This computer was build by grade 12 students at Vancouver Technical School. I still have one. I ordered the keyboards from a surplus electronics magazine.
@@gesnowCool. We had a Netronics ELF II with 8K RAM, a video card and ascii keyboard at my school in the electronics shop. We used a dial up phone with a handset modem to connect with a BBS where we could chat with VanTech. That was around 1981. We also built an 1802 board that if I recall correctly was a joint project between PW and VanTech for robotics.
This system looks like the 1802 computers we built in VanTech in Vancouver. We also used a commodore tape drive for back ups and restore. That planner board should have 5 to 6 rows. The other slots are for memory upgrades.
The intention of the additional slots may have been for additional peripherals such as a serial port. I designed and built my own 1802 based computer about 40 years ago. I still have it and was recently getting it up and running again. I still have an original copy of the RCA 1802 User Manual from back in the day, photocopies of the articles about the COSMAC ELF, and several program listings in 1802 assembler.
Thanks for an excellent video that brings back memories. I built a NS SCMP sbc back in the 1980s that used switches for data input, a push button to Write a value to RAM and another push-button to advance the program counter. Oh how things have changed.
This thing is gorgeous. This esteemed graybeard was able to get over the DIY enclosure hump so effectively, the PCB's are double-sided and gorgeous, mass storage, keyboard, and video out. So many hurdles, even for modern homebrew hobbyists, accomplished maybe in the mid 1970's. Hats off to the architect of this!
beautifully built - TBH, with 1k of RAM there won't be a BASIC interpreter for the machine that would be workable. I completely agree that it was built by somebody who knew what they were doing, with a lot of pride, and was likely a stepping stone to the next step. You will find that the ROM contains a full monitor. potentially a disassembler. Likely has tape read / write routines Great find!.
I’m really surprised they wasted 5 whole bytes of ROM on that “READY” prompt. The “>” would have been quite enough. Or, if it was a boot-up self-test, a simple “OK” would have sufficed.
@@lawrencedoliveiro9104 The old TRS-80 model I had a "READY>" prompt implemented also, and this machine was released for sale somewhere before 1977; so it's possible whomever built this home brew computer, may have been trying to emulate the TR-80 to some degree Saving 3 bytes wouldn't make a difference in a 2k EPROM. No matter how efficiently a machine language start-up program and support routines (e.g. Keyboard address to ASCII conversion, ASCII to pixel character matrix video conversion, Mass Storage I/O) may be; there will always be several bytes left over anyway. So whatever would've "sufficed" is a moot point and subject to the system designer's prerogative. If you'd like to show us a system that saves a few bytes with prompts and messages that "suffices" please feel free to produce a video showing your completed home brew 1802 based computer project. Please enlighten us by mentioning what additional and useful routine you implemented with the few bytes you saved, that also fills up the entire 2048 bytes of EPROM memory space, with zero left over bytes. You don't need a CRT based video display, as many modern wide screen TVs have analog RF and composite video inputs. The "READY" prompt wouldn't necessarily be copied from the EPROM into the 1k system RAM. This prompt could be copied as an ASCII character string from the EPROM, encoded as a matrix of 5 x 7 pixels (or 5 x 8 if using lowercase descenders) straight into the video RAM. Where did the missing 15 bits go? They were diverted into a bit bucket placed at the bottom of the desk under the computer.
15:42 Hard to believe someone would have “moved on” and simply abandoned such a lovingly-put-together piece of work. I have feeling the creator of this thing could have died, and then the family, going through their stuff, saw this as little more than “junk” and got rid of it ...
What a beauty! I hope you make a follow up video very soon when you found out how to load a program. I see that the Yuotuber 'Tr Een' in the comments below appears to know what MACBUG is. I hope you solve the puzzle and keep us posted.
Wow! It’s a beautifully designed machine that’s for sure! Also the 1802 world did have a “killer app” in some respects with the “CHIP-8” programming language and its variants.
But CHIP8 was in many ways tied to the CDP1861 graphics chip, which this computer does not seem to have. An antiquity now, but a small wonder in the days when you could not buy computers in a store yet.
@@CDP-1802 Very true. There is some 64 x something graphics mode in the 6847. Taking apart CHIP8 and learning how drawing sprites worked was one of the first things I did long ago.
Neat computer. The Netronics ELF II was my first micro in 1980; it had a 1/4 K RAM, which I expanded with a 4K board. Tiny Basic takes up 3.5K RAM if I recall correctly. The first online action I had was in 1981 on a BBS. Dial up, Penny Whistle modem. Then I got a Sharp pocket computer with a Z80, Basic interpreter and 24K RAM. That pretty much left the ELF II in the dust. I've still got a couple of ELF II's and Super ELF's a COMiX-35 (reproduction) and a couple of others with a disc operating system called ELF/OS. I've got a RCA Studio II Game Console as well. Hindsight is always 20/20 but I really think RCA missed the boat. The RCA VIP2 was advertised but never released and it would have been similar to a VIC-20 but a year before the VIC-20 was released. If RCA had taken on the challenge, they could have been a contender. The downside is that the Basic interpreters for the 1802 are pretty slow with loops and some math routines. The ability to change the address register on the fly makes it very efficient for programming but it was pretty sluggish. That said, when 6502's were humming along at 1MHz or 2MHz, an 1802 will do 6MHz without breaking a sweat and an 1806 will do 8MHz - 10MHz.
My dad used a heavy printer case to build an Apple ][ into. My very first experience with computers in the late 1980s, with a Zenith green phosphor monitor. My mind was blown in the 2000s when he hooked that up to the same television set that we were using in the 1980s and I saw that it had color.
Gorgeous. Something similar I dream of building when I retire in a few years, but first have to live that long (thanks to USA Social Security moving the goal post). Decided on the RCA 1802 over Moto 6502 because of the architecture. That and I'm a loner, Dottie. Well done.
Excellent video, very interesting and informative. Your production quality is always very good and I hope you can succeed on youtube, you put in the effort and it shows!
@@Nf6xNet In a way, yes. Microcontrollers often have Harvard architecture, that internal memory is a huge set of registers and that internal ROM often has weird word sizes, so that it can fetch one complete instruction in one go. Microprocessors are all Princeton (aka von Neumann) architecture and rely mostly on having both program and data residing in external memory. The 1802 simply extends the use of its registers to also function as memory pointers. The one you make the program counter then is simply used to fetch the instructions. I think that uniting these two architectures and trying to use the advantages of both actually is what RISC is all about.
I am a sucker for all things: A: Homebrew... B: Wide format... C: Vintage computing... This checks ALL my boxes! Good Lord, this machine is a real beauty! I love it! :D
Amazing find, amazing build. I happened to find an 1802 homebrew computer a few years ago, I'll go find it later and add a comment about which one of the 3rd party ones it is specifically, but there was a standard for expansion board connectors that was common (if anything was ever common with that rare processor) for homebrew. In my case the backplane had the spots to add the slots but it had the 1802 & VIP onboard. The slots themselves were never added originally and are absolutely impossible to find these days.
Netronics ELF-II is the one I have. I'm pretty sure I saw in my research about COSMAC stuff back when I got it there were two 'competing' (and i say that with big idiot air quotes because how can there be competition in such a small market) connector standards, one of which was the Netronics slot connectors and the other was based on dupont style pins, which is what you have. I would be very, very surprised if any of this was truly home brew and not something purchased via mail back in the day. Which is still home brew, but not 'mf had a vat of acid in the back yard to etch boards' home brew. I also really doubt that case was totally homemade as well, the holes are way too regular.
@@treennumbers Which is weird because I went through searching for mail order options as well but never found anyone selling kits or empty boards in a similar configuration, or even people just posting etch layouts. Also I am seeing the other comments with the URL links to Tiny BASIC that are being auto-moderated. I can't control that from my end but thanks for the links as well.
The CDP1851 is a dual 8 bit parallel port, very similar to a Z80 PIO. It's quite obvious why it sits right next to the keyboard connector. And don't call little old CDP1802 weird. It' was already CMOS back then and the 'weirdness' is mostly what's now called RISC. Got one going at 7.3 MHz on a breadboard. And you are wrong: The 1802 can traditionally do without a UART because it has one instruction controlled output pin (called Q) and four separate input pins (called EF1 - EF4). Looking at the board of that tape interface, EF2 was used on this computer for the cassette input and deserialisation was done in a small software routine. My breadboard computer talks to a terminal (actually a PC running Tera Term) at 19200 baud that way, using Q as output and also EF2 as input. The only other part inbetween is a MAX232 to do the RS232 level shifting back and forth. A UART was expensive in these days and today on a breadboard thatÄs one less thing you need to wire up. Overall, you would find that most 1802 based computers are very similar at the core. The ROM probably is addressed at 0000, RAM and video memory somewhere above. With RAM at 0000 the computer would have nothing to execute at 0000 after a reset and this computer does not seem to be designed to use the 1802's LOAD mode, like the Elves did. The only remaining question would be to which ports the 1851 is mapped and how the EF1 - EF4 are used. Adapting Elf software should not be too hard.
Just learnt from other video that it had 4bit = 16x separate program counter so basically "16 slot syscall slots" right? Were these EF1-EF4 only flags for IRQ or was that direct "hard selector" of program counter? Q was sort of 1 bit programable port for "acknowlage". Id had freaking DMA! Just WOW. Normally had one single address space so it was possible to build computer. What a neat little monster. What i reed it executed one instruction in 8 to 16 cycles that dont make it "RISC" but transfering data internally was RISC like LD, ST and that data bus 8bit what a bottleneck. 6502 had exposed all 16bits. Im nobody to to make judgment. But overall it's impresive design extept 8bit bus and lack of registers like Z80 im not surprise by quick look that Z80 was so popular
@@KabelkowyJoe Q and the EFs (external flags) had nothing to do with DMA. Q was a simple output bit that could be set or reset with instructions. The EFs were input signals that were controlled by external devices and could be tested by the 1802 with instructions. With only little extra hardware could turn Q and one EF of your choice into a simple serial port and communicate with a terminal. Everything was done in two small and carefully timed routines. That's what I have on the breadboard right now and I have it working with up to 19200 baud with an old PC running terminal emulation on the other side. And DMA? That worked a little differently than it did on other processers at the time. The 1802 never really yields control of its bus to anyone that asks politely. Instead, it adds another bus cycle to the current instruction in which it does the memory addressing itself and signals the external device that now is the time to read or write its byte to the data bus. If there was more than just one byte to transfer, the requesting device would simply keep requesting DMA and wait until the 1802 serviced that request with the next instruction. The 1802 would use register R0 (one of its 16 general purpose registers) as pointer to the DMA buffer and increment it after each DMA request. This way of handling DMA allowed even early DIY computers to have graphics by simply adding a single IC to the system. That was the CDP1861, my first way ever to put pixels on a screen. And why am I still a fan of that primitive old thing? Because it does the job in a smart way, instead of throwing an expensive board full of memory and logic at the problem.
Initially, I thought the home brew RCA 1802 CPU based computer you demonstrated, may have been the one called FRED (Flexible Recreational Educational Device) built by Joe Weisbecker c.1975. Once the TH-camr went through the hardware, I realized it couldn't be the Weisbecker design because it used a Motorola MC6847 Video Display Generator chip, which was released commercially some years later. Joe Weisbecker built the prototype of what was going to be first the two microchip 1801U and 1801R CPU commercially released in 1975 and a year later they were combined into the 1802. Weisbecker built his prototype using 100 discrete logic chips in October 1971. The 1802 CPU was marketed by RCA to Chrysler, as a car engine spark management system to improve fuel efficiency. The senior management of RCA did not share Weisbecker's vision, that an affordable microcomputer for home use for educating a child and entertaining people at home, was feasible with the 1802. spectrum.ieee.org/chip-hall-of-fame-rca-cdp-1802 Two hour video: RCA 1800 Microprocessor Family Oral History Panel th-cam.com/video/pmGkgoH3kO0/w-d-xo.html
Love that LaserJet 4. I sold a lot of those back in the day. Back in 1993 HP offered us a 75 dollar spiff for each one sold. I think I sold ten that month.
I know some computers from a little before this one's time would use the tape drive to record programs as key codes, but then basically just emulate the keyboard as it entered the program "by hand".
I was just playing with my RCA Cosmac Elf that I built back in 1976. It has the ceramic 1802 processor. That's the original rad hard version like was used in the Voyager probes that are still out there working.
Hi, The RCA game console is called the Studio II. What an amazing find! I would love to get detailed pictures, schematics and a ROM dump to post on the COSMAC forum.
Regarding the uses of the 1802: I understand the ELF and similar systems were popular among hobbyists in the 70s because they were a really cheap way of getting started with computers. But the 1802 found its real niche in space systems, as you mention, because it was the first readily available radiation-hardened microprocessor that could withstand the radiation levels in space.
Another reason was unlike the original 1970's CPUs based on PMOS (e.g. National Semiconductor's SC/MP aka "SCAMP") and later NMOS fabrication processes with dynamic registers such as Intel's 8080 and Zilog Z80, the 1802 was the first CPU that used CMOS transistors and had static registers, this meant it could be put into a timed stand-by mode simply by gating the clock to an inactive state. Whilst the 1802 was in a standby state, the CPU drew very little power and the static registers meant the registers' bits' were retained. This made it ideal for Satellite and Space applications, where power budgets are always tight. On a side note, the static registers of the 1802 also simplified the implementation of single-stepping for the CPU machine language code, to enable debugging of the programs. The implementation of single stepping on the 1802 could be easily implemented in hardware, with a flip-flop, perhaps one or two spare logic gates and debounced switches, one to enable single-stepping and one to step through one of 16 registers designated as the Program Counter address register.
I could be wrong but I was under the impression the RCA1802 was developed for military and space in ceramic packages but then was packaged in plastic to try and get into the consumer market. I was involved in thick film development in the early 80s and I know the 1805 was one of the only bare chip CPU's available.
That fits with my recollection of the CDP1802, which is the first "computer" I had contact with. I seem to recall seeing a schematic for it in Popular Electronics magazine, but I was so new to electronics my attempt to produce it failed miserably. Eventually I bought a Cosmac SuperElf and greatly enjoyed assembling it. However I was not yet exposed to computer programming languages so I didn't really implement anything with it. Later I obtained a commercial desktop computer and also gained experience with microcontroller so I lost interest and gave it away to a small electronic museum collection. Regretted it immediately, it would be fun to play with today. Now if my old Radio Shack PC1 and PC2 computers still worked....
I was a fourth year electrical engineering student at the University of Toronto in 1978 when a group of us each bought an 1802 kit from a fellow named Eugene Tekatch. It was a single board computer similar to the ELF. It had a whopping 256 bytes of RAM, output was 8 LEDs to be read as hex. Input was through a hex keypad using the DMA feature. There was no ROM. Not particularly impressive to my family and girlfriend. So I expanded it. I wire wrapped two boards with 4k RAM each. An S100 connector could fit onto the side of the board (although obviously the pinout was different). Being an engineer and just wanting the thing to work, I didn't build a beautiful case. I took a thin sheet of plywood and built a backplane by cutting slots and screwing in S100 connectors which were tied together underneath by ribbon cable. Then I got Tom Pittman's Tiny Basic on Cassette, built a cassette interface using the Q line for output and one of the EF lines for input to the computer. Since I had a Ham radio license, I was eligible to get a working surplus KSR35 teletype from the telephone company. Built a 20 ma current loop interface and connected it to the computer again using Q and EF (I got a lot of help on all these from articles in Popular Electronics and a local newsletter created by 1802 enthusiasts called Ipso Facto). The teletype was controlled by software. Since there was no ROM, a cold start was a bit of a chore. I had to enter a 25 or so byte bootstrap program in Hex to read Basic from the cassette. If all went well, I was rewarded with a greater than sign prompt on the teletype! Then I was away to the races! I could program in Basic and store the program on the cassette. *That* impressed everyone. I thought about but decided against getting the 1851 chip which would output composite B&W video on a TV because I didn't think it was worth the cost. The graphics were very chunky and not impressive. I still have most of the hardware. I had to give up the KSR35 when I moved overseas 10 years ago as it weighed more than 60 pounds. I donated it to UCLA who were looking to recreate what the room where the first internet message was sent. The original room had a KSR35. That unit was a tank! Our basement flooded twice, drowning the teletype. Each time after a week or so drying out I turned it on and it worked fine!
The Galileo space probe used 4 RCA1802 CPU's. "The spacecraft was controlled by six RCA 1802 COSMAC microprocessor CPUs: four on the spun side and two on the despun side. Each CPU was clocked at about 1.6 MHz, and fabricated on sapphire (silicon on sapphire), which is a radiation-and static-hardened material ideal for spacecraft operation. This 8-bit microprocessor was the first low-power CMOS processor chip, similar to the 6502 that was being built into the Apple II desktop computer at that time." -- Wikipedia
RCA 1802 is an interesting, if a bit weird, CPU. It features SOS (Silicon on Sapphire) chip production technology which basically make it nuclear bomb proof, meaning you can keep X-raying it with something brighter and stronger than dental X-ray machine, it will continue to march on without any hiccups. Thus, RCA 1802 was also a decent choice for avionics and satellites that must survive the punishing shines of Gamma-rays from Sun. It's kind of wild that childrens and adults alike got to play with the military technology, occasionally in form of ELF single board computer kits which wasn't too expensive to procure back then.
A real nice setup, perfect monitor, now all you need to do is soft out that keyboard cable, then maybe design a memory expansion board for it. Pity there is no serial port, it would make for a very retro terminal.
@@CelGenStudios You don't need to add a UART. The 1802 can do that itself on its Q signal for serial output and any one of the EFs for serial input. You just need to take care of level shifting these signals from and to RS232 levels (using a cheap MAX232) and add two short subroutines to send and receive bytes. That's it. The tape interface works in a similar way. Look at the names of the signals etched onto the tape interface. Q and two of the EFs. Things like this made building an Elf the easiest and cheapest way to build your own computer in the 70s. If only memory had not been so expensive...
@@CelGenStudios Be a little careful with the MAX232. One of the external capacitors must be connected to VCC or GND, depending on the manufacturer. Some even don't seem to care what you connect it to. I fell into that little trap when wiring that up on the breadboard. I used the TI datasheet and my MAX232 was from MAXIM. And of course they connected the capacitor differently. It took me a second to spot and correct that. The only other trap is adapting the two small subroutines you will need to fit your setup. They may be designed for setups that run Q and the EF through additional spare inverters or buffers to give the processor a little more separation. So you will have to negate some instructions in those two routines. SEQ (Set Q) might need to become REQ (Reset Q) and vice versa. The same may be needed for the receiving routine and the EF you use.
I had a design a number of years ago where I left the second input on the MAX232 floating and it would cause the chip to randomly go into latchup and cook itself. That one took some research and a pulldown resistor on the unused input fixed it. @@CDP1861
Oddly, I think that the computer might have been made by/for someone for music. The case looks more like a synth case than a computer case. Perhaps they wanted to experiment with early MIDI?
Use a CDP1802. Really. A simple 1802 based computer is nothing more than a 32k ROM, 32k RAM, the processor itself and two or three simple logic chips, And that simple thing can talk to a terminal without a UART.
Very curious. The videos subsystem sounds a little like something that Steve Wozniak wrote about playing around with when he was experimenting with video output.
I have a bit of 1802 computer experience. I can't say what your computer is based on but I'm pretty sure it is not an RCA design. RCA had their own video and ROM chips so they would not have used a 2716 and 6847. Also, all of the RCA monitor software that I know of is named UTnn where "nn" is some number. RCA also had a bus for their Microboard system that used a 44 pin edge connector. So you seem to have a very unique bit of hardware. Very nice!
My best guess: 1- Demo stand for some app, just needs video out and data in (cassette) 2- mid way, the dev realized they can't go around some limitation, resulting in them scrapping the project 3- cheap brilliant dude with a lot of ambitions, makes its own computer.
There were companies that did cases like these(some just did them at home tho) in 70s (that you'd fill with your own industrial automation professional thingy or whatever hobby project or whatever you needed the case for. Junking a synth back then would've been expensive just to get the case, tho there were dyi synth kits and plans and stuff like that for which you could buy such cases for. You'd have to cut the holes for the kb etc yourself
Excellent videos, superbly well researched and quite obviously made by someone who really loves, and is enthusiastic about the subject they are talking about. I hope you are a radio ham too. Stephen G7VFY
Look in the description for a link to a thread I made on the vcfed forums. there you will find an image of the ROM, a disassembled version of the ROM and an emulation of the system.
Would there be a way for you to dump the memory contents from the ROM and publish it (on Github maybe?), so we can dive in to reverse engineer it and see what it does?
In the video description is a link to a forum where I discussed the technical details of the system. Copies of the ROM in binary and disassembled form can be found there.
I wonder if anyone at the Chilliwack Retro-Computing Club would know, or know of, the original builder of this system? I was going to build an ELF, and bought an 1802 chip. But I designed and built an 8085 system instead: 4 KB RAM (8 x 2114), 5 I/O ports (5 x 8212), 8 toggle switches and LEDs, a 9-digit LED display from a calculator, and a surplus PET 2001 keyboard, all wire-wrapped on a prototyping board. I used the "dirt cheap bootstrap" idea (from Byte) to load code/data. (I was in high school, and couldn't afford an eprom programmer. )
Not very popular? I’ve been given to understand that the RCA1802 was used by NASA due to its reliability. It also was my first computer on the ELF II board.
Don't poo-poo the 1802 se quickly. It was many peoples' first foray into computer. It was featured on the cover of Popular Electronics (August 1976) . I know it started me on my 40 year career in computers. It is a power processor and is even better than my two other favorites... the 6800 and the Z-80.
Oh, I'm saying nothing of the likes. While I was researching this I did see there was a LOT of hobbyist groups, projects and other one-off machines built around it. It was clearly popular.
It could possibly be a debug monitor for Motorola, although I'm not sure if there are any Motorola chips inside. That's what Apple had when they were running their Macintosh line on the 68000 CPU series.
This looks more like a prototype for a commercial system than your typical homebuilt. Could it be that whoever built this actually intended to produce and sell it, maybe from home, but the business never took off? There were lot of small computer manufacturers in the 70s and 80s that started as hobbyist ventures in somebody's garage.
That's my guess as well. As Vancouver densifies there are a lot of older homes being redeveloped as the estates are sold and a lot of it ends up in the region's second hand stores.
MACBUG is not rare or weird, idk what research you did but you didn't look hard enough. MACBUG is the most common 1802 operating system back in the day, Tiny Basic was what you wanted or if you got the magazines, CHIP-8. If you were really hardcore about the 1802 you might have found the COMX-35: en.wikipedia.org/wiki/Comx-35 Sorry for several comments but youtube throws an error when I try to edit to add to my previous comments.
This computer was build by grade 12 students at Vancouver Technical School. I still have one. I ordered the keyboards from a surplus electronics magazine.
the year was 1983
WOW!
@@gesnowCool. We had a Netronics ELF II with 8K RAM, a video card and ascii keyboard at my school in the electronics shop. We used a dial up phone with a handset modem to connect with a BBS where we could chat with VanTech. That was around 1981. We also built an 1802 board that if I recall correctly was a joint project between PW and VanTech for robotics.
Wow so did the school design and buy the casing or was it built at a school workshop?
This system looks like the 1802 computers we built in VanTech in Vancouver. We also used a commodore tape drive for back ups and restore. That planner board should have 5 to 6 rows. The other slots are for memory upgrades.
Interesting! So it was kind of like a class assignment to build?
The intention of the additional slots may have been for additional peripherals such as a serial port. I designed and built my own 1802 based computer about 40 years ago. I still have it and was recently getting it up and running again. I still have an original copy of the RCA 1802 User Manual from back in the day, photocopies of the articles about the COSMAC ELF, and several program listings in 1802 assembler.
Thanks for an excellent video that brings back memories. I built a NS SCMP sbc back in the 1980s that used switches for data input, a push button to Write a value to RAM and another push-button to advance the program counter. Oh how things have changed.
My first computer was a hand wired ELF, and it led to a long and successful career in computers. Thanks for an entertaining trip down memory lane.
That is an amazing piece of vintage hardware. Your channel continues be so unique and insightful. Thank you for this.
This thing is gorgeous. This esteemed graybeard was able to get over the DIY enclosure hump so effectively, the PCB's are double-sided and gorgeous, mass storage, keyboard, and video out. So many hurdles, even for modern homebrew hobbyists, accomplished maybe in the mid 1970's. Hats off to the architect of this!
Well, whoever built this computer did a fantastic job !! This is a real work of art !!
beautifully built - TBH, with 1k of RAM there won't be a BASIC interpreter for the machine that would be workable. I completely agree that it was built by somebody who knew what they were doing, with a lot of pride, and was likely a stepping stone to the next step. You will find that the ROM contains a full monitor. potentially a disassembler. Likely has tape read / write routines
Great find!.
I’m really surprised they wasted 5 whole bytes of ROM on that “READY” prompt. The “>” would have been quite enough. Or, if it was a boot-up self-test, a simple “OK” would have sufficed.
@@lawrencedoliveiro9104 The old TRS-80 model I had a "READY>" prompt implemented also, and this machine was released for sale somewhere before 1977; so it's possible whomever built this home brew computer, may have been trying to emulate the TR-80 to some degree
Saving 3 bytes wouldn't make a difference in a 2k EPROM. No matter how efficiently a machine language start-up program and support routines (e.g. Keyboard address to ASCII conversion, ASCII to pixel character matrix video conversion, Mass Storage I/O) may be; there will always be several bytes left over anyway. So whatever would've "sufficed" is a moot point and subject to the system designer's prerogative.
If you'd like to show us a system that saves a few bytes with prompts and messages that "suffices" please feel free to produce a video showing your completed home brew 1802 based computer project. Please enlighten us by mentioning what additional and useful routine you implemented with the few bytes you saved, that also fills up the entire 2048 bytes of EPROM memory space, with zero left over bytes. You don't need a CRT based video display, as many modern wide screen TVs have analog RF and composite video inputs.
The "READY" prompt wouldn't necessarily be copied from the EPROM into the 1k system RAM. This prompt could be copied as an ASCII character string from the EPROM, encoded as a matrix of 5 x 7 pixels (or 5 x 8 if using lowercase descenders) straight into the video RAM. Where did the missing 15 bits go? They were diverted into a bit bucket placed at the bottom of the desk under the computer.
Since the MC6847 only needs 512 bytes to completely fill its display in text mode, I'm willing to bet that the rest was used as general-purpose RAM.
15:42 Hard to believe someone would have “moved on” and simply abandoned such a lovingly-put-together piece of work. I have feeling the creator of this thing could have died, and then the family, going through their stuff, saw this as little more than “junk” and got rid of it ...
I have to say that is a very amazing build someone did. they spent huge amount of time and really did it properly. Thanks for sharing!
What a beauty! I hope you make a follow up video very soon when you found out how to load a program. I see that the Yuotuber 'Tr Een' in the comments below appears to know what MACBUG is. I hope you solve the puzzle and keep us posted.
I think @TechTimeTraveller would like a gander at this.
Well if he would like to take a stab at it he knows where to reach me at! :)
Wow! It’s a beautifully designed machine that’s for sure! Also the 1802 world did have a “killer app” in some respects with the “CHIP-8” programming language and its variants.
But CHIP8 was in many ways tied to the CDP1861 graphics chip, which this computer does not seem to have. An antiquity now, but a small wonder in the days when you could not buy computers in a store yet.
@@CDP1861 you can always adapt the interpreter to your particular video chip, CHIP-8 isn’t that complicated, it’s only 512 bytes total :)
@@CDP-1802 Very true. There is some 64 x something graphics mode in the 6847. Taking apart CHIP8 and learning how drawing sprites worked was one of the first things I did long ago.
Neat computer. The Netronics ELF II was my first micro in 1980; it had a 1/4 K RAM, which I expanded with a 4K board. Tiny Basic takes up 3.5K RAM if I recall correctly. The first online action I had was in 1981 on a BBS. Dial up, Penny Whistle modem. Then I got a Sharp pocket computer with a Z80, Basic interpreter and 24K RAM. That pretty much left the ELF II in the dust.
I've still got a couple of ELF II's and Super ELF's a COMiX-35 (reproduction) and a couple of others with a disc operating system called ELF/OS. I've got a RCA Studio II Game Console as well.
Hindsight is always 20/20 but I really think RCA missed the boat. The RCA VIP2 was advertised but never released and it would have been similar to a VIC-20 but a year before the VIC-20 was released. If RCA had taken on the challenge, they could have been a contender. The downside is that the Basic interpreters for the 1802 are pretty slow with loops and some math routines. The ability to change the address register on the fly makes it very efficient for programming but it was pretty sluggish. That said, when 6502's were humming along at 1MHz or 2MHz, an 1802 will do 6MHz without breaking a sweat and an 1806 will do 8MHz - 10MHz.
My dad used a heavy printer case to build an Apple ][ into. My very first experience with computers in the late 1980s, with a Zenith green phosphor monitor. My mind was blown in the 2000s when he hooked that up to the same television set that we were using in the 1980s and I saw that it had color.
If you want straight composite out of the MC1372, that chip has a test mode that will put composite on the RF out described in the datasheet.
Those hand done boards are freaking beautiful...
Gorgeous.
Something similar I dream of building when I retire in a few years, but first have to live that long (thanks to USA Social Security moving the goal post).
Decided on the RCA 1802 over Moto 6502 because of the architecture. That and I'm a loner, Dottie.
Well done.
Excellent video, very interesting and informative. Your production quality is always very good and I hope you can succeed on youtube, you put in the effort and it shows!
That is a lovely machine, and the 1802 is a neat processor.
CMOS and RISC. The world was not ready for that yet. 🙂
@@CDP1861 Aside from not having internal RAM and ROM, it strikes me as a microcontroller ahead of its time.
@@Nf6xNet In a way, yes. Microcontrollers often have Harvard architecture, that internal memory is a huge set of registers and that internal ROM often has weird word sizes, so that it can fetch one complete instruction in one go. Microprocessors are all Princeton (aka von Neumann) architecture and rely mostly on having both program and data residing in external memory.
The 1802 simply extends the use of its registers to also function as memory pointers. The one you make the program counter then is simply used to fetch the instructions. I think that uniting these two architectures and trying to use the advantages of both actually is what RISC is all about.
I am a sucker for all things: A: Homebrew... B: Wide format... C: Vintage computing...
This checks ALL my boxes! Good Lord, this machine is a real beauty! I love it! :D
Amazing find, amazing build. I happened to find an 1802 homebrew computer a few years ago, I'll go find it later and add a comment about which one of the 3rd party ones it is specifically, but there was a standard for expansion board connectors that was common (if anything was ever common with that rare processor) for homebrew. In my case the backplane had the spots to add the slots but it had the 1802 & VIP onboard. The slots themselves were never added originally and are absolutely impossible to find these days.
Netronics ELF-II is the one I have. I'm pretty sure I saw in my research about COSMAC stuff back when I got it there were two 'competing' (and i say that with big idiot air quotes because how can there be competition in such a small market) connector standards, one of which was the Netronics slot connectors and the other was based on dupont style pins, which is what you have. I would be very, very surprised if any of this was truly home brew and not something purchased via mail back in the day.
Which is still home brew, but not 'mf had a vat of acid in the back yard to etch boards' home brew. I also really doubt that case was totally homemade as well, the holes are way too regular.
@@treennumbers Which is weird because I went through searching for mail order options as well but never found anyone selling kits or empty boards in a similar configuration, or even people just posting etch layouts.
Also I am seeing the other comments with the URL links to Tiny BASIC that are being auto-moderated. I can't control that from my end but thanks for the links as well.
The CDP1851 is a dual 8 bit parallel port, very similar to a Z80 PIO. It's quite obvious why it sits right next to the keyboard connector. And don't call little old CDP1802 weird. It' was already CMOS back then and the 'weirdness' is mostly what's now called RISC. Got one going at 7.3 MHz on a breadboard. And you are wrong: The 1802 can traditionally do without a UART because it has one instruction controlled output pin (called Q) and four separate input pins (called EF1 - EF4). Looking at the board of that tape interface, EF2 was used on this computer for the cassette input and deserialisation was done in a small software routine. My breadboard computer talks to a terminal (actually a PC running Tera Term) at 19200 baud that way, using Q as output and also EF2 as input. The only other part inbetween is a MAX232 to do the RS232 level shifting back and forth. A UART was expensive in these days and today on a breadboard thatÄs one less thing you need to wire up.
Overall, you would find that most 1802 based computers are very similar at the core. The ROM probably is addressed at 0000, RAM and video memory somewhere above. With RAM at 0000 the computer would have nothing to execute at 0000 after a reset and this computer does not seem to be designed to use the 1802's LOAD mode, like the Elves did. The only remaining question would be to which ports the 1851 is mapped and how the EF1 - EF4 are used. Adapting Elf software should not be too hard.
Just learnt from other video that it had 4bit = 16x separate program counter so basically "16 slot syscall slots" right? Were these EF1-EF4 only flags for IRQ or was that direct "hard selector" of program counter? Q was sort of 1 bit programable port for "acknowlage". Id had freaking DMA! Just WOW. Normally had one single address space so it was possible to build computer. What a neat little monster. What i reed it executed one instruction in 8 to 16 cycles that dont make it "RISC" but transfering data internally was RISC like LD, ST and that data bus 8bit what a bottleneck. 6502 had exposed all 16bits. Im nobody to to make judgment. But overall it's impresive design extept 8bit bus and lack of registers like Z80 im not surprise by quick look that Z80 was so popular
@@KabelkowyJoe Q and the EFs (external flags) had nothing to do with DMA. Q was a simple output bit that could be set or reset with instructions. The EFs were input signals that were controlled by external devices and could be tested by the 1802 with instructions. With only little extra hardware could turn Q and one EF of your choice into a simple serial port and communicate with a terminal. Everything was done in two small and carefully timed routines. That's what I have on the breadboard right now and I have it working with up to 19200 baud with an old PC running terminal emulation on the other side.
And DMA? That worked a little differently than it did on other processers at the time. The 1802 never really yields control of its bus to anyone that asks politely. Instead, it adds another bus cycle to the current instruction in which it does the memory addressing itself and signals the external device that now is the time to read or write its byte to the data bus. If there was more than just one byte to transfer, the requesting device would simply keep requesting DMA and wait until the 1802 serviced that request with the next instruction. The 1802 would use register R0 (one of its 16 general purpose registers) as pointer to the DMA buffer and increment it after each DMA request. This way of handling DMA allowed even early DIY computers to have graphics by simply adding a single IC to the system. That was the CDP1861, my first way ever to put pixels on a screen. And why am I still a fan of that primitive old thing? Because it does the job in a smart way, instead of throwing an expensive board full of memory and logic at the problem.
Initially, I thought the home brew RCA 1802 CPU based computer you demonstrated, may have been the one called FRED (Flexible Recreational Educational Device) built by Joe Weisbecker c.1975. Once the TH-camr went through the hardware, I realized it couldn't be the Weisbecker design because it used a Motorola MC6847 Video Display Generator chip, which was released commercially some years later.
Joe Weisbecker built the prototype of what was going to be first the two microchip 1801U and 1801R CPU commercially released in 1975 and a year later they were combined into the 1802. Weisbecker built his prototype using 100 discrete logic chips in October 1971.
The 1802 CPU was marketed by RCA to Chrysler, as a car engine spark management system to improve fuel efficiency. The senior management of RCA did not share Weisbecker's vision, that an affordable microcomputer for home use for educating a child and entertaining people at home, was feasible with the 1802.
spectrum.ieee.org/chip-hall-of-fame-rca-cdp-1802
Two hour video: RCA 1800 Microprocessor Family Oral History Panel
th-cam.com/video/pmGkgoH3kO0/w-d-xo.html
Love that LaserJet 4. I sold a lot of those back in the day. Back in 1993 HP offered us a 75 dollar spiff for each one sold. I think I sold ten that month.
It's showing its age, but still a good printer.
Very cool! Someone definitely put a lot of work into building it.
I know some computers from a little before this one's time would use the tape drive to record programs as key codes, but then basically just emulate the keyboard as it entered the program "by hand".
I was just playing with my RCA Cosmac Elf that I built back in 1976. It has the ceramic 1802 processor.
That's the original rad hard version like was used in the Voyager probes that are still out there working.
Hi, The RCA game console is called the Studio II. What an amazing find! I would love to get detailed pictures, schematics and a ROM dump to post on the COSMAC forum.
Check out the vcfed thread link in the description for a link to the dumped ROM and better photos.
@@CelGenStudios I did and I'll definitely take a deep dive into this project! Thank you!
It's an artefact from alternative history. Very impressive.
Regarding the uses of the 1802: I understand the ELF and similar systems were popular among hobbyists in the 70s because they were a really cheap way of getting started with computers. But the 1802 found its real niche in space systems, as you mention, because it was the first readily available radiation-hardened microprocessor that could withstand the radiation levels in space.
Another reason was unlike the original 1970's CPUs based on PMOS (e.g. National Semiconductor's SC/MP aka "SCAMP") and later NMOS fabrication processes with dynamic registers such as Intel's 8080 and Zilog Z80, the 1802 was the first CPU that used CMOS transistors and had static registers, this meant it could be put into a timed stand-by mode simply by gating the clock to an inactive state.
Whilst the 1802 was in a standby state, the CPU drew very little power and the static registers meant the registers' bits' were retained. This made it ideal for Satellite and Space applications, where power budgets are always tight.
On a side note, the static registers of the 1802 also simplified the implementation of single-stepping for the CPU machine language code, to enable debugging of the programs. The implementation of single stepping on the 1802 could be easily implemented in hardware, with a flip-flop, perhaps one or two spare logic gates and debounced switches, one to enable single-stepping and one to step through one of 16 registers designated as the Program Counter address register.
I could be wrong but I was under the impression the RCA1802 was developed for military and space in ceramic packages but then was packaged in plastic to try and get into the consumer market. I was involved in thick film development in the early 80s and I know the 1805 was one of the only bare chip CPU's available.
That fits with my recollection of the CDP1802, which is the first "computer" I had contact with. I seem to recall seeing a schematic for it in Popular Electronics magazine, but I was so new to electronics my attempt to produce it failed miserably. Eventually I bought a Cosmac SuperElf and greatly enjoyed assembling it. However I was not yet exposed to computer programming languages so I didn't really implement anything with it.
Later I obtained a commercial desktop computer and also gained experience with microcontroller so I lost interest and gave it away to a small electronic museum collection. Regretted it immediately, it would be fun to play with today.
Now if my old Radio Shack PC1 and PC2 computers still worked....
I was a fourth year electrical engineering student at the University of Toronto in 1978 when a group of us each bought an 1802 kit from a fellow named Eugene Tekatch.
It was a single board computer similar to the ELF. It had a whopping 256 bytes of RAM, output was 8 LEDs to be read as hex. Input was through a hex keypad using the DMA feature. There was no ROM. Not particularly impressive to my family and girlfriend.
So I expanded it. I wire wrapped two boards with 4k RAM each. An S100 connector could fit onto the side of the board (although obviously the pinout was different). Being an engineer and just wanting the thing to work, I didn't build a beautiful case. I took a thin sheet of plywood and built a backplane by cutting slots and screwing in S100 connectors which were tied together underneath by ribbon cable.
Then I got Tom Pittman's Tiny Basic on Cassette, built a cassette interface using the Q line for output and one of the EF lines for input to the computer. Since I had a Ham radio license, I was eligible to get a working surplus KSR35 teletype from the telephone company. Built a 20 ma current loop interface and connected it to the computer again using Q and EF (I got a lot of help on all these from articles in Popular Electronics and a local newsletter created by 1802 enthusiasts called Ipso Facto). The teletype was controlled by software.
Since there was no ROM, a cold start was a bit of a chore. I had to enter a 25 or so byte bootstrap program in Hex to read Basic from the cassette. If all went well, I was rewarded with a greater than sign prompt on the teletype! Then I was away to the races! I could program in Basic and store the program on the cassette. *That* impressed everyone.
I thought about but decided against getting the 1851 chip which would output composite B&W video on a TV because I didn't think it was worth the cost. The graphics were very chunky and not impressive.
I still have most of the hardware. I had to give up the KSR35 when I moved overseas 10 years ago as it weighed more than 60 pounds. I donated it to UCLA who were looking to recreate what the room where the first internet message was sent. The original room had a KSR35. That unit was a tank! Our basement flooded twice, drowning the teletype. Each time after a week or so drying out I turned it on and it worked fine!
The Galileo space probe used 4 RCA1802 CPU's.
"The spacecraft was controlled by six RCA 1802 COSMAC microprocessor CPUs: four on the spun side and two on the despun side. Each CPU was clocked at about 1.6 MHz, and fabricated on sapphire (silicon on sapphire), which is a radiation-and static-hardened material ideal for spacecraft operation. This 8-bit microprocessor was the first low-power CMOS processor chip, similar to the 6502 that was being built into the Apple II desktop computer at that time." -- Wikipedia
There was a version of Tom Pittman's Tiny BASIC for the 1802. I have a copy of the assembler code.
RCA 1802 is an interesting, if a bit weird, CPU. It features SOS (Silicon on Sapphire) chip production technology which basically make it nuclear bomb proof, meaning you can keep X-raying it with something brighter and stronger than dental X-ray machine, it will continue to march on without any hiccups.
Thus, RCA 1802 was also a decent choice for avionics and satellites that must survive the punishing shines of Gamma-rays from Sun. It's kind of wild that childrens and adults alike got to play with the military technology, occasionally in form of ELF single board computer kits which wasn't too expensive to procure back then.
I hope you find the story on this machine! I remember my old RCA game console, I had bowling, black jack and like there other cartridges for it.
Maybe MACBUG has something to do with COSMAC (ELF)?
A real nice setup, perfect monitor, now all you need to do is soft out that keyboard cable, then maybe design a memory expansion board for it.
Pity there is no serial port, it would make for a very retro terminal.
It least it has an expansion bus to you can add a serial port, but you will then have to write all the support for it as well.
@@CelGenStudios You don't need to add a UART. The 1802 can do that itself on its Q signal for serial output and any one of the EFs for serial input. You just need to take care of level shifting these signals from and to RS232 levels (using a cheap MAX232) and add two short subroutines to send and receive bytes. That's it. The tape interface works in a similar way. Look at the names of the signals etched onto the tape interface. Q and two of the EFs. Things like this made building an Elf the easiest and cheapest way to build your own computer in the 70s. If only memory had not been so expensive...
That's a crafty way to reuse a port provisioned for cassette. I know I got a small pile of MAX232's here so I like this idea! @@CDP1861
@@CelGenStudios Be a little careful with the MAX232. One of the external capacitors must be connected to VCC or GND, depending on the manufacturer. Some even don't seem to care what you connect it to. I fell into that little trap when wiring that up on the breadboard. I used the TI datasheet and my MAX232 was from MAXIM. And of course they connected the capacitor differently. It took me a second to spot and correct that. The only other trap is adapting the two small subroutines you will need to fit your setup. They may be designed for setups that run Q and the EF through additional spare inverters or buffers to give the processor a little more separation. So you will have to negate some instructions in those two routines. SEQ (Set Q) might need to become REQ (Reset Q) and vice versa. The same may be needed for the receiving routine and the EF you use.
I had a design a number of years ago where I left the second input on the MAX232 floating and it would cause the chip to randomly go into latchup and cook itself. That one took some research and a pulldown resistor on the unused input fixed it. @@CDP1861
I love the Video Professor music at the end :)
Oddly, I think that the computer might have been made by/for someone for music. The case looks more like a synth case than a computer case. Perhaps they wanted to experiment with early MIDI?
this is very interesting. I feel if I was born at a different time I would have made something like this. I feel like making a Retrofuturism project.
Use a CDP1802. Really. A simple 1802 based computer is nothing more than a 32k ROM, 32k RAM, the processor itself and two or three simple logic chips, And that simple thing can talk to a terminal without a UART.
Very curious. The videos subsystem sounds a little like something that Steve Wozniak wrote about playing around with when he was experimenting with video output.
I have a bit of 1802 computer experience. I can't say what your computer is based on but I'm pretty sure it is not an RCA design. RCA had their own video and ROM chips so they would not have used a 2716 and 6847. Also, all of the RCA monitor software that I know of is named UTnn where "nn" is some number. RCA also had a bus for their Microboard system that used a 44 pin edge connector. So you seem to have a very unique bit of hardware. Very nice!
Great video! Thank you for sharing!
It’s to bad TH-cam will not allow me to upload some photos.
Have you tried decompiling/dumping the ROM?
It's been dumped and analyzed. Check in the description for a link to the code.
My best guess:
1- Demo stand for some app, just needs video out and data in (cassette)
2- mid way, the dev realized they can't go around some limitation, resulting in them scrapping the project
3- cheap brilliant dude with a lot of ambitions, makes its own computer.
The case of that computer looks like it started out as a synthesizer, then the enclosure was repurposed.
There were companies that did cases like these(some just did them at home tho) in 70s (that you'd fill with your own industrial automation professional thingy or whatever hobby project or whatever you needed the case for.
Junking a synth back then would've been expensive just to get the case, tho there were dyi synth kits and plans and stuff like that for which you could buy such cases for. You'd have to cut the holes for the kb etc yourself
Excellent videos, superbly well researched and quite obviously made by someone who really loves, and is enthusiastic about the subject they are talking about. I hope you are a radio ham too. Stephen G7VFY
I saw the edge connector on that daughter board and thought C2N, then immediately dismissed the thought. Wow, cool!
Who built this?
Can you provide a copy of the EPROM (2716), if so, it could be disassembled to determine the MACBUG commands?
Look in the description for a link to a thread I made on the vcfed forums. there you will find an image of the ROM, a disassembled version of the ROM and an emulation of the system.
Would there be a way for you to dump the memory contents from the ROM and publish it (on Github maybe?), so we can dive in to reverse engineer it and see what it does?
In the video description is a link to a forum where I discussed the technical details of the system. Copies of the ROM in binary and disassembled form can be found there.
5:17 Are those greasy fingerprints all over the inside of the case?
Do you have ads on your videos yet? I really want to see you make a living off of this excellent work.
I've been holding out, but I've been getting a lot more people asking that I should.
@@CelGenStudios Once you start getting paid, you'll be more prolific. I'll be there for every video.
Its video mode really feels similar to TRS-80 CoCo, but in monochrome
They're both using the same video display generator chip - Motorola 6847.
I think the monochrome is just the TV. It is the same VDG as the CoCo 1 & 2.
The 1802 was used in a few pinball machines (Playmatic from Spain made those using the 1802).
I have a phone that uses an 1802 as its CPU.
I wonder if anyone at the Chilliwack Retro-Computing Club would know, or know of, the original builder of this system?
I was going to build an ELF, and bought an 1802 chip. But I designed and built an 8085 system instead: 4 KB RAM (8 x 2114), 5 I/O ports (5 x 8212), 8 toggle switches and LEDs, a 9-digit LED display from a calculator, and a surplus PET 2001 keyboard, all wire-wrapped on a prototyping board. I used the "dirt cheap bootstrap" idea (from Byte) to load code/data. (I was in high school, and couldn't afford an eprom programmer. )
maybe related to Abacad Inc fx-14?
looks very close..
I've seen a backplane like that in an RCA 1802 homebrew machine before...
Not very popular? I’ve been given to understand that the RCA1802 was used by NASA due to its reliability. It also was my first computer on the ELF II board.
Don't poo-poo the 1802 se quickly. It was many peoples' first foray into computer. It was featured on the cover of Popular Electronics (August 1976) . I know it started me on my 40 year career in computers. It is a power processor and is even better than my two other favorites... the 6800 and the Z-80.
Oh, I'm saying nothing of the likes. While I was researching this I did see there was a LOT of hobbyist groups, projects and other one-off machines built around it. It was clearly popular.
Awesome video, subbed!
That uses a Motorola CRT controller like the Color Computer
MACBUG reminds me of SWTBUG for the soutwest technical stuff, but that's most likely a coincidence.
It could possibly be a debug monitor for Motorola, although I'm not sure if there are any Motorola chips inside. That's what Apple had when they were running their Macintosh line on the 68000 CPU series.
@@minty_Joe That was MAC*S*BUG, Motorola Advanced Computer Systems Debugger...
New Sub
Great Video, Thats sweet as hell
This looks more like a prototype for a commercial system than your typical homebuilt. Could it be that whoever built this actually intended to produce and sell it, maybe from home, but the business never took off? There were lot of small computer manufacturers in the 70s and 80s that started as hobbyist ventures in somebody's garage.
If I built something like that, it wouldn't ever go to a charity shop. Maybe the maker passed and family was trying to clear out their estate.
That's my guess as well. As Vancouver densifies there are a lot of older homes being redeveloped as the estates are sold and a lot of it ends up in the region's second hand stores.
You had basic, we only programmed using 1802 hex.
MACBUG is not rare or weird, idk what research you did but you didn't look hard enough. MACBUG is the most common 1802 operating system back in the day, Tiny Basic was what you wanted or if you got the magazines, CHIP-8. If you were really hardcore about the 1802 you might have found the COMX-35: en.wikipedia.org/wiki/Comx-35
Sorry for several comments but youtube throws an error when I try to edit to add to my previous comments.
Beautiful design but shocking soldering.
(By the way, that's 'soldering' with an 'l'. 🇦🇺🙂)
OwO
i appreciate the sound of zee-80 instead of zed.. 6:49