Plexus P/20: This is the end.

Added: 2026-05-04

[00:00:00]Well, hello everyone and welcome back to Adrian's digital basement. On today's video, I'm getting back to something that is one of the most requested things I get back to, I think, ever in the history of my channels, and it's this, the Plexus P20. This is an 8 user dual processor multi-user Unix system running AT&T Unix system 5. This computer came out in the early to mid 80s and was targeted at kind of the small office situation where you'd have eight terminals hooked up to it for some collaborative working. This is probably the only known example of this machine left and the only working one as well.

[00:00:32]And when I got this thing, it was not working. So, there was a five-part series, I think it was, where I go through how I managed to get this thing working and it was it was a lot of work and a big community effort because there was data on the hard drive we were able to recover and there were engineering documents that we found on the hard drive. And the story is frankly amazing.

[00:00:51]So, I highly recommend going back and checking out those original videos. And I even made a super cut if you just kind of want to sit down and watch it as one long video and that will kind of catch you up to speed on all the absolutely amazing revelations that came out of working on this machine. Now, I gave myself a deadline for getting back to this because I'm bringing this to VCF Pacific Northwest in Seattle in the beginning of May. And that gives me like three weeks to get this machine prepped and ready for the show. And the goals in this video are a to see if this thing still works. Hopefully, it is still working and not broken. I haven't turned it on since I last worked on it back in 2024. But I also have a pile of hardware here that I want to show off and test for this machine. And also, I want to talk about the fact that there's some cool software developments that have happened with this thing. So, with all that said, I think it's time to get working on this. So, without further ado, let's get right to it.

[00:01:58]As I mentioned in the intro, I haven't worked on this thing since 2024. And because of that, I've worked on how who knows how many things since. And I've kind of forgotten a lot about all the details of this machine. It'd be interesting to go back through my videos and actually count up the number of distinct things I've worked on since working on this. But even the AGA Compraphic 9000 PS project I worked on right before, I spent so much time working on that thing. And I have so much of that in my brain right now that a lot of the details of this are kind of foggy. So I have a little print out here to help me remember all the details.

[00:02:28]Now, one of the things that I know for sure by looking through my document here is that this thing boots up to a console mode and there's a little bit of a boot prompt and you have to hit enter to kind of get Unix to boot. Inside of this thing has a full height hard drive which is an SD506 or MFM type hard drive and there's a SCSI interposer which then converts that to work with the main motherboard which is giant. It takes up the whole side of the case here. The interposer also handles the floppy drive and then there's a tape drive that's normally used for loading software onto it that is hooked up to the scuzzy bus directly. So, I think the first thing I need to do with this thing is I need to fire it up. I need to hook a terminal up to it and I need to see if I get signs of life to make sure we get that boot prompt. If it does power up, uh, we can try using the original hard drive, but I really don't like to use the old full high hard drive. I'd just leave it in there. I want to use a blue scuzzi and have this installed permanently in here.

[00:03:17]This is one of the new blue scuzzies which I think uses a Raspberry Pi 2350.

[00:03:21]And this runs a bit newer firmware than the one I was using back in 2024. And to even get the Blue Scuzi to emulate the hard drive that was in here, the Scuzi hard drive interposer thing, we had to write some custom firmware. So, I'm really hoping this Blue Scuzzy when I load up the hard drive image onto the SD card will just work in here without any modifications. If it requires modifications, then I think the Blue Sci team will be able to help me and then this will just live inside this machine so that I never have to worry about that spinning hard drive again. So, I think the first thing I need to do is clear off all the stuff off the bench. I have a bunch of like serial port adapters and stuff like that. Get all the stuff out of here. Get the top cover off. Take a look inside. Make sure everything looks like it's connected properly, and then see if this thing comes to life. Oh, and of course, get a terminal connected to it. One thing I did after working on this machine in 2024 is I went ahead and I found and ordered this uplogic 5000 terminal server. I got this because it has a couple cool features. I can SSH into this box and then access any of the serial ports on it. And all these RJ45s except for these two over here are just normal serial ports. This would allow me to connect up to all of the serial ports on the Plexus and, you know, open up multiple SSH sessions on my computer and interact with all of them simultaneously. I don't have eight terminals to plug into this thing. So, there's no real way to test it outside of using this. And even though I haven't touched this thing since 2024, I did fire this up and it still works and is still configured on my network and I was able to SSH into it. I found the IP address and therefore I think I'm ready to do testing. So I don't even really need to get a terminal physically plugged into this. I'll just SSH into this box and we'll do the work from my lab computer. I really forgot how big this computer is. It's uh quite large.

[00:04:57]There are these covers and these are for expansion cards that can go in this thing. I have no idea if any were ever made, but I think there were plans for additional serial ports and potentially a network interface. From my recollection, we'll find out when we open this up. I think the expansion cards go into the front of the case. But these two knockouts that are on here follow the same physical dimensions as the expansion port holes, whatever you want to call it, knockouts on the older Plexus models, the larger, more expensive version. I think they were trying to make it somewhat compatible with the older one. I don't know if the cards are compatible, but certainly the way that the ports go on the back are.

[00:05:34]Now, to remove the cover, there are two Phillips things. You just turn them uh counterclockwise. Looks like they're already in the unlocked position. Yeah, like that. And now the whole cover should slide towards me. Yes, it does.

[00:05:48]Now, the one caveat I remember on the other side where the motherboard is, I have a clock battery replacement because the original battery leaked all over the motherboard and caused some damage. So, I replaced the original clock chip with a Dallas module that has a hot glued CR2032 battery sitting on top of it and it sticks out a little bit and actually can catch on the case. So, I need to make sure when I pull this off, I very carefully do it so that the case doesn't take the battery off.

[00:06:14]All right, here we go.

[00:06:18]There we go. I cleared it. No problem.

[00:06:20]And there we have it. That is the inside of the Plexus. This motherboard is much larger than even the AGA Computraphic motherboard. There's our two 6810 processors. There's that Dallas clock with the battery attached to it. This is the power connector from the power supply. And I think this is this SCSI connection which is just not connected right now because I had the blue SCSI plugged into that probably last time I used this thing. Okay, we're going to do a little shaky camera cuz I got to show on the top side here. So, there's a few things going on that I recall at this point. There are two power supplies in here and I think the second one is generally for running the card cage for the expansion cards and I currently have that disconnected. I could see down there the the cord is not plugged in. I remember doing that specifically because if you're going to run that other power supply and there's no cards installed, there's this hilarious card just filled with resistors that is designed to load up that power supply, which of course, well, if I don't run the power supply, we don't need this installed. And I did all my original testing without this.

[00:07:21]So, I don't even know if that original power supply works. Of course, doesn't really matter if it works or not because we don't have any expansion cards anyways. So, it's all just installed in here and I left it uh for someone in the future if they ever want to make an expansion card. So, I have a little note from myself from 2024. PSU reset output bad. That's pin 4. I cut the trace to start the system. Pull pin 18 at this chip. U24 to 3 volts. Resistor blah blah blah. Okay. Well, hopefully Oh, I see there's like a bunch of notes to myself here. Oh, I have the serial pin out. That's right. Cuz I remember it was being it was a little non-standard.

[00:07:56]It wasn't using the regular PC one. All right. Well, hopefully I don't need this stuff here, but I'm glad I stuck it in there just to kind of remind me how things work. So, these are codes. I think power fail. I don't know. I don't really remember, but hey, it's something. And then finally, we have our SCSI devices. So, it looks like floppy drive is unplugged, which I remember this floppy drive, which is just a normal 80 track 5 and a quarter inch drive. This drive is bad. So, I had this unplugged because I was just testing with another floppy drive. I was testing with a 3 and 1/2 in drive. And then it looks like this is a SCSI cable and it's unplugged from what? Looks like everything is kind of unplugged here, isn't it? All right, I remember now.

[00:08:39]This is the archive tape drive which has a SCSI interposer that then connects to that edge connector. And this was the connection to the tape drive. So, I just have the tape drive unplugged cuz I don't have any tapes to use with it and it might be broken anyway. So, there's no point to even have that connected.

[00:08:51]It's kind of funny in a way. I'm basically refamiliarizing myself with this machine. So, this is the full height hard drive. This is one of the power supply modules. And yep, even though they look a little wonky, that's just the way they were on here. I have something down here in a bag. Okay.

[00:09:07]Looks like original capacitors that I took out of the power supplies. So, I guess I just saved those. And then there's a board that's inside the back side of the motherboard. I have my finger on it. This connector. This is the RAM expansion board. This board has 2 megabytes of RAM on it. Well, it's not even a RAM expansion. This is the DRAM for the board. We will revisit this RAM board later in the video, but I'm glad to see that it is installed, the original one, and that's what matters.

[00:09:31]And then, as for the card expansion slots, um, that go to those knockouts on the back of the case, there's an assembly that screws onto these four screws here. And the cards live in that space there. They go backwards that way.

[00:09:44]And that board with all those resistors was originally stuck in there. And this is the power supply that powers the card cage. And I think there's a dedicated connector. I think this plugs in around the top right here into the top of those card slots. And that's what powers that board. But as I mentioned, this power supply is not even powered up at all.

[00:10:01]It's unplugged from the output from the other power supply. The other power supply from my recollection feeds 300 volts DC into this one. And then this one creates whatever the necessary voltages to run the cards. All the bodgege wires. This board is covered in bodgege wires. I mean, there's like six or seven running in parallel right there. But this PCB is so big, it would have been so expensive to make. So, for a company like Plexus, which not a huge company, to just throw away PCBs this big is not economically viable. So, it's worth it to have someone manually fix all of the boards that have some kind of a production problem, which this one obviously did. And the serial number on this unit was like 3000 something. And for all we know, the serial number run on these started at 3000. So this could be like the 100 something machine actually made. Anyway, things are looking pretty good here. I'm thinking this system is pretty much ready for testing. So I think I should get the terminal server, the uplogix's fired up, connected up to the console port, which I'm pretty sure is this top one. Or is it this bottom one? I can't remember.

[00:11:07]It's either one at the bottom or one at the top. and then we should see if this thing comes to life. All right, here we are at the bench. I I remember I have an OBS profile set up for this. So, I have one camera pointing at the status LEDs.

[00:11:20]We can just see on the edge of the motherboard. This is the This is the Uplogix. Let's see if I can remember how to connect up to it. Yes, there it is.

[00:11:30]Okay, so two ports are configured on here. Plexus TTY6, Plexus TTY7.

[00:11:37]I don't need to look at the rest. How do I connect to one of these ports?

[00:11:42]I forgot how to use this thing. I think the way this works is you say port and you name the port. And now we type connect. No, terminal.

[00:11:54]Yes. Okay, there we go. So, I'm opening up another session. We're going to do port one two. And now we type terminal.

[00:12:02]So, when I power this up, hopefully one of the two ports it's connected to is actually the console port. And we should see it flow by maybe.

[00:12:11]I hope. I should have taken more notes when I was working on this in 2024 just so I could remember how to actually use the thing. All right, next step is getting the Plexus actually connected to power. Now, one of the thing about the Plexus is the power switch is a key switch that I don't have. So, I drilled out the lock so I could actually turn it with a screwdriver. You know, like a car you were stealing.

[00:12:33]And here we go. It's the moment of truth.

[00:12:40]Whoa, it's all come to life. There it is. Plexus self test. All right, so we're on port one. I need to look at like which physical port on the Plexus it's connected to. But this is good.

[00:12:54]This is freaking good. Oh. Oh no.

[00:13:02]Oh no. We're getting a failure.

[00:13:06]Prox TF F20 or FO26.

[00:13:11]I think there's a reset mode with the key switch. I'm going to switch it to reset and then out cuz I I vaguely recall that what I was talking about with that note on the on the thing about the power supply reset circuit is the power supply itself creates the reset for the motherboard. And if you turn the switch all the way to the reset position and back, then everything is going to be good. So, I'm going to try doing that.

[00:13:34]Okay, there we go. I just switched it to reset and then back to the run position.

[00:13:39]So, the reset position is like it's like starting a car, you know, you can turn it to start and you can kind of feel it go into that position and I think it might supposed to spring back, but it doesn't spring back on its own. But then you turn it back to the run position and then there's an off position.

[00:13:53]Will that make a difference or will I need to really dig into Oh, no. Oh, I just thought this thing would be working and um well, I'd be showing off some of the new stuff like the new blue scuzzy and whatnot and we wouldn't be having to try to do a repair. I mean, at least it's it's doing this. This is better than nothing. But I don't know what procs failed. FO26 means. Oh dear. I might have to watch my own my own supercut video of this thing just to kind of see well like if I ever saw this error before and what it means.

[00:14:28]No, it's still failing.

[00:14:30]Okay, let's just think about this for a second. I do not have the hard drive plugged in. Maybe this is the failure of it just saying that it can't find the interposer.

[00:14:43]Yes, I bet you it's that. Okay, what I mean by that is the SCSI interposer, which converts that SCSI to the SD506 for the hard drive and for the floppy drive. I bet you it's complaining that that's not found because there's nothing connected to the scuzzy bus right now because I don't have anything like the blue scuzzy plugged in. So, I'm going to just set it back to stock and see if the original hard drive works. And let's just see if we can get this thing even booting. Just it's all coming back to me now. But it's so funny how you you kind of forget stuff. And I didn't take copious notes last time. Although, actually, I did somewhere around here.

[00:15:15]Here it is. I do have a bunch of notes all hand typed. Chain of events.

[00:15:22]Okay, I do have stuff. Huh? Okay, check this out. In my notes here, I had a thing that says, "For the diagnostics, try pushing the backspace key or control question mark. This prints a Dell message and then exits and the test fails." So, now we're at the DMA menu.

[00:15:37]Can I do anything here? It says, "Push the exclamation point, which gives you the boot menu." Yes, there we go. Okay, well, at least my notes are good for something here. Scuzzy not found. Okay, there we go. I think that's our problem.

[00:15:51]I'm going to plug the interposer back in with the original hard drive. We'll see if we can just boot this thing off its stock hard drive. So, 100% bone stock configuration. First thing you always have to do with the Plexus is unplug it because when you have your hands inside, there's 300 volts DC up here on these pins and I'm going to have to figure out how to reconnect stuff. Okay, so the SCSI cable to the interposer is reconnected and the hard drive, the power was unplugged from it. So, that's reconnected and I think that's it. Let me plug this back into power now. I think I have it turned off. So, it should not turn on. There we go. And now I'm going to reset the system. And I'm bring it up. Turn it into reset.

[00:16:30]There we go. All right. Let's see what happens now. Oh, I should see if the hard drive is spinning.

[00:16:36]Sounds like it is. So, I think according to my notes at this point, I can push the delete key and that will exit out of what it's doing here. I'm just going to do that. E pushing exclamation exits to boot. And we hit enter.

[00:16:53]I can hear the hard drive doing stuff.

[00:16:56]Yes. Yes. I was thinking we'd have to do a repair. I just forgot how this thing worked. And I forgot that this cryptic error proc failed. Is it talking about the fact that it can't find the interposer? Are the lights doing anything? Yeah, there's some blinky blinky going on. Oh, it's running. It's running everyone. System 52 release 8 single user mode. I do have a little bit of stuff on the in my notes here about this. So, we have to go into multi-user mode. I'm very disheveled. I thought I had my information in order to run uh to make this video and uh clearly clearly not. The good thing though is that this thing is definitely working.

[00:17:39]Okay, so we are in single user mode and we got to do init to go into multi-user. Ah yes, what's awesome here is that the uplogix just allows me to connect remotely like this.

[00:17:54]It makes it so much easier.

[00:17:56]Theoretically, once we enter multi-user mode, we should see a console appear on this other port as well. Theoretically, at least. It's quite amazing, isn't it, that that old hard drive that's in there still just works perfectly. Now, it's fully image, though. So, if this hard drive just gives up the ghost and dies right now, it doesn't matter. We're not going to lose any data. Nothing that's was on that drive hasn't already been archived. So, we're all good there. H looks like in my printout I even had the uh original document kind of describing what this machine is. The P15 and the P20 multi-processor design. The high performance of the machine is made possible by a special multipprocessor architecture optimized for the Unix operating system. The architecture links a powerful 16/32-bit job processor with a second 16/32-bit IO processor of equal power to remove IO processing overhead from the job processor. So even though this is a multipprocessor system, the reality is one processor is handling there we go we got our second log prompt. One is handling actual running of execution code and the other one is entirely handling IO. But that's cool in a way because if you are say doing disk operations, it doesn't freeze the entire system because it uses the other processor to go service that request meanwhile the job processor can run all these user tasks and it doesn't slow down just because of disk operations or serial operations. Okay, so I think I have an Adrian account.

[00:19:20]I don't know what the password is.

[00:19:21]Adrien. Adrien.

[00:19:23]Oh, that was it. Cool. So there we are.

[00:19:26]We are logged into two sessions at once on the Plexus and it's up and running.

[00:19:31]So, I don't need to do a repair. I can move on to trying all this other stuff.

[00:19:35]All right. So, I need to think about what I need to do next. I think the first order of business is I need to try to get this new blue scuzzy working because this is the thing that if this doesn't work, we're going to have to make firmware changes on the blue scuzzy itself, which is going to require some time. Here is the Blue Scuzzy Ultra I'm going to be putting into the Plexus.

[00:19:54]This is the new version of the Blue Scuzzy devices that uses an RP2350 as opposed to the old 2040. It's got more power, is much faster, and even though this is only an 8 bit wide version of the Ultra, there's also a 16- bit wide version if that's what you need. There's a little module that plugs in. This is the wireless module. I don't know if this is quite ready for prime time. There's also another adapter that you can connect up, and this gives you audio output, both 3.5 millimeter and an internal audio connection. This would be used for instance to emulate a CDROM if you are going to have the audio feed into your sound card into the motherboard something like that. Now last time I was using the Blue SCSI on the Flexus because that thing uses that very old interposer that's from the early 80s actually predates the SCSI standard. Some modifications have to be done to the firmware on the old Blue Scuzzy to allow it to work. Now those changes have been forward ported to this Blue Sci Ultra and I have the firmware links that Eric sent me. I'm going to try to load that onto here. I don't have the original firmware that works with Plexus on my other Blues Scuzzy anymore because I need to use that thing with other stuff in the meantime. I'm sure I could dig it up if I need to. And I guess let's just not worry about that right now. And let's get the firmware loaded onto this and see if this works.

[00:21:04]Updating the firmware on these is really easy. You can go into bootloader mode by holding down this button, plugging in the cable, or you can do it over the SD card, but I'm just going to do it this way since it's the way I normally do it.

[00:21:16]When you use the bootloader mode with that button, you get a drive letter H in this case. And I can copy the specific version of the firmware for this device, which should be this one. Blue Scuzzy Ultra. This is the patched version that has the Plexus patches UF2. If you're going to do the SD card update, you just use the BIN file. Put that on your SD card. Drop this in here. And this should just copy the file, and then it will reboot itself if it works, I guess.

[00:21:44]Yeah, there we go. It's doing the firmware flash itself as far as I'm aware. And the drive letter should disappear. And there we go. I think this thing is booted up now. It's not going to be unhappy because there's no card in here. So, we'll unplug the power. And this is my other blue scuzzy, the one that I was using with the AGA. So, this is the card. It does have the Plexus images on it. So, I'll just reuse this.

[00:22:04]So, there's the HD0000 AGA RIP file. I'm just going to put that into drive images. Oh, it's already there, but I'm going to replace whatever's in there.

[00:22:14]And there is HD0ero fixed plexus which if you notice the date 2024 that would have been the last time that I booed the Plexus off of Blue SCSI. So the idea will be I'll be able to plug this Blue Scuzi directly into the motherboard. So I'll disconnect the interposer entirely and then we will see if the computer gives us that proc error that we were getting when the interposer was disconnected. I'll just let the blue scuzzy hang just like that. I don't know if I even need to wait for the diagnostics to finish. So, we're going to push backspace and then I'm going to type exclamation point and hit enter.

[00:22:50]Oh, no. Unknown controller assuming default.

[00:22:55]You know, I'm just thinking, there it is. Proc failure and fo 26. It's exactly what we got last time. Okay, so that firmware on this new blue scuzzy does not work. I suppose what I need to do is take this one and try to find the firmware from 2024 and load it onto this one, which is definitely what I was using on here. And maybe we can get it booted. All right.

[00:23:19]Well, unfortunately, that was not successful. I tried a few different things. I turned on the debugging. The definitely the blue scuzzy being seen by the Plexus, but the Plexus just not happy with the way the Blue Scuzy's presenting itself. The Plexus at this phase in the boot process looks for the OMT 5200, which is the SCSI interposer.

[00:23:37]It looks specifically for that. You can't even plug in a normal SCSI hard drive and have it work. It has to be the Interposer. And unfortunately, it seems that the new firmware that I loaded onto the Blue SCSI Ultra isn't quite giving the computer what it needs. And I even found the original uh Blue Scuzi.ini file that talks about the quirks you need to enable. And I did that and it didn't really make a difference. I looked on my download directory, which I still have files from 2024, to see if I had a copy of the firmware for the older blue scuzzi because I would just flash that back onto there and then I could keep going with this video. But unfortunately, I wasn't able to find that either. I must have saved it somewhere else on my computer, maybe the desktop, and then I, you know, periodically clean it up and I probably deleted it thinking I don't need this anymore. So, for now, so I can keep making this video, I need to put that stuff aside and I'll try to get back to it. And I guess I'll put back the original interposer with the original hard drive and keep my fingers crossed that it survives until I get the blue scuzzy working. Next thing I'd like to do is I'd like to hook up all eight of the serial ports to the terminal server so I can just test out to see if all the ports work. I don't think I've ever actually tested them. I have a bunch of these types of adapters that go from 9-pin to RJ45 and these were sent in by a friend of the channel and supporter Vax 11. And the cool thing about these adapters is you can reconfigure the pins any way you need. And I already had one of these that was configured to work and that was the one I was using with the console port. So I just opened it up and looked at the color coding and then I just replicated all of these other ones to be exactly the same. That means I have a total of six of these nine pins that should work on the Plexus. Now the reason why I had to do all this is the 9-pin serial connection is not standardized. And this machine's from 1985. They probably developed it in 1983 or 82 or whatever. And it's quite possible they use these 9 pins on some of the older Plexus machines as well.

[00:25:21]So, they were just following their convention. But when IBM came out with the IBM PC, they had a 9-pin standard. I don't Well, I don't even know where that actually came from, but I think it was IBM with the PC. And in 1981, that just sort of became the deacto standard that everything kind of uses, but the Plexus did not use the IBM standard, which is why I need all these adapters. And then it did just use the normal 25 pin standard for RS232. That was already well established by the time this machine came out. So, let me spend a moment just to get all this hooked up and I'll configure the terminal server and then we'll see if we can get eight SSH sessions open to the Plexus all at once. Okay, everything is all hooked up.

[00:25:56]The Plexus is so loud it's so noxious to have on in the house, but I have all eight serial ports connected up to the ULLogix 5000 here. And all of the little adapters are plugged into the Plexus, including an extra one for the 25 pin since I didn't have one for that. So all eight ports are now available through the terminal server and let's get connected up with eight SSH sessions.

[00:26:20]You see how sluggish this seems and really this again is the terminal server. It's just really slow. Plugged into the console port directly, it wouldn't be this laggy. But of course, the nice thing is I can open up multiple sessions here. So I duplicated the session. Let's log in. And now I well I've reassigned the Plexus to be on different ports. So you can see we have TTYS0 through S7 available on ports 21 through 28. So if we go to port 22, I type terminal. Here we are. I can log in as myself, not root.

[00:26:54]And as this logs in slowly because the computer is not very fast and we're using the original hard drive. If I use the hoot command, we can see Adrian's on TTY1 and root is on console which is TTY0. And there we go. What took just seconds on video. actually took me quite a while to get set up. Just getting all those little adapters ready, configuring the terminal server, but I am now configured and logged in on all of the ports. So, we're using the Plexus for well, its capabilities of running all of these sessions all at once. Now, what can you do with multiple logged in users? I mean, you're going to do the first thing that, you know, everyone thinks of doing, using the wall command to spam everyone. Stop eating all my candy.

[00:27:36]And I think it's control D. And there we go. It sent that to all the users simultaneously. So we saw it on all of the session windows.

[00:27:48]Now unfortunately there's no quick and dirty chat app to talk in between these users. At least I'm not aware of that, but I think we can use the wall command even from a non-privileged user. So let's see. You ate all your candy. Stop blaming us. Yep, it works.

[00:28:04]broadcast message from Adrian.

[00:28:08]I have a feeling like back in the day, multi-user Unix systems probably had limitations on the wall commands. Only the privilege user root could do it and you didn't have this kind of spamming going on by regular users. Now, I talked about this a little bit in the series on the Plexus, but the Unix that's on here is this old version of Unix and it is so limited. There are so few commands on here that you're used to. Like for instance, you might be used to using dssage to see console messages. Nope, that doesn't exist. How about using exclamation point to access the last command you used? Nope, that doesn't exist. You can of course push up or down. It's not like we're using bash or whatnot or even ash or any kind of shell that's capable. I think this might be shell actually now I think about it. No.

[00:28:51]Well, I'm not sure. Let's see. I think there's a process command pse. Yeah.

[00:28:55]Okay. Yeah, it's just regular shell is all we're using on here. There hardly are any processes running on this machine. You know, it's funny. You look at a modern Unix system just like nothing's logged in or one user's logged in and there's so much stuff going on.

[00:29:06]All this kernel things. Well, you can see there's like hardly anything happening on this machine. Just it's sitting here. There's init. There's a swapper. I guess that's for virtual memory. There's the lo the interactive login for the console. And we have a cron job that is running. So, there is cron on here. And this must be for printing, I'm assuming. LP sketch. Is that for like LPR maybe? Either way, not a whole lot available on the Plexus. So, it's kind of interesting to use just because it's it's so limited. But, I guess you have to think about the fact that in 1986 when I think this machine was finally released, it was kind of cool and advanced to have something that wasn't a big mainframe or, you know, mini computer that could just have multiple interactive Unix sessions. And I have to think that the Plexus B20 was trying to be cost-effective for your small office to allow you to kind of do this uh shared work stuff somewhere on here. And I forgot where it is. There is actually like a word processor/off suite I think with a spreadsheet and stuff like that. But at this point I am not 100% sure where that is. Like there's this tutor directory. What is this? You can ignore the date 2017. This stuff is not accurate. In fact, what does this think it is right now? Now, I think it's 1998 right now. I'm pretty sure people who did analysis of the hard drive determined that that date there is not accurate. It wasn't like someone was actually using this computer in 2017.

[00:30:26]This was just uh left over from I think when they untarred these files. I think it just got assigned this date incorrectly. I can tell that at some point this machine was connected with a modem to other Plexuses or might not have been modem might just been a serial connection and it was using UUCP which is Unix to Unix copy. It's a scheduled process that can run that allows email to work because if you notice here, this thing does have a check interval and it has mail. That email was made possible by this UUCP process that was running in the cron job. Um, I don't even know if we could use the normal commands. Okay, I guess there's nothing in the cron job right now, but at some point there was and this thing was transferring email and I think files between other machines using this UUCP. Now, I bet you there's people watching who want to know, hey, like how can I use this Plexus? Well, one of the things I mentioned, I'm going to be bringing this to VCF. So, if you want to play with it, I'm going to bring a VT220 terminal to be plugged into it so you can poke around and, you know, play on the real thing. Now, there is another way though, if you can believe it. Back when I was first working on this, the incredibly smart people who were helping me and we were reverse engineering how the board works so I could get the repairs done, they created an emulator. And even better, it's a web-based emulator and it's open source.

[00:31:42]So yeah, you can play around with this at home. Of course, I'll have a link in the description if you feel like poking around on the Plexus, but let's check it out. Here it is right here. So, we're booting up. I think it's actually booting right now. I don't know if this emulator runs at full speed or not, but here it comes. And it is running the same hard drive image that came off the real machine. So, everything that I was just showing you um on this thing is on here. So, hitting the about button here.

[00:32:10]This is an emulator for the Plexus C20, which is a Unix server from the 80s. It runs system 5 on a dual 6810. This is all stuff we know. Adrian Black from Adriation did some videos on the topic of getting it to work. And there are five links right now. The emulator is written in C. While it can be run as a native program, the version you're looking at right now is compiled with web assembly using emripton that uses XtermJS for the fancy terminal which is connected to emu script via xterm-pt.

[00:32:34]Plexus emulator copyright 2024. Sprite TM and contributors licensed under the MIT license and the source code is available on GitHub. Let's just go take a quick look at that. Stuff will be in the description so you can play around yourself with this. I mean, you could even build a little replica Plexus, like a miniature one with a Raspberry Pi inside. Just running this and there you go. You have a full Plexus. Now, with this particular web version of the Plexus emulator, I think one of the problems is the disc is ephemeral. So, as soon as you leave and go back, like anything you've done is gone. It's not like it saved it. So, you're really going to need to compile a local version if you want to run it and actually do any kind of software development on it.

[00:33:06]There are a couple folks I've spoken to in the leadup to VCF when they bring in this computer where they want to develop a little bit of software and then try to get it running on the real computer.

[00:33:13]Now, there's no reason why it wouldn't work on the real computer because really this emulator appears to be completely faithful versus what we have here other than potentially the speed not being exactly accurate. It's pretty slow and kind of similarly slow to um what's happening on the actual Plexus. Are we in in it two right now? Yes, we are. Run level two. Okay, so let's see. Yeah, it's pretty snappy, especially compared to what we were getting compared to the terminal server. I mean, I don't know. I guess it's like somewhat similar.

[00:33:42]Definitely the typing is smoother.

[00:33:44]Here's the Unix kernel on the real machine, November 27th, 1985. 180 28,000 bytes. And that is exactly what we have on the emulator. So, it really is running exactly the same code. So, to me, it's absolutely fantastic that this Plexus emulator exists and it has existed for a couple years now. But if you're just seeing it now and you want to play around with this thing, then yeah, use the emulator and download that source code and make a local port of it and do whatever with it. It is so absolutely mind-blowing to me that we have a Plexus emulator. This machine, like I said, almost certainly the last one left on the planet, was obscure and forgotten, and hardly anyone had one of these. And now there's a full emulator that anyone can run, fully open source, and it was all made possible by the fact that the hard drive in this thing survived. And we were able to image it and then get that out there to the general public, which has resulted in emulator. If the hard drive had died, we would have never been able to probably get this thing working just because all the technical documentation for the way this computer works was on the hard drive. So once we able to recover that information, I mean, I don't just mean the operating system was on there, but there were actually design files for the Plexus that gave us all the information we needed to reconstruct the electronics and figure out how the motherboard works so I could repair it. Again, that's all back in that original series, and I don't want to rehash it too much in this video, but I highly recommend you check out the supercut or just watch through the individual episodes and just kind of follow along with the discovery process of figuring out all the stuff there was to discover on this machine. It's absolutely mind-blowing. Now, one thing that a viewer shared with me that I think is so freaking cool is this. This is a 3D printable ATX or a mini ATX or maybe microATX case that is modeled after the Plexus P20. It's called the PC20. And it looks like an extremely faithful reproduction of this machine, which is amazing to me because I'm the only one who has this machine right now.

[00:35:40]No one's ever seen this in person except for myself and a few people who have been over here to the basement. This recreation to me uh really just captures the absolute feel of this machine. And and of course, you can put a regular mini ATX motherboard in here. And this is the motherboard tray for your modern computer. It would be amazing if someone took this 3D case and even shrunk it down more, put a Raspberry Pi in there, and then ran the emulator um with actual serial ports. So, you could have a miniature replica of the Plexus. How freaking cool would that be? Anyhow, I really have to give a huge props to Tank Girl who designed this case. It looks absolutely fantastic. And yeah, I'll put the link down in the description. So, another problem with the Plexus is that it's really hard to get software on and off of it. It was really designed to use the tape drive, which is a scuzzy tape drive for loading bulk programs onto it.

[00:36:29]So, there's whatever that word processor that's somewhere on here was certainly loaded off the tape. And if you're going to be doing a hard drive recovery, say you put new hard drive in here, you need to reload the OS that would have come on a tape as well. In fact, the ROM has provisions for loading directly off the tape drive and loading it onto the hard drive. So, it kind of treats the tape drive as a file system. So you're able to run standalone programs that's outside of Unix and then that will format and image the hard drive. In the video series I showed a little bit of that where I was using those standalone programs which are available on the hard drive as well by the way. So you can access those in the bootloadader and then you can do stuff like reconfigure the nonvolatile RAM which is in the little clock chip the Dallas clock chip that's on there to set up you know which SCSI drives exist and stuff like that.

[00:37:07]So that's all there and it is possible actually to access those standalone programs on a floppy disc as well because that does have a disc drive in it. Problem is the SCSI interposer which is what converts the MFM ST506 hard drive to work in SCSI also is connected to the regular MFM style Sugarart floppy drive interface. So if you take the interposer out of the equation you're using the blue SCSI then you don't have access to the floppy drive anymore. And it's possible this can emulate the floppy drive because I think it just shows up as like a block device a scuzzy block device under a different ID or maybe it was a logical unit number alone. But either way, I was never able to get the Blue Scuzzy emulating the floppy drive. And that was when I was running the Blue Scuzzy firmware that could at least emulate the hard drive.

[00:37:47]Right now, I still don't have the the correct firmware to run on the Blue Scuzzies to get them to emulate the hard drive. So, yeah, I'm still running on the spinning disc. But one thing we have figured out is this machine does have a few utilities on it that allow you to at least transfer over the serial ports.

[00:38:00]Back here on the console port, we actually have Yuyu Encode and Uyu decode available to us. What that lets you do is take a binary file and convert it into something you can send over email or like an asky connection. So let me do an example. I think I can cat a program like Vue and pipe that into you encode and uh yep, there we go. So I hit control C. So what this does is this is the conversion of the binary data into a way that's transmittable through something that doesn't support binary data. Well, obviously if we're copying and pasting into a terminal session like through Putty here, I can't I can't paste binary data, but I can paste all printable characters. Now, of course, if you're going to take a a 1K file and you're going to enuode it, it becomes larger. I think it becomes like 2K or something like that. In fact, let's just test that because I I forgot. I haven't thought about this stuff in so long. So, we'll start with a small file. I don't know what this 300 is. So, we will cat 300, which is 8K or so. It's 8464 bytes and we're going to output that to 300- UU. So, we shouldn't see it on the standard out. Okay. So, now if we take a look at 300, let's see how much larger the UU version is. Okay, there we go. It went from 8,400 bytes to 11,000 bytes.

[00:39:12]So, it's larger, but it's not like it's massively larger. Now, I think what we do have is we have the pack command. So, let's cat the 300 again, and let's pipe that to pack, which I think should compress it if I recall. And we might need to output this to 300.pact.

[00:39:30]Okay, let's see what that did. I am not 100% sure. Uh that made a zero byte file. So that clearly did not work. So the idea would be you would take a program or a file or whatever. You would compact it or compress it and then you would encode it then send it over that connection and then you would reverse that process on the other end. Now the idea is I can send and receive this stuff through my putty session here. You don't even need to use a protocol like Z modem or anything like that. And it should theoretically work. Although that doesn't seem right.

[00:40:03]Is there a tail command on here? I don't know why there's a dollar sign in here.

[00:40:08]I thought every record has to start with an M. You know what we could do? Let's just try decoding this file that we encoded just to make sure this whole process works because I don't know actually. First, we'll just move it all to the temp directory. Okay. Okay. So, first I'm going to rename the original file to be 300.

[00:40:25]Now, we're going to uyu decode 300.Uu, which should have made a new 300. Oh, it did. And it's exactly the same number of bytes. Either way, the UU encode decode process does allow me to transfer stuff on and off the computer, which is excellent. So, at a recent retro meetup here in Portland, I was talking to one of my friends, Ian, who runs the YouTube channel Tales of Weird Stuff. So Ian was telling me he wanted to come up with a multiplayer Tetris that he could show off running on the real hardware up at VCF and we could potentially have some head-to-head Tetris action on this old multi-user Unix system. So Ian, who has a full-time job and is super busy, isn't quite ready with this multiplayer Tetris, at least not at this point, and hopes to have it ready for VCF. But I was chatting with another viewer who was able to whip up kind of a quick and dirty multiplayer Tetris, which actually freaking works. It's not really super playable. Like the way it works, it's um the RNG doesn't work properly. And with multiplayer Tetris, you want to be able to send junk to other players and stuff like that. None of that's really working. But using UU encode UD code and transferring the source code over was C file and then compiling it, I was able to get this quick and dirty metric we're calling it working. Keep in mind that this is quick and dirty and really not working properly yet. But it's just sort of a proof of concept that this sort of stuff is possible.

[00:41:44]And let's run it. And I want to have a second session open and I can show this in operation. I I do think it's really cool. I'm really looking forward to a much more capable version of this game uh once Ian gets his all kind of worked out. You can pick color or monochrome.

[00:41:59]And um it's using VT100 code. So you have to have a VT100 terminal, but the the mono mode is obviously if I'm using like an original terminal doesn't support color. And now we should be able to type a name. And there's the game.

[00:42:11]And it's running. It's waiting for players. So, let's run metrics over in this session. And there it is. I'm going to call myself John.

[00:42:22]And it didn't work. This was working earlier. Honestly, that is super hilarious.

[00:42:33]I wonder if it's not working because I am logged in as root on one and Adrian on the other one. And maybe the shared memory is only between users.

[00:42:43]This shows how little testing I've actually done with this. Very, very little. You did get an idea of what the color looks like. I mean, it looks good, right? Using the color anti code. So, this is like PC style antsy versus VT100. Okay, let's see if this actually works now. So, I don't know. I'll put this in color mode as well. And I'll call this one John. And maybe now we'll get two player mode. No.

[00:43:05]What a fail. What an absolute fail.

[00:43:08]Let's try running it again. color major in earlier this was working but now it doesn't work so I don't really know what's going wrong here but what would happen is the other players board would appear where it says waiting for players and then well it's not really working properly so even if you completed Tetris it didn't send the junk over to the other person which is kind of the whole point I really think multiplayer Tetris should work like game boy link cable Tetris did where it would send you the junk with like an empty space so you could clear those lines and then you could send junk back to the other person and you were playing headto-head with the same random number generator. So that way you got the same order of pieces. So it was really like a skilled test and then whoever topped out first that would like end the level. Anyhow, proof of concept not really working.

[00:43:57]Hopefully by VCF in Seattle we'll actually have a playable version of this that really will work. I guess the one issue is I only have one terminal to bring. So hopefully people at the show maybe can bring other terminals so we can have a multiplayer Tetris with Plexus. Theoretically we'd have eight players with like a tournament style, you know, where you're always playing headtohead with another person and there's bracketing and stuff like that.

[00:44:20]And I doubt the designers of the Plexus ever thought people would be playing headtohead Tetris that totally failed on video when honestly um this was working earlier. I just don't know what happened here. Players online shows zero. So let's start this.

[00:44:40]Okay, players one waiting for players.

[00:44:43]Now we'll try this one over here.

[00:44:46]Players online. It shows one as opposed to zero before. So that that kind of tells us tells us I don't know. I was just getting stuck on the username. Oh, there it's working. Okay, so now it works. Adrian versus L1. Tell us. Okay.

[00:45:02]So, um, if we hit space, that drops the piece. Yeah. So, you can see this is this works. Yeah. I really thought that demo was going to crash and burn, but in the end, I was able to actually show it actually working. And it really just was a matter of copying and pasting the source code for this game. Um, well, it was Uyu encoded and then I pasted that in and then I was able to compile it.

[00:45:24]Now, you might be asking, why did I have to encode source code? Shouldn't that be human readable? Like it was typed by a human. The problem is is the terminal server the uplogix the escape character to get out of it is the tildy and that is not changeable unfortunately it's hardcoded so if you type the tildy it doesn't get sent to the computer and it eats things now I can demonstrate that here this is still just sitting here running I can hit the tildy and hit enter and now we're at the prompt so I can type terminal to get back and I think it's going to try to keep running the game I we better just exit out. So, here we are at the prompt. Same thing over here. And yeah, so the TD if I hit it twice, then it prints it and actually sent it to the computer, but one time it doesn't. And there's a bunch of escape sequences you can do to like change settings on the port and stuff like that and it's hardcoded and you can't change it. So unfortunately with a lot of source code, a lot of times you end up with tildies in the actual source code and while you're copying and pasting it in, it breaks it. But UyU and Code does not use the TY ever. So therefore, copying and pasting it completely works.

[00:46:29]So this is strictly a limitation of the uplogix's terminal server and wouldn't affect other terminal servers. Probably they would have a different escape characters or you could disable it altogether. And if I plugged in directly with a serial cable to my PC and used putty, it would be fine as well. Moving on, there are still more to look at and talk about when it comes to the Plexus.

[00:46:46]One of the things about this, and I alluded to this earlier when I was looking at the hardware, is this thing is limited to 2 megabytes. And that's shared across the entire operating system and all logged in users. Not great. The designers of this thing had envisioned that there would be an 8 megabyte and 4 megabyte RAM card for it.

[00:47:01]So it would bring it up to a total system memory of four or eight megabytes. It is only a 6810 processor.

[00:47:07]That means the entire memory space is only capable of having 16 megabytes total. But that includes all the IO space and everything else. So the designers had envisioned 8 megabytes would be the amount this machine would be able to support. It does have a custom memory mapper as well and that's because 68000 6810 doesn't have a built-in memory mapper. So there is a custom TTL logic with SRAMM memory mapper and of course the page size and all that stuff matters for how much memory it can support and I think that memory mapper does support mapping in the entirety of megabytes but no RAM card was ever made. There were actually emails on the hard drive that were to the designer of this computer that said hey please design an 8meg or maybe it was a 4meg RAM card but they never got to it and I think the project was canceled like this machine was discontinued before that ever happened.

[00:47:54]So, no card was ever built, or at least we don't think any card was ever built.

[00:47:58]That's not to say we don't have one now, because I have potentially an 8 megabyte RAM card right here. This card was designed and built by a friend of the channel, Super SVGA, was one of the contributors that was really working on helping reverse engineer the computer early on in the Discord channel on Usagi Electric's Discord server. And this card has never been tested, and we're going to test it today. Super SVJ is local here in Portland, but they actually never been over to my house to look at the Plexus and made this completely based on the design files that we found on the hard drive. There was pinouts for the expansion connector. And one day at one of our meetups, Super EGA came up to me and said, "Oh, here you go. Take this." And I was like, "What? What?

[00:48:35]Wait, Lexus RAM card? What?" So, this may not work. And that's because, of course, it has the DRAM on here. It has level shifters, but it also has some glue logic, which is a CLD that takes all the convoluted logic through this connector and then does the refresh and I don't know whatever else is going on on the RAM card. So, this thing replicates all of that down to this small little board. And we might plug this in and might short out the machine and it might not boot at all. I mean, we don't know what's going to happen, right? This may work or it may totally not work. So, really, the only way to know if it's going to work or not work is to well, give it a try. This is a programming header for the CLD if we need to actually update that glue logic.

[00:49:17]Back side, there's nothing really to see. And on the front side, there you go. RAM prototype. And it is open source. All right. So, I think what I need to do is get repositioned. And I'm pretty sure I have to take the main motherboard out of that computer just to get the RAM board out. And then we can try this and see what happens. To get this out, I never remember the exact process. I took all the connections off the back. There are two thumb screws, one up at the front and one at the back.

[00:49:45]Now, I do have all these little gender changers attached to the serial ports, and I'm really hoping that I can get them through. I think they'll fit through the case. Just motherboard is free. Oh, yeah. They fit.

[00:49:58]There it is.

[00:50:03]There is the back of the motherboard.

[00:50:05]Oh, I completely forgot that I installed a custom reset circuit on the back. So, the battery used to be in this position, the one that leaked and caused some damage. I don't think the problem with the reset has to do with the battery, though. It came from the power supply connection. I think it was this trace right here. I have an arrow pointing to where I cut it. So, I cut that connection. That is power going to the AT tiny. And this just generates the reset signal when it gets powered on because this thing has an automatic reset. And that fixes the problem. So, you don't actually need to put this computer into reset when you first turn it on, like my notes said. So, I completely forgot about this. I probably would have remembered I did that if I just rewatched my videos. Here's our RAM card. So, this is what we're going to be replacing. What I do need to do first is before I power this thing up, I'm going to plug in the new Super SVGA RAM card once I get this off, of course. And then I'm going to check polarity just to make sure it has a little voltage regulator.

[00:50:57]So, we'll just make sure that 5 volts is going to the correct place because obviously if the polarity is reversed, like say the connectors wired up backwards or something, then that would just catastrophically damage the RAM card and we don't need to do that. Now, the Plexus won't work at all without the RAM card, right? This RAM card is all of the system memory for the whole computer. So, this thing is essential for this machine. So, here is the new card. Look how tiny it is. It's just so small.

[00:51:26]We're in continuity mode. It would be completely irresponsible of me to to basically power this up without at least just checking that we have ground and 5 volts going to the regulator properly.

[00:51:35]So, I'm going to use the AT tiny since uh pin one is down here bottom left corner and I know that the pin opposite that pin 8 that is 5 volts. So, this is 5 volts off the power supply and ground is coming off this bottom connector because nothing is labeled on the back side here. So, I don't really know what's what. So, first we'll start at 5 volts and I'll go on the regulator.

[00:51:54]Okay. Actually, you know what? No, there's a problem there. I'm not sure you can see the multimeter, but 5 volts, we have 6 ohms. And the ground pin 6 ohms as well. My assumption is this little card is causing a short of some kind. Let's remove it. And we'll check for a short on the AT tiny now. Cross the ground M5 volts.

[00:52:19]Uh, it's 8 ohms. So that I could totally expect because all the TTL logic on here, even though the power supply is not plugged in, 8 ohms seems reasonable between the 5 volts and the ground, but certainly not less than 1 ohm. Yeah, 8.2 ohms. So on this regulator, I was checking the outside pins and they were both short. But also when this was plugged in, even on the AT tiny, it was showing as a short between those two pins. And I'll just plug this back in just to double check I wasn't making a mistake or something. But I'm pretty positive that we had a short between the 5 volt and ground with that board plugged in. Yeah, we're at 02 ohms now with this board plugged in. So powering up the Plexus with this connected would probably be catastrophic. It would probably destroy something on this board, whatever is causing the short or it would just cause the power supply to shut down or it would burn traces on the board. Let's take a look at this on the bench and maybe I can see a problem.

[00:53:18]Maybe there's a bridge or something.

[00:53:20]Well, one thing I can do is I could just compare the new card to the old card and see where the short might lie. And I think I figured it out. So, these two pins, these are connected together, but it's also connected to this pin. This thick trace here, this is actually the 5 volts that goes up and over to the voltage regulator and also is powering up these level shifters. Well, the problem is if we take a look at like say this pin here, this is 5 volts on that chip, which is the same as 5 volts there. So, I'll just pick it up right there on that corner. So, this is 5 volts going directly into that IC. And it looks like these three pins are all directly on the 5V rail. But this pin, this pin is not. This is probably on ground. Let's go to the ground pin. Yes, it is. That is ground. And if we match that up to what we're looking at here, uh I think let's see here. So it's these three that are 5 volts right there, which should match on this, although it's not connected there. Interesting. But yeah, this pin is erroneously connected to the 5 volt rail when that is actually a ground pin. Now, what I'm not sure about is for instance, this thick trace going to there. Is this ground or is this 5 volts? Cuz obviously this is going to be connected to here. Yeah, there's no actual short on this board when it's not plugged into the Plexus. Yeah, that beep you heard was just the capacitor charging. The 5V pins and the ground pins that are on here aren't shorted together over here. I think this might be the only place that's causing that short. So, I kind of have nothing to lose except for just using my engraving tool here and just cutting through this trace like there and also right there.

[00:55:05]My assumption is there's going to be ground pins elsewhere on this board.

[00:55:08]This thing has a flood fill which is probably where the ground is being picked up. So, these chips here, yeah, probably picking up ground on the part of the flood fill, I'm imagining. It's hard to say exactly, but this this is a problem right here. This this needs to be corrected. Something else I could do is I could just remove this pin entirely.

[00:55:30]That's probably the easiest thing to do because it should be one of these top pins here. So, I'll just desolder this and then I'll just pull it through because if it's not connected to a flood fill on this side, then well, yeah, let's try that. Okay, there we have it. That pin is loose. So, on this side, there it is. I'm just going to peel this out of there so it can't possibly make a connection. So, with that out of there, the question is, if I put this back in the board, we shouldn't see that short anymore. Board is back in. Let's just double check. We don't have a short anymore, which I'm sure we do not.

[00:56:10]It's We still do. Yeah, we still did.

[00:56:14]0.21 ohm with this plugged in. So, there are other lines that that are shorted.

[00:56:20]So, that kind of makes me think that the pin out that Super SVGA was working on with this board maybe wasn't the actual pin out because he never looked at the actual board. He didn't have the RAM board or the motherboard to work from.

[00:56:35]It's quite possible the documentation he worked from was faulty. And there must be more connections like that where the ground and the 5volt are connected together on here. And unfortunately, what that kind of means is that I bet you a bunch of the data connections may not be correct as well. But I'm going to spend just a few minutes to kind of do what I just did and look for maybe where there's a the 5 volts and the ground are swapped around on here and maybe I can do the same thing where I can lift the pins. I don't know. It's possible, right? Well, unfortunately, the RAM card is not going to work. There were too many other connections that were not correct. And I just have to assume that the information that we found on the hard drive that described this connector didn't match what's actually on the hardware. It seemed like some of the information on the hard drive was preliminary and I think things had changed on the actual final production board and that's probably why Super SVGA ended up making this and so many of the connections are different. That's not to say there will never be an expansion card for the Plexus because it could be with some reverse engineering of the actual RAM card that's on that machine.

[00:57:38]Then maybe a replica could be made. Part of the issue is I'm a little reluctant to let that card go because obviously that computer doesn't work at all without that RAM card and it's the only known example of that card and it would be terrible if it somehow got lost because then the computer would be 100% useless. Now, one thing I've never actually done with the Plexus is I've never actually run it with the front card cage installed. And I wanted to just sort of put this thing back together fully. So, I have the card cage installed and I have this big resistor board installed as well. What I didn't do is I did not reconnect the power supply connection. That's just tucked up here. And that secondary power supply, this one right here, is also disabled.

[00:58:19]Hopefully, this computer can run in this configuration because this whole time I've had this machine together, but this has not been installed. And I was worried that just having this card installed would somehow create issues and stability problems on the motherboard. But I can see right here a bunch of transceivers. So hopefully those are just turned off when you're not addressing the cards, and that shouldn't cause a problem. Now, what I did do is I broke out my one terminal that I sort of keep in stock as opposed to ones that are big and I give away.

[00:58:48]So, this one is the one that I've kept and it's a deck VT220 amber screen.

[00:58:53]Nothing special about it, but one of those TTY ports is plugged into the terminal. And really, I'm just preparing for VCF because I want to use this terminal with this computer. And I think that's a perfect combo. So, let's see if I can get this all working and make sure that the Plexus itself works with this card cage installed and the RAM cards installed and, you know, nothing got damaged in me taking it apart again. I'm always reluctant to take this thing apart because I'm just worried that you take it apart and then it won't go back together and it won't work again. And then yeah, anyways, there's not enough time left before VCF for me to go into an in-depth repair on this. So, just please work. Please work. Let's power on the Plexus.

[00:59:29]There it goes. And I don't need to worry about resetting it because there's that reset circuit now. Awesome. Just want to make sure the LEDs look correct. I think that's normal. It's doing its uh power up sequence. Okay. Well, that's not a good sign that it has a PL and then it's stopped. That is exactly what it does when it's crashing or something. So, let's try resetting this.

[00:59:56]So, that's reset. Unreset. At least now it says the full plexus.

[01:00:03]Okay, good. So, maybe my reset circuit not so good.

[01:00:08]This is normal again. And it might be crashing because I have that card cage thing installed on the front.

[01:00:14]Oh, multibus failure. Okay. Well, multibus is that card cage thing on the front. Obviously, it does not work if that's attached. Oh, that's really annoying. Actually, that means I have to take that thing off just to run this.

[01:00:29]Okay. Well, it is what it is, right?

[01:00:31]Self test failed. Yeah, of course. I don't really want to leave this just sort of loose inside of there because the RAM card is there and I don't want this, you know, potentially touching that and shorting out. So, I just going to have to keep this separate, I guess, and just make sure I don't lose it. All right, let's power this thing back up again.

[01:00:49]I think I had the camera flipped earlier. Sorry about that. So, it's running the self test again, so that's a good thing. I could just abort out of the self test and I probably could have done that and just left the multibus failure as well, but it I don't know.

[01:01:01]I'm just going to let it run its full diagnostics. Make sure everything is happy here. There we go. Self test complete. No errors. And at this point, if I hit enter, should start booting.

[01:01:12]And there's a hard drive LED, but you cannot see it as obscured by sort of the gray plastic there that's covering the hard drive bezel. All right, we're in single user mode. But before I start this up the rest of the way, I'm going to see if I can get this terminal to a flashing cursor and see if this like pops up. Because once you run in to it, it starts all those login prompts. Let's see if I can get this terminal to show a login prompt. Okay, I know it's really glary on the screen there, but it is sitting there at 1900 bits per second, so should be ready. So, we'll start booting this up. Ah, yes, our login prompt. Oh, and our login prompt appeared on the terminal as well. So, cool. Well, this video is nearly at its end. I'm kind of bummed about the fact that the RAM card doesn't work. That's not to say that uh Super SVG can't fix this. It's just it's not going to happen in time for this video. Obviously, I never got the blue scuzzies working in time for this video, but we know these can work. They just need the right firmware, and unfortunately, that's just not something I can get in time for this video. PCF is still a few weeks off, so hopefully by the time that happens, I can get the Blue Scuzi working. I'd rather not rely on the spinning hard drive for the entire weekend. Just not 100% confident that it's going to last.

[01:02:22]But to be honest, it's lasted this long, so maybe it'll keep working. I don't know. There's one more thing I'd like to talk about with the Plexus before I end this video. And I alluded to this a little bit in the original series. When we were working on this machine, I actually managed to get in contact with the original designer of this exact machine, the Plexus P20. And there's a little bit of interesting backstory as to why there was engineering documents on this computer. And I can't really remember if I've talked about this in the previous series, so I'm going to talk about it here just to kind of put the record out there. Essentially, I was told by this person that when they left the company, it was near the end. Plexus didn't last much longer from after when this machine was designed. It was absorbed by Motorola. They wanted to have a a Unix system design in their portfolio. So, they bought Plexus. And I have no idea if they ever went anywhere with the intellectual property. But the person who designed this computer left the company shortly before it all dissolved. And as a sort of partying gift, they were allowed to take this computer. this exact computer was the one they were allowed to take with them, take home that is because the computer they took was the one they used in the office. That's why all the design files were on the hard drive. But what this person ended up doing is they wanted to repurpose the machine and not just see it like, you know, go to waste. So they decided to actually donate the computer and a bunch of terminals to their local church that they were attending. And they ended up going into the system and kind of erasing those engineering files and setting up like a word processor kind of office system, the one that's on there, so that the church could use this computer for like bookkeeping and recordkeeping and you know this administrative tasks and stuff like that. Well, obviously at some point the church decided to upgrade probably to newer computers or whatever happened. I don't know. And someone ended up with this computer and it was disused and then I ended up with it. Well, the original designer when they had given it to the church, as I said, they sort of reformatted the hard drive and reinstalled the OS, I think, from scratch and tried to set up to be as easy and maintenance-free for the church as possible. But when that formatting was done, it didn't actually erase the part of the hard drive that had all of those original design files on it. So, that is why they were preserved. It was just a mistake in the formatting process that left that part of the hard drive untouched. And when we did a disc image, we were able to recover that data. Now, how the machine got from that church to me is completely unknown. A friend of mine rescued this computer and gave it to me and didn't really have any backstory as to where it came from or any information whatsoever about it.

[01:04:57]When they reached out to me and said, "Hey, I have this really cool computer.

[01:05:00]I think you're going to like it. It's really weird and interesting, and you look online, there's not really anything to find about it." Well, yep, it was Plexus P20 and the rest is history.

[01:05:08]There's a whole bunch of videos on it.

[01:05:09]So, what makes this whole story about the Plex B20 so wild to me is that this could have easily just been a machine used in a small office like a church.

[01:05:17]They that would have been a perfect use case and then it got, you know, taken out of service and ended up to me. But if that were the case on the hard drive would just be some church files and that's it like the operating system and but none of those technical documents.

[01:05:32]The only reason why we have all the technical documentation of this that we can make a perfect emulator for it and potentially even things like RAM cards, well, that's just because this was used at Plexus's headquarters by the person who designed it and that's why it had all those design files on it. But what are the chances that the hard drive that was, you know, had all those files didn't get fully erased before it got repurposed for church duty? And then what are the chances that it didn't go to, you know, a dismantler and just get shredded and ended up at my place where I had the right tools and abilities to actually dump the contents of the hard drive and get the computer fully operational again. It boggles my mind all the puzzle pieces that have to come together to make this all work. And besides me and you know getting the hard drive image, all the folks on Discord and my friend David who did so much of that analysis of the lost data on the hard drive so we could get back all the data that was on there which resulted in the emulator which resulted in all the documentation that exists which resulted in this machine becoming something that's alive again. And what makes me kind of sad in a way is that for every one of these stories like we have here with this full circle story, there's probably 50 more like lost computers that companies were, you know, were small, they made it, it was cool, it was actually sold to people, never went anywhere, and that was that. Maybe those computers are sitting ready to be recycled or they already have been recycled or they are sitting in a warehouse of a collector who doesn't have the time to get to them. And hopefully they do because if you know the stuff will just rot if it gets rained on and left out in the elements or maybe the collector won't will pass away before they get to it and their family won't know what to do with it so it just gets recycled as well. I mean there's so many things that have to go right for something like this to be saved. And this one is saved. And I'm going to bring it up to VCF like I said.

[01:07:33]And the big announcement here is that I'm actually not going to keep this machine when VCF is done. I'm gonna actually donate this machine to the interim computer museum up in Seattle, which has an incredible collection, including some really cool artifacts that came from the Living Computer Museum, a now defunct computer museum in Seattle. And I thought it'd be perfect to give this machine to them because that museum has a bunch of amazing cool old obscure Unix systems that are actually online and working. So members actually have access to be able to log in and use those computers and potentially the Plexus will be one of the machines that you'll be able to log into. I can't give any timelines as to when the machine will be available to say look at or log into, but that's what's going to be happening with the machine at well in less than a month's time. So that's the full story of the Plexus P20. And while it's not the most interesting machine in the world, it's still something that was almost lost to time. It is definitely one of the many forgotten machines and it is no longer forgotten. It has been preserved. You can emulate it. There's a replica case.

[01:08:42]I mean, there's so much stuff happening.

[01:08:44]And really, when I look at it in retrospect, I'm only a small like piece of the puzzle that was bringing this computer back to life. community played such a huge role in its rediscovery and allowing us to see this thing run just like it's running right now, making my basement hot and noisy.

[01:09:04]The community is absolutely amazing and I I'm humbled by the outpouring of help that comes from so many people and yeah, it's a beautiful thing. It's it it really warms my heart and uh brings a little tear to my eye to be honest kind of think about it. So, thank you very much for watching. Thank you very much to everyone who's helped out with this project. And I hope to see you at VCF up in Seattle and the other VCFs are going to go to as well. And hopefully this machine at its new home at the ICM, the interim computer museum up in Seattle will have a future where other people can use it, tinker with it, and do what do things to it, and it won't be just sort of stuck in my basement down here where it'll never be seen again.

[01:09:46]So, that is going to be it for this video. Thanks very much for watching.

[01:09:49]And if you liked it, thumbs up. If you didn't, you know what to do. Huge thanks to my supporters. And that's going to be that. So, I guess stay healthy, stay safe. I'll see you next time. Bye.