• Please review our updated Terms and Rules here

Any ideas what this project could be?

C

cliffmiller

Member
Joined
Aug 12, 2021
Messages
12
1974 or ‘75 - As a Senior in Electrical Engineering I was permitted to join a graduate class in Computer Interfacing. We studied digital logic and the instruction set of the PDP8/E (?) - do not remember. We were assigned a task to design and construct an interface to connect a teletype or serial terminal to the computer and to program it to echo back characters typed to it.
I have no records or materials from the course. We used wire-wrap construction and 7400 family ICs and all this was way over my head - the grad students did a bit better on the project than the undergrads as I remember. I had a chance to swap some emails with my professor from back then and was disappointed that he didn’t remember as much as I did - it was only fifty years ago!
Any PDP8 experts have some idea what we were interfacing to? The part that’s fuzzy is I’d think the computer would already have a serial interface board so wouldn’t need any additional logic.
I also welcome comments on which model 8 might have been employed in that time frame in a university course. It definitely had the blinking lights and switches for programming.
 
Could have been a PDP8/L.

Teletype interfaces thru M706 receiver/M707 transmitter dual height modules populated with several 74xx chips each.

PDP8/E uses one quad height module, typically M8650, for serial interface which has many more chips as it includes all the logic to connect to the 8E omnibus. I doubt you would have been asked to build a module that actually connected to the omnibus.

Perhaps you were given task of re-creating the M706/M707 logic and testing it in place of the original DEC modules or maybe you converted the current loop interface into RS232 as I recall doing on my tech school's PDP/8i back in the 70's as we had a brand spanking new Hazeltine 1500 terminal to attach (also included bumping up the baud rate IIRC).

Dale
 
In 1974 and 1975 there were three possibilities for 8 models that would have been TTL compatible. The 8/e would be the most likely as there were many more of them manufactured by that time than the other models. The 8/i and then 8/l would be next. Both of those were out of production by 1974. The 8/a came a couple of years later and by then colleges seemed to be buying PDP-11's. Also the 8/a has a programmer panel but it used 7 segment LEDs to display octal and pushbuttons to let you enter octal. That makes it even more unlikely to have been what you used.

The interface that goes in an 8/e would have been a single card with I/O address decoding logic, latches, buffers, and a UART. For the 8/i and 8/l it would be more likely that you would be interfacing to the Posibus which is a set of I/O cables that comes from the CPU. They are essentially TTL logic level signals. The cables loop through your controller and continue on to the next I/O device. The older Straight 8 and the 8/s used Negibus. The Negibus uses negative logic levels and single sided paddles on the cables so there were twice as many cables. DEC made converters that would go both ways so that peripherals using the other kind of bus could be matched to your machine.

Interfacing to Posibus is not difficult but would have required a way to terminate the cables and a chassis to hold your interface cards. The 8/e would have needed a single card to plug into the Omnibus backplane of the computer. DEC made wire wrap cards with sockets and breadboard areas just for this purpose. As Dalmill mentioned, both the 8/i and 8/l have built in console interfaces. The trick would be adding a second serial interface that would respond to different I/O device codes. The PDP-8 has a 6 bit device code so that up to 64 devices could be attached. This is not actually true since some of the codes are used for CPU operations like turning on and off interrupts and selecting different banks of memory. Most machines had a lot of unused device codes.

I could pin it down a little more if you can remember any more details. Maybe something mentioned will knock off the memory cobwebs.
 
We were provided a prototype board to wirewrap our design. There was a chassis outside the computer used for the connection of our project board. A lab assistant plugged it in so that action must have been viewed as outside our expertise. We did have to enter the program for testing.

I recall the computer as being surplus from previous projects rather than new and purchased for this purpose-the class had the air of being hastily constructed - xeroxed handouts and a copy of the Small Computer Handbook and lectures.

All was great fun and exciting but involved hours of construction and experimentation. One or two grad students finally got it to work and then helped the rest of us.

My recollection aided by your replies is the computer looked like an ‘E or an ‘F so that means Omnibus…the task was constructing a second serial device for a system with an existing console. By the end of the class we were loading a common program in on paper tape then demonstrating the operation of the interface we’d constructed.

Thanks for the help! It would be years later the PiDP-8 was offered and Simh developed so I could explore the PDP8 again.
 
Last edited:
From your new info I am changing my mind and think it must have been an 8/i or an 8/l. This is because of the chassis outside the CPU. This would be the device controller chassis connected to the CPU via a set of Posibus cables. This does not exclude the 8/e because there is an interface card for omnibus machines that interfaces to Posibus. That card is an M8350 which you would not have interacted with directly.

Of the three likely candidate CPU, here are short descriptions that might allow you to distinguish from memory.

  • Rotary knob just to the right and below the two rows of blinken lights. Paddle switches. This is an 8/e,f,m. There were several different color schemes sold depending on how the machine was marketed. Color other than orange and brown would be a giveaway that what you used was an 8/e. But the most common color was the orange and brown. The rotary switch was used to select between displaying the AC (ACCumulator) and the MD (Memory Data) on the 2nd row of lights. You can display other things also but those are the most common selections.
  • The front panel would have three rows of blinken lights with paddle switches. A second smaller cluster of status lamps was on the right side of the panel. The switch colors are a yellow and orange. This is an 8/l.
  • Four rows of blinken lights and vertically mounted rocker switches are distinctive to the 8/i. Again this has a smaller array of lamps on the right for statuss. The switch colors on the i were brown and white.
All the panels on these machines look similar and there are reasonable photos of all of them online. But if you just go look at them this will affect what you remember so think about switch shapes, colors, number of rows of lamps and if there was a rotary knob before you go look.

Not mentioned earlier but the possibility exists that you used a PDP-12 although I doubt it. The 12 is an 8/i all dressed up that has an integrated LINC processor. The console uses rocker switches like the 8/i but due to the integration of the LINC has a lot more lamps and switches. Also a lot more rare and it seems unlikely that you would have thought this was an 8.

It is also possible that the CPU was either a Straight 8 or an 8/s. These are Negibus machines but could have been fitted with a Negi to Posi bus converter. However this seems almost as unlikely as the PDP-12.

The least expensive of these machines when being sold new was about twice that of a high end luxury sedan with most being 3 or 4 times that as a minimum. My 1967 Straight 8 as sold was north of 120000 dollars. Great effort would have been taken to find continuing use for these machines.

Your echo program probably looked something like this:

Code: 
        *200
KLOOP,  KSF             /WAIT FOR A KEY
        JMP KLOOP
        KRB             /READ THE KEY
        TLS             /ECHO THE KEY
TLOOP,  TSF             /WAIT FOR TRANSMIT TO FINISH
        JMP TLOOP
        JMP KLOOP       /AND START OVER

Your instructions for KSF, KRB, TLS, and TSF would have been different mnemonics defined for your device codes but that is most likely what was used. Took me over 40 years to realize the TSF and the JMP back to it are totally unnecessary for the echo program. Of course if the lesson was to do this utilizing interrupts then the program becomes quite a lot longer and more involved.
 
I may have biased myself over the last 50 years but the colors of the ‘E stick with me as well as paddles over rockers. Do not remember the dial switch being involved at all.
Your comments about the M8350 board as an Omnibus connection for outboard logic seem applicable. I enjoyed the code you shared but do not believe they let us near an assembler, rather I think we looked up the op codes directly and entered them the hard way at least once. The paper tape came late in the quarter as a time saving method to get as many boards tested in as short a time as possible.
The focus of this course seemed to be the logic design which may have been necessary to avoid the appearance of poaching on the Computer Science department —-
 
Last edited:
You must log in or register to reply here.
Back
Top