How Many PDP-11s? All the PDP-11s!
Вставка
- Опубліковано 4 чер 2024
- Ooooh boy, this one is exciting! Not because we dive into the PDP-11 we picked up in the previous episode (well, that’s pretty exciting too), but because a kind viewer reached out and hooked me up with a ton of QBUS PDP-11 stuff. The collection is absolutely staggering, and in this episode, we take an in-depth look at each card I received. Next, we need to think about what kind of system we want to build, because with all the new cards, we have some options. So check out the list of cards and let me know how you would build up your ultimate PDP-11!
List of PDP-11 Cards:
i.postimg.cc/4dThyhzy/AllCard...
If you want to support the channel please hop over to Patreon:
/ usagielectric
Also, we now have some epic shirts for sale!
my-store-11554688.creator-spr...
Come join us on Discord and Twitter!
Discord: / discord
Twitter: / usagielectric
Intro Music adapted from:
Artist: The Runaway Five
Title: The Shinra Shuffle
ocremix.org/remix/OCR01847
Thanks for watching!
Chapters
0:00 Introduction
2:11 Removing and cleaning the cards and chassis
4:13 But wait, there’s more!
7:48 H9275/H9278 Backplanes
9:01 M9407 Grant Continuity and M9405/05 Extender
10:13 M3107 8-Line Async MUX
11:17 M7555 RQDX3 HDD/FDD Controller
12:31 M7513 RQDXE Extender
13:32 SRQD11-B HDD Controller
15:15 M7546 TK50 Tape Controller
16:10 Mystery Floppy Controllers
17:50 Bomem DEB1800L
18:51 Matrox QRGD-G and Bomem DEB4100L
21:21 M8637, DMA5400 and Q-RAM 44B Memory
22:52 M8190 and M8192 QBUS CPU
24:25 A mystery memory card
25:40 What am I going to do with 6 CPUs?
26:07 A portable PDP-11
27:10 A mega PDP-11
29:37 Bunny! - Наука та технологія
Back when I was in college, we found a PDP-11/45 that had been banged up and left on one of the school's loading docks as trash. Over the next several years, a friend of mine managed to completely restore it to functionality. He repaired the front panel, repaired or replaced many of the boards, and even had to custom-kludge a new power supply for it out of what he could scrounge up.
In the process, he also came across several full books of schematics for the machine, and was blown away at just how complicated its design really was. (He went into the project thinking it was an old enough computer that he could actually understand how everything worked, and that turned out to not be the case.)
Please read regarding the Carpet!
It is industrial carpet on top of concrete, very thin, no static buildup. The humidity in the room is also kept at a good level. I have never had any static buildup with anything at all ever in that room.
Give me a little credit guys...
Welll..... unfortunately, "I've never felt any shocks" isn't a good criteria. You're thinking of a good "SNAP!" when you touch a doorknob. That could be up to 25,000V. It doesn't take that much to destroy a CMOS chip, especially vintage ones which don't have much ESD (electrostatic discharge) protection built in. Your body is a 250 pF capacitor. Think what would happen if you charged one of those up to, say, 35 V and touched it to a pin on a chip. A carpet, even a think one, is chock full of nice juicy synthetic fibers that love to build voltages. "Static free" carpets or mats have conductive fibers woven through them. Just *blowing air* across a board, or using a non-conductive solder sucker, can build up sufficient charge to kill a chip. Oh, and I've seen recommendations that repair labs should have *50%* humidity levels.
Folks say "I've never had any problem". Oh? Never found a dead chip? How do you know it was never *you* that killed it? And a chip doesn't necessarily die right away. It can be damaged but still work for a while and *then* die.
I was more wondering how much of the dust in some of your dusty equipment is Usagi fur. (I know, cats are MUCH worse than rabbits in that respect. I'm vacuuming cat fuzz out of fans all the time.)
Na, I've never had a problem with those thin carpets. I work in a basement myself. You get more static buildup with those collapsible ABS tables you get from the ULINE or Costco (found out the hard way that it wreaks havok on laptop touch pads, they go all over the place!).
We used to build dyeing machines for use in the carpet industry. Collins & Aikman manufactured carpet 'squares' that were made from a yarn that had stainless steel fibers imbedded. I believe the carpet squares were mainly sold for computer rooms back in the day. The plant I installed the equipment in was located in Dalton, Georgia.
@@georgehacker6939 we should have had them where I worked, we had a left over hotel candy stripe carpets laid cheap in the offices. I got zapped all the time on the door knobs. The computer was in the basement. That was OK till it flooded ( not my Idea to put it there ). Worked somewhere else in portacabins, wet steps outside in the rain were lethal, as was the lightening strike that took the multi user computer out of use.
Just a suggestion - start with getting the original PDP up and running in it's original configuration. Get that working and then you can work on adding changes with the knowledge that the rest of it works. Tackling the "ultimate" up front without knowing if you have working boards, the right switch settings, good cables, proper software settings, etc, etc. is just asking for frustration. Baby steps.
I think that is the intended path, there is a lot more on line and in YT land on PDP series than the centurion. The fact that effectively custom firmware was built by others for specific problem solving is interesting as I thought these kind of computers were a single solution with the software being the variable.
Oh absolutely! Totally the intended path!
But, the main thing I'm trying to work out right now is what to hang on to and what to get into the hands of others. Once i get at least a basic idea of how I want to build out my ultimate system, I can start setting aside the cards I know I need for that, set aside some extras for just in case, and then start getting the rest out into the world.
I think the first step for my system though will be the 11/83 CPU and a memory card. That should get me to at least some life on a terminal. Then we'll start adding in peripheral cards for storage and get to a booting working system, then start expanding storage and finally end up with the graphics capability.
@@UsagiElectric i think get the basic system up, buld out the ultimate system, build the portableDP/11, then worry about getting the leftovers out into the world
@@inerlogic Saving everything so as to possibly have more spare parts sounds wise to me too.
@@UsagiElectric What will really help is if you can get one or more DEC handbooks. The handbooks were published yearly (I believe) and had a synopsis of all the Q-bus boards and their DIP switch settings. You can figure out what is essential for building up a working system. Unfortunately my DEC books were destroyed several years ago by a water leak, but there must be old-timers out there with handbooks. The other thing you really need is the installation medium to build the operating system so that you can run the script to link the correct drivers for the cards you have and the features you have enabled. I recommend starting with a fairly basic system with all DEC cards if you can - you should be able to find all the configurations in the handbook and the installation medium will have support for all of the cards.
This is a a great project and best wishes for success. There is so much flexibility, that the hard part is figuring out what you want to have in the system. It will be very helpful if you can get the installation medium for an operating system for one of your bootable devices because you may need to configure the system for the cards you have. You should be able to run a flavor of RT-11 (single user real time lab work) or RSX-11 (multiuser). From my recollection, bringing up the O/S, you needed to bring up the boot device and have a 9600 bps terminal with xon/xoff protocol attached to a DLV11 Q-bus board with the dip switches set to 9600 and RS-232 (not current loop) mode. The handbook would have those settings. On our system, we had RL02 bootable cartridge drives and we could boot one drive and build the O/S onto the second drive. DEC supplied us with the disk for building RT-11. It is a pretty long script of questions to answer and the link process that follows takes a long time, like an hour. It is handy to have something like an LA-120 DECwriter to save your script answers or maybe a PC running Procomm Plus to record your session to a file. It has been close to 40 years since I worked on a PDP-11, so apologies for not having more detail. You could probably also bring up Sys-V UNIX on your 11/83, but it would probably be more of a challenge than RT-11 or RSX-11. UNIX would give you a lot of nice tools and some that are still familiar today, such as the vi editor and troff for formatting text. AT&T also made a software suite called Writer’s Workbench which had a lot of tools for spell check, readability index, grammar check, etc. I recall running that on a 3B1 Safari desktop I had. UNIX also has UUCP which allows you to exchange files and e-mail with other UNIX systems via serial communication.
One thing I remember is the 5.25 inch dual floppies were pretty squirrelly. The spindles turn in opposite directions driven by one really long belt, so both diskettes go in the slots with the labels facing away from the center of the mechanism. We found kits for replacing those DEC floppy drives with two regular half height drives. The DEC floppy drives looked cool, but they were very unreliable. We had a lot of DEC Rainbows which used those dual drives and found the drives troublesome. If you have a mounting kit to allow you to use two standard 5.25 inch half height drives, I would go that way.
That Zilog chip on the mystery board is not a Z80, it's a Z8002. Introduced about 1980, it was basically a 16-bit machine, but its general registers could be harnessed together in pairs and quads, so it could do 64-bit arithmetic with single instructions. The Z8002 was a version on a compact form factor that broke out only 16 address lines (limiting it to 64 kB of memory, unlike the 8 MB of its bigger siblings).
The NONSEG version is a also a simpler version. The bigger simplings in the Z8000 line, with the Z8010 memory controller, did paged, segmented virtual memory. You could run a timesharing system on it, because you could isolate one process's memory from another's and didn't need to reboot every time a program crashed.
That's quite a powerful processor for a peripheral controller from the late 1980s (I see the 8705 date code.) I'm wondering why the designers put in something that big, and I wonder whether this wasn't some fancy thing like a video processor, or some sort of emulator board.
The date code is way past the popularity peak for the Z8000 series. The Motorola 680x0, the National 16032 and 32032, and the Intel 80x86 pretty much overtook it within months of its release.. The only system manufacturers that I recall using the Z8000 to any great extent were Olivetti and Onyx.
I wonder if this guy ccould tell you more: www.smbaker.com/scotts-z8000-cp-m-8000-clover-computer
I think you mean 64K on 16 address lines.
The (googleable) August 1980 edition of Kilobaud Microcomputing magazine featured the Z8000 on its cover. (The Zilog Z8001, Z8002, Z8003 and Z8004 were all versions of the Z8000.)
Later on, there was also a Z80000. This was mostly unsuccessful in the marketplace.
PS: And you "wonder whether this wasn't some" what? - edit: Never mind the post-script; OP has edited and corrected their post.
Commodore 900 CPU?
@@yosi1989 According to Wikipedia. Yes, it was used in the never realized Commodore 900. Well, Z8001 on the Commodore prototypes.
The Z8002 was used in Namco's Pole Position arcade machine. Well, two of them.
I used to work on a camshaft measuring machine that had 2 PDP-11s in it. The software was written in Fortran.
One of the PDPs was still in use in the 2000s!
There's a (possibly apocryphal) story that the entire Melbourne (Australia) train network, ran on 4 PDP-11's up until 2015...
@@SomeMorganSomewhere If it ain't broke, don't fix it. :-)
@@OZ9HEP Problem is less "being broken" and more "how do I get parts when there's an inevitable hardware failure".
It's less of an issue with a PDP-11 than some legacy stuff but they do use some really oddball logic chips that aren't produced by anyone anymore.
On a related note, if you can source/test/fix/etc for example, VAX parts there's a very lucrative market for them... Mainly because there are still a few floating around in the bowels of banks and the like.
Why did it need two computers?
@@b43xoit At a guess, as that sort of thing requires "proper" real-time control one was probably controlling the machine and the other was probably for the user interface.
The fact that "One of the PDPs was still in use in the 2000s!" suggests the one doing the user interface stuff probably got replaced with a modern HMI at some stage leaving the other to do the machine control.
Im probably the youngest viewer here at 16 years old but i think this kind of stuff is amazing. Seeing what computers were like back in the day.
Me 16
I'm 33 and these are old even for me, haha.
@Jason Grim Well, I am 14 So I don't
This is good stuff to be interested in at any age. If you check out my post above, I mention how this (and DEC) are historically linked to most server, desktop, and laptop computers in the use today. And when I first got into it, I was 14.
54! 🤩
The advice given by Russell Hltn below about getting a minimum system up and running before expanding is correct. You will save yourself mega time by starting small and then expanding. I used to have a PDP11/34 with RK05s followed by 1103, 1123 and finally an 1173. Was also a DEC engineer for 10 years and still have my original badge. You made no mention re what operating system you are hoping to run with so this is going to be an epic journal for you. I got pretty good at sys-gening RT11 before I gave it all up. I still have an 11/40 front panel and half a 16kb core memory board as wall art. I also had all the schematics for half the boards you have there but all gone now. Best of luck.
I worked as a Research Scientist (Robotics) on the PDP-11/23 during the 1980s, working with both RSX-11M & RT-11. Do you remember the vast number of large red DEC manuals available for both OSs? They both covered a large shelf, and were updated regularly for a not-so-small fee! Now, you're lucky to get a leaflet with your comp!
That's absolutely and actually is the intended path!
But, the main thing I'm trying to work out right now is what to hang on to and what to get into the hands of others. Once I get at least a basic idea of how I want to build out my ultimate system, I can start setting aside the cards I know I need for that, set aside some extras for just in case, and then start getting the rest out into the world.
I think the first step for my system though will be the 11/83 CPU and a memory card. That should get me to at least some life on a terminal. Then we'll start adding in peripheral cards for storage and get to a booting working system, then start expanding storage and finally end up with the graphics capability.
As for the operating system, it's an interesting world I'm stepping into here! The Centurion had but one OS and we lucked out finding a backup of it, but with the PDP-11 there's a ton of OS options. Initially, I plan to boot from the Maxtor drive that came in the 11/83 system (if it still works) and see what they used, and hopefully get backup or image of the drive and OS that's on there. Then, as we expand the system, I want to get into RSX as that's what Mike used and what a majority of the 8" floppies I got from him are geared towards. I would like to get as much of the unique FTIR software that Mike used up and going to experience it!
@@UsagiElectric I wouldn't start giving away anything until you are totally sure that you have working boards. And the second working board is your spare. And maybe your carpet is okay but I don't forget the paranoia about making sure all our spare boards were kept in anti-static bags. I still have a collection of the bags.
@@ChuffingNorah Are you sure they were red ? My recollection is they were orange or blue. And a set did occupy a shelf. I may even have a couple of the folders left somewhere. Were a great source of 3 ring binders instead of those bothersome 2 ring types.
@UsagiElectric At 6:11 (approx) when you were laying out the boards on that carpet, I was cringing. Is it by any chance an anti-static carpet? Or do you keep the humidity well in the mid range to avoid any kind of static buildup? PDP cards are fairly sturdy and static resistant, but the CPU cards and disk controller cards have some complicated (for the time) IC's so there is still a risk of blowing them.
This all over
I manufactured PC products in Central Texas for a spell, and I cringed too.
People don't understand that it's very rate for ESD damage to kill a part outright; service life reduction is the normal effect.
On the PDP, the interrupt lines are daisy chained. So the highest interrupt priority goes to the card that is closest to the CPU. If you have an empty slot, the cards that come after that will not receive the IRQ. That is why you fit a bus grant card in any empty slots. Cards that do not use IRQ have the interrupt.pins just shorted together on the PCB.
Looks like that's what those boards he called blank boards were.
You can see the pins shorted together on those pcbs as he removed them..
@tradde11 been there, done that! 😁
@@MichaelClark-uw7ex YUp. easily identified as "bus grant cards" They were so common that many PDP-11s came with a bunch that had to be removed for any peripherals added. A few interfaces were designed to be put in the last bus slots and included termination resistors that sometimes had to be removed if the interface were used anywhere other than as the last interface on the bus. (it made it easier for lab bench tests that would have bus extensions added for use by customized devices - such as dasychaining the backplane to add more interfaces making it easier to put test probes into the device during development. I briefly worked on such a system to develop software, the only problem I had was that it was necessary for me to turn on three power supplies (one for a fan, one for sthe floppy disks, and one for the system... and one day I forgot the fan - 3 hours later the Power supply exploded while I was at lunch. (that was what the fan was for). When the engineers replaced the power supply they connected everything to a single switch so I couldn't forget again.... (sorry about that, guys).
Congratulations David on your note taking ability plus your knowledge in general. It was such a pleasure talking to you a couple weeks ago. You found me just in time, before I started cleaning house of the old stuff". It was such a pleasant experience for me. Bringing back memories of 50 years ago almost brought tears to my eyes several times. You were so generous in discussing my wife, Joann (model). I lost her exactly one year ago to COVanything I can ever do to help, I'm here.
Absolutely brilliant! I was a field engineer for dec in the 80's here in the UK - I worked on Unibus mostly, 11/34s, 84s etc.. but sometimes got called out to fix Q-bus machines out of hours. A lot of our systems were used in heavy industry and I've seen some systems that were dirtier than yours 3 months after a complete service - I was always amazed how they continued to work under those conditions. I spent a heck of a lot of time changing fans!
I love that there are still people like you out there keeping these old machines alive and producing content like this for the younger generation (myself included) who weren't alive when most of these were actually in production. Even if I can't understand the appeal of machines like this from a usage standpoint (I've never been fond of CLI as my UI) I certainly appreciate the opportunity to see how modern computing has developed from the past and machines like this were a vital step in that process. Very cool.
Also, that kind viewer absolutely hooked you up. Even if it wasn't going to be used in an ongoing project, those are some beautiful boards he gave you. Even just putting them in a shadowbox on display would be nice. There's something deeply appealing for computer nerds about so many PCBs just laying out in the open like that. It's circuit board pornography, that is.
The form factor of that mystery board looks a lot like a MULTIBUS.I card, very old school.
Certainly is, I am sitting next to a Multibus I system right now! Even have a card sitting on top of it...
That was my immediate thought too.
My thought exactly, for a MultiBus Industrial Control Computer or an Intel Intellec.
That unknown memory card is a Micro Memory Inc MM8086D-512 Multibus Memory Card. At the moment, there's a guy selling one on eek bay with a manual.
That was so awesome of Mike to let you have so much stuff! I have no idea what kind of combination would be a good build, but I can't wait to see what you do with all of it!
Back in the day, my dad had a "compact" PDP-11 system in a cube shaped case made by Unbound in California, you might want to try and find one of these cases for the "portable" PDP-11 project.
Those 8 smaller chips on the serial board are differential line drivers. The huge plastic chip probably has the serial UARTS.
Nooooooooo! Not on the carpet! 😢 I’m not an ESD nazi, but that’s a massively bad idea! 😩
On my yes i wanted to scream carpet spares !!!! no no no
Where is the static mat? He got the cards on a table ?!?!
You all worry too much.
@@brettcarcio4013 maybe so. But those cards are probably not replaceable should they get damaged by esd! I have had to repair newer cards that had better ESD protection on our production floor at my old job. Trying to fix cards that we don’t have circuit diagrams for would be a nightmare
Always nice to see those characters strung together like that.... PDP-11... Worked for an analytical instrument company years ago that used DEC cages (for 11/23 and 11/73) to control things, running RT-11. Most people have never heard of such a thing.
After the relative obscurity of anything Centurion; it’s kind of nice to see an over abundance of parts for a minicomputer!
And I’m sure there’s much more PDP-11 info floating around out there for troubleshooting purposes.
In 1978 my high school in Connecticut had a PDP-7 with small tape drives, card reader and a Teletype printer. We upgraded to a PDP 11/34 with 6 VT100 displays. My Uni was also a DEC shop, with two VAX 11/780s with over a hundred terminals and an 11/750 running BSD Unix. We also had a PDP-10 and a grad student found a bug with an assembler opcode that no compiler ever used. To fix it, DEC sent a tech out, he climbed inside with a soldering iron and fixed the opcode. Ah the good ole days. Now I have more computing power, memory and storage than all of that combined in my phone.
OMG!! This video brought me tears. I did work for DEC as a tech for almost twenty years, and I know the PDP'S . Beautiful video. Thank you.❤❤
Looks like my other comment got lost, but before you power up anything you want to check those H7864-A/B power supplies in the BA23 chassis. They are notorious for having RIFA brand caps on the input as a noise filter which after 15 years or so tend to crack and then explode and leave a really nasty smoke and smell all over. There are 3 such caps in each H7864 P/S. Two on the A/C input filter and one on a separate board. All three are of the X2 variety for safety when connected to main inputs.
The good thing is that if they blow, they rarely take out anything with them, just leave behind the residue and stench. You can safely just remove them if you don't mind allowing some noise to escape onto your power main.
@@johnreinhardt2662 They can also burn hair when they blow - (had one blow up in my face once, only singed hair as the rest of the powersupply shielded me from the electrolytic splash.
6:30 - I just shed a whole bunch of ESD tears..
what i like most in this video, is the excitement in his voice
The first CAD system I ever trained on was run on a PDP-11. It was donated to my college by Shell. It was obsolete then.
I'm kind of late to the party here, but my first job out of Tufts (which ran DecSystem 10) was at Amecom Division of Litton Industries in Calverton, MD. I was hired as the engineering OS support person. I used to do RSX-11M sysgens for a small 11/23 in the lab where they developed ATC communication systems. I believe one I worked on went to Saudi Arabia. Their main business, though, was passive threat detection systems for the military. I helped with the development of the EA-6B threat detection system (the Navy plane with the big horizontal radar dome on top). I also did VAX VMS OS upgrades. Brings back memories of 1982.
One other project they gave me was to get AT&T Unix running on an 11/34. Back then it was distributed on 9-Track tapes. It would not load and nobody could figure it out. I told them I could do it, so they gave it to me to solve. I discovered that the AT&T bootstrap loader initialized the tape drive interface in the wrong sequence. So I got a copy of Dec Unix and it loaded no problem. My boss told me I cheated. :-)
Thanks for rescuing these old systems.
I was a sales rep at DEC. sold a lot of those bad boys. PDP 11 heaven 😂
Yup, very nostalgic. Fun Stuff! I was a DEC employee from 1978 until 1998 (when we got purchased by Compaq) and spent most of that time as an in-house Telecom/Network guy working on getting VT100 users attached to TTY Ports as well as handling Ethernet to get all these servers ( VAXes and such ) connected to each other. One of the fun things I would work on around 1982 was converting *idle* PDP-11/70 chassies that had a full complement of Async Serial RS232 Interfaces in them into functional DEC LAT Protocol Terminal Servers.
My first job as an electronic apprentice was wire wrapping backplanes for custom PDP-11s. I then moved on to soldering custom PCBs for them. It was millitary stuff iirc, some kind of satellite signal processing unit maybe. But we were never told exactly what they were for.
By the way, those Seagate hard drives are actually CDC Wren drives. After Seagate tool over Imprimis, the CDC disk drive business, they continued manufacturing the Wren and Swift drives and gave them an additional Seagate ST- number. They should also have the old CDC model number (e.g. 94155-XX) on the label.
I've seen worse dust. You'd be surprised how much you can do with just compressed air and a paint brush. Worst case scenario, hot water and a little dish soap, even on PCBs, does wonders (just be careful of porous components, and be sure to let the boards dry after rinsing!). My vintage computing niche is stuff from the 80's and 90's, but while I wouldn't try collecting a system like this (unless it was just to rescue it from scrappers so I could pass it along to somebody else), I can definitely appreciate people who are able to work with such systems as this!
Those are not really early MOS-based circuit boards sitting on ESD generating carpet are they? :facepalm: You very likely zapped at least a few chips on those boards or created some intermittent issues that combined with the age of these boards, may make diagnosis very difficult...
This should be capable of insanely impressive graphics for the time, it would be interesting if someone made a demo for it.
Hi Usagi. The 11/73 is a pimped out /23 running at 15.2mhz. The 11/83 is running at 18mhz and is a lot faster because it has a private memory interface (PMI). Some 11/73's have the J-11 (jaws-11) out of the /83 but run it slower. This always confused me a little. There's also the 11/53 and the never-made 11/53+ which is a hacked DEC Server 550 with resistor and ROM hack to turn it back into a full 11/53 with 1.5mb RAM and a serial interface, which used the same J-11 processor. Lovely boards. Your collection rivels mine. Really loving watching you go through this discovery process. Wonderful videos and wish you well.
that NEC multisync in the stash is a good monitor so keep it around, it can do pretty much every TTL and Analog mono/ RGB mode up to 640x480 vga (or a grumpy 800x600 56hz) I had one as my first VGA monitor back in the early 90's and used it for just about every hobby computer since until the neck snapped in a moving accident about 7 years ago. That sucked but its CRT was half burned out 30 years ago and was imprinted with autocad so I made my piece with it being gone
Sweet! I'd love to see you get your hands on a Symbolics machine. I remember seeing photos of some of those cards and the systems looked fascinating.
Everytime I am about to go to bed on a Sunday!
Lol, who am I kidding, I stay up late just for those sweet Usagi Electric uploads.
17:27 (Let me see if I can remember without looking at any manuals.) There were a few variants of the Z8000 processor. It was sort-of-16/32-bit processor, competing with the Motorola 68000, but while the instruction set had 32-bit addresses, only 23 bits of each address were used. And those bits were split up into a 16-bit part and a 7-bit part, which were not even contiguous. So effectively it was a segmented-addressing scheme, which if anything was even worse than the one Intel invented for the x86 architecture.
The “non-seg” version would be one that only looked at 16 address bits, ignoring the rest. There might have been addressing modes that only had 16 bit addresses, so you could save some memory by leaving the unused bits.
Just got to love all this PDP stuff. I got to play around with a PDP 11/23 about 30 years ago, making it work and everything. Many fond memories of that. Unfortunately we had to give it away because we didn't have anywhere to keep it, and eventually it was scrapped. I think some parts of it is still around.
Unlike 30, or even 20 years ago, these days, there's lots of vintage computer collectors and groups on Facebook, and other platforms, so even if you've got no use for something, there's lots of people who are interested in taking it off your hands!
These old cards remind me of working on the P-3C Orion, with it's PDP-11 -esque main mission computer (ASQ-212, and then ASQ-222)
Still used in 2020, if you could believe that. Now, the P-8A has taken over.
I remember scolding sailors for carrying the old MUX and EGPC cards (we had full graphics for plotting radar contacts) in their fleece parka liner, as it was an ESD risk and we would occasionally have cards go bad that way. More often, it was our techs not knowing what they were doing.
Looking forward to the build.
Your mystery card looks like part of a older Bomem high speed vector processor. It was a possibly 4 card unit that did the FFT and sat in a separate cabinet between a DA3 FTIR and a PDP11. Later systems fitted in the same chassis
I just found your channel a few days ago and wished I found it sooner. Seeing what you do to preserve old computers is amazing! Thanks also to Mike for donating parts for this project and sharing with you his knowledge. It's people like you and stories like his that keep these awesome machines alive!💯
I have a PDP-11/35 which has a hardware debugger attached to it. This computer was one of the development nodes at CompuServe's R&D lab. If you or one of your viewers is interested in this system, let me know how we can get in touch with one another.
I'd absolutely love to get my hands on that. I used to have a decent collection including a DECSystem 5400 and an IBM System/34, I've never laid hands on a PDP but I'd guess it's a similar architecture to the 5400.
@tradde11 Haha... very rare indeed!! My bad.. 11/35 -- made the correction!! 😊
I was actually IN that lab off Sawmill Road during my time at CompuServe. I think I remember you too Jorge!
@@bigdaddympd Indeed Mike, many moons ago!! This 11/35 is from the R&D Lab in Tucson. While in Columbus, I was in Building 4 at the Arlington location. Cheers Mike!!
@@JorgeOchoaLions My memory is fading, but there was a project I was on that had me there for some meetings, perhaps with you. I do remember the days pre Cisco and other Ethernet routers, where CompuServe used the pdps to build our own node network using b-protocol. You tell some kids about this stuff, and they look at you like you were cavemen! Funny story, I am consulting and working with a large North American logistics and supply chain company that still uses VAX COBOL and Sterling EDI software to process freight bills. They are amazed when I can talk about thos systems and how to migrate certain functions off them!
The PDP-11 design influenced so many computers, including the 6502. I have a PDP-11/93 but not enough bits to run it. I also have a VAX 6000 series machine which I can't afford to power on and a MicroVAX 3100/95. I still play with the MicroVAX ocassionally, with proper OSes of course. None of that *nix stuff.
without unix stuff you don't have a proper OS. VMS was so buggy it had trouble running more than 3 days without a crash.
@@jessepollard7132 Uh-huh.
Dude, honestly. I absolutely love that you love this stuff. The amount of effort and time you dedicate. The sacrifices you make in life for this channel. It's truly amazing. Thank you
I used a PDP-11 for IR imagery processing back in 1988 for AFRL at Wright Pat. Loved it!
Hi, Usagi! Cool! Your need an Ethernet adapter DEQNA-AA (M7504) or DELQA-M(M7516) and run UNIX on that PDP-11.
or RSX-11M+, which has an awesome modern TCP/IP implementation by Johnny Billquist
@@vihapuu That's the guy. Johnny Billquist knows his stuff.
My dad said he used to have PDP-11s in his office back in the '80s. He was in charge of the computing equipment there at the time and had mentioned to me that he regretted scrapping them when he eventually replaced them with PCs, but they didn't have space to keep them and at the time nobody was interested in buying them...
I've seen PDP-11s in all kinds of applications: a point of sale system in a pharmacy, controlling a cold roll mill, even a machine to test wire rope, like the cables for bridges or drag lines. It was sort of the all purpose machine of those days. DEC machines were solid hardware.
Palm to face, I hope that rug is "grounded" being winter time.
About the TK-50 drive: I have around 2 dozen of DEC tapes with software on them but no means to dump them. Can you maybe create proper (archival-grade) dumps of these with that drive when your PDP is finished?
Those are some of the most beautiful PCBs I’ve ever seen! Especially the M9405!
Usagi: It can support up to 32 serial devices at the same time.
Me: Oh nice, half a stack of serial termials.
On a more serious note, it totally blows me away that you don't only manage to get the hardware but also the documentation and even software for it.
I'm looking forward to see what you come up with, the lego style of being able to just through together an almost random amount of cards and scale your system seems super cool, especially for such old hardware.
I'm thinking that the card with the Z80 might be a tape drive controller. It looks a bit like cards for a TS-11 9-track drive that I've saw long ago. The controller cards for these often used the Z80 to drive the logic. There were Q-BUS and UNIBUS versions like this.
Funny to see your info on x25 as we, in the Netherlands, used PDP11-72's to manage the national X.25 network Datanet1.
The PDP 11s didn't handle any of the actual traffic but managed the 3 regional data exchanges and the national management centre. Each center had a cluster of 2 PDPs in active/standby setup where the actual traffic handling hardware (BTMC/Alcatel DPS1500) was directly connected to the PDP bus. Brilliant system and was far more flexible than its replacement based on Nortel data nodes managed with Sun workstations.... The new "improved" network lacked features that we were used to with DPS1500 + PDP11s...
I'm amazed a PDP-11 is still running here in Aus, but not that surprised - they were a great machine. When the RAAF were getting the F111-C aircraft, they had a simulator which was run by a PDP-11, and that machine, I'm told, used drum storage rather than disk because they could get multiple, independent heads on it. That was 1971... Russell
My first computer love -- worked on pdp-11 in various material handling environment where the pdp-11 supervised other hardware - started at a petro-industry controlling the loading of fuel truck -- then to running several AS/RS (automated storage and retrieval system) and AGVs (Automated Guided Vehicles), Automated Monorail System ... even worked on a converted fire alarm system used to monitor assembly line operation. -- common among these all systems -- RSX-11M+ highly recommended!
Oh - -and for Unix ppl -- the proto-type for RTOP -- is RMD on RSX-11M
The chip density on those boards is amazing.
Wow you are really gonna build a pdp11 freaking awesome I can't wait to see more of these, I was always fascinated by those pdp systems and seeing you build one it's gonna be fantastic!
I absolutely love this stuff! I always look forward to the next video on your channel.
Those sexy RX50 dual 5 1/4" floppy drives were also used on the DEC Rainbow CPM desktop computers, which might help in finding one. The pair of disks share a single spindle motor so the second disk needs to be inserted upside down.
They were also present on some Microvaxes.
@tradde11 also on the DEC Professional 350 (and another model number?), which was kind of a "big brother" to the Rainbow. I used a Rainbow to learn Pascal in 1982-84, and a Pro at uni in 1988 to learn assembler with MACRO-11. PDP-11 is just a beautiful instruction set! And the sound of the RX50 drive is just unforgettable. (As is the sweet _smell_ of these DEC machines - probably some sort of toxic evaporation, but never mind that...)
@tradde11 It was the 8088 - 8080 wouldn't have made much sense. I'd love to have a Rainbow 100 as well, but I am fortunate to have one of the actual Pro 350 machines I used in 88.
I meant bigger brother in the sense that it was supposed to be a "serious" machine, whereas the Rainbow was intended as a personal computer. I think it was a better design than the IBM PC, in particular with the dual CPU architecture that worked transparently in the "hybrid" CP/M-80/86 OS.
Around that time I read about a different "pc", the Dimension 68K (can't remember who made it, not a big company), it could be equipped with both 68000, Z80 and 8086 iirc. It's a shame the PC and clones killed almost all diversity within a few years...
DEC made 3 lines of PC-class machines:
* DECmate, where the CPU was a PDP-8 on a chip.
* DEC Professional, with an LSI-11 CPU (PDP-11 on a chip).
* DEC Rainbow, with two processors, a Z80 and and an Intel 8088. So it could boot up two versions of CP/M (8-bit on the Z80 or 16-bit on the 8088), and also MS-DOS (on the 8088, of course). I think while running MS-DOS, the Z80 still handled I/O duties, but beyond that it could not really be considered a “multiprocessor” machine. (Must have been fun writing drivers for that, though.)
As you can see, the Rainbow had quite an interesting architecture. I remember our CS department upgraded the RAM to 896kiB. I also remember arguing with a Microsoft guy who came to talk to us about this new “OS/2” thing that they and IBM were working on. For some reason he thought the well-known 640kiB RAM limit was built into MS-DOS itself, rather than being merely a hardware limitation of the original IBM PC. I told him there was this machine upstairs that proved him wrong.
Talking about smells, yes, I remember the smell of the DEC Rainbow. I thought it was just me ...
And yeah, as you can probably figure, all three lines were flops.
Should be a great project! Some of those boards are familiar, as they are also used in my MicroVAX II. Obviously a later model PDP-11. I had a PDP years ago, but the CPU was over 2 boards (I think), and using Unibus, not Q-Bus. It used RL-02 drives (very similar to the Centurion), not those new-fangled RD drives.
@tradde11 Yes, I think it was a 34 with programmer console. Unfortunately I gave it away as it was taking up too much space (2xRL02 as well)
Amazing. From a personal perspective, this was one of the most exciting episodes and I already know it'll get even better.
Once during the mid-90's I got a fully working MicroVAX II with VT220 monitor, LK201 keyboard, external TK50 (+ a bunch of tapes) and external RD54 (Maxtor XT2190 HDD) for free.
Those external cabinets for the TK50 and RD54. Oh man Ultra quality.
I knew nothing about the machine or how it worked, but eventually I got the VAX-VMS OS working and managed to mount the external units.
The MicroVAX II itself looked exactly like a BA23 with a beige plastic/metal enclosure, 4 button + power front panel and was equipped with the DEC 5.25" dual floppy and a RD53 HDD, so imagine my excitement watching this video. "Hey, that's more or less my old MicroVAX II".
I guess it's not far off to aniticipate, that a MicroVAX II is a fully working PDP11. Almost portable in fact. Amazing.
A mere couple of years later (1996-1997) I swappeded it right over for one IBM 5150 PC and one IBM 5160 PCXT with 5154 EGA monitor and card.
Even if a MicroVAX II might be a stripped down PDP11, it was still a very capable machine, servicing around 24 people doing their work on serial terminals based around software from WordPerfect (for VMS), the Shell menu system, the WordPerfect itself, DrawPerfect, some database-registration system and electronic mail system.
It's truly amazing to think of that his was effectively enough for an entire workplace to function on a daily basis.
In some way, items from the living DEC era still lives on to this. The user management, file flagging and security from VMS was carried into Windows NT, because the Microsoft hired the software team.
And the TK50 lives on, in form of the LTO Ultrium tape storage backup system.
Just found out that the large radio telescope in Australia is still running off a PDP 11
Yes! Tom Scott just did a video about that! It's the Parkes Radio Telescope: ua-cam.com/video/6o38C-ultvw/v-deo.html
There are multiple large radio telescopes in Australia. Tidbinbilla is half a days drive south of the Parkes dish, and is even bigger. (But I don't know what systems run it)
I hope sitting on the carpet didn't ESD-zap those boards.
When he zooms in, it looks like they are sitting on conductive foam. If so, it will reduce the risk greatly….but still not “ESD safe” without a grounding strap.
26:18 Here’s a visual of that happening😂
PDP-11 is multi purpose computer (though not marketed as such) from instrumentation in laboratories or in manufacturing to Office based system. DEC made many Operating Systems from Standalone RSX-11 to multi-user RSTS-E and even a near real time RT-11 OS plus UNIX was built on it. PDP-11 is one of the most iconic system with 16bit and loads of 3rd party devices to plug in the UNIBUS.
sThat was while DEC was prowd to document their equipment. too bad they didn't last.
In college day early 80's my 1st system I used was a PDP 11/70 system that was fully decked out to support over 70 users running UNIX. After that I was a admin systems operator for PDP 11/30-40-70 and VAX 11/780 & 11/785 and VAX 8650. as well a handful of MicroVAX's. All running VMS versions 3-7. Great times with these systems since I had clustered them. Sadly they where all decommissioned in the late 90's ... due to PC's.
That's quite a collection of cards for the new machine. I worked at Mostek back in the 70s (on Crosby Rd, about a mile up the road from the DFW Retrocomputing site you were at Saturday). We had scads of PDP11s and a VAX. The test floor computers were primarily 11/08s. I worked in a systems group with an 11/70 used by product engineers for production analysis. I hubbed all of the floor PDPs to the 11/70 using 11/34s and Decnet to download test programs and collect device test data from the floor machines. Our 11/70 had a couple of the 'washing machine' hard drives and a 5MB removable disk pack on it, with a massive memory card built inhouse. Since we were a memory manufacturer it seemed natural we would build our own. Our VAX was installed about a year before I left the company, and wasn't production ready before I was gone. BTW, it was cool hear your Centurion talk and nice to meet you afterwards.
PDP-11 goldmine! It will be very cool to see the stuff Mike did and how he did it using the PDP.
In the Coast Guard every morning I use-to boot program our PDP-11 from memory before I made coffee for the LORAN shop (ooooooh, to the good-old-days ;)
>that mystery memory card
pretty shure that goes to an 8008 (or was it 8080?) industrial automation (was a standardized backplane) system, theres a youtuber under the name "hello world" that showed some cards that looked similar to that memory card
AH! its intel multibus
Ah yes, the easiest username to search for.
Your 11/73 talks to its memory over the Qbus. The 11/83 uses a private memory interconnect, which allows the CPU to talk to memory over a private connection without bothering the Qbus - it uses the CD interconnect for this. Now there's a trick: if I recall correctly, the memory actually needs to be "in front" of the CPU, so if you have two 2MB memory cards (maxing it out at 4MB), you would install the memory in slots 1 & 2, the CPU in slot 3, and peripherals after that. (Only slots 1-3 have the CDI.) If you don't do that, the memory will still be accessible over the Qbus, (that's how DMA over the Qbus works) but you won't have the performance boost that the PMI gives you. (The Qbus VAX machines which used the same chassis had their PMI on a ribbon cable, and put the CPU in slot 1 - I think they used the CD interconnect too.) IIRC, the CDI is basically a bunch of links from the top of one slot to the bottom of the next,, so you'd similarly lose the PMI if you took out the a memory card from slot 2; if you have only one memory card, you'd put the memory in slot 1 and the CPU in slot 2. Oh, and don't put grant continuity cards in the CD-interconnect slots. You don't actually need to fill the chassis with grant cards; you can put the extender up against the other cards (as long as there's no gap along the serpentine grant chain). And you don't need them in the second chassis (as long as there are gaps in the grant chain). But you have them so they don't do any harm.
I am fan of PDP11 and LSI11 systems since beginning of 80s. Thanks!
While getting the spindle drives to work (floppies and hd drives) to work for all these old museum pieces would be awesome. I think that getting them working with floppy emu (flash) and many of the st506 (flash) emulators would be even better. Reasoning that avoiding mechanical wear on the spindle devices would be best for preservation and using less power by using modern PSUs for just being able to have a mega pdp11 and centurions that use less power is better.
I expect that backplanes will be a very minor issue, at least if you can source the edge connectors. As far as I know, they are completely void of anything but electrical paths (maybe some termination resistors?).
You can always make a bespoke new one if needed, either for yourself or for the guys you are hooking up with cards.
yup watch for the termination resistors as they would only work for the end of the bus.
I hope you have a tape drive .. a mini computer just can't be one without (grin) ! I always loved my register / memory display on the front of my 11/45. (The good ol' days running RSTS/E!)
Oooo, I have a multibus I system sitting next to me I am desperately trying to get working. It has an Intel iAPX 432 CPU card installed in it from High Integrity Systems.
I immediately recognised that card type when you picked it up :D
looks like a couple of commenters have multibus. Should be a forum for users on line somewhere
@@highpath4776 I'm in the long process of making a video about my system, but yes, I agree. There seems to be a significant lack of commentary about it. Which is unfortunate, it's quite a capable bus!
DEC is my favourite computer company of all time. I learned COBOL on a then already antique DEC VAX in 94.
The PDP-11s we used at Fugro AS to drive our navigation systems actually supported both serial terminals and full colour RGB.
The CPU received and processed data from multiple positioning and directional systems (including Magnavox GPS, RACAL Pulse 8 and Sercel Syledis) and a Sperry MK37 Gyroscope, as well as bathymetry data from a SIMRAD EA800 echo-sounder.
While processing (and saving to tape) all of this data from these devices, it would also generate positional navigation fixes, and drive THREE full-colour helmsman's displays (via three RG58 RGB coax cables). Oh, and while all that was going on, you could also post-process the data and generate navigation coverage charts, with the PDP driving a massive HP, six-pen, A0 plotter.
A truly amazing system, and it still breaks my heart that I watched it being pulled a part and thrown into a recycling dumpster (back in 1997) - without saving as much as a keyboard.
I sure love the nonlinearity of mr Usagi’s time.
The "QD" bus is an interconnect between two neighboring slots. It's used for various option cards that come in pairs that need to be interconnected. For example, the RLV11 controller set is a pair of boards that are used with the RL01 and RL02 hard disks, and these two boards are interconnnected with the "D" side of a QD backplane.
I'm fairly sure the "Jaws" processor chip had the floating point processor built in for all revisions. Not sure what that 40 pin socket was for, it might have been to connect a user programmable micro code storage unit (writable control store). Some DEC CPU's allowed user defined extensions to the instruction set.
That unknown memory board looks like an Intel Multibus style card.
I'm looking forward to seeing what you do with this machine. And an episode or two with your own Cray XMP-4, or a Thinking Machines
Not exactly sure why, but I find the PDP-11 quite fascinating. I think it's because its described in wikipedia as being quite enigmatic in it's uses - complex traffic light systems, ATC radar processing, heck even nuclear power plants. It's great to see you making a video like this especially of one of the more modern and power variety. I saw there is an emulator for such systems but I'm not sure what I could do with it. The bulky minicomputer aspect seems a huge part of the experience too! Looking forward to seeing how your ultimate setups pan out and what you decide to do with them!
PDP machines were also extensively cloned and used in the USSR, and a decent bit of their home computer architectures (as much as home computers were a thing) earlier on were PDP derived.
It was relatively easy to program in octal.
They got everywhere.
Here's another use case: control system for a radiotherapy machine.
The software for this Therac-25 (written in PDP11 assembly iirc?) was critically bugged and the bugs have been examined in great detail because several patients died. ua-cam.com/video/7EQT1gVsE6I/v-deo.html (skip first hour or so)
If i remember correct so the empty socket on the 11/73 boards are microcode for the Commercial instruction set. The J11 has floating point as standard.
Are the Weitek chips on the Bomem DEB1800L the same as the Weitek floating-point coprocessors they released as drop-in replacements for the 80387?
on the topic of a small portable pdp for exhibitions, it would be neat to have a bunch of tiny terminals all attached and working. retro game consoles made of an arduino and lcd are common and cheap, i bet an inspired fan could 3d-print some fully functional micro terminals that way. that 70s æsthetic is awesome and small terminals are convenient for testing stuff.
PDP-11's are such great machines. This is going to be huge fun.
I'd say you had xenomorph eggs in those PDP-11s! An explosion of dust indeed. Good that you have some space outside the shop where you're able to blow it all out :)
Nicely recombobulated. I'm impressed by your collection. It's a great look inside that legendary computer. Just look at those boards... a joy for ever!
Cool gfx cards indeed. I've never head about 1024x512 resolution :)
Waiting for that Mother-Of-All-PDP-11s build :)
22:49 I think the latest/highest-end PDP-11 models only had 22 bits of physical memory addressing. That’s a maximum of 4MiB of RAM (a bit less, because I/O devices are also addressed in the same memory space).
Yeah. When that topped out, DEC decided they had to go with a bigger bus AND a bigger addressing model, while keeping at least some similarity in the instruction set, and came up with the Virtual Address eXtension. Did you know that's what VAX stood for?
I used VAXes when they were new. Yes, I knew.
I also thought that “VMS” (“Virtual Memory System”) was a very boring name for an OS. Until I found out how powerful it really was.
and only in the 11/70 versions. From memory the 11/73 was the only dual processor 11/70 and understand DEC never sold any (partly due to the release of the VAX-780, and ithe 11/73 could out perform the VAX, but memory could be faulty.
Never played with the AMD 2901 but did build a VME rack board using the next generation 29000 series bit slice processors as part of my MSc. You can’t get much closer to the metal than designing your own instruction set!
I used a PDP-11/73 in a Mfg. Engineering department in the early 80's.
I loved this I loved the PDP-11. I owned an LSI-11 with RSX OS!
16:12: Damn, those TTL chips are tightly packed, especially on the "mystery" card.
That last card you showed has the same board edge connector as many boards that I have found over the years at my uncles. Most of them already hard parts salvaged off of them unfortunately so that's what I kept them for. I still have one or two of them around. I have no idea what they went to.
My first os was running DR'S CP/M. I ran Basic, COBOL, Pascal, Fortran, PL/I on a Z80 desktop(Morrow). The experience I had on the pdp-11/34 with its vt100 terminals I got the use of at Queen's really got me excited back in the late 70's. We had been using the IBM 029 Card Punches to punch out our WATFOR/WATFiV UCSD FORTRAN was entertaining and educational but ohhthe typos lol don't forget the commas or periods lol
CP/M was initially cross-developed from a different DEC machine, a big PDP-10.
@@lawrencedoliveiro9104 Back in 1980-1981, when I worked for ADL Printed Circuits in Concord, CA. I remember running CP/M on a Multibus machine, with an INTEL 8080, 48K ram, and 2 - 8" floppy drives. The thing was in an enclosed rack-mount chassis the size of an under-the-counter refrigerator!. I worked in the office as the 'engineer' (no formal training) it was my job to make sure that the orders the sales team brought in could actually be made at our plant and there were no errors in the plans or production film. I used it to develop Panel Sizing software for the sales team, so they could accurately quote jobs. It would take into account the finished PCB size and whether it had gold plated edge connectors and would orient them so as to get the optimal panel size for the part and our machines. I also developed the Carbide Drill Bit - Speed and Feed charts for the Excellon Automation CNC printed circuit (PCB) drill machines, and the Cutter/Router Bit - Speed and Feed charts for the PDA (Paul Dosier Associates) CNC PCB Router machine. Each type of PCB material had to have a unique profile on each machine to prevent delamination and/or glazing of the holes. I had a blast! CP/M and Benton Harbor Basic. In fact, I still have a copy of all my programs somewhere on an 8" floppy.
Some of the quirks of CP/M were actually copied from DEC. For example, the name “PIP” for the file-utility program, and CTRL/Z denoting end of file (actually only supposed to work on terminals, not in files). Also DEC OSes had multi-character device names, but these were simplified down to a single letter for drives in CP/M.
Unfortunately, this feature of single-letter drive names was carried over by Microsoft into MS-DOS, and persists from there into Microsoft Windows to this day. A holdover from the 8-bit era, and an OS in the 32-bit and 64-bit eras still suffers from the same restriction!
@lawrencedoliveiro9104 oh I know pip intimately as I now do backend python dev. I was heavily into the early MS era about the same time supporting both hardware and software installs/software support. I dabbled in everything at the time. the b/w -> green phosporus -> orange phosporus screens/terms/monitors. Thank you so much for it tickled the memories back. I think I might have my Morrow Z80 box around here somewhere in amongst my Pet/ TRS-80 machines.
Man, I'm jealous of those matrox/bomen video cards. I hated having to use the master console printer when needing to do maintenance or repairs lol.
The 5 chips with the big "S" logo around the ribbon connectors in the M3107 MUX board are NE5180A, which is a "Octal differential line receiver" for RS-232/422/423. This means that the other 3 could possibly be transmitters.
yeah, they're the MAX232's of the olden days like the MC1488 and MC1489
11:09 Plenty of pins on each connector, enough to operate a modem. Because DEC terminals only really needed 3 pins (2-TXD, 3-RXD and 7-GND).
And typically all the RS232 pins were present and functional, and could be used not only for modems but for other serial interface applications.
The cables we had in the walls - to the printer and so on, were thickly populated.
Have you ever thought about buying an old NCR 500? I worked on them in the army in the early '70s in Vietnam and Germany. Love your show. I also so worked on the PDP11/83 - IBM 7083-1401-360/20,30,95-370/155,158,165,168,195,3033.....
Cool that you got some more Qbus cards. The ultimate system you may want to run on your pdp11 might be 2.11BSD which is a authentic unix flavour with even network (TCP/IP) support. I used to run a webserver on a machine similar to yours. I may even have some QBus cards if you are interested let me know.