed at 100khz by hand :P
I saw this at the VCF and thought it was wonderfully bizarre. At the time the 6502 was popular, most people were trying to get a form factor like the Apple 2 – motherboard, expansion slots, keyboard, CRT emulating a TTY. But it was cool to see someone putting the time and effort to build a retro computer that was so fully realized. This guy was sitting next to a table with the guy with the https://monster6502.com/.
Funny thing, friend of Sam Spade, I saw the thing at VCF East as well. And did not see the Monster 6502 project sitting next to it in the second booth.
I found it to be an amazing use of technology I was familiar with, I got my start on the 6502 system, and did splendidly in assembler for it.
I did bump into the others from the group[ but not you Brian.
I built a 6502 computer with a front panel a few years ago: http://www.update.uu.se/~bjarni/brum/
It uses an original NMOS 6502. You don’t need to stop the clock, just halt the CPU. There’s a pin on it for that.
Beautifully done!
where is power on/off
emergency mushroom button
look at cosmos 1999 and old film, why this is still small led. change it to rectangle or other shapes
There was a nice Forth (FIG Forth from original Bill Ragsdale listings) IIRC. I did a ton of scientific analysis code and instrument control with that Forth, an OSI Challenger, and an AMD9511 floating point processor on one of those gigantic OSI proto-cards. Actually, that who system is now in a storage container I have been “planning to get to”. FIG Forth would be perfect for this. : FORTH LOVE IF HONK THEN ;
: HONK .” Honk!” cr ;
I found a rack-mounted box labelled “Vacscan” a few months ago, with a small CRT, parallel and serial ports, and other stuff in it. It’s some sort of embedded control system for.. something. I took out the ROM and dumped it. After a little while of scratching my head I realised that it contained compiled Forth code. It turned out to be FIG Forth. I want to try to “disassemble” the firmware to see how it interfaces with the hardware, so I can write my own firmware. It’s a pretty neat little rack-mount microcomputer with a built-in CRT after all.
Here is a friendly Honk Honk, right back at ya, clown.
pla all day
I think that should be
HONK Honk!
OK
It’s a residual gas analyzer (RGA) and there is software out there to interface to it with the existing firmware that might help you to reverse engineer the firmware and hardware interface.
Do you have a link?
excellent build.
flipping and tappin is my favorite way to learn and experiment with assembly
thank you for adding beauty to such a fun architecture.
cactus cactus
hi from roy new mexico.
now hang up the phone (& tell those kids in the background to shut up)
Hey Brian, his last name is spelled “Pierson”
Fixed, thanks.
It’s a cool project, but doesn’t seem like anything has changed since HAD last featured it. Not sure why a second post about it was needed?
Would have been nice to at least get some newer pictures of the internals. The front panel at different degrees of rotation doesn’t strike me as worthy of a gallery…
Work was done on a new serial card, iirc? But the main draw this time around was that the booth next door had the MOnSter 6502, and Alexander created a special CPU card designed for it… as a result, the Cactus is the third computer ever (and the first made by someone else) to run the MOnSter 6502! See proof here: https://twitter.com/EMSL/status/1026199308163379200