National SC/MP micro

Aug 16, 2025 Last reply: 11 months ago 32 Replies

A bunch of stuff about our late regular, Don Lancaster too



formatting link
Cheers



Phil Hobbs



-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC / Hobbs ElectroOptics Optics, Electro-optics, Photonics, Analog Electronics


Yikes. A Raspberry Pi Pico costs $6, in degraded modern dollars.

My first computer used SC/MP

formatting link
taught me everything I needed to know about what RAM was, what ROM was, what firmware (known as "the monitor") was, how a microprocessor was interfaced to external devices, what an address and data bus was, how to code in machine code, what pointer registers and accumulators did, etc. Programs were stored on my reel to reel tape recorder. I still have it in a box here somewhere, along with an MK14 VDU.

My favourite was the 8749 ( UV erasable version) and later the NSC 800

My first computer was a DEC PDP-11, with a teletype and 4K words of

12-bit core memory. I used it to simulate 32,000 HP steamship propulsion control systems. It cost $12K, the price of a modest house in New Orleans.

I was a freshman EE student at Tulane.

VAX/VMS had taken over the mainframe world when I was a student. Using a computer meant sitting in a room which had maybe 20 VT100 terminals. Some students amused themselves by writing command procedures which would make a terminal look like it was available for login. An unsuspecting student would then come by and after trying to log in would be told that there was an issue with the terminal so please use another one. The student running the program now had that student's login credentials. I never did anything like that myself of course.

This problem was subsequently addressed on other platforms by the now mostly forgotten ctrl-alt-del before logging in.

Someplace deeply buried I still have a few i4004 chips...from the days of program entry from switches & paper tape!

PDP-11? That was a 16-bit machine, you must be thinking of the PDP-8.

Unusually the UK version of this - Clive Sinclair's MK14 was a year earlier actually cheaper £40 and had integral display and keyboard.

formatting link
It was sold as a kit. Putting it together wasn't entirely trivial.

I recall a similar kit "ELF" at about the same time using the slightly more esoteric RCA 1802 COSMAC device which in one form was space hardened and in another became a staple of various games machines.

formatting link
Then there were a host of 6502 based machines with video all at once. The quirky 6502 inspired ARM's design of RISC chips still around today.

The 6847 (NTSC) and 6845 memory mapped display chips made it a lot easier about then. Acorn Atom used the former and BBC Micro used a custom chip that was sort of like a 6845 with 640x480 maximum resolution (at two colours). Acorn managed to cobble the 6847 NTSC timings to work OK with UK PAL TVs somehow on the earlier Atom product.

I always had a bit of a soft spot for the TI9900/99k and support chips. I didn't appreciate at the time quite how good they were for realtime context switching until we tried to do it on a 68k (on paper faster).

The 9918 graphics controller was impressive for the time with 256x192 resolution, hardware sprites and relatively affordable.

I later guessed horribly wrong backing the NEC 7220 against IBM's ropey E/VGA. Japanese market standardised on it for domestic PCs NEC9801 - essential for Kanji displays. It was the first true graphic coprocessor.

There was an ancient PDP-7 with vector graphics display available to play with. It was totally unsupported and kept alive by the goodwill of graduate students.

Papertape input and you had to bootstrap with a magic start address set on front panel switches. Asteroids was a favourite when it was working. Vector displays and joysticks were still rare back then.

Sorry, it was a PDP-8. I got an 11 a couple of years later.

The PDP-8 ran Focal, a nice interpretive language, in 4K words of core.

The 11 ran RSTS, a wonderful time-sharing system.

It's tragic that Intel and Microsoft won.

We are developing a new product line of PoE boxes based on the RP2040 chip. I was going to use a Pi Pico as a surface-mount component, but it wastes too much board area and has a goofy USB connector and not enough i/o pins.

We have defined a tiny 20-pin ribbon cable connector that will run to a custom adapter board and then to a Pi200, so we can compile and load and test code.

formatting link

The middle board has scope connectors to make it easy to time routine execution and irq rates and such. Most programers don't have any idea of how long it takes to run their code.

The obvious c code sequence

set port pin high set pin low etc

makes 7 ns pulses every 14 ns on a 2040. That's amazing.

We have a cute circuit on the blue boxes. There's a pushbutton available through a hole, with a toothpick or something. A short push resets the uP. A very long push resets it and enters USB memory-stick mode. That makes it easy for our customers to upgrade code; long push and then drag/drop one file.

Mine worked almost first time after assembly. It's hard to remember all the details now but the keyboard was useless so I built my own. And when I added the VDU I had to change the crystal from 4.43 to 4 MHz.

There was one little issue with the processor's SA and SB inputs being open circuit on the edge connector. For some reason this caused programs to fail to execute when pressing Go. It taught me a valuable lesson about checking for unused inputs and using pull up or pull down as needed.

The 6502 felt very restrictive after using SC/MP because of no on-chip 16-bit registers. I liked the 6809 because like SC/MP it had two accumulators and on-chip 16-bit registers. After that I got into 68k. The only thing I disliked about the 6809 and 68k was that they are big endian. Intel got it right with little endian.

After the MK14 I never bought another Sinclair machine, so no ZX81, Spectrum, or QL. The QL looked tempting but my experience with the MK14 said wait and see.

SED is increasingly about connecting computing boxes and coding them, less and less about designing circuits. That's the trend in academia too; kids these days type and don't solder.

AI is wiping out the coding jobs, so lots of ee grads are unemployed, and $100K starting salaries are uncommon lately.

I didn't "design" it, I designed it. No PIC is going to make a kilovolt.

This is better:

formatting link
The LTC chip is inside my control loop and the output can't go above

650 no matter what the user wants. That's important when driving GaN fets. Some things move around in GaN at high fields.

An all analog design doesn't need code to be written and loaded and documented. We do all that when it's necessary.

Typing thousands of lines of mostly unreadable punctuation is silly. Something more like LabView is the future of programming.

Are youws calling me a religious finatic? That's wrong and obnoxious. I'm finatic about electronics, women, and skiing.

The history of Europe is warfare. Nothing new there.

They were skilled at that before Columbus arrived.

Obviously bogus.

"The US military has minimum enlistment standards at about the IQ 85 level."

He's a bilionaire and the President, and you're not.

I didn't see anything about Don at the above URL.

Is there a schematic somewhere?

Before:

I somewhat cleaned up the above schematics using ChatGPT.

formatting link
formatting link
formatting link

  1. Go to:
    formatting link
  2. Upload your illegible and faded pencil scribbles to ChatGPT. Just click, hold and drag to the AI.
  3. Ask the AI to clean up the mess with something like this: "Enhance the lines and text. Make it look neat and readable."
  4. Download results. It took about 4 minutes per file.

There are some errors where the AI couldn't read your symbology or scibbling. To be fair, I couldn't read them either. You can probably fix the text and symbols with a vector graphics editor such as Inkscape or LibreOffice Draw. I haven't tried these on your schematics yet.

I hope this helps.

Follow the ‘TV typewriter’ link near the bottom.

Cheers

Phil Hobbs

An arcade 'game' named Bio-Rhythm III by Compunetic Devices, from the late 70s, used the I4004, along with 4201 (Clock), 4001 (ROM), and 4002 (RAM).

Odd machine, and folks here may recall the Bio-Rhythm craze of the 60s and 70s. About 2/3 of the way down this page:

formatting link
(remember when .htm(l) pages could have only 8 characters for the name? I never change page names if I can possibly help it...)

On a side note, I have something like 16 of these machines still packed in their factory boxes in my warehouse. Trying to figure out what to DO with them!

John :-#)#

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required