Writeup here: http://www.glitchwrks.com/2022/10/25/dec-scholar

I picked up two of these last year because the lot came with the power adapter, which most of these seem to be missing. It's a triple-voltage wall wart (pinout in the writeup). I was expecting just another 2400 BPS non-Hayes command set modem, but found this inside:

(image uploaded separately, apparently no IMG tags here?)

It's a Rockwell R65F12 FORTH chip! I dumped the ROM, TangentDelta is working on deconstructing it (he's getting the second modem). The ROM dump is linked in the writeup, too.

So, yeah, it's a non-Hayes 2400 BPS modem, but it's neat to see a mass market product from a large manufacturer that includes a Rockwell FORTH chip!

Very nice! I have a vague recollection that these kinds of modems were at one time a major volume of 6502s in the field.

Wouldn't be surprised! A friend who used to run a printer repair business said he's seen a lot of Rockwell QUIPs in old line printers, too.

Wow, I read your web page and it was a blast from the past.

I worked at DEC and we often joked that DEC was one of its best customers. Employees were loaned obsolete equipment to use from their homes. My take home equipment was a VT-220 and one of those Scholar modems, so seeing your setup was nostalgic.

I used that gear pretty frequently when I needed remote access or to telecommute on snowy days. Like most DEC equipment it had a proprietary command set, and I no longer remember it because it has been over 3 decades.

I didn't know it was using Forth and a 6502 internally, which would have increased the cool factor at the time. I kept it on my desk right beside my Macintosh SE30, and you would think I would run a terminal emulator rather than a bulky terminal and modem. But at the time a hardware terminal was a better option, especially when using 132 column mode.

Very cool! I can understand wanting to use a VT-220 instead of emulation on the SE/30 -- you do end up having to use a LK201 though The Scholar definitely has some features that makes it nicer to use with a dumb terminal, too.

TangentDelta has been working on disassembling the ROM, and it's looking like mostly 6502 machine code. He says a lot of the code doesn't make a ton of sense from an optimization point of view, and looks like it may be output from a higher level compiler. We're thinking maybe their RM65 FORTH development system, which according to some of the FORTH chip datasheets will output machine code.

It's been a while, but I seem to recall multi-session support by both the terminal, modem, and terminal server.

Like many users I accessed different environments concurrently. In my case I needed to get email on a production machine, code on a development machine, and test on an integration machine. In a modern setting you would have multiple terminal windows open and switch between them. But in that era a workstation like that cost a fortune, so terminal was a more common solution. It took a bit of getting used to, but you could pretty easily juggle two or three sessions that way.

I'd be curious to see the code, especially if it's can be turned back into the Forth source code.

Later Rockwell modems used a chipset with one of the chips being a 6502 based microcontroller, this started with the R65C19, which was similar to the R65F12 but with an added host port that could emulate a 16450 serial chip, and added Forth instructions. The last of the line was called the L28 and could run at up to 35 MHz. So a lot of old Rockwell modem cards have a very nice 6502 based micro-controller on them. These microcontrollers also have a hardware bank switching system which allows access to up to 512k of RAM. A lot of the modems had 128k ROM and 128k RAM. I haven't found any documentation on the software side, but since they went to the effort to add special instructions for Forth it would be logical to think that they did use Forth in the modem firmware.