• Please review our updated Terms and Rules here

Intel Intellec 8 MCS

I found a nice addition for my machine. Not really necessary but fun to have I think.
An experiment board with a serial port. All the chips date from 1973/1974. So it might
be used for an Intellec with an 4004 or 8008 also.

intellec board 2.jpg intellec board 1.jpg

There are a few chips missing, but the numbers are on the back of the board.

Regards, Roland
 
rgcCsrFl.jpg


Well. This a 8 year old thread, but it isn't much discussed on the Intellec 8 systems so I just continue here. I just started to get my little Intellec 8 working. I bought it three decades ago at a flea market for around 10 euros.

I have dumped the firmware and taken photos of all the cards. The firmware and photos: https://drive.google.com/drive/folders/1NOUxgHN0HNzZujf-OtU8pmPDZS40e1zJ?usp=share_link

Apart from the 8080 CPU card and the front panel controller It has two 4k RAM boards, one 4k PROM board and two I/O boards. One interesting thing is that someone upgraded the EPROM programmer to handle 2704 and 2708 instead of 1702. The old PROM programmer card is still sitting in the machine but not used.

8LTmvHll.jpg


This machine is not switchable to 230VAC so there is an external step-down transformer. It has s/n 178. Most of the ICs in it has date codes form 1973 or 1974. Except for the CPU which is dated 1979 so I wonder if that got replaced at some point in time?

My next step is to check out the PSUs. Rather difficult to get them out of the chassis actually. Then we'll see if it does something when the power is applied.
 
rgcCsrFl.jpg


Well. This a 8 year old thread, but it isn't much discussed on the Intellec 8 systems so I just continue here. I just started to get my little Intellec 8 working. I bought it three decades ago at a flea market for around 10 euros.

I have dumped the firmware and taken photos of all the cards. The firmware and photos: https://drive.google.com/drive/folders/1NOUxgHN0HNzZujf-OtU8pmPDZS40e1zJ?usp=share_link

Apart from the 8080 CPU card and the front panel controller It has two 4k RAM boards, one 4k PROM board and two I/O boards. One interesting thing is that someone upgraded the EPROM programmer to handle 2704 and 2708 instead of 1702. The old PROM programmer card is still sitting in the machine but not used.

8LTmvHll.jpg


This machine is not switchable to 230VAC so there is an external step-down transformer. It has s/n 178. Most of the ICs in it has date codes form 1973 or 1974. Except for the CPU which is dated 1979 so I wonder if that got replaced at some point in time?

My next step is to check out the PSUs. Rather difficult to get them out of the chassis actually. Then we'll see if it does something when the power is applied.
Do look at the power supplies. I didn't mean to say that the unit had a switch to select 240/120, I meant that the transformers, usually, had 4 primary connections. These would be wired in parallel or series for 2 different voltages. They did use a lot of different power supply manufactures so your machine may be different.
Anyway, look at the transformers. for the power supplies.
The 2708 setup can often be modified to program 2716s. It does require some software mods though.
Although, Chuck mentioned the series 3000 bit processors, I don't recall them being used on an Intellec 8. That was used on the later MDS800, ( the next in the development machines ). I worked on the 1036 card for that setup. I don'r think a 2 card setup would fit in a Intellec 8, without an external controller. I suspect it would have looked like the setup I have for a IMSAI with a simple bus interface card. It the time I didn't have the disk interface made by John Torode. Now that I have it, I see how it would have easily been connected to the Intellec 8. I'd originally thought Gary Kildall used a MDS800 looking at his software interface but now realize it must have been a Intellec 8 because of the time period he did this, the MDS800 didn't exit.
If you can still find a Torode drive setup, you can make a simple interface with a wire wrap interface for the intellec8. It was designed to be a dirrect DMA without needing a DMA controller on the machine. The software is real simple. You tell the controller what address to send the sector data and it does the rest. It also has the ability to load a boot track on reset. My IMSAI has no ROM in it. It is a full 64K of RAM. I just set the address for 0000 and hit the reset button. When the drive finishes, I hit run. I wrote a simple boot sector code to load the rest of CPM.
I believe it was used with a teletype ( ASR33 with punch/reader ) and not a disk drive.
I suspect you'll see a current loop serial as well as a RS232.

Dwight
 
Last edited:
I checked the PSUs. They are quite "hard coded" for 115 VAC. One single winding on the primary side. No big deal. I have the step down transformer so it solves it.
While starting to check the PSU I discovered that the crowbar for 5V triggered at 8.5V. I tuned it to trigger at 5.7V instead.

I checked the dumped firmware and could find a string "8080 v4.0" in there and then also some assembly mnemonic strings.

I think I will just run standalone software on this machine. Not any OS. I was thinking I can load the micro-chess or the Altair 4k BASIC. After all it only have 8 k memory. Not much. Would CP/M even run in 8k?
 
Hi Mattis, nice machine! In the meanwhile I found a few 8008 cpu boards for it. I hope to find the rom images for the 8008 version some time... My 8080 monitor is on github. I wonder if your version matches mine...
 
I checked the PSUs. They are quite "hard coded" for 115 VAC. One single winding on the primary side. No big deal. I have the step down transformer so it solves it.
While starting to check the PSU I discovered that the crowbar for 5V triggered at 8.5V. I tuned it to trigger at 5.7V instead.

I checked the dumped firmware and could find a string "8080 v4.0" in there and then also some assembly mnemonic strings.

I think I will just run standalone software on this machine. Not any OS. I was thinking I can load the micro-chess or the Altair 4k BASIC. After all it only have 8 k memory. Not much. Would CP/M even run in 8k?
Years ago when I worked for Intel, My room mate had purchase a bunch of pieces from Intel's scrap sale ( something they don't do now days ). He put together enough pieces to make a working series II machine. We had a problem to solve, the disk drive we got for it, was 50 hz and 220V ( 8 inch drive ). I found a transformer used for one on the SDK setups and it had a three tap primary transformer. I used that as an auto transformer to get 220V from 110V. For the frequency difference I used a file and modified the motor pulley to the right speed for 60 Hz. It required also elongating the mounts for the motor. As I recall, the motor pulley and fly wheel were a set to 50/60. Anyway, we got a working series II computer running with all scrap parts ( most all the boards required some repair parts ). I'm trying to but a date on that but it was in the 1980's sometime. For the scrap stuff,we had to sign an agreement to never sell any of the stuff.
Fun times.
Dwight
 
Here's an Intellec 8 EPROM card modified to use 2716s. The code is contained in one of the 2716s (2k). The other contains only a jump instruction.

The Intellec 8 MOD 8 version 3.0 (April 1975) 8008 monitor code was released by Intel in their publication 98-022B.

The 8080 monitor, also version 3.0 and April 1975, was published as 98-007C.
 

Attachments

  • ModifiedEPROM2716_Intellec8.jpg
    ModifiedEPROM2716_Intellec8.jpg
    3.4 MB · Views: 6
  • ModifiedEPROM2716_Wiring_Intellec8.jpg
    ModifiedEPROM2716_Wiring_Intellec8.jpg
    4.7 MB · Views: 6
Cool! I found the document you mentioned. So in that case I can recreate these monitor roms :biggrin:

Do you have images of the rom content from your board?

I'm not sure but I have two or three of these 8008 cpu boards. So the monitor makes it possible to actually do something with the machine in combination with the monitor.

Thanks for pointing me there! When I have a bit of spare time I want to make these roms. I have plenty of 1702 roms for it...
Regards, Roland
 

Attachments

  • 9800022B.pdf
    913.1 KB · Views: 6
Someone directed my attention to this thread, seeing as my name and interests were discussed here. This Web page explains what I mean:


There's a few loose threads I can pick up on from this discussion.

1) Mattis contacted me about his intellec 80 ROMs. After some work by me and Bill Beech, results are on this Web page:


Turns out Mattis has version 4.0 of the MON-80 monitor, plus some kind of disassembler. I'm not aware V4.0 is available elsewhere. If anyone follows up on understanding the disassembled disassembler, or further disassembling the V4.0 monitor, please contact me so I can add that information to my Web page.

2) Years ago in this thread, Chuck G. mentioned some kind of Intel floppy controller with Intel 3000-series bit-slice chips in it. I believe one of Intel's earliest floppy controllers was of that sort. But I believe it was Multibus based, and this early intellec 8/80 is not Multibus. By the time floppy drives were available, Intel had just moved to the Multibus and away from that older intellec architecture.

3) It's certainly possible in principle, to construct some kind of DMA interface for a Torode floppy controller, that might be intellec 8/80 compatible. I don't know if there's particular "gotchas" that would prevent it from operation. The MON-80 monitors run at 3800H, but they could be removed or reassembled to a higher address (if you can get there from cold boot). So there's operational issues, but someone might resolve them. *Why* is entirely up to such persons. It's not vital that everything with an 8080 in it, "run CP/M".

In fact: the 4040 and 8080/08 intellects, ran on paper-tape software. If any of that paper tape survives, please please please, make the binaries available.

4) It's been mentioned, that Kildall and Torode first ran CP/M on some Intel box. On my Web site, one of the Torode links is to my accounting for that event. Most of what is known, was from Kildall's account as published in Dr Dobb's Journal. (Details on my Web site.) Kildall was vague.

What is not clear, is which model of Intel box was in use. It's *likely* it was an upgraded intellec 8 with an 8080 CPU board. But it's *possible* Kildall obtained an early Intel MDS-800 which is Multibus based. Kildall was vague about a) when Torode produced his prototype controller and b) when Kildall ran it with his prototype CP/M software and c) what box it ran upon. My Web page tries to bracket the likely date. However, my investigation was done years ago. It's possible more information became available since. I asked Torode, he didn't recall; please don't bug him about it.

I'd appreciate, if anyone has specific information on either Kildall's Intel hardware, or when Kildall likely ran with Torode's controller.

That's a lot of stuff to consider. And that's one reason I do my work on the Web, where stuff is findable and stable, rather than discussion groups which are pretty busy with discussions that come and go. But, here we are, maybe these responses will be useful. and the Mattis ROMs are resolved.

regards, Herb Johnson
retrotechnology.com
 
Here are the contents of the two 2716s I mentioned earlier. These are from an Intellec 8/MOD80.

I'll also attach the 8/MOD8 code which I found on the website Nj7P.org of Bill Beech at the 'Monitors' page - I think it wasn't on his website when I looked recently.
 

Attachments

  • Int8-80_2716B.txt
    9.1 KB · Views: 2
  • Int8-80_2716A.txt
    9.1 KB · Views: 1
  • mon8_8v30-3800.txt
    2.2 KB · Views: 2
Intel document AP-7 (Application Note), dated 1975 and with Order number 98-031A, has the title:

"Disk Controller Design Uses New Bipolar Microcomputer LSI Components".

I'll post scanned images of the title page and of the controller shown on Page 2. The controller has a 100-pin interface which may - or may not - correspond to the arrangement of the Intellec 8 backplane. The same document includes schematic drawings.

This was included in the collection of papers prepared for a 'Computer Fair' held in London in 1976. The first part of the papers summarised talks to be given alongside the fair. The bulk of the collection consisted of Intel brochures relevant to the products at the time.

Another of the products was 'Intel memory systems in-481', which provided 16kb in 2107B DRAMs and was expandable to 64K x 8 by the use of four memory cards. I'll attach the two-page document. To highlight a couple of points:
- this was stated to be 'IMM 8-82 and IMM-83 Compatible', however
- "While the in-481 and in-481-1 are designed to work with the IMM-82 and IMM-83, they are not intended for use in the INTELLEC 8/MOD8 or INTELLEC 8/MOD80 since the current requirements of the in-481 and in-481-1 exceed the 60mA capacity of the INTELLEC +12v power supply".
 

Attachments

  • AP7_001.jpg
    AP7_001.jpg
    3.5 MB · Views: 10
  • AP7_002.jpg
    AP7_002.jpg
    6.2 MB · Views: 11
  • In481_001.jpg
    In481_001.jpg
    6.6 MB · Views: 12
  • In481_002.jpg
    In481_002.jpg
    2.8 MB · Views: 11
http://www.nj7p.org/Manuals/PDFs/Intel/9800031A.pdf - the AP-7 document

the Intel in-481 document is listed on this Web page

via https://www.intel-vintage.info/intelmemory.htm

look for the in-481 reference; it's to a Google drive with a long URL that may be time-limited

Regarding the AP-7 document and design. There's no evidence so far, that controller was ever sold to anyone. This was an Intel development project. My guess is, management looked at that card and their heads exploded at the cost and size of it - it's 11 inches in height!. Maybe it led them to abandon the intellec bus and led to development of what became the Multibus. But I'm guessing about what did not happen.

I've been asked for details on the Torode (Digital Systems) controller. It is described on my Web site. Please go there for more information. My Web pages are there for precisely. that. purpose. Others have these controllers, and some have Web pages and other resources about them. These were designed in 1974, when there were few processors which could handle the data and computations needed for floppy formats *in real time*. In my opinion, it's a triumph of microcoded logic design - that's why I bothered to create my Web pages about it! And, if you look, there's another Web page, on the *other floppy controller products* that existed before and during the Torode / DS controller era. Many companies provided floppy controllers in the microcomputer and *minicomputer* era.

The fundamental problem with the intellec 8080, for floppy disk work, is that it. is. slow. 8080's just don't run very fast. Even with FDC chips like the WD 1771 (as used on the first Tarbell floppy controller for IMSAI/Altair) it barely kept up with single-density 8-inch (or 5.25 inch). A DMA controller was also needed to put floppy data in memory faster than an 8080 read/write loop. The Z80 ran faster, and had looping instructions available; of course the 8088 was better still.


This Web page has *way* too much information, about a Versafloppy II FDC for S-100. But there's work by Bruce Jones, who shows multiple efforts to operate a FDC chip with 8080 code versus Z80 code.

If I had to design a floppy controller for the intellec, I'd look at old S-100 single-chip floppy controller designs, like the first Tarbell, like the SD systems, others. The 8080 code is there! The bus is similar, running the same 8080! But the smart move, may be to put a Z80 on that FDC card, let it do the heavy lifting; then copy the results from RAM to RAM with the 8080. Maybe a 6MHz 8085 could do the trick, with an Intel processor. Please don't use an arduino, but that's just me talking.

Regards Herb Johnson
retrotechnology.com where the work is

PS: I private-messaged Roland Huisman. I hope he contacts me to confirm receipt, it's an unfamiliar tool to me.
 
The controller has a 100-pin interface which may - or may not - correspond to the arrangement of the Intellec 8 backplane. The same document includes schematic drawings.
Hi Jon,

I don't think it is intellec

As far as the fingers go, the ground (3&4), +5 (99/100) and +12 (49/50) are in the correct locations and the left and right shoulders adjacent to the fingers seem to be asymmetric and the correct length to fit the card profile. What bothers me is that on the intellec 8 & 80 fingers 47 and 48 are -9 but on this card the top/bottom fingers are not joined with a via but rather they are completely separate traces. same for -12. probably more telling, if this were a disk controller that was going to do DMA, the /HOLD_REQ (51), HOLD_ACK (46), and /BUS_BUSY (53) should all be utilized and all three appear to be unconnected to traces.

I suppose those are 3601s PROMs on the right, can anyone make out what the chips are on the left side?

regards, craig
 
Last edited:
Turns out Mattis has version 4.0 of the MON-80 monitor, plus some kind of disassembler. I'm not aware V4.0 is available elsewhere. If anyone follows up on understanding the disassembled disassembler, or further disassembling the V4.0 monitor, please contact me so I can add that information to my Web page.
It took me a while to find the right link for the rom dump, but eventually I got it. Here's my quick disassembly, and yes that stuff is clearly for a disassembler.
 

Attachments

  • mon80_v4.asm.zip
    5.9 KB · Views: 6
It took me a while to find the right link for the rom dump, but eventually I got it. Here's my quick disassembly, and yes that stuff is clearly for a disassembler.
Thanks Bruce, good to hear from you. I sent Mattis a disassembly some days ago, of the disassembler and what seems to be a "V4.0" of Intel's MON-80. (It's possible someone modified a 3.0 or earlier and called it "V4.0".) And that disassembly is on my MON-80 Web page with other versions of MON-80. The disassembler part is not part of the MON-80 code but calls to it (presumably for I/O). I also sent Roland, privately, some quick disassembly of his monitor code., which is V2.0 of MON-80; his ROMs have yet other code not disassembled.


In due course Bill Beech and also I, will release versions of Bill's disassembler. No big hurry, there's lots of 8080 disassemblers around. It's a good exercise to write one or modify one. - regards Herb
 
In due course Bill Beech and also I, will release versions of Bill's disassembler. No big hurry, there's lots of 8080 disassemblers around. It's a good exercise to write one or modify one. - regards Herb
Yes, like mine that I used, which is a manually controlled tracing disassembler (links in sig). Except that when I tried to get it to generate Intel mnemonics, I realized that I had never actually made it (or its predecessors) generate Intel mnemonics! I still have testing on my to-do list for some edge cases, but it took me a couple of hours to add a new opcode table.
 
A bit of an update on my Intellec 8/80 project.

The first thing I recognised when checking the power supply was that the small crowbar add-on board to the 5V PSU didn't at all trigger anyway near 5V. It triggered at almost 9V. Adjusting it downward made it trigger at slightly below 6V which was much better.

Then next problem was the rest of the 5V supply. It refused to regulate at all so I had to temporarily remove the crowbar to continue investigating the PSU. A capacitors were checked, measured just fine. All the carbon composition resistor checked - within tolerances. The semiconductors had proper readings when testing the PN-junction forwards voltage. Could it be the LM723 chip. Gave it a try and replaced it. Nope no difference. Then I checked the driver transistor a bit more and it looked like it simply was very leaky. A new power transistor in place (and the old LM723 back) and the PSU worked just fine.

The entire Intellec had been dismantled into pieces to be able to clean off many years of dust and it had now to be built back together. With everything in place with and just the front panel controller, CPU and memory I woke it up for the first time in probably 40 years. LEDs were glowing red and some switches had noticeable action. But I couldn't get it work fully. It was possible to examine and deposit into the memory, but I was unable to single step a program.

I needed to debug the frontpanel. The LED board was checked and all LEDs were OK. Then also a quick check of all the switches which also were working just fine. It looked like the problem could be on the Front Panel Controller in the backplane.

But it was very tricky to reach this board when installed. I needed an extender board. So I sat down a Saturday an draw one in KiCAD. I also added test points so I could attach a logic analyser if it turned out to be necessary. After some delivery time for the PCB and connector I was able to assemble it a few days ago.

qulnBwCl.jpg


With this in place it was much easier to debug the frontpanel card. I ordered five boards from JLCPCB (no just HASL, not ENIG, cannot really justify spending the money on something I will most likely use just once) . I have two bare boards left if anyone want just a board. Just need to add a proper connector for it (100 pin, 0.125" division). I'd like 10 Euros for bare board. All the design files and gerbers are on Github: https://github.com/MattisLind/Intellec8

Fs8Oeyml.jpg


Now I could probe around much easier. In run mode it seemed that a lot of signals were toggling. Interestingly the FETCH signal toggled but the LED was not dimly lit. It was all black. It turned out that the OC buffer, A3, a 7417 chip was dead. Replacing that made the front panel a lot happier. Not just the FETCH LED was glowing but everything worked much better. I could toggle in a small loop at address 0 and step through it.

The machine seemed to work. But the monitor startaddress would be at 3800h and there were just nothing valid at all at that address. At least not the same compared with the memory dump I did earlier. I pulled out the EPROM card and immediately saw the SIP-switches. I have had trouble with old DIP-switches not making contact before and sure enough they were several switches that were closed that read open circuit. Trying to exercise them didn't help. Trying deoxit made it worse so I had to replace them. And voila the correct content showed up at adress 3800h.

Putting a simple jump to 3800h at address 0 and press RESET indeed made it running some code in the monitor. The IO LED was dimly lit indicating that it was probably polling the non-existing serial port. To get further I now have to create a quick little RS-232 to current-loop adapter and see if the serial port is happy to communicate with me.
 
Back
Top