EEVblog
HTML-код
- Опубликовано: 21 окт 2024
- A mystery Mailbag teardown item!
Forum:
www.eevblog.com...
HDSP-2000 LED Display Application Note:
chronix.pl/down...
EEVblog Main Web Site: www.eevblog.com
The 2nd EEVblog Channel: / eevblog2
Support the EEVblog through Patreon!
/ eevblog
EEVblog Amazon Store (Dave gets a cut):
astore.amazon.c...
T-Shirts: teespring.com/s...
💗 Likecoin - Coins for Likes: likecoin.pro/@...
I KNOW THAT PCB TYPE!
When I was a technician in the US Navy in the late '70's we had lots of custom electronics systems implemented as assemblies of small to medium-sized wire-wrapped boards. Of course, the vibration, temperature and humidity cycles on a warship wreaked havoc with wire-wrap, even when perfectly performed.
The kind of PCB shown in the video was rolled out as upgrades for systems that were expected to remain in service for a long time, such as propulsion control systems that needed to last as long as the ship.
The key factor was to avoid the need for field modifications, which were easy to apply to wire-wrapped systems. But even the most stable systems could incur upgrades when other parts of the system changed. The extra isolated PCB pads were there to permit (and simplify) the addition of future bodge wires!
Most often, we'd first cut obsolete traces, next install a wire-wrap pin at designated locations, then run jumper wires to the desired locations, with one end at the part pin and the other at the wire-wrap pin. This permitted urgent changes to be tested and re-worked in the field (on a rocking ship at sea) minimal trauma to the PCB. It also supported easily adding generic DIP parts at intentionally vacant areas.
This retrofit technology became obsolete as the ships involved were retired, and short-run PCBs decreased in cost and improved in quality. The main improvements being (IIRC) PCB design tools running on workstations (instead of mainframes), and pick&place assembly technology.
I don't suppose you remember if such a technology had a name, do you? I'm very curious to look up more info about this
As a former IBM inventory controler I can testify to the power of the IBM part number system. If it didn't have a part number it didn't exist. Everything had documentation with the in's and out's of a duck's arse. Suppliers were queuing up to be official recognised by IBM, it was a massive payday if they were willing to jump through the hoops.
I worked in a Hard Drive fab pant in Havant in the UK. I never saw one of these but the Kestrel line was just down from my desk. Kestrel was old school storage drives for mainframe computers still used by banks at the time. I started on Aladdin and 540Mb 3.5 PC drive through to 6.5Gb beasties from the mid to late 90's.
The displays are Hewlett-Packard HDSP-2000 series, originally designed for the HP 9825A. They came in a number of different colors and minor variations. The metal can ICs are all custom IBM hardware. The odd board construction is standard for IBM equipment from the mid-1970s into the 1980s.
Yep, the app note is cool: chronix.pl/download/HDSP2000_Application.pdf
EEVblog RS still sell them, search for HDSP. I used them in the 80's, real nice daylight readable. Expensive.
Nice hat. You should always wear it.
Merry christmas
oh shit they are spendy yes!
you have to do something with the led matrix now, it's way too cool to ignore
I want Dave to do something cool with that display. Dave, please!
The nixie tube display is well along the design stage, but if there's ever a Mk II, then yeah, use these.
How to get Dave to open anything: write "Test Equipment" on the outside.
The "S Loop" connector is a 1970s style American telephone jack. These were fairly common before modular connectors arrived.
Searched that part number, found a page saying "This device was used to diagnose and repair floppy disks/drives. "
My favorite bit: Those 8" floppy go in SIDEWAYS. Wow!
And another page: "It assists with the repair and/or adjusting of IBM floppy drives. A CSE would bring this to diagnose and repair the drive. The test diskette has was written on a highly accurate writer and almost certainly has various tracks written at offsets from the normal locations to allow diagnosis, such as head position misalignment and tracks with a constant signal to permit diagnosis of the read amplifier and associated circuitry."
I doubt that's correct. The only floppies on 80's mainframes were the 8" IPL/Diag drives. If you had one of those out, no CE is going to spend downtime diagnosing it, they would have simply replaced it and gotten you up and running as quickly as possible. The DTE/DCE connectors on the side make me suspect it's for diagnostics on a communication or terminal controller like a 3274
FC-CPGA- a technique developed originally at IBM for cheaply and effectively packaging high-performance chips. They were the only ones to use a crimp-on aluminium cap. This technique would continue to be used at IBM into the mid-90s.
IBM System/370 was all made of these tiny boards with regular array of pads and four connectors at the edge. Hundreds of boards and several 100 ampere powersupplies, weighing tons. This design just looked absolutely crazy in comparison with Nova and DEC computers, which had large boards, two feet wide.
Part of their mainframe strategy was to sell a high end mainframe but cripple it with a suite of daughter boards, when the customer was ready to pay for an upgrade an engineer went out and removed one or more boards and flipped a few circuit traces in others and presto-changeo the customer had faster throughput without IBM have to manage a change over to a new setup - which they naturally charged for regardless - they didn't get as big as they are by missing an opportunity :)
There is a lot of technology packed in that suitcase. I worked on mainframe design in Poughkeepsie in the early 80's and am very familiar with the process. You may want to peek at en.wikipedia.org/wiki/IBM_Solid_Logic_Technology for an overview of the board and chip designs. Like you guessed, most of the logic on testers like this was not custom designed just for a particular machine, but used some common designs that were placed in many of the mainframes and peripherals. Although I'm not familiar with this specific test equipment, it probably uses one of IBM's "Universal Controller" 16-bit minicomputers, which is the four boards on the backplane. IBM had a huge catalog of IC's that were similar to the TTL logic of the day, but many of the mainframes used much faster logic such as ECL.
And you are also right that these boards were autorouted. But more than that, the circuitry was mostly constructed using very sophisticated CAD/CAM software and the board layout was just a byproduct of the software. The grid system on the boards also offered a big advantage; engineering changes could be done manually since almost every net on the board had a soldered via. Our development boards would sometimes have hundreds of bodges..
The one megabyte 8 inch floppy disk would contain a "standard" operating system. There would be additional floppies that contained the diagnostic programs for specific machines. The "S loop" and "R loop" are most likely an early version of the IBM token ring networking architecture or something similar. Again, the interfaces tried to be "universal" in design. These testers were used in place of a "front panel" on many of the IBM industrial-type computers since they were meant to be installed in unfriendly locations. In addition, one field engineer could support many machines with a single box.
As for the mechanical design, these were made to be tossed in the back of your car and survive all the punishment that a field engineer would render.
Merry Christmas, Dave. Nice of you to spend a bit of your holiday with us.
I feel like this should have been put in a museum. haha, it looks crazy.
IPL means Initial Program Load. The IPL command tells it where the boot up software is stored, so, tape, disks, floppy etc.
stonent In other words, a boot loader.
+Chaplain Dave Sparks Nope, IPL is the correct term. You really need one of those before accessing the DASD when setting up DCAF over APPL in your SNA environment ...
HDSP shift-register displays.. awesome.. those are so uber cool.. you should make a project with them!
"not gonna have time to tear it down ". haha Merry Christmas Dave man.
Dave is likely my favorite. No bs. Only teardowns. Reminds me of me only he actually knows what he's talking about. :)
Wow, what a great teardown! Merry Christmas, Dave.
I vote to see that LED matrix display working. Looks fascinating!
The S Loop connector is the old style US Bell telephone plug, so that's probably got a modem inside it. Your brand new snazzy Princess Phone would plug into a socket just like that.
IPL=Initial Program Load, i.e. Boot.
Those were very common circuit boards found in IBM equipment. Also, IBM was the (or one of the) largest manufacturers of integrated circuits at that time. And those metal cans is what all their ICs looked like. The mainframes were full of cards just like those, even my IBM 5100 has that style of printed circuit board in it. I have no idea why, no one else did boards like that I am aware of, but IBM churned them out by the thousands. Those mother board connecters were used everywhere in IBM systems too.They liked to design everything to be easy to replace in the field. Thus the very specialized test equipment!
The IBM silver square packaged SLT modules with the square PCB board pattern date all the way back from 1965 with the introduction of the IBM 360 mainframe - that was totally revolutionary at the time. Then they just kept on going! This technology made it to the very first IBM 5100 "personal" computer in 1975 (just look up a RUclips video for the IBM 5100).
But what is the deal with that square PCB board pattern? Why?
It would make sense if they could manufacture them at mass and then specialise later like a Gate-Array or Sea-of-Gates, but I do not see how.
Are the pads vias and pads at the same time and they laminated multiple boards? Did it give them the chance to alter the circuit later?
+ 0MoTheG I'd need to ask my fellow IBM history specialists. However I would guess that this arrangement provided for a higher board density than other available PCBs at the time (1965), allowed for a lot of opportunities for top to bottom plane connections, that the very regular arrangement was simpler to deal with using primitive CAD (remember, the CAD they used would have run on an IBM 7090 in the early 1960's or something of that ilk, as they didn't have the luxury of a 360 while they were designing it!), and was easy to produce with early CNC tooling. Remember, IBM used CAD and CNC unbelievably early: for example, the ALD "schematics" for our 1959 IBM 1401 are all computer generated, and the giant wired-wrapped backplanes of the same machine were laid out by computer controlled (CNC) wire wrapping machines. Probably all on 709 or earlier tube computers. When it came to manufacturing, IBM was ahead of the pack by a very large amount. Looks like they stuck to it for a very long time ;-)
My IBM specialist colleagues confirm the above. Its roots are historical and based on SLT technology. The basic square pattern is made to accommodate the SLTs contact pattern (the IBM square silver proprietary modules, which started as multichip modules flipped on ceramic carriers, then evolved to single IC carriers as you see here). The overall pitch was progressively reduced as generations of denser SLTs and PCBs unfolded (which allows to roughly date PCBs and SLTs by pitch). The signals run mostly East-West on one side and North-South on the other side, the empty holes are used for vias. The very large number of predictably placed vias, regular grid, and simplified unidirectional routing pattern allowed CAD algorithms that could run very efficiently on early computers.
perfect video length for a 5 minute teardown. more of this kind pleeeeeaaaaseeee
Yes this is a Maintenance Device use to troubleshoot both IBM mainframes like the 3081 and 3083 as well as connected devices like the 3380 DASD, 3274, 3705, 1174 Terminal Controllers and many other devices. The MD could also be used to duplicate microcode disks for the devices that still booted from micro code load on disks. I still have the MD that was given to me when they quit using them for onboard diagnostic terminals around the year 2000. I have repaired many an IBM machine or changed its configurations with this device over my twenty plus year career with IBM Mainframes.
I have always loved IBM PCB like these, just simply amazing that IBM was manufacturing these in the 70. I also have a couple of ceramic packages like those, but way bigger, and each one has 9 "flip chip"s in them. Blows your mind.
Happy Merry Christmas.
Knowing how IBM likes to over engineer their internal equipment it's probably a pocket calculator.
I would like to see when someone sends him the test equipment, for the test equipment
I was gonna say a PAT tester.lol.
I know that place. "Puh - kip - see". :-)
en.wikipedia.org/wiki/Poughkeepsie,_New_York
Dave pretty much nailed it. From Wikipedia: IBM has a large campus in the town of Poughkeepsie, once referred to as IBM's "Main Plant",
Those displays look like HP matrix led displays, namely QDSP-2289 ( I've got some from a military surplus). You can drive them using high speed serial line and transistor keys - to feed in the data (row by row, as far as I remember) to be displayed, they latch the data and patiently wait.
That sort of autorouting and grid layout of the board, sort of makes sense for a prototype board, as it would allow for easier bodge wires, or even adding on of components.
Yes, it might be a form of combined PCB and protoboard, the free pads would be useful if one wanted to add a component or two.
Other than that, the only thing I can think of is that they might be used for tracking in some automated manufacturing system. Such a system might use either optical sensors or electric contacts to count pads to the correct location for a component or soldering.
It could also be for some standardised but limited testing system (i.e. you feed in a table of integer coordinate pairs and wether they have to be connected/ aren't allowed to be connected or maybe what resistance there's supposed to be between a pair of coordinates or maybe even simple instructions like "apply voltage X between pad A and pad B" and "measure the voltage between pad B and pad C")
It looks to me like the boards might have been made by etching the copper traces as normal but applying the solder pads using a completely regular screen printed grid rather than a mask. Also it looks like they're mounting through hole components as surface mount, which has got to be an interesting (Or possibly highly frustrating) assembly job.
That PCB layout would make it easy to transfer from a wire wrapped breadboard.
IBM token ring PC network boards used this construction, so not limited to small volume products
Yep seen that style a number of times,
- I remember trying to desolder the square arrays as an 16 year old youth at the Olivetti electronics lab. I seem to remember the AT&T 3B2 of the same era had the same style as well.
This video is the best christmas present I got. Thanks Dave! Happy holidays!
You really need to salvage some of those LEDS to send to Fran for her collection of leds
This was used by ibm service men to diagnose and repair issues with floppy disks.
I wonder if those PCBs included a dielectric layer for decoupling between planes
Nice teardown, and I would love to see that display in action!
Awesome teardown and device! Merry Christmas and thank you!
As we're approaching EEVblog 1000, I'd wonder if anyone ever played along at home
Wow. I just discovered this video. I worked on this device for IBM Toronto Canada in the late 1970s. It was called the Maintenance Device. Developed by IBM Toronto and first manufactured in Toronto. It was intended as test equipment for field service personnel for internal use only and thus was not subject to full scale qualification testing (as was equipment shipped to customers). My role was to to ensure the technical release from development to manufacturing. I had to devise test equipment for the keyboard/display and for the final assembly. In so doing. I had to work closely with a very dedicated hardware and software team. Looking back this was an incredible development effort (remember this was prior to the IBM PC release). On the personal side I hand carried (driving through a snowstorm in winter through New York State to Toronto) the first development keyboard/display (designed in Oswego NY) that is shown in this video. After a year or so of manufacture in Toronto it was transferred to IBM Raleigh. Many thank for this video....brings back so many memories.
You should send a few of those chips to electronupdate to decap, he could get some HD pics.
On the first chips you looked at you just looked at the wrong spot.
Bottom left corner, 84/18, 84/40, 84/48, 84/46 - and then the later chip with 85/02 that you used.
Thank you for the most interesting channel on youtube, and thanks to everyone who mails him stuff to teardown.
The auto insertion machine was part of the tool chain for the design of the boards.
I haven't seen that many bare cans since I visited Sunny Hills Nudest Camp 🤤 "He who does not learn from the past is condemned to repeat it". Old School Rocks.
Thank you Ausi Man 👍😁
6:55 "because I dont want to take it all apart ... its Christmas eve" 20 minutes after and I thank you for one nice and entertaining video :)
Along with the grid array and the chips in cans those resistor and capacitor arrays are items I've never seen used anywhere else by anyone else. Only on the IBM boards so another in house part. And even in the day finding information on them was harder than finding hens teeth.
Very interesting Dave. That light display in the controller is crying out for a repurposing project, but I dont want the super computer project to suffer! So hopefully you will find time for both. Thanks for another great year of EE related content, hope next year is a good year for you and the family.
all of the old ibm CCT boards used that PCB technique.
I had an old ibm system 32 RPL data processing system mainframe from 1976, and it was full of boards that looked very similar!
Nice device. Lot of designing for test device.
And absolutely you need some project for that nice matrix display.
cool. I'd like to see you get that weird led display running
You have GOT to figure out some project to use those LEDs in.
I bet they made lots of these. There's one on ebay now with SN 33795. This one is 39251. As soon as I saw the black plastic box I thought it was a disk drive field test unit for hard drives. We had a CDC one at the TV station I worked at in the mid 80s, and we only had six drives. Any big maintenance department responsible for 8 inch floppy drives probably had one. Back then electronics were too expensive to throw away. Radio/TV repair shops were very common. We used the HDSP-2000 displays in video router control displays. I remember throwing out hundreds of brand new ones because we no longer needed them.
first 6m summarised: "this is an electronic thing, that does stuff"
IBM stuff is just awsome.
They invented so much crazy manufacturing process that are more close to wizardry than engineering.
This boards reminds me of the LVDC flight computeur from Saturn V booster, those kinds of package are evolution of technics dating back to these days (encapsulated ceramic flip chips).
I have few of that things in my long forgotten pile'o'junks, I may check in back that graveyard one day and make a necklace out of these chips.
Those are really cool LED displays! Awesome. I just found some on Ebay, holy crap they are expensive!
These IBM "Tester in a suitcase" units are very useful for troubleshooting older IBM tech. We have sever at the museum I work at (The Large Scale Systems Museum, near Pittsburgh) With the proper cables and test discs, they where used with everything form the 1970's 4341 Mainframe, to the 3380 disc drives and 3480 tape backup subsystem, but they are also use with early AS/400's and their associated peripherals like the 9480E tape subsystem
Holy shit those boards are the most gorgeous veroboard-like circuit I've seen to date. And IBM certified. Dayum. What a beauty!
On the PC board - IBM was an almost totally vertically integrated company, they made their own IC modules, had their own CAD software for designing PC boards and so on. Even their own style of drawing schematics. Those custom square IC's are called "MST" which stands for monolithic systems technology.
I happen to know a guy old enough to know this stuff, and he said that's a 3380MD2, a testing device for IBM's first big disk array 3380DASD, with 2,4-10 GiB in 1980ish.
When I was 10 or so I had a giant IBM machine with dual 8" floppies. When I finally took it to bits it was loaded with boards of that era with IC's contained in little aluminium box's like that. This really took me back. Thanks. ;-)
Apparently that machine I had was a IBM System/23. Same crazy PCB design!
That would be pronounced as "Per - Kip - See" (upstate New York IIRC).
That looks like a four layer board with internal power and ground layers.
When I worked at DEC, they bought two of IBM's "Series One" mini computers, and took one of them apart to reverse engineer it. It was FULL of boards that looked just like that with all the square solder pads and square metal BGA chips. (The DEC engineers even took the chips apart, and XRAYED them!).
Maybe the capacitance between the power and ground layers of the board was all that was needed for bypass?
I thought it was Puh-kip-see?
Matrix design: for easier bodgewiring?
That connector labelled S loop is a really old-school US telephone connector. They used to use those huge things back in the 50s and 60s before the day of the RJ-11.
That's a tool we used to use to debug and diagnose IBM CPU's. All the old panels full of lights and switches disappeared and instead we plugged this thing in to run diagnostic routines. It kinda took the fun out of repairing computers.
Had he watched Friends, he would have known how to pronounce Poughkeepsie :D
An IBM MD (Maintenance Device) !
Long time ago I saw one of these. Its used to run tests on Mainframe computers and their peripherals like disk controllers and such.
Each mainframe controllers would have had 8" diskettes in a special holder on the inside of the door. You loaded the disk and it would start diagnostics for that device
"Do not touch diskette when running"
Maybe they had boards predrilled with thix matrix-type layout to avoid one more step in prototyping?
It looks like Dave has a new eevblog branded multimeter model. Is it also a brymen?
Pronounced Poe-kip-see. Thanks, Dave. Very interesting. Very hard to tell what is going on when all of the parts are discrete and proprietary. Imagine your flat screen being built with that LED tech.
The weird grid PCBs might be a quick prototyping thing.
Possible that they have a whole stack of copper clad boards with a grid if plated vias in them and then they use some sort of printing process to pattern it and etch in acid to get a finished board. Acid etch is quick and easy while via plating is a complex and dirty process.
Wrecked an AS360? main frame years ago. It had a lot of gear that looked like that inside. It was truly engineered, even had little vacuum 's on the big hard discs. They were probably 14" or so.
You gotta do something with the led-display! Great mailbag btw!
I saw a PCB like the odd one in the video when I was a kid. My dad brought it home in a lot of electronic junk he bought at a state auction for me to take apart. It was also an IBM product, with devices in metal square metal cans like the ones on your board. I believe it was some kind of early networking card, possibly (though I'm not sure) for a PC.
Lovely ! Merry Christmas & Happy new year ..
I recall working at a company who had an IBM "mini" computer system, the IBM AS-370, upgraded to AS-390, back in the early 90's.
Awwwwwwwwww you didn't bother to fix it :/
It would be quite a challenge with the lack of documentation tho.
The displays were the most interesting part, i reckon.
I love this Xmas vibe man! Dave just looks like a kid ! :DD
18:15 how else would you rout it, if you have to avoid the row of pins in the middle?
Jesus von Nazaret Yeah I was curious about that, how else could they do it?
I mean, for higher frequency applications you don't use 90° bends, but that's mid 80ies gear with a couple of MHz max.
So worst case scenario they would have to curve the track around the pin in as smooth an arc as possible. I don't think it was that critical in this instance.
Maybe he thought that the traces connected to the middle pins.
Firecul42 With a simple 45-degree segment at each end (or beautiful sweeping hand drawn curves if one prefers), but, seriously, why did anyone need to use an autorouter with its rigid right angles to lay out that board? That is the real silliness, as my 4 year old nephew, who can barely color within the lines, could have handled the routing for that board himself without the need of an early autorouter application to do the work for him.
That display is *awesome*
Be sure to snag those 7 metal can Teledyne relays off the board that had them. Don't know about you, but TO-99 is a pretty adorable way to package a relay.
Towards the end of the video did you leave the "super secret multimeter" in shot?
Never heard of an LGA circuit board, but it's used to describe a certain type of chipset on CPUs today
This would have been used alongside one of the IBM System/370 mainframes. The metal cans and grid array are found on boards from the IBM System/370 mainframes of a similar vintage.
Those resistors/capacitors packages
I think, the ones who are marked with a polarization dot and/or small bar are resistors and/or multi package
the ones without a dot/bar marking are non polarized caps I think. Like @18:06 there are some packages in the upper corner without a mark but with a code '372' which would be 3.7nF?
"Pa-KIP-see". Amazing how a pronunciation that's almost instinctive to 'muricans can be troublesome for a Cornstalk. Thanks for playing!
This was entertaining as always. LOVE it!
Is that "S loop" connector really a repurposed valve rectifier socket? I'd be surprised if it isn't!
Don't know about the viability of using a matrix board for production gear (especially with traces running between the pads), but it's still the best game in town for one-off and bespoke circuits. (Unless you're dealing with valve voltages, in which case you want tag strips - none of that turret or eyelet board rubbish!) Of course it would have higgledy-piggledy jumpers and bodge wires, but the electrons aren't going to fly out.
15:08 Oh Dave, you and Lawrence Krauss with your Force references. lol
Bha ha ha... Dave destroys whatever you send him :-) You should now desolder those displays and definitely keep em. Rest to the scrap pile! :)
Pronounced : Po kipp see (NY) IBM was huge back in the day in that city. Installations and buildings covering acres and acres. Now pretty much gone.. Take Care
I'm thinking the grid board approach just gives them more flexibility while modifying the circuit after it's been spun. All of the extra pads allow for attaching new components, pads can be bridged if need be, unused pads can be connected to existing traces, and traces can be easily cut. Say you need to reroute a trace, you can cut it where it comes near an unused pad, connect it to that pad with a little touch of solder, and then just solder in a mod wire to connect it to some other pad(s).
Seems like a cool solution to me, actually.
I had some sort of telecom signalling interface thing once (made by STC!) I bought from an auction for $2, The front panel was covered in little LED hexadecimal displays similar to that in little clear DIP packages!
that's a nice case for your RaspberryPi cluster project, Dave ;)
hope you had a nice crimbo, mate, and have a good hogmanay too!
It would be neat if you could get that display going or at least pop one of the display ICs out and get it going with a microcontroller. Those old LED matrix displays are the ducks guts.
It's "puh-kip-see. How hard is that? :-)
The matrix design PCB was used for even large production runs in many different systems: mainframe computers, terminals, terminal controllers, test gear, etc. This was a design philosophy. Everything was designed to fit on a .10"- inch grid, including many custom components. It eliminated the question of where a through-hole was to be placed; IT HAS TO FIT ON THE GRID! Various width boards can plug into one or more of the 24-pin motherboard connectors, so the mechanics of that are just as modular as the daughter boards themselves. Some boards were short and some were big, so I would guess the depth of the boards were a part of the modular design scheme, as well.
I don't think the loop connectors have anything to do with telephone, even though they look like it. IBM had loop-based communications schemes for interconnecting various devices, so those loop connectors are probably used to interpose the tester between devices. (BTW, large IBM banking-industry terminal systems were interconnected with gold-plated 1/4" stereo phone plugs and jacks throughout a building, so IBM was used to re-purposing standard [and cheap] connectors. Of course, other companies' equipment [modems, etc.] designed to be used with those systems also used those same re-purposed connectors along with compatible signal levels, etc. in order to be out-of-the-box compatible. Imagine a crew of non-nerdy, cable-pulling installers plugging in a bunch of phone plugs. It's hard to screw up.)
You can see the super secret meter in the top right. Looks like a new eevblog meter, it has a red button!
Many years ago I had some boards like that in my junk box. No idea what they came out of,but there were some with big potted bricks on them. I tried to break one open,to see what was inside,it seemed like mostly passives,perhaps some kind of R/C filters? No idea.
I NEED THAT DISPLAY.
The other thing to note; no XTAL oscillator cans! (Unless that's what the round ones are, none of my boards had any round cans.)
It looks like the board construction is a late example of IBM SLT (Solid Logic Technology), which was a technology stack incorporating the boards, IBM's own pre-IC custom hybrids and the related construction techniques. Vintage 1960s tech, for sure.
The dots on the board could be for easy and clean modding and bodging.
The only thing i can think of was for them to save on drilling when manufactured in house, using a grid to take in all the ICs and do all the vias by manufacturing the grid on eatchable PCB and eatching in house
10:24 The date codes seem to actually be 8445, 8446, 8440, and 8418
Are you sure there isnt some sort of automated sandwiched wire point to point board present in the image at 9:50. It kinda reminds me of CPU boards in PDP-8/I. There were machines around that would wire up a board for you from a netlist, attaching the wire ends to metal tube via/pads. You just needed a blank board with rows and columns of plated holes. Changing a design was just a matter of letting the machine know where the chip/part pins were (x and y) and loading a new netlist.
My guess for S and R Loop would be Send and Return Loop.