The PDP-11/44 Lives!

this is starting to feel a bit like Groundhog Day this is now the third time I've done this intro talking about what this machine is what this drive is and why I'm so excited about it but the fact of the matter remains that I'm still doing this intro because it's all still broken this is my PDP 1144 it is the last non LSI pdp1 ever made which means it's chalk full of 7400 series logic chips and am 2901 4bit bits nice alus it's a pretty dramatically cool machine and I really want to plug this drive up to it this is a Fujitsu Eagle hard drive has 10 and 1 12 in platters in the most epic sci-fi looking Arrangement ever but uh well the reason I'm still doing this is because we can't test any of that out because we can't get the power supply on this thing fully working we got pretty close in the last episode there are six power rails on this power supply it's actually built kind of like like two separate power supplies three of the Rails are + 15 + 5 and minus5 for the logic those three are coming up fine they've been working perfectly since we started this whole ordeal that we've been going down the other half of the power supply is generating + 12 + 5 and minus 12 these three rails are dedicated to the memory and these three have been given us some grief in the previous episode we actually managed to get two of those up we got the uh + 5 and min-2 up but that last + 12 is really starting to elude us uh so we're going to dig in deep to the plus 12 today there's some pretty basic stuff that we want to check right off the bat and uh then if that doesn't do it we'll just keep digging into it I'm not doing another one of these intros I'm tired of doing these intros so the next time I do an intro for this machine it will be working so that means that we're going to get the power supply f working in this video I absolutely promise it not because I'm Clairvoyant or know the future but because this footage will never see the light of day until that power supply is working so I'm just going to pull this out put it on the table and get to [Music] work let's go ahead and get the memory voltage module out again I'll just remove the screws holding the 300v DC flat Cable in place then unscrew and remove the grounding cable then it slides right out the front and with it on the bench I'll remove the two mounting screws the two capacitor screws and the three Transformer screws and then the whole thing splits into and gives great access to pretty much all of the junk inside and we're getting plus 35 volt generation correctly now but each individual regulator has its own fuse so let's start with the most obvious and work our way down to more complex difficult areas the fuse in question is F1 here which is a 5 amp fuse and after finding it on the PCB I'll check it with the multimeter and yeah that's not looking right I should be seeing Zero ohms instead of whatever that is it's in a tough spot to get to and it doesn't look like your traditional fuse but the probes are definitely on it it's a little as we would say in Japanese aashi so let's pop it out and see what's what I'll just add some solder then use the iron itself to poke the fuse out through the through holes with it free I pulled a spare fuse out of a different board and sure enough the spare is measuring correctly but the original is showing open circuit so the fuse was definitely popped we'll toss the spare into place and solder it down properly then check it in circuit one more time and yeah that's looking good all right the fuse has been replaced it's all back together what we're hoping to see here is for all six voltage rails to come up that would be fantastic and I want to see the DC on light on the front here stabilize uh I don't know if it'll do that without a heavy load on it but I think it might because we have a heavy load on the 5vt which is going to be our little Automotive light bulb over there so when I flip this switch that light should come on all fan should spin up and we should get all six rails the multimeter is on uh the plus 12vt rail the one that was dead so hopefully we can see that one and and the other ones we already measured as fine so let's do it fan spun up light bulb came on still don't have 12 volts though dang all right so it wasn't just the fuse or it popped the fuse immediately which means we've got a fault someplace else all right I think I've said this probably about 15 times but I think we're starting to zero in on something I'm taking a look at uh the control signal that's coming out of our 555 here this should be a little pulse signal that is turning our 3904 on and off that is controlling the pass transistor up here but nothing is passing through the fuse has been replaced Q8 and Q7 have been pulled and tested to be known good so something down here is incorrect if we look at what's coming out of the 555 we have our little pulse there that's excellent but if we take a look at the next step along the [Music] way we can see that it's just just stuck low this is looking at just after r112 which is a 4.7k so something here is pulling it to ground blanking is not the problem because it's used by Min - uh 12 and + 5 and they're totally okay plus 12b shutdown is also not the problem I've double checked that it's okay so q23 has to be what's causing the problem here if we look at the base of q23 here it's kind of hard to see see if I change the scale way down here it's super noisy but I think that's enough for q23 to be turning on and if q23 is turning on that's going to be pulling this to ground so let's take a look at what's controlling q23 here and if we go back over to this one right here that is the emitter I think is the correct terminology of q27 it's the leg on the bottom here uh and we can see that that's at uh 510 15 Vols so we've got a full 15 volts passing through q27 now let's check the base of q27 here and you can see the base is also 510 15 volts the base is pulled straight up to 15 that's not great so uh it seems like we've got 15 volts on every corner of q27 here which might mean that q27 is unhealthy okay it's been a couple of days since that last round of Diagnostics with the transistor that I was looking at I've been pretty much over this schematic beginning to end and I think I'm finally starting to zero in on what might be the fault I've said that far too many times but uh initially what I thought was happening was that I thought that something wasn't controlling our past transistors up here so we replaced F1 and we have 35 volts coming to the other side of F1 the problem is is that Q8 is getting pulled low and not allowing anything to pass so I checked Q7 Q7 looks good I went down and checked r31 and then checked Q10 the 3904 and this is where we kind of came into our little bit of trouble and coming out of the 555 we do have a proper pulse coming out of pin 3 so that means that something here is pulling that pulse to ground before it gets a chance to meet the 3904 q23 is kicking off and it's pulling that uh to ground and that's not allowing the 3904 trigger so we followed q23 back to q27 and that's what we were just looking at and uh turns out that is actually working fine the difference between plus 15 volts and what we were seeing on the base is to be expected for a P&P transistor that's enough to turn it on we're just shutting this whole thing down so we had to follow it even further back and if we look at E2 here I looked at the inputs of it and the output of it and it is not causing the problem so it's not a current sense problem so we follow it back over in the other direction and that brings us to the Crowbar circuit and uh in checking out the Crowbar circuit all of the components looked fine and that was when it somebody told me that I should be looking not at the steady state after it's on I should be looking at what happens from the moment I flip the switch until the power stops and that scope shot was very Illuminating because it says that we are actually passing voltage through Q8 all the way out the plus 12 volt rail it just runs away though we are getting too much voltage and that is kicking off the Crowbar which is shutting everything down so the question now is why are we getting too much voltage the regulation happens with this little bit down here and the 555 and this is a very weird setup the opamp over here on the right has an interesting voltage divider set up and it will turn on or off depending on what the relationship between the two voltages coming into it are and I checked most of the components around there and most of them check okay but they're in circuit so they're probably totally fine now we know that the 555 is mostly working cuz we are getting a uh pulse signal out of it but it is this setup on pin six that is a little interesting we have q24 which is a P&P transistor then we have r111 uh d34 and r R17 now r111 and r117 measure bang on perfect in circuit uh d34 I can't measure in circuit I get something in both directions on it but it could be fine could be bad it's kind of hard to tell while it's in circuit uh and the same goes for q24 I measure some funky things on q24 and one of them is actually a little bit more of a red flag than a d34 and let me zoom in and show you what I'm talking about now again this is in circuit so this could be totally wrong but I'm going to try and check this little transistor right here and one way of checking that is you can use the diode test on your multimeter and with a PNP I should be able to put the black Probe on the center pin here and I should be able to go to either of the other pins and see a diode drop if I go to that one I get 651 that's kind of what I'm expecting and if I go to this one I don't get that diode drop that's a little strange to me there should be a diode drop between Bas and collector as well as base and emitter and uh we're not getting that um we can also check it the other way here by putting the red Probe on the base here in the center and I can check in this direction and again I don't get uh any diod drop that direction but I do get diod drop in this direction of one volt now again this is in circuits so that could be totally expected uh but the fact that we aren't getting a diode drop here is a little that's a little strange to me so maybe this PNP has given up The Ghost and what that would mean is that we're not getting the appropriate voltage on the threshold pin of our 555 which means that it's not going to regulate appropriately with the signal that's coming in on the control pin pin five so if that P&P is bad it could absolutely cause the problems that we're seeing all right PMP has been replaced with a 2907 I think it's just kind of what I had on hand it was a generic PMP I think that should work just fine uh I'm skeptical that that was actually the problem I'm betting that there's something deeper going on but I'm kind of running out of avenu so let's go ahead and give it a test we'll flip the breaker on here then we'll flip the main power switch on here that was totally it 12 volts we have 12 volts [Laughter] yes what a relief man just all right I've been going hard on this uh power supply for this thing and boy it feels amazing to just climb that hill and get to the top at least for now it all seems to be working but what an interesting failure we didn't have anything wrong with the big pass through transistors or any of the regulating ic's or anything like that it was just a simple PNP transistor that as I understand it is giving a regulated voltage to the threshold pen of a 555 this is a an incredibly low current uh situation there should be absolutely no current moving through this at all and yet somehow that PNP went down what a what a journey all right so so what's next well we're going to start plugging Cards into it and see how it handles a proper load and uh well we're going to do just the CPU and memory cards and this is just one of the CPU cards this is going to be about eight cards in total when I said this was a non LSI machine I absolutely meant it so we'll take a look at each card as I plug it in and maybe we can figure out what they're doing then we'll plug them all in flip the switch and hopefully uh nothing goes up in smoke and it all good okay we're going to go for the bare minimum setup here for initial testing just the CPU cards and two memory cards uh and we're going to start with this one this is an m793 it's called the data path module and the manual says it does arithmetic and logic processing shifting of 8 16 and 32-bit data formats bite swapping and sign extension of data General register storage and Status information and we can actually kind of confirm that with some of the ic's on here there are four 74s 181s these are 4 bit bit slice alus we have some uh 93422 DC ceramic ic's up here these are 256 WX 4bit high-speed static Rams uh and then down here we've got three real big n82 s100s and these are all pla or programmable logic arrays and hopefully those are all okay because those are probably going to be the most impossible things on this board to replace next up is the m 7095 control module this one contains the control store Proms and logic to decode and execute the pdp1 instructions it also contains the system clock power fail autor restart boot control and trap handling logic and looking at the far left here we have a bunch of am27 S1 181s or equivalents and these are 1024x 8 bit ROMs and then uh right next to it we've got more of these n82 s 101s these are again PL there's a lot of Hope riding on all these being good this is the m796 multifunction module it contains an 8085 CPU which I find absolutely hilarious has two serial lines a line clock and all the related Logic the serial lines are for the console terminal and the t58 tape drive and the 885 enables execution of all the console commands and we can actually see that 885 right up here in the top left uh this is a full-fledged 8bit microprocessor uh right below that we have a bunch of dm87 S1 185s these are 2048x 4bit programmable ROMs or readon memories and then all the way over here on the far right we've got some 6402 these are going to be our uarts or universal asynchronous receiver transmitters these are the two chips that are doing the serial communication to the console terminal and the t58 tape drive next up is the m797 cache memory module this contains contains 8K of high-speed static Ram cache memory to decrease processor to memory read times the memory is located in the most commonly accessed memory locations and it's a directly mapped memory cache with right through facility I got all that lingo out of the manual I don't necessarily understand what all that means but I do understand that over here on uh the far left we have a bunch of 2114 uh chips these are not socketed but these are all the 1024x 4bit static RAM chips that make up the cach memory and this should make the whole machine move a lot quicker but uh other than that it's just full of a bunch of 7400 series logic tips to I guess interface all of that with the rest of the CPU rounding out the CPU here we've got the m798 unibus interface module this provides all of the logic which allows the processor to access the unibus and it also has bus arbitration Logic for interrupts and nprs boot circuits and buffers for packs data lines to in the processor and it contains a map that allows for dma between main memory and peripherals and it's uh mostly just a bunch of 7400 series stuff except for this row of big fat chips that we got right down the center there are two different types of chips in here there are 93 459 that's going to be these three over here those are more pla hooray uh and then over here we've got 93 419 and everything I found on these say they are 64x 9bit high-speed static Rams which is a really strange configuration but uh the access time on these is like 45 NCS so those are booking and then finally I'm going to put in two of these These are M 8722 ECC 256k memory modules and uh they're pretty much what they say on the 10 uh there's a boatload of 4116 Dynamic RAM chips over here these are 16k by 1bit dams uh access time on these is a lot lot slower around 500 NS on these than those high-speed static Rams that we were looking at but uh these are a lot cheaper and you can fill a whole lot more of them on one board than those big honken static ramps now this board uses the +2 plus 5 and minus 12 of the memory portion of our power supply and so this is really going to put a bit of stress on that power supply we'll see how it holds up to it with two of these in there that's going to pull some current I think okay processor and memory fully in I did check uh the memory for shorts across all of the power rails because there's a lot of little tanum capacitors on there there were no shorts obvious but that doesn't mean that one of those tanms won't short out as soon as it gets power hopefully though that doesn't happen what we're hoping to see here is for all the fans to come on and for dcon to stabilize down here so uh here goes all of fans are on DC on looks like it's stabilized says run I think we're okay no smoke came out of anywhere I think that's powered up if we had a terminal connected to it we might be able to see something happening the final piece of the puzzle is this card right here this is the m790 console interface module and it does exactly what it says on the tin it interfaces with the console the front panel plugs in here the t58 drive plugs in here remote Diagnostics plug in here and this plug here is for the console terminal everything else on here is just kind of uh buffers and level shifters and everything to support both rs232 and 20 mamp current Loop uh now you can see that this is not a standard de9 plug it's just it's a specific ribbon cable that's got to plug into that and I dug through my stash and I think I found the right cable for it this is it here it says this side up has the correct key in it and this should just plug straight into here like that and uh the other end of that has a db25 on it so maybe this will work I don't know we're going to plug it in and give it a shot to give this a test I'm going to use my trusty ADM 3A here I've got it set to 96008 in1 I have no idea if that's what the pdp11 is set to but uh we're about to find out so we'll go ahead and flip the switch here hopefully we see something show up here if I flip the nope nothing at all all right I've confirmed the setup on the 7090 and the 796 these are the two boards that control whether it's doing rs232 or current Loop it is set to rs232 96008 in1 which my data terminal is also set to but I didn't have a bus Terminator card in there last time so I've got a uh m932 bus Terminator in there and I've got a bunch of knucklebusters down here that should be moving Grant continuity lines over to that so hopefully that's the problem but there also very well could be something more Sinister lurking in these cards here that's preventing it to come up but let's flip the switch and find out yeah our run light does not come on and stay on it just comes on and then kicks off uh if I hit boot it turns on and kicks off again dang we got we got another problem somewhere something deep is going on here right I think I may have found a uh pretty important problem even though it's not really that big of a deal I'm pretty sure I need a null modem cable to go from uh the 970 up there to my data terminal and uh that cable was not that and I don't have a null modem adapter so I had another one of these cables which is built in the same way and I just modified it to switch RX and TX around and that should get the job done uh I've got my Terminator card in there and I've got my continuity cards in there this should work please work I'm going to flip the switch here and we'll see what happens console v3.4 C program it works yes well my ADM 3A decided to die so we've moved over to my rainbow 100 uh this is actually the in use case here this is the terminal that I'm going to use with this 1144 so this is good I guess so we'll go ahead and hit T here and uh I just want to make sure that we are set up correctly for the appropriate speed here and we are not uh there we go 9600 bod uh so we'll come out of setup here now let's turn the power supply on for the 1144 there we go console v3.4 C program I can't do anything from here no matter what I type nothing happens but if I hit contrl P then we can get into the console here and uh there's a couple interesting things we can test here which is Illuminating a different fault that we have uh first of all I can do a t here for a self test I don't know I guess that passes it just does that over and over again uh but if I try to do anything else like I try to boot from here which you can do by typing B and then whatever you want to boot from like say dx0 here it says halt CP so if I flip the halt switch on the front of the machine and I try that again dx0 still says halt CP it should be halted if I type h to halt the CP here I get failed to Halt so I think my switch on the front that is supposed to Halt the processor is not functioning and then that means that the amount of things that we can do here at this prompt is very limited if I do I to initialize it says halt CP if I do s for start halt CP so I think what we're going to have to do next is we're going to have to dig in deep into that front panel and figure out why it's not halting the processor or why we can't halt the processor at all it might not be the fault of the front panel so we got an insane amount done today we actually saw life out of this machine we are so close to getting this thing actually running software off of some kind of Drive floppy or hard drive like that big Fujitsu Eagle so in the next episode we're going to try and get it halting properly we're going to try and get it doing stuff trying to boot from things and then maybe we can start plugging more cards into this to see if we can get it to actually do some pretty fancy tricks but this this is epic we started this episode with this thing totally dead not being able to run anything and now check it out it is properly running that is awesome so this would not have been possible in the slightest without a massive amount of help from everybody over on the Discord namely Coy he's been a huge inspiration cuz he's battling through an 11:05 and an 1134 restoration of his own and he's been really inspiring me to dig in deep on this also Mitch if you're watching thank you so much you hooked me up with a lot of really epic stuff and because of that we're able to get some really exciting stuff going and we have a huge number of people that have been contributing to help me get pointed in the right direction on this thing and it has been an absolute Journey so thank you all so so much and thank you guys so much for watching and I hope to see you all in the next episode e

Channel: Usagi Electric

Views: 35,302

Keywords:

Id: KCoO0YV4Vo4

Length: 27min 18sec (1638 seconds)

Published: Sun Jul 14 2024