IBM PS/2 Part 5: Motherboard Repaired for Good

[@rjy8960]

There is absolutely NO WAY to diagnose this in retrospect. You guys put a lot of time into this repair. Respect!

[@bengelman2600]

Seriously brute force

[@VintageTechFan]

@@bengelman2600 Only after they circled in the fault be careful, tedious elimination. That's what most people l see looking for faults on the Internet lack. They jump into ideas they got somewhere (usually bad caps as he said .. there is only one case where you can be relatively sure about those, and that are wax/tar-sealed paper foil types from

[@Fredjoe5]

@@VintageTechFan Everyone wants it to be the "quick fix". Everyone wants to say "it's bad caps!" as if that's even the most common failure for these kinds of functional issues. A good knowledge of the schematic for a particular board, plus basic tool skills, plus a lot of invested time, can give you meaningful results beyond concluding it's one of the "easy fixes".

[@jeremiahrex]

As you were saying “make your bet ” I was thinking “guesses suck, I hope they stick a scope on it”. Never give up and measure stuff, a great way to solve the problem. Great work.

[@CuriousMarc]

Well put. That is indeed the proper way to do it.

[@Castaa]

Eric the true PS/2 MVP. Just wow. Congrats guys.

[@SidneyCritic]

Schrodinger via, as soon as you probe it affect the state - lol -. Reminds me of my leaking plastic petrol tank. When you pumped it up to 0.5psi it would weep, but you couldn't see the exact point of the leak. So you raise the pressure to 1psi expecting it to leak faster, but instead it presses the crack shut and stops the leak. Took a few days to work that one out - lol -.

[@federicodidio4891]

This mechanism would explain a (different) fault I'm experiencing right now. Thank you, kind stranger on the internet! :)

[@efroymson]

Careful comparison of this episode with the flashbacks reveals the truth: The key to the repair was Eric getting a haircut.

[@IBM_Museum]

I'm glad I wasn't the only one that noticed that...

[@TubeTimeUS]

haha my hair has been all over the place in this series

[@IBM_Museum]

@@TubeTimeUS: Hair "all over the place"? Have you seen my beard? I think there was even loose hairs on that salvage planar I sent - certainly they have been on the other planars I've taken photos of!

[@XMarkxyz]

It allows superior airflow to cool the brain thus making it able to work faster as for all processor

[@AndyGoth111]

Tin whiskers, I tell ya!

[@prairiedawg792]

My favourite way pf proving a bad connection was to use a scope in differential mode on the source & destination ends of the trace which shows up really well. Mind you, that was way back when using DIP with 0.1" pin spacing so a lot easier!

[@ImmortanJoeCamel]

That's really clever! Going to have to try doing it that way

[@zyeborm]

I'll steal that. But you do have to suspect signal integrity as the cause before you get there too lol.

[@iamdarkyoshi]

Ah, the repair is typically the easy part. Finding the fault however? That's when you've got to enlist some help. Great series!

[@AsbestosMuffins]

unless the fault was a bad power supply then the repair was equally tricky, but that's another story

[@russellhltn1396]

It's always simple - once you find the problem. But finding the problem can drive you bonkers.

[@IBM_Museum]

Success! A great diagnostic process to find something that is so hidden. I'm thinking that an issue of a couple of "Bermuda" planars not holding their configuration may be similar - are you (and Eric) ready to check a sample that shows that fault?

[@joelmarcott3282]

This is what I would be eager to see! I have two 77 Planars that have this issue. it's an annoying failure preventing a complete system startup. Hello Dave by the way!!

[@IBM_Museum]

@@joelmarcott3282: "Santa" seems to be my new nickname, although I hope people aren't expecting me to bring them something. Only one of my "Bermuda" planars is affected, although I know now to start pulling out my USB microscope and voltmeter. We will get this solved!

[@TubeTimeUS]

just write a custom bios 😉

[@joelmarcott3282]

@@TubeTimeUS It may not be that simple. Could be a bad connection or faulty Dallas Clock.

[@kaliban4758]

@@joelmarcott3282 if it is the Dallas clock chip Necroware has a replacement for that

[@graemedavidson499]

Well done finding that broken via! I had a real head scratcher with late 80s Philips car radios where the radio would come on when the car was left unattended and flatten the battery. Philips used conductive silver paint vias/upper circuit board traces and there was an issue where water would drop through a hole in the top of the radio (condensation), leach out some silver, dry out and leave a variably conductive invisible electrical leak that would trigger the radio out of standby!

[@samio3907]

I have an IBM Model 70 PS/2. This thing is a nightmare to work with and I can definitely put myself into your shoes. I would never recommend these things to retro enthusiasts, unless you like unusual hard- and software configurations.

[@IBM_Museum]

Which can vary by PS/2 level - a desire to run the microchannel P/390 board that was shown is definitely both an "unusual hard- and software configuration". What is the submodel of your 8570? There are adjustments that can be made to reduce the "nightmare".

[@BlackEpyon]

Aww, but fixing them is half the fun! But you're right, I certainly wouldn't recommend them as a retro-gaming machine. Space hardware can be a bit hard to come by sometimes, never mind finding sound cards for them.

[@samio3907]

It is actually running fine 🙂 after a lot of fiddling. 80386 at 20mhz (FCC ID: ANO9SA8570-D) / Kingston KTM 16000|386 memory expansion / XGA Adapter 69F9704 / a replacement Sound Card from (who guessed it) TubeTimeUs, because the originals are unaffordable, + his floppy adapter for a Gotek Floppy drive 😅. Btw. most of the problems I had, were (not kidding) capacitor related. Especially the original floppy drive was a nightmare. Just the harddrive (whopping 120MB and pretty rare as well) is giving me some headaches. Because what do I do when it reaches its end of life. No MCA slot is free anymore for a SCSI Controller and the ESDI drives are way too obsolete to find replacements for!! 😢 Hopefully someone is working on an adapter for this issue too. ESDI 2 SD TubeTime !! I count on you 🤗

[@jimmy2drinks]

It is immensely satisfying to me that you guys have fixed this board. +1

[@AshtonCoolman]

Eric transformed his look so many times over this series that I barely recognized him each time.

[@macgvrs]

I confess I didn't even know that vias could do that. Embarrassed to admit that. Definitely learned something today. Thanks for a very interesting series of videos. Very nice to see it repaired and working. I suspect you are way happier than I am. Great that you have such a talented friend who was willing to spend so much time getting it going again. Kudos Eric.

[@mrmobodies4879]

11:35 I remember those cdrom drives, the one I had was in an external encloure with a parallel interface, manufactured by Matshita in 1998.

[@yadt]

@CuriousMarc the playlist for this series has part 1/2 swapped. Made for a confusing catch-up experience until I realised!

[@waterslidexplorer]

Wow, I am amazed by the level of knowledge, persistence and professionalism that you have put into this troubleshooting.

[@radarmusen]

Happy that’s not was ending in gluing the finger on the chip.

[@Wobblybob2004]

Well it's taken three months, five episodes and a lot of hard work and dedication, but we finaly got Eric a decent haircut! well done everyone. :-)

[@voneschenbachmusic]

As someone who is as old as this machine, it feels great that you'all devoted so much time to fixing it and giving it new life!

[@AlainHubert]

Not all caps go bad with age. I've got some electrolytic caps that are 40 years old and show very low ESR readings and good capacitance, still. It depends on the quality, ambient temperature, and electrical stress.

[@Hans-gb4mv]

I want to swap out all modern caps with quality old ones. Have more faith in them.

[@mikesradios]

Capacitor failures have been the bane of electronics repair techs since they were invented. For sure, there are caps that have lasted better than others. But ask anyone who repairs vintage electronics (tube gear) and they will not waste time trying to reform old electrolytic or paper caps. Are there still some that work? Sure. Are they reliable? Not likely. Of course an IBM PS/2 is not a 1931 Radiola... I would not jump to shotgun replace all the caps on a motherboard. But I wouldn't be surprised when a tantalum or two popped either.

[@senorcapitandiogenes2068]

It really depends. I have never seen a bad ceramic cap, but I have seen shorted tantalums, foil caps with drastically decreased capacitance and a fair few old electrolytics that vented a few seconds after the device being turned on. I guess the application of the cap as well as the quality is the key here.

[@video99couk]

I fix Philips video recorders from 1975-1980 on my channel. Virtually never had a bad capacitor on those. A few bad ICs though. JVC DVD recorders and Panasonic professional studio recorders from the early 2000s, now they suffer bad capacitors all day.

[@rocketman221projects]

@@senorcapitandiogenes2068 Ceramic caps can fail, but it's usually from the board flexing or something being dropped on them causing them to crack and short out.

[@RetroAdrianBlack]

Epic repair! I voted for bad trace inside the PCB, so how nice the fix was easy once the fault was found. Awesome.

[@BryanI00]

Eric was really fantastic. He's incredible and a lot of fun to watch work!

[@prillewitz]

I have seen such a machine in the past and people who were busy with these machines told me that they weren’t documented, that’s why I remembered! Strange!

[@15743_Hertz]

life is not about the destination but the journey. In the process of getting an interface to vintage tech, you've discovered a small treasure trove of very interesting IBM facts. Well done!

[@blueskunk9163]

I spent too much time babying OS/2 2.0 to Warp. I gave up after realizing IBM would never devote the resources necessary to make it prime time. Windows nailed it with 95. And IBM OS/2 had the dreaded single input message queue. When that froze (and it did frequently), the UI would also freeze.

[@BlackEpyon]

Eh.... I've got a couple retro machines running 95 because it's period appropriate, but 95 could be very buggy sometimes. By 98SE, they worked the kinks out.

[@LMB222]

Windows nailed it with 95? From my perspective, it was the worst OS in history. What do you like about it?

[@SianaGearz]

@@LMB222 Worst OS in history? Not even close. It's both more robust and a million times more capable than 3.x, and non-MS OSes or NT weren't really an option for home users with limited hardware. It introduced preemptive multitasking, replacing the cooperative model, which means unresponsive applications (at least as long as they don't trample over some shared memory, or interrupts, or chipset IO) didn't halt the whole OS. There was also (a limited form of, with substantial issues) memory isolation between processes with individual per-process memory maps, and while NT did it better and actually well, at the cost of drastically higher system requirements and reduced compatibility, 3.x didn't have it at all, so a transitional system was needed. The taskbar was revolutionary. Windows 95 did exactly what was needed at the time when it was needed. Teething pains with driver quality and just bugs with something so radically reworked are kind of expected... that's why there was a Service Pack and an OSR2 release. And yet, it could run perfectly fine as it came out, it wasn't badly tested or released unfinished. I think the worst bane of the time was the RAM module quality, complete garbage. You go upmarket, things get better, they get actually good, but there was a lot of garbage. Windows 95 forced people who wanted it to get more RAM, probably bought unbranded or shady kits, which is where some of the associated problems may have started. RAM quality issues lead to the worst experience in 95-98-Me. But there's nothing the OS could have done at the time to be more resilient against bad hardware while keeping performance and compatibility in mind, so why would you count that against it?

[@MrCommodorebob]

Usually tantalums from that period have a low failure rate anyway, most of the bad ones that I have seen are from the early to mid 80s. If a tantalum goes bad... you'll know it. About 80% of the time I have run into that issue it's because one of the little suckers exploded.

[@RingingResonance]

tantalums throw tantrums to terrify timid technicians

[@Bobbias]

Yeah, a bad tantalum usually let's you know in dramatic fashion.

[@andrewbarnard3229]

every amd motherboard from the 90's...

[@MarcelHuguenin]

Brilliant !! That's dedication and another saved piece of tech. Great stuff.

[@DangerousPictures]

my dad told me a story where a circuit didn't work but worked with the oscilliscope probe attached. turns out it just needed that extra little capacitance of the probe

[@ComputerHistoryArchivesProject]

A very fascinating video! Great clarity and good explanations. You guys have great patience too! Thanks for posting this. ~ Victor

[@nutsnproud6932]

Well done everybody involved to get the PS/2 working.

[@72polara]

I almost forgot why you guys were getting this machine going. Looking forward to the future videos!

[@BlackEpyon]

As a retro enthusiast myself, I find fixing the damn things to be just as much fun as playing with them.

[@cpt_nordbart]

As you can see, you can repair almost everything. Though if I would have had to pay professionals they'd have said it's broken you need a new one. Or if I said to repair it it would have been prohibitively expensive. Still good to see it working again.

[@ferrari2k]

Well if you consider the man-hours that went into diagnosing the problem, it quickly becomes way more expensive than the machine itself is worth, so this kind of repair definitely is only for the hobby user type ;)

[@reasonablebeing5392]

Finding a bad via is inversely proportional to the time it takes to repair it. Been there. Congratulations!!

[@GameTechRefuge]

Really cool to see the bad via up close. Great job. EDIT: That NEC Monitor is sweet.

[@240p_is_enough]

I wish you and Eric did more retro PC/Mac videos together! Great work

[@AnonyDave]

Well that's me excited that you found the cause in this one. Without access to all those tools, I guess I might just have to hope that its similar on mine and poke at the vias around that chip one day. If it turns out to be a via on mine too, then that'll suggest a really old issue with their design

[@CuriousMarc]

Tell us what you find. I would not be surprised if it is a generic issue. But it could be anywhere on the board, on any side of it. That’s many 1000 of vias to poke at…

[@Shmbler]

You never told us how long it took Eric to find that broken via. Most other guys would just have given up.

[@ErikZarth]

What a saga! Congratulations Thanks for this series I think we all gained a bit of knowledge from this.

[@informativt]

Probing all the pins. Patience is truly a virtue.

[@DextersTechLab]

Amazing repair, total respect for you all putting the time and effort into the project!

[@joelmarcott3282]

Fantastic troubleshooting work congrats! It's the painstaking probing and checking the smallest connections that won the day!

[@russellhltn1396]

Excellent work! You deserve all the praise the internet sends your way today.

[@michaelmiller641]

It wasn't that long ago, maybe a couple of years ago, that I discovered a hole in the wall bank teller machine was still using OS2 warp! Fascinating video! Marc, thanks!

[@prairiedawg792]

In a former life I worked quite a bit with ICL. Walking through one of their facilities I noticed some ATMs being commissioned were booting *Microsoft* OS/2

[@afsjeff]

You are very funny Marc, but the best joke you ever told is that there are commenters who would tell YOU GUYS that it was a bad cap. I guess it's a board level repair. The one you try right after "did you turn it off and back on again?" and you can do it if you only have a soldering iron. Brains optional. I guess anyone can type. Thank you for sharing your brilliance with us all and keep an eye on those caps!

[@Zerbey]

Once again I'm just dumbfounded how much work you guys put into getting these machines back online and I don't think I'll ever reach this level of expertise. Either way, it's incredible to watch the process.

[@HawaiiScenicDrives]

THANK YOU FOR THE UPLOAD MARC!!!

[@jtwhite2084]

Dielectric constant of air strikes again! Congratulations on fixing it as those are often the hardest problems to find. In my experience, broken vias are a lot more common than a broken trace on an inner plane. Whenever I found one there ware often more in the same area of the board so soldering a wire through all the vias in that area might be a good idea to avoid future problems.

[@Kae6502]

Too bad about Amazon. I remember them from back in the day. So sad, I had high hopes for them. Oh well. ;)

[@peteroneill404]

Intermittent faults like this are pain and time consuming to find. Well done in tracking it down. It's easy to speculate and/or over complicate the issue but 90% of the time for intermittent faults at least, connections are the cause. The capacitor changers would have come up short here.

[@NaoPb]

Congratulations and well done! I am very impressed!

[@wacholder5690]

Eric fits in your team. Same approach finding errors. Good to know that there are still people out that look for the real root of a problem. Thanks for sharing ! And you should look for a 1.2GB SCSI drive. Preferably a power-saving IBM DCAS3120. The system has a "natural limit" on the size because it was supposed to hold a "convenience partition" (with the content of the reference and diagnostic disk) for immediate access when required. The limit is 3.94GB and most modern SCSI drives are either larger or don't have the "narrow" 50-pin connector any more. It is possible to use 68-pin F/W drives with a converter plug - but the primary / boot / ID6 drive may not exceed 3.94GB. Once booted an additional drive can be larger. I had machines with 60MB (MegaByte) capacity as boot drives and 68-to-50 converted 18GB Ultrawide SCSI drives as second. The small boot drive is just used for the reference partition, but everything else was on the bigger drive.

[@BlackEpyon]

I had some old full height 5.25" SCSI drives years ago. 9GB capacity, 50-pin connector, and bloody noisy beasts! The server tower they came in, I've repurposed as my main i7 rig, but sadly, those drives failed and went for recycling years ago.

[@wacholder5690]

@@BlackEpyon From my "times with PS/2" I must have dozens of DORS, DCAS, DGHS and suchlike in stock. Along with some Seagate Elite Full Height drives. Wonder, how many of them would come up after years of being not used. There are 2 DCAS and 10 DGHS (or DDRS) in the one Server 520 alone ...

[@dentakuweb]

Great work. I also had a hunch about a bad via.

[@SkyOctopus1]

Tremendous job all round; I'd honestly never have found that. A trivial fix in the end, but the skill is in finding WHAT to fix. Anyone downvoting this video has to be a capacitor replacement aficionado ;) My assumption was a design flaw in the chip where they just went bad over time. Happy to be proved wrong!

[@pparadigm]

A great series - thank all.

[@Evergreen64]

Good job finding this. What a pain!

[@compgeke]

I look forward to seeing the 370 parallel channel working! It might encourage me to go about replacing the SCSI array in my PC Server 330 P/390 with the SSD and reinstalling OS/2 to get that running again. Has everything, including the PCI parallel channel adapter and Y cable :) .

[@Damien.D]

That was a tedious troubleshooting. Well done. And isn't the via next to the cracked one also looking suspicious?

[@UReasonIt]

This brings back a lot of old bad memories! ;) The PS2 line has to be the homeliest computer of the era, and MCA could drive you nuts. I ported some SYSV and Vax code to Warp. It was a lot of "new wheel building", and I think I started smoking cigarettes in 1988 ;)

[@SteelHorseRider74]

"No one will ever find a broken via..." CM and friends: "..."

[@jackflash6377]

OS2 Warp. That brings back memories. NEC Multisync! I had a 19" version. Beautiful monitor.

[@GadgetUK164]

Excellent work! Would have been hard to work out it was that via for sure!

[@fedorkovkov9337]

It took me almost a 80K of subscribers later to realize that Marc's channel picture is an R2D2 head. I always thought it was like an igloo viewed from inside.

[@danricho]

I like the link to “Quantas” among the other airline links!

[@Dust599]

Nice work. Glad it wasn't some really nasty inner board via.

[@albertsandberg]

Thanks guys. Fun to see and also a learning experience.

[@borayurt66]

This happened to me with a Sinclair Spectrum. When I put my finger on the certain pins of the ULA, the board worked, and stayed working until next power off/on cycle. Same with a cold sprey too, as long as I kept it very cold, it worked. My initial thought was a bad solder joint, so I reworked all of them, no result. Then I tried pulling the certain pins up or down with a 10K, produced intermittent results. Lastly I tried capacitive coupling on those pins with a 10n and that killed the ULA for good. Replacing the ULA (took 3 months to get one) fixed the board for good. So, my final diagnosis on it was that the original ULA was very tired and only working marginally with a magic finger on it.

[@julianrichards9509]

That's so cool guys!!Well done Eric!!! You guys should keep him around he's really cool ,a great addition,hmm,the 5 musketeers..I was convinced when eric did the finger thing that it was either a hairline crack,in one of the tracks,or a tiny crack in a via,it was small enough for a thermal change to cause expansion,and to complete the connection,so i was chuffed when i saw that many patreons were focused on those two possibilities,lots of talent out there Marc!! I thought you guys had given up,really impressed with you guys,you can fix anything!! Looking forward to you firing up the IBM 360,want to really see you all put it through is paces,i had some tinkering with old DEC PDP and VAXES back in the day,,never IBM though this a terrific channel Marc. I'm sure that ken could write some really nifty programs to run on the 360,machine code would be nice! Thanks you guys..

[@Epictronics1]

I don't think Jeff is likely to abandon Amazon any time soon, haha. That is one funny PS/2 you'v got there :D. Well done finding that via!

[@AKATenn]

I had that exact same monitor in the mid-90s, i kept it with my 400mhz k6/2, gave it to someone, wish i still had it now...

[@user-hj5nr3wy5w]

Well done team, congratulations. Eric your blood is worth bottling mate. Marc good on you for persisting to the end. Hopefully there will be many more needles in haystacks to come.

[@EngineeringVignettes]

All evidence pointed to a break in the circuits somewhere. I'm glad Marc & team found it. A break in the via plating itself is a weird one, to say the least. Cheers,

[@ZenwizardStudios]

Love to see the hard ones! Those are the fun ones to learn tricks from.

[@tinygriffy]

I remember when I got my first DX2-40, I could upgrade my graphics cards ram with 16 pin dip packages to incredible 512K .. that machine was solely running final frontier first encounters (and doom sometimes) though I think I had win 3.11, but that gobbled up too much space on my 20 MB shoe box sized hard drive, dos and nc were sufficient ;) It's incredible how small all that became, that processing power and storage might now look like a spec of dirt under your fingernail :D

[@525Lines]

My first IBM PC was a 30286 PS/2 which I later found out wasn't even a 286. It was an 8088. In 1988, I bought a 8088 when I could have bought a 386 tower system for a LOT less money.

[@vinatron8075]

Question when you attached the B/T card to the PC what programs will you be using? I’m trying to attempt this however slightly different. I want to put a ESCON card in it and use a channel converter to test a printer I have and a couple other peripherals that have B/T connections. Although I do have a mainframe system unit I don’t have enough ESCON cables to connect everything now so I thought this would be an interesting way to expatriate. I have a couple A50 tape controllers and they are RS6000 P43 based and I’m also wondering if they can be adapted to present other drives such as DAT and possibly a 9348. Speaking of the 9348 I’m looking for a proper SCSI card to drive it.

[@richardkaz2336]

I think my early guess was a break in the multi layer board and that I would have just frisbeed the board sadly but lifes too short and customers have no desire to pay me for my time. If one via has broken chances of another breaking over time with temprature expansion and contraction is possible. Well done finding the fault.

[@IBM_Museum]

@11:17 - Hehe, the missing bezel piece for the CD-ROM drive is far more "rare", even though it is a minimal amount of plastic. Maybe Eric can 3-D print one, as I don't have any extra.

[@drewscruis]

Fantastic job on finding the bad via. I collect pinball and arcade games and repair th e boards, finding damaged vias is a nightmare. I also get hit with "it's the caps" in my hobby alot, these boards were made with the lowest quality parts possible so sometimes it really is the caps lol.

[@CuriousMarc]

Yes, truth be told, I have bad crapacitors too in my more recent consumer electronics (and a few "modern fails" videos to go with it). And occasionally on my beloved vintage HP equipment too!

[@AlexFr80]

Very impressive videos! Thank you all

[@MarkyShaw]

What a ride!!! Great job!!

[@tony359]

Amazing! What's the story of that digital clock with days of the week in Italian?

[@CuriousMarc]

Aha! My Italian 1957 Solari Dator 10 flip clock. I have an old playlist about that! kzbin.info/aero/PL-_93BVApb59Xs5gCX_Sch1LoIm8TcWQ2

[@scarpaz]

Fantastic work!

[@erictourangeau3651]

Good job guys I would have give up a long time ago !

[@Fredjoe5]

Sometimes such problems in manufacturing would fail in bare board testing, but suddenly disappear under the vacuum used on bed-of-nails testing. Very frustrating.

[@neypimentel479]

Great job !!!

[@sragga]

os2 warp. there is a blast from the past. i recently pulled an old pc i found buried under piles of junk. plugged it in and found its a ms5169 ali board with an old intel chip, and was running os/2 warp. trying to figure a good use for it

[@TheOddVideoChannel]

This growing of tin whiskers is especially a problem in a vacuum. I was told that it is more pronounced on lead-free solder and that ESA, therefore, insists on a minimum percentage of lead in all soldering for space applications.

[@BlackEpyon]

I prefer 60/40 lead solder anyways. Toxicity aside, it simply works better, and is easier to work with, than complete tin.

[@TheOddVideoChannel]

​@@BlackEpyon I completely agree. Besides that, I think that in a hobby environment the toxicity isn't much of an issue. With a bit of ventilation and washing you hands afterwards everything should be fine. If you're soldering 40 hours a week then it's different story of course.

[@ydonl]

Apparently it was a common failure mode in early transistors, too -- inside the can. If I remember correctly, you charge up a capacitor and zap across the junctions to fix it... or something like that. :)

[@TheOddVideoChannel]

@@ydonl My old physics teacher used to repair his broken light bulbs like that. Cap across the connector and wiggle the filament around until it welds itself. I always thought this was more for the challenge than for cost savings...

[@ydonl]

@@TheOddVideoChannel Pretty crazy! Yes, I would say... for the fun of it!

[@herauthon]

questions..questions.. OS/2 seems particular on CDROM drives - i think .. unless there is some new driver around.. Does that imply one must keep the 2speed mitsumi around to install OS/2 ?

[@leonardoliveira]

Absolutely outstanding content

[@DrFrank-xj9bc]

Great repair. The next via, third from left, looked as bad as well.. Probably all of the vias are inherently affected.

[@IBM_Museum]

And there are known instances of the "Bermuda" planar not holding a configuration (the Dallas clock chip and Extended CMOS chip are nearby) - it may be something to the layers delaminating or stress from the riser nearby. I have an affected unit in a Model 76 case, but know of two reported in a Model 77 case - the riser does have a metal frame on the 77, but maybe it bends the planar too much.

[@paulloveless4122]

This was SO fascinating!!!! I watched the whole series. This whole process reminds me of the book by Bunnie Huang: Hacking The Xbox Used very similar capture and analysis tools and strategies.

[@JeffreyGroves]

Ugh, you have brought back bad memories of far too many OS/2 installs.

[@jeffreyplum5259]

I recall hearing the PDP 10 at MIT had a odd switch with two settings " Magic " and " More Magic. " I have no further details.. God bless you all.

[@wonderbars36]

Tough find. Kudos on your repairs, gentlemen. Had to deal with this sort of issue on an 8 ch. mixer. Took what seemed like an eternity to find that kind of failure. Board was tweaked/twisted just enough to crack a via.

[@chrissavage5966]

Back in pre-history, we had some pretty reliable stock-fault diagnostics we could rely on. Link 110 TV camera - look for the dead tantalum bead capacitor. Always an easy find, they stink when they explode and are easy to spot. On that device they would also take out the inline chokes on the power rails. Some years later, TV camera again, this time a Thomson 1530 Almost every fault I ever had with those was a short 0.1UF ceramic capacitor...unfortunately, they were almost always embedded inside a hybrid module - Thomson called them PLs so all you could do was look at the theoretical circuit, work out which cap was bad then replace the entire module. Later again, into the PC era now and of course, Dell computers and the infamous bad electrolytic caps. Had a really spooky one in my early days as an engineer on a Studer tape machine. When first turned on for the day, it would go into fast forward and nothing you could do would stop it. Wait 5 minutes and it would then behave. Finally tracked down to a tantalum bead cap that would self-heal. So, yep, I'm a subscriber to the belief in evil capacitors. :)

[@zyeborm]

Oh they are evil, just there's other gremlins too. I had a board I was making for uni around 2003 or so. Stripboard, my one and only all nighter before flying to woomera to put it on a rocket. Hundreds of dollars in ADC chips on this board. And there was a short. I couldn't find it. My dad couldn't find it. Eventually he had a suggestion. There's one way that will 100% show you where the short is. Hook the 5V rail up to a car battery. Just before I'm about to do it with my face stuffed into the PCB to find this damn short father says "Oh hey, aim that away from your face", "good idea" says I. Dutifully flip the board over and engage my safety squints. Connect alligator clip to battery. BANG! Fire goes spewing across the desk, red hot glowing coals sizzle into the laminex. The same desk I'm using now funnily enough. It's still got the burn holes in it. God damn tantalum cap was shorted from new. I've avoided them like the plague ever since lol. I'm getting pretty excited about the giant MLCC caps that are getting around now.

[@casalambretta]

Funny how at 13:47 the web business and shopping application spelt QANTAS as Quantas.

[@yuppiehi]

People are always saying capacitors because it makes them feel better about themselves, as they feel that they're contributing to the conversation without providing any intellectual value whatsoever. They say "change the caps" because 1) they may be right once every dozen or so times, and 2) even if you do change the caps and it doesn't fix the problem, it doesn't ruin anything, and yet they still get away with making the suggestion. To me, it is like watching a golf tournament, where on every single stroke of the golf club, you hear many people in the audience scream "In the hole!!!!" Yes, I believe "change the caps" and "in the hole" are one and the same mindset.

[@carlklitzke9455]

Wow! Great job!

[@dougholtz]

Warp 3 and 4 were great OS's. Had a free copy of Warp Server once, sold it when I realized I didn't need it.

[@ikocheratcr]

Not sure if you have considered it, maybe replace the hdd with an SCSI to SD or some other flash device. A simple google search shows that they are available on ebay $95.00 thou.