This was gut wrenching to watch. It suddenly made me consider installing crowbar/overvoltage/... protection circuits in the irreplaceable things I own.
If we can revive it, it’s getting a crowbar protection circuit for sure. Actually HP themselves later made one that could be retrofitted, I’m planning to copy that.
@@CuriousMarc I used to work for a company called BFI electronics. BFI is a term that I learned just before starting in their employ, and it is what you use in last- resort "mechanical" situations, as it stands for "Brute Force & Ignorance" . A crowbar may be handy in such situations. I can only imagine. I wonder if Dave has covered on Eevblog.
Not exactly the same but not too dissimilar to when an original Commodore 64 power supply fails and starts sending unregulated voltage into the 5V DC rail... Big difference is a C64 is not too hard to repair and source parts for...
OK, this is scary to me. I was working at HP when this machine first came out. There was the 9815, the 9825 and the grand daddy and my personal favourite the 9845. Did a lot of work on the 9845. I also like the HPIB display modules you have. Did a lot of work with HPIB back in the day. Jes,, I don't feel that old. Thanks for posting.
@@stampeder1686 The 9845 is such an amazing machine, so very far ahead of its time! But it is so freaking huge and heavy. What a monster. But it's such a beautiful sleek design, you wouldn't know from pictures of it. I have resisted getting one so far because it takes so much space and it's so hard to move around, I wouldn't really "use" it. The 9825 on the other hand fits everywhere, handy dandy to this day when you need to do simple control an HPIB instrument, with style.
@@CuriousMarc Aha. The trick with the 9845 is to put is on it's own roll around cart. I had one that just fit the 9845 with 19" rack mounts below for all my test equipment. It worked like a charm. I used it to run our final validation specs after repairing equipment. I used to do all the microwave stuff plus a few counters and DVM's back in the day. Had a nice little calibration setup. So don't chicken out go ahead and splurge for a 9845. You be glad you did. You can then slave the 9835 to it..... The 9825 was the basis of design for the first apple boards.....Steve Wozniak (sp?) worked on them.
Very sorry to hear about your HP. When something like this happens to me with a piece of pretty much irreplaceable vintage equipment, I feel almost as bad as when losing a beloved pet. Hope you are able to repair it. What a headache. I look forward to the excellent troubleshooting video(s) though.
Had this happen to my beloved HP 5328A that I had for 23 years. One of the 15 volt pass transistors had shorted collector to emitter and basically rendered it a parts unit. I hope you can get this beauty going again, one has to wonder why the engineers at HP didn't include crowbar protection or even a internal fuse like the ones found in the HP 853A.
Good morning guys. I follow you from Italy and I have a lot of admiration for your work, pure electronics. I wanted to ask you, since you have now opened the power supply card, Isn't it better to prevent and replace electrolytic capacitors too? Congratulations again, the Apollo electronics videos, I loved them. Thank you
Some people feel strongly about replacing every electrolytic they see. We don’t. We measured ours, they were fine, we left them in. Not saying we find them good all the time, we do change some occasionally, but I’ve had *way* more problems with modern electrolytics from the late 1990s and 2000s, where every single one goes bad. Also if you work with older tube equipment with much higher voltage, like Mr. Carlson, it’s a different ballgame, as you have to be much more worried about capacitor leakage and will end up replacing them much more often. Here is a video on the subject: ua-cam.com/video/GMNQDoFpYQY/v-deo.html
Old HP kit was awesome. I used to operate an HP9830 in the day. Apart from being serviced once a year by HP, it was used day and night for 14 years. Very very reliable.
So happy to see the notification for this and that you're back doing your thing. Can't say I hold out much hope for the machine but if any one can you lot can and it'll make good watching
Yes, this is serious stuff with power supplies that can damage valuable older gear. I was watching a video on a PDP power supply. It has crowbar protection but it required some very careful testing to prove that it would work under load and not overvoltage. Always interesting videos Marc.
I remember when I was looking at a old cnc there has been in a 5 volts over voltages. The EPROMs has stuck bits, it was making sense, when looking under microscope there was wire fused to see in the window.
Best of luck. I still have not dared plugin my pair in, I will have the confidence to check the supply out after watching your video. I wonder if a clamp protection circuit could be added?
Maybe its just the lighting, but it looks like the bond wires might have actually been glowing. Can see what appears to be oxide layers close to the feed-thru's and on the ends of the wires after the bonds on the die
I hope the processor itself hasnt failed, as it seems it maybe an unusual beastie. Do you know,i love Marcs repairs on older equipment becos,with discreet logic,you have a good chance of availablity,and being able to understand the circuit topology. Nowadays you open a box and,oh joy,,Big FPGA's stare back acha! Yep,, in the pooper. Less mystery if you were staring at a singularity..I love the calming music!!
gee, i was just getting started in computers and could of used this info... our company had kept the manuals as a big secret and now you make it logical... better late than never.. thanks...:)
Marc! I'm so sorry to see this beautiful machine has died on you. I can't tell you how I was cringing every time I watched you power it on and off and on and off. I was so fearful of seeing the magic smoke come out somewhere. I really hope you can revive this wonderful machine.
I remember watching the previous videos where you were powering off and on that HP 9825T a whole bunch during testing and thinking "that's gotta be tough on that old machine".
Oooh, an HP Thinkjet printer! The first removable ink cartridge/disposable printhead printer! You should do an episode on this groundbreaking device. It is a real masterpiece of industrial design.
wouldn't a zener diode on the output of the power supply help against any potential jumps in voltage? (zener conducts past its breakdown voltage and the load blows the fuse before any chips are exposed for long)
This is exactly what I'm fearing of. I'm very sad for your beloved HP. I hope you can fix it.. I'm pretty sure you can do it 😃 ( probably, you have already repaired it )
Sad day. I think I would put a fuse before that buck regulator, in case it happens again. The fuse might blow before the magic smoke is let out. But it's still a design with a serious weak point. It really needs to be corrected. 😢
Man..... I'm both excited to see a new video and sorry for such a catastrophic failure. I'm no expert in old computers but the resistance measurement mode usually isn't high enough voltage to overcome the forward voltage of ICs and all you should be seeing is pull up resistors, etc in the 100s of ohms. I'd go with an approach of applying a current limited supply to the boards one at a time and with a thermal camera seeing any abnormally hot chips. A crowbar circuit would have prevented this but I guess we can't have it all.... Good luck with the repair
HP’s definitely had the best kabooms of my life on the bench, when they blow, they blow good. Look forward to seeing the next episodes, one mans bang is our good fortune! P.S. what looks like a bottle of Japanese sake doing on your bench? 6:40
Ouch! I've had that happen with various things. Not fun. Not fun at all. When my dad was designing power supplies like that in the 1970s, he would put a husky zener on the south side of a fuse on the output, or alternately an over-voltage crowbar to short the power supply output and blow the input fuse if it went out of spec OV. It sounds like an OV protection circuit might not be hard to add, and worth the extra parts to protect against future failures.
I had known of this issue with the 9825 and I'm waiting for the parts to arrive so I can fit a crowbar circuit. After seeing Marc's video I'm not switching on my 9825 till it has a power supply crowbar.
Always fond of microscope autopsies of failed components. (I don't suppose you'd want to cut open a good one of those and check what it's supposed to look like?)
Great to get the 5v up and running again but what current is being drawn from it? I see you measured the resistance but things can easily go wrong once powered up. It may be giving 5v ok now but may still be getting hot over time but I guess you also have other fish to fry.
Surprising these linear regulators are not fitted with a crowbar. A lot of the HP gear of this vintage has. Maybe a high power 6.3V Zenner would at least have a chance of going short and blowing the 5V fuse?
The CR5 diode is not a shottky diode but this is a zener diode. Have you check it ? Have you replace it ? It a good solution to remplace it by a transil diode like 1.5KE6V8A or 1.5KE6.8A with a case different. A transil diode does not need to be cooled because in normal operation it is not active.
It is a bit surprising that they did not add some protection after the power supply. E.g. some beefy suppression diodes. Perhaps good to do this anyway?
I did something similar whilst repairing an audio mixer console... the +48 volts for the phantom power was one pin over from the +15 for the one hundred or so op-amps on the mainboard... As luck would have it, I plugged in the ribbon connector with one pin off to the left, and sent +48volts to every op-amp... needless to say, that board was now like Rover... Who was dead all over... :(
I fried my homebrew LSI-11 when the power supply failed (same way). I knew the minicomputer power supplies in the 1970s all had crowbar protection circuits, but not my supply. 😕 They seem to be a lost art.
I'm a bit surprised HP didn't fit a crowbar type protection circuit in the PSU? Even my cheap Italian made 13.8V 7Amp CB and ham radio PSU from 1980 has a thyristor crowbar circuit. Which triggers the thyristor across(crowbars) the supply in the event of an over-voltage fault and blows the fuse, like if the regulator transistor lets go.
With the expense of serious hp gear, you'd think they could afford a protection zener. It's really an essential part of a series regulator in a complex device. When I used to service TVs, they were at the output of almost every supply and worked well as a crowbar circuit. We're all hoping you get it up and running again.
Going to add an SCR crowbar circuit for overvoltage protection? A couple of resistors, a bypass cap, a Zener diode and an SCR are pretty cheap insurance. (There are no guarantees in the world, but I've put SCR crowbar circuits on a lot of supplies I've designed.) No need even to heat-sink the SCR since it needs to last just long enough to blow the fuse.
The other power supply having 2 stages is for a completely different reason though. The isolated dc/dc stage is unregulated and probably runs in a resonant mode for lower losses and EMI, the buck stage before it provides regulation of the output voltage. Here, the isolation is provided by the 50Hz transformer, so nothing more is needed. A crowbar would have been nice though.
For the future, I would suggest installing a crowbar circuit, I used MC3423 in the past. I would short everything and trip the PSU fuse, for just 1$ in parts.
There should have been an SCR crowbar circuit. So that if the 5v went past 6v the SCR would fire and short out the 5v rail. This would in turn blow fuses thus protecting the rest of the machine. I wonder why there wasn't one?
Today I locked myself out of my phone and then blew up the battery (magic smoke) while trying to repair it. I have ordered a new battery but I may have blown up the phone too. I'm thinking of live-streaming the repair. I feel your pain.
I think that would be the first time I've seen a Motorola TO-3 transistor in its original packaging. Let's just hope the 5601-3xxx hybrid CPU isn't dead...
Although it may be considered sacrilegious by some, when it comes to old equipment like this I always put a simple SCR crowbar on the supply rails just in case something old decides it wants a rest and causes damage like this. I also do it for new cheapie switching power supplies, but that's another story.
You'd think HP could have at least stuck something like a 5.6V zener in there to clamp the rail down and blow a fuse in the event of transistor failure.
Time to add in the crowbar circuit, which uses a 6V8 zener and a high power SCR to pop the fuse on transistor failure. Unfortunately those fail open circuit and silent, so YMMV. Had that happen on a system, and the only damage to the entire system was a single failed TTL hex inverter, ironically a 5406 30V open collector device, which removed it's top cap. Every other IC , including the EPROMS and the unobtanium Intersil DAC in it's purple case, was fine and unaffected, after having run for a few hours on 16VDC, and the only symptom was that it was not showing a 5 second self test lamp. Said lamp provided by the inverter, using a RC delay to make intermittent faults show up for 5s, and fed from a power good signal from the power supply, which just detected all rails were greater than half voltage. Power supply fix I replaced all parts, down to the resistors with new ones.
I'm really surprised that HP didn't provide any over voltage protection anywhere ? But then again, it has happened to me back in the day with my C64 when the 7805 regulator that was supposed to have overheat protection failed and passed unregulated 9VDC and fried most of the computer. But a C64 was not an expensive and groundbreaking HP product like this!
Reasons why I don't like buck converters without any overvoltage protection.. 😕 Similar situations I see on modern PC mainboards - CPU power rail switching transistor fails short and CPU gets killed by direct 12V.. Best thing, if ground side transistor fails first.. Very sad situation.. Can imagine how many old chips were killed due to transistor short..
In that configuration it would be just a simple clamp. A zener diode wouldn't be able to handle the much current though. Looks more like a schottky diode in a standard buck converter configuration.
About 25 years ago I got a call from a service tech working on the million $ CT scanner machine at the local hospital. He was looking for a TTL chip to repair the one of the main boards. I said I would bring him the chip if he let me have a look at he inside of the machine. Once I got there I got involved in circuit tracing and quickly discovered 9V on the main 5V power rail of the board with the failed chip. This supply went to all the boards in the machine. A power supply swap got 5V again but the poor CT machine had a lot more than one TTL chip blown!!! A few years before I rebuilt the night sky planetarium lighting controller for the projector at Iowa State University that suffered 120VAC finding it's way through a shorted TRIAC in a very poor design onto the 5V supply....not pretty!
Ouch! Hurts to see that. But as long as there are not too many unobtainium parts, I am certain that you and your partners in crime will make it live again. Bon succès !
This was gut wrenching to watch. It suddenly made me consider installing crowbar/overvoltage/... protection circuits in the irreplaceable things I own.
If we can revive it, it’s getting a crowbar protection circuit for sure. Actually HP themselves later made one that could be retrofitted, I’m planning to copy that.
@@CuriousMarc I used to work for a company called BFI electronics. BFI is a term that I learned just before starting in their employ, and it is what you use in last- resort "mechanical" situations, as it stands for "Brute Force & Ignorance" . A crowbar may be handy in such situations. I can only imagine. I wonder if Dave has covered on Eevblog.
I thought the very same thing !
@@CuriousMarc will a crowbar react fast enough, is the question.
@@CuriousMarc I recently rebuilt an HP 8672A 18ghz signal generator which has a SCR protection on 5v rail. It may be of some help looking at it.
Not exactly the same but not too dissimilar to when an original Commodore 64 power supply fails and starts sending unregulated voltage into the 5V DC rail... Big difference is a C64 is not too hard to repair and source parts for...
coool to see you here!!
@@thegodgen It's always cool to see familiar friends like Adrian. Love his channel too. :)
Maybe Marc should add some TVS diodes to the 5V rail as we already do in our C64's ;-)
Unless one of the hard to source chips die ofcourse :) Luckily there are a number of projects on the way for replacements.
That's the first thing I thought of when he described the fault, the infamous failure of the Commodore 64 power supply.
Nightmare came real! Maybe adding a simple crowbar circuit would be appropriate?... May the force be with you!
Nice suggestion, Mr. Freeman
8:18 Poor computer lets loose the wail of the digital damned.
Marc: OK, it's unhappy with something.
Marc's gift for understatement continues. :-D
I love how the gentle happy music makes the seemingly impossible task seem like a walk in the park for Marc and his team.
Ouch! Best of luck with the repair - here's hoping it can be saved.
I can't wait to see the repair attempt. I hope you guys get it going again!
OK, this is scary to me. I was working at HP when this machine first came out. There was the 9815, the 9825 and the grand daddy and my personal favourite the 9845. Did a lot of work on the 9845. I also like the HPIB display modules you have. Did a lot of work with HPIB back in the day. Jes,, I don't feel that old. Thanks for posting.
I should note I spent 10 years with HP.
@@stampeder1686 The 9845 is such an amazing machine, so very far ahead of its time! But it is so freaking huge and heavy. What a monster. But it's such a beautiful sleek design, you wouldn't know from pictures of it. I have resisted getting one so far because it takes so much space and it's so hard to move around, I wouldn't really "use" it. The 9825 on the other hand fits everywhere, handy dandy to this day when you need to do simple control an HPIB instrument, with style.
@@CuriousMarc Aha. The trick with the 9845 is to put is on it's own roll around cart. I had one that just fit the 9845 with 19" rack mounts below for all my test equipment. It worked like a charm. I used it to run our final validation specs after repairing equipment. I used to do all the microwave stuff plus a few counters and DVM's back in the day. Had a nice little calibration setup. So don't chicken out go ahead and splurge for a 9845. You be glad you did. You can then slave the 9835 to it..... The 9825 was the basis of design for the first apple boards.....Steve Wozniak (sp?) worked on them.
Very sorry to hear about your HP. When something like this happens to me with a piece of pretty much irreplaceable vintage equipment, I feel almost as bad as when losing a beloved pet. Hope you are able to repair it. What a headache. I look forward to the excellent troubleshooting video(s) though.
Mine is in the exact same state! I've talked to the helpful folk on the vintagehp forum and we've had some guesses but nothing so far. Best of luck
Had this happen to my beloved HP 5328A that I had for 23 years. One of the 15 volt pass transistors had shorted collector to emitter and basically rendered it a parts unit. I hope you can get this beauty going again, one has to wonder why the engineers at HP didn't include crowbar protection or even a internal fuse like the ones found in the HP 853A.
This is the nerdiest drama/sitcom ever! An emotional roller coaster. (and i love it)
love that you are maintaining/repairing this historic equipment, much i used in my younger days !
My thoughts and prayers are with you and the hp's family, my it get well soon. RIP dear friend...
Good morning guys.
I follow you from Italy and I have a lot of admiration for your work, pure electronics.
I wanted to ask you, since you have now opened the power supply card,
Isn't it better to prevent and replace electrolytic capacitors too?
Congratulations again, the Apollo electronics videos, I loved them.
Thank you
Some people feel strongly about replacing every electrolytic they see. We don’t. We measured ours, they were fine, we left them in. Not saying we find them good all the time, we do change some occasionally, but I’ve had *way* more problems with modern electrolytics from the late 1990s and 2000s, where every single one goes bad. Also if you work with older tube equipment with much higher voltage, like Mr. Carlson, it’s a different ballgame, as you have to be much more worried about capacitor leakage and will end up replacing them much more often. Here is a video on the subject: ua-cam.com/video/GMNQDoFpYQY/v-deo.html
Good luck I know you can fix it. You are the Mr. Scott of electronics. A miracle worker.
Good luck Marc! You'll get it!
Wow. If you manage to fix it, I hope you do something to prevent against the fault
Old HP kit was awesome. I used to operate an HP9830 in the day. Apart from being serviced once a year by HP, it was used day and night for 14 years. Very very reliable.
So happy to see the notification for this and that you're back doing your thing. Can't say I hold out much hope for the machine but if any one can you lot can and it'll make good watching
Glad to see you back :)
You had me at "power supply meltdown". :D I look forward to seeing the troubleshooting. Good luck!!
Yes, this is serious stuff with power supplies that can damage valuable older gear. I was watching a video on a PDP power supply. It has crowbar protection but it required some very careful testing to prove that it would work under load and not overvoltage.
Always interesting videos Marc.
I remember when I was looking at a old cnc there has been in a 5 volts over voltages. The EPROMs has stuck bits, it was making sense, when looking under microscope there was wire fused to see in the window.
Hope you can get it working again, its a beautiful machine. sad to see this stuff die but if anyone can get it working again its you guys
Best of luck. I still have not dared plugin my pair in, I will have the confidence to check the supply out after watching your video. I wonder if a clamp protection circuit could be added?
Maybe its just the lighting, but it looks like the bond wires might have actually been glowing. Can see what appears to be oxide layers close to the feed-thru's and on the ends of the wires after the bonds on the die
I hope the processor itself hasnt failed, as it seems it maybe an unusual beastie.
Do you know,i love Marcs repairs on older equipment becos,with discreet logic,you have a good chance of availablity,and being able to understand the circuit topology.
Nowadays you open a box and,oh joy,,Big FPGA's stare back acha! Yep,, in the pooper.
Less mystery if you were staring at a singularity..I love the calming music!!
Bummer, bad luck Mark, good for us though as we get to watch you as you skilfully repair another bit of HP gear...DA.
I am looking forward to this class in HP diagnostics
Heartbreaking. I hope you can bring her back. Maybe with some over-voltage protective circuitry once it's up and running again.
gee, i was just getting started in computers and could of used this info... our company had kept the manuals as a big secret and now you make it logical... better late than never.. thanks...:)
we believe in you! After watching you repair an Alto i have confidence you can repair this.
Marc! I'm so sorry to see this beautiful machine has died on you. I can't tell you how I was cringing every time I watched you power it on and off and on and off. I was so fearful of seeing the magic smoke come out somewhere. I really hope you can revive this wonderful machine.
You were not alone. I was cringing inside too!
I remember watching the previous videos where you were powering off and on that HP 9825T a whole bunch during testing and thinking "that's gotta be tough on that old machine".
Oooh, an HP Thinkjet printer!
The first removable ink cartridge/disposable printhead printer!
You should do an episode on this groundbreaking device. It is a real masterpiece of industrial design.
Keen eye, and episode already done, right here: ua-cam.com/video/UiHNymmxKWs/v-deo.html
wouldn't a zener diode on the output of the power supply help against any potential jumps in voltage? (zener conducts past its breakdown voltage and the load blows the fuse before any chips are exposed for long)
On the positive side: we can look forward to many interesting videos.
This is exactly what I'm fearing of. I'm very sad for your beloved HP. I hope you can fix it.. I'm pretty sure you can do it 😃 ( probably, you have already repaired it )
Sad day. I think I would put a fuse before that buck regulator, in case it happens again. The fuse might blow before the magic smoke is let out.
But it's still a design with a serious weak point. It really needs to be corrected. 😢
Man..... I'm both excited to see a new video and sorry for such a catastrophic failure. I'm no expert in old computers but the resistance measurement mode usually isn't high enough voltage to overcome the forward voltage of ICs and all you should be seeing is pull up resistors, etc in the 100s of ohms. I'd go with an approach of applying a current limited supply to the boards one at a time and with a thermal camera seeing any abnormally hot chips. A crowbar circuit would have prevented this but I guess we can't have it all.... Good luck with the repair
I really hope you get it back up and running
HP’s definitely had the best kabooms of my life on the bench, when they blow, they blow good. Look forward to seeing the next episodes, one mans bang is our good fortune! P.S. what looks like a bottle of Japanese sake doing on your bench? 6:40
Ouch! I've had that happen with various things. Not fun. Not fun at all.
When my dad was designing power supplies like that in the 1970s, he would put a husky zener on the south side of a fuse on the output, or alternately an over-voltage crowbar to short the power supply output and blow the input fuse if it went out of spec OV. It sounds like an OV protection circuit might not be hard to add, and worth the extra parts to protect against future failures.
What an awful failure mode!! 😢 Really hope you can bring that lovely bit of hardware back to life.
Great one man.
Appreciate your work and content.
Always keep it up.
I had known of this issue with the 9825 and I'm waiting for the parts to arrive so I can fit a crowbar circuit. After seeing Marc's video I'm not switching on my 9825 till it has a power supply crowbar.
I’m sure eventually you or one of your other buddies will get those pixies moving in the correct direction! I have faith!
Was CR5 a Schottky diode? The symbol is for a *zener* diode - perhaps intended to be overvoltage protection?
Maybe an addition of an SCR crowbar if you get it working again.
For sure. It is weird that it wasn't implemented in the first place.
Not only did that go to 13V, it went to about 20A. :)
Yay, CuriousMarc is back!!!
Always fond of microscope autopsies of failed components. (I don't suppose you'd want to cut open a good one of those and check what it's supposed to look like?)
And upgrade the power supply to prevent this from happening again!
Hey Marc, love your videos. This one made me think perhaps using a lavaliere mic would improve the audio. Anyway thanks!
Entertaining and informative, thanks! I am always looking forward to your next video.
Great to get the 5v up and running again but what current is being drawn from it? I see you measured the resistance but things can easily go wrong once powered up. It may be giving 5v ok now but may still be getting hot over time but I guess you also have other fish to fry.
there are a if that is a 2N4398 there is a few out there that is an upgrade that can handle more currant..
Surprising these linear regulators are not fitted with a crowbar. A lot of the HP gear of this vintage has. Maybe a high power 6.3V Zenner would at least have a chance of going short and blowing the 5V fuse?
The CR5 diode is not a shottky diode but this is a zener diode. Have you check it ? Have you replace it ? It a good solution to remplace it by a transil diode like 1.5KE6V8A or 1.5KE6.8A with a case different. A transil diode does not need to be cooled because in normal operation it is not active.
I wonder if you'll ever get to fix Ray Ozzie's Plato terminal?
That's a good one too !
Hang in there little fella.... you're in good hands.... they'll get you fixed up.
I was just wondering what you are up to the other day.
Good luck with this old HP!
Could you not bring up the other boards individually on a separate PSU with a current-limit set?
Love these things! First used one in 1976 at General Dynamics automated test equipment group. Anyone know where I can find one to buy?
It is a bit surprising that they did not add some protection after the power supply. E.g. some beefy suppression diodes. Perhaps good to do this anyway?
If anyone can you guys can, I have every confidence in you. Thumbs up.
I did something similar whilst repairing an audio mixer console... the +48 volts for the phantom power was one pin over from the +15 for the one hundred or so op-amps on the mainboard... As luck would have it, I plugged in the ribbon connector with one pin off to the left, and sent +48volts to every op-amp... needless to say, that board was now like Rover... Who was dead all over... :(
I fried my homebrew LSI-11 when the power supply failed (same way). I knew the minicomputer power supplies in the 1970s all had crowbar protection circuits, but not my supply. 😕 They seem to be a lost art.
I'm a bit surprised HP didn't fit a crowbar type protection circuit in the PSU? Even my cheap Italian made 13.8V 7Amp CB and ham radio PSU from 1980 has a thyristor crowbar circuit. Which triggers the thyristor across(crowbars) the supply in the event of an over-voltage fault and blows the fuse, like if the regulator transistor lets go.
Transistor lost its mojo.
Love the phrasing.
Aw man. :( I hope you can get it working!
With the expense of serious hp gear, you'd think they could afford a protection zener. It's really an essential part of a series regulator in a complex device. When I used to service TVs, they were at the output of almost every supply and worked well as a crowbar circuit. We're all hoping you get it up and running again.
Happy troubleshooting!
Ouch, that's a bad design but at least they learned their lesson in later ones. Looking forward to seeing the repair videos.
Going to add an SCR crowbar circuit for overvoltage protection? A couple of resistors, a bypass cap, a Zener diode and an SCR are pretty cheap insurance. (There are no guarantees in the world, but I've put SCR crowbar circuits on a lot of supplies I've designed.) No need even to heat-sink the SCR since it needs to last just long enough to blow the fuse.
The other power supply having 2 stages is for a completely different reason though. The isolated dc/dc stage is unregulated and probably runs in a resonant mode for lower losses and EMI, the buck stage before it provides regulation of the output voltage. Here, the isolation is provided by the 50Hz transformer, so nothing more is needed. A crowbar would have been nice though.
The DC resistance on each board has me a bit worried, as these aren't tube filaments. I really hope all those rare IC's haven't become resistors.
For the future, I would suggest installing a crowbar circuit, I used MC3423 in the past. I would short everything and trip the PSU fuse, for just 1$ in parts.
There should have been an SCR crowbar circuit. So that if the 5v went past 6v the SCR would fire and short out the 5v rail. This would in turn blow fuses thus protecting the rest of the machine. I wonder why there wasn't one?
His own equipment keeps providing new content 😂 But seriously, I hope you get it repaired, love your videos.
Can you tell us about that gorgeous bottle of sake at 6:42?
Sadly, it’s only a small bottle of tap water for the sponge of my soldering iron. Looks cute though.
Today I locked myself out of my phone and then blew up the battery (magic smoke) while trying to repair it. I have ordered a new battery but I may have blown up the phone too. I'm thinking of live-streaming the repair. I feel your pain.
I think that would be the first time I've seen a Motorola TO-3 transistor in its original packaging.
Let's just hope the 5601-3xxx hybrid CPU isn't dead...
If it is (which is likely), the machine is completely f**ked. No chance of repair.
Tragedy! But come on Marc, if you and your friends can resurrect an AGC this has to be a walk in the park... even if an overgrown park. Good Luck
Although it may be considered sacrilegious by some, when it comes to old equipment like this I always put a simple SCR crowbar on the supply rails just in case something old decides it wants a rest and causes damage like this. I also do it for new cheapie switching power supplies, but that's another story.
maybe now would be a good time to make sure there is some kind of protection mechanism for a future such event! When one goes...
You'd think HP could have at least stuck something like a 5.6V zener in there to clamp the rail down and blow a fuse in the event of transistor failure.
Very unfortunate failure.
If anyone can fix it, it's you guys! Best of luck!
Time to add in the crowbar circuit, which uses a 6V8 zener and a high power SCR to pop the fuse on transistor failure. Unfortunately those fail open circuit and silent, so YMMV. Had that happen on a system, and the only damage to the entire system was a single failed TTL hex inverter, ironically a 5406 30V open collector device, which removed it's top cap. Every other IC , including the EPROMS and the unobtanium Intersil DAC in it's purple case, was fine and unaffected, after having run for a few hours on 16VDC, and the only symptom was that it was not showing a 5 second self test lamp.
Said lamp provided by the inverter, using a RC delay to make intermittent faults show up for 5s, and fed from a power good signal from the power supply, which just detected all rails were greater than half voltage. Power supply fix I replaced all parts, down to the resistors with new ones.
When you do get it running again, stick a "crow bar" in with the power supply to blow fuses at 6V.
I'm really surprised that HP didn't provide any over voltage protection anywhere ? But then again, it has happened to me back in the day with my C64 when the 7805 regulator that was supposed to have overheat protection failed and passed unregulated 9VDC and fried most of the computer. But a C64 was not an expensive and groundbreaking HP product like this!
Fingers crossed you can bring it back to life!
Reasons why I don't like buck converters without any overvoltage protection.. 😕
Similar situations I see on modern PC mainboards - CPU power rail switching transistor fails short and CPU gets killed by direct 12V.. Best thing, if ground side transistor fails first..
Very sad situation.. Can imagine how many old chips were killed due to transistor short..
2:55 Schottky diode? Looks like a Zener symbol to me. It might have been a early attempt at a crowbar circuit.
In that configuration it would be just a simple clamp. A zener diode wouldn't be able to handle the much current though. Looks more like a schottky diode in a standard buck converter configuration.
About 25 years ago I got a call from a service tech working on the million $ CT scanner machine at the local hospital. He was looking for a TTL chip to repair the one of the main boards. I said I would bring him the chip if he let me have a look at he inside of the machine. Once I got there I got involved in circuit tracing and quickly discovered 9V on the main 5V power rail of the board with the failed chip. This supply went to all the boards in the machine. A power supply swap got 5V again but the poor CT machine had a lot more than one TTL chip blown!!!
A few years before I rebuilt the night sky planetarium lighting controller for the projector at Iowa State University that suffered 120VAC finding it's way through a shorted TRIAC in a very poor design onto the 5V supply....not pretty!
I've had a similar fault happen on a homemade buck converter. Absolutely heartbreaking.
If someone can fix it is you Marc!
Yay new Marc vidjeooo! Thank you!
Ouch! Hurts to see that. But as long as there are not too many unobtainium parts, I am certain that you and your partners in crime will make it live again. Bon succès !