HP 9825 Repair Part 3: HP Logic Analyzer to the Rescue 

Unusual times require unusual solutions

Uuh, hp 9825 not hp 9824, and this is a 9825T to be exact

Indeed. As Tony Duel put it in his 98xx repair manual, "No board swapping!" Swapping boards between a good unit and bad one is cheating. It's also dangerous b/c you can break a good board by putting it in a bad machine!

if I didn’t have a second Fairlight CMI to check my other one, I would be at a loss,

I used to do a lot of production repair. You do tend to have the best information (nip over to development to get a ROM listing) but the downside is that, when repairing a failed machine, there's rarely more than one fault. In production, the machine has never worked, so it might have all of one part upside down or any number of errors. Repairing Marc's machine is in-between - you haven't got manufacturer support and there could be any number of faults. But it did work once.

These machines have a sense for drama!

It's technical ER...Marc is George Clooney!

Fingers crossed it's not going to turn into an episode of Quincy, M.E.

Yes it's exciting. I hope the Alto isn't having an affair..!

I watch these more frequently than medical soaps. Seems like a rather quirky idea to pitch to a TV studio, but I'd be down for it.

I think that the curiousmarc has already found a solution, but he releases its videos with an about 3 weeks delay so most likely it will not be relevant anymore

@@michvod Still useful for a lot of people reading this. And very interesting.

I never heard of this approach before. If I am ever in a situation where it might be helpful, I hope I will remember it.

@@benbaselet2026 It is very interesting :) My experience with failed ICs was usually almost zero ohm shorts

And the ALE pin signals on the 8051 had such sharp edges that it radiated every 0.5MHz up well into the UHF band

There are people that do board-level maintenance every day like this. Many facilities have their own engineering departments that repair damaged units such as this.

I still like the 1630G best. The manual is written in terms of how it works rather than how it's used. A great help for a nerd.

The danger is that you will blow up the good machine / board so you have to be really careful about knowing what the fault is before you do this otherwise more magic smoke might emerge.

It was called a crowbar modification right?

@@zomaardan Yep, simple self-kill switch, when voltage goes too high.

@@zomaardan You could use a fuse in line with a simple TVS Diode to ground or you could implement the active crowbar circuit that was implemented later.

I can't remember the 9825 having a reset button but it was a long time ago since I had one on my desk. It was the late 1970's we brought several as part of a project to improve the reliability of the Victoria Line Automatic Train control system.

*disassembler :)

@@catalinbadalan4463 Normalkt, yes but not in this case. For whatever reason, HP's documentation refers to disassemblers rubbing in their logic analysers an inverse assemblers. There is a difference in that this code runs in real time not after the fact so that you see mnemonics on the analyser screen instead of just hex

Agree on everything! Tek scopes and HP LAs (at least in this vintage era - not so true anymore). BTW I just got a 7D01 for my Tek 7854, maybe if we make it work I can do a video on it. It’s so small and cute!

Exactly. That’s why we pulled both chip N and N+1.

Great idea: we could end up with a bonus repair of the second one ;)

I think he's intentionally avoiding that kind of shortcuts for making the video more interesting. (Forces the flow of the video into deep troubleshooting and circuit analysis)

@@leonardoliveira You wouldn’t want a working board to be damaged by fault in the broken machine, then you would have two broken units. Maybe a semiconductor is now a dead short to the power rail, or a capacitor is a dead short to ground. That chip at the end of the video probably didn’t leak its magic smoke due to its age unfortunately...

If you like it; just start doing it! Get a computer (calculator) and mess around with it. You break 10 things you fix 1.

I don't think the ttl chips loading the CPU are the problem so much though,it was the stonking great 13V rail that caused problems to all the chips that were on the 5V supply rail,,, weren't expecting that!! Some would have gulped too much,and died in their bedrooms,,some may have just held on before the supply failed,its a question of evolution now.The chips they replaced weren't loading the CPU bus really,this looks like its now a dead chips hunt.. The FPGA route is certainly a possibility i agree,if the worst has happened and CPU is damaged ,but,,,,,,a TON of work..Karl looks dead!! What's happened to ken?Hope he's okay.

Like his old self!

Facial hair makes you smarter

That a good point,i think Marc is too focused on the damage that's been done,i really fear that the processor has sustained damage,and will be unobtainium, so pretty much game over.BUT....if Karl REALLY fancies it,(he looks resigned to defeat though!)he could build an FPGA board that emulates the processor,it will need quite a bit of work to do,but is absolutely doable,that would be really fun,and would impress us all with the never say die mindset..

Risky...CPU may well be damaged,and it could damage a unit its swapped into,they have to hang it out,we'll get there.Fault finding is 50% patience,when you try to short cut,you often make things worse,i know,,ive learnt to my own cost!!!

@@julianrichards9509 Do you really think so? Maybe I've just been lucky, but I've never had issues with a faulty CPU frying things. It's not outputting weird voltages or anything.

@@briangoldberg4439 Sorry to be so long Brian,it is my mindset to be very cautious mate.Did you ever see the movie Apollo 13?The side of the module blew,and all of a sudden,the telemetry went absolutely crazy back on earth.The crew reported power supply failures,the AGC rebooted,there was a loss of pressure on fuel tanks,and during one point in the mayhem,you hear this cool voice of mission director Gene Krantz say"okay lets work the problem people,lets not make anything worse by guessing" He was damned right.I used to repair multiplexers used by the broadcasting industry in their newsrooms here in England the BBC ,ITN, amongst others were our clients.Whenever i tried to shortcut, sometimes i got away with it,but sometimes i made things worse,and regretted it.Working the problem takes longer,and is painful whilst you're going through it.But when you get out the other end,you feel that you now understand the problem,and can deal with it confidently. You maybe right,swapping it may tell us quickly,that the processor is a veggie,although it looks as if it wont be easy to remove these CPU's,and you'll have to do that twice. But i think Marcs approach is sound.He's getting readings on the Logic Analyser which don't make sense,and he's verifying he's not being a jerk,by setting his analyser up on a good working machine.He will then know he's correctly setting it up,and can attack the dud machine.I used to sometimes have crazy readings on my hp logic analyser,i was faultfinding Intel 286 mboards. It usually turned out,i wasn't setting things up correctly,and when i corrected my setup,the bus activity made sense. I hope i havent been too wordy here Brian!!!

@@julianrichards9509 No worries Julian. I asked a serious question, and I got a serious answer! I think it's refreshing that the viewers of this channel seem to be calm an intelligent, and capable of respectful answers. :)