They had a variac which gives more flexibility. But that doesn't really limit current on a short either, so a light bulb is often the weapon of choice around here too.
Ahh, the joys of SMPS repair. Most painful I have ever repaired was from my own Sony Profeel KX-27PS1 TV Monitor circa 1981. It was a strange self oscillating power supply with a epoxy coated hybrid control board that would control the supply's output by injecting a signal into a control winding on the isolation transformer/inductor. It would randomly blow the chopper on startup once every few days to every few weeks. After replacing half a dozen expensive sony transistors I modified the soft start circuit which cured the problem. Soft start was very odd too consisting of a huge current limiting resistor and relay that would switch the resistor out of circuit as the secondary voltage would rise. Cheers.
This really gives you the chills like a proper horror movie..."noooooo please don't go in there sweet Fairchild transistor, it's not safe...aahhh and in you went".
Power supplies have become too cheap to justify repairing them for our customers. Replacement is generally the better economic bet. But when we did repair switchmode supplies back in the bad old days, an isolation transformer was absolutely mandatory. It breaks the galvanic path from the AC mains input to the SMPS, and permits you to choose any arbitrary point in the primary circuit as your reference ground. Makes a 'scope more useful. Also prevents the "exploding ground clip" hazard. Kinda takes me back. Great work!
Loving this series Marc! Being able to journey with you through the testing, diagnosis, repair and final testing is great and I’m learning a lot. You guys have some excellent test gear too, which comes in handy. Good luck on the next power supplies, hopefully they won’t be so painful to repair.
Excellent. I really enjoyed your repair video. Whenever I do PSU repairs for industrial customers, I replace ALL caps, re-seat each and every IC, and check every resistor for proper values, in prior to any measurements. C20 is very likely to fail soon, even though, I can see you have checked it already.
Some of the greatest videos on all of RUclips here. Covering so many realms. Thank you is all I can offer. Set up a Patreon and I’d be happy to chip in.
that was interesting. my suggestion would be: a) during testing put really low rated quick fuses, this may save transistors b) for regular operation, put sand filled fuses. it's a lot more secure if they blow. Or alternativly put silicone caps on them c) get a transformer for propper earth separation plus an easy accessible switch. looking at your "extention cord operation" really makes me shiver. this is not safe.
Very good. I worked on switching power supplies for the AUTOVON system in the US Air Force. The power transistors would blow even when you did everything right. Tough to diagnose and repair!
I love to see this. Always a joy to see you work on this really old machines. It really nice to see you bring them back from the dead. They are finally undead again ;-)
Great job of repairing a PSU without any supporting documentation. I don't envy your task with the other PSU's - let alone the rest of the Alto! I'll be following your endeavours at every stage. Good luck!
Very interesting vid, as always. Only a handful of repair vids that go to this detail. My first thought about the multivibrator is that it has a feedback path, so will probably inhibit the chopper signal if the output is missing. Anyway - still watching. lol
Thanks! And the depth or detail requires more effort on the part of the viewer. That's why the retr0bright restoration vids gets 500M subscribers and mine get 8k. Nothing wrong about that though :-)
CuriousMarc lol. True. RUclips seems to be flooded with Nintendo Switch vids atm, too. Easy ingestion for the masses. Magic smoke was from a thermistor, no? (for in-rush current?)
only a handful of repair vids ever go into this much detail, is right. FFCossag is another I know of that will indeed get into this much detail, and the occasional MaxxArcade, when he wasn''t swamped with work and put out plenty of audio amp repair videos some years ago. I think FFCossag gets more special mention since he's one of the no-frills electronics guys on youtube that not only repairs stuff in alarming detail, but also will bend stuff to his will and "improve" the design of commercial products to make them less prone to inevitable failure otherwise. Off that, I'm very pleased the replacement caps were Cornell-Dubilier. I don't think any restoration job on something such as this deserves anything less.
+Jordan I agree, 8-bit is very entertaining! His videos are very well made, well researched and instructive. And he works on vintage hardware that is quite accessible to most folks.
I like 8-bit guy / 8-bit keys as well, but he's not so experienced in electronics (self-confessed). Don't get me wrong, David makes some great vids. I just like to watch the more in-depth stuff, where possible. The Alto stuff is right up my street, as I spent many years helping to restore a classic Quantel Paintbox from the BBC. ;) (which, for various reasons still isn't booting.) Actually +CuriousMarc - do you know of any SMD HDD readers / writers / emulators that don't cost a fortune, or any FPGA stuff that your colleagues may want to open-source? We're still trying to recover the (possible) data from a Fujitsu SMD HDD with 14" platters. I've done most of the PCB layout for the differential transceivers, and will need to write the FPGA logic soon for the MFM encoding / decoding.
From what I can remember, it was considered 'common practice' to set the 5V output voltage somewhat above the nominal 5V to help account for cable resistance. eg: If you set it to 5.1 volts (+2%), then there's a decent chance it will still be within tolerance (normally +/-5%) even when it's fully loaded down.
I was very scared when you started to test the PSU. Especially when you couldn’t find the source of the buzzing. I’ve had multiple caps blow up in my face and so maybe now I’m a bit traumatized. This turned into a thriller video for me haha
One trick I use is to use and incandescent lamp instead of a fuse during debugging. The lamp will limit the current to a safe value in case too much current is drawn. Also as the resistance of the lamp when cold is about 1/10 of working resistance the voltage drop is not so big at low currents. This has saved me many blown power transistors.
Here in April 2020, that OEM transistor is $40 used on Ebay. Multiple vendors selling NOS have quantities under 100 - probably none have sold since you made this video. You sure do have perseverance!
Excellent video, thank you very much to make this stuff, it's pure gold i learned a lot!. Keep the good work. Ohhh man that scope... really cool gear (and brains of course) to restore this piece of history! it totally deserves it.
Love switching power supplies in a sick kind of way. Currently in the back of a telly and it's perhaps the most challenging thing I have ever tried to fix
I work in aviation qualification. Unfortunately, for the qualification of equipment, the exact spec of the configuration is essential. Any old components from the zenzyg market could result in unsafe performance. That makes for $+++ and there is profit to be made, sure, but don't underestimate the work in qualifying aerospace equipment. Just look through DO-160! Snakes on a plane are the least of your worries.
Use a an underrated voltage isolation transfomer. Probably 50VA to 100Va will be enough to kick the PSU in. Two good points. First, it protects your scope. Second, you CANT blow a fuse nor explosions, nor sparks since the isolator drops very gently (when underrated). I repairing ATX PSUs, I always use a 50VA isolator, never got a serious event, its is not powerfull enough to blow the fuse. I put a led on the output so I can tell when the isolator is shorted or overloaded (to avoid toasting it). The isolator started most of the PSUs. Strugled once with a POE PSU from a CISCO switch, because it has active PFC and it need to start working even with in stby and takes a nice amount of current. But eventually kicked in.
Great job! It's never a good idea to open the control loop of a SMPS; you learned that the hard way ;) I've modifed some small 20 to 50 watt SMPS to turn them into constant current power supplies for LEDs, and I've always double checked that the control loop is closed before powering it on. SMPS using the flyback converter design (common for up to like 100 watts) will blow up instantly when starting them with an open control loop.
This Power Supply appears to be an early switch-mode power supply, that resembles some aspects of older Linear Designs. Given that these power supplies are ~35-40 years old ... leaking Electrolytic capacitors are common. Blown diodes suggest an external surge voltage for their failure.
I’d be very nervous leaving the IC in a socket - should the socket develop a bad connection over time (thermal cycling etc) and the regulation go haywire, I shudder to think what damage would happen to the the Alto. Well done on getting it up and running though. Switch mode power supplies can get quite entertaining when they blow up their switching transistors...
Had a desktop power supply let loose the magic smoke once. Was one hell of a bang then loads of smoke. Took out a cd rom drive as well as a hard drive. Never want to experience that again
If you used at least a dim bulb tester, I think you would know immediately if there was a short without popping the fuses. Of course the lamp(s) would start out at a minimum wattage threshold way below the power supply's normal consumption then work your way up in steps.
The Antique Radio Forum helped me find standard cross reference for transistors with proprietary 1960's 3M part numbers stamped on them. After a couple months someone came forward with access to a database which showed the cross reference (for which it was a very common ) transistor. If interested, the thread is here: antiqueradios.com/forums/viewtopic.php?f=13&t=319927 I suspect these may have been manufactured by someone else and a proprietary p/n was placed on them for the power supply manufacturer, in order to keep someone from easily repairing it, therefore keeping the manufacturer and/or their service techs employed.
Seems a LH research domain was created for Pioneer Magnetics which builds replacement power supplies. Probably a lot less expensive than $600 for those two transistors. www.lhresearch.com/
Excellent suggestion. I always use a variac and lamp limiter when powering up any bit of kit that hasn't been used for some time, saves fuses and tears.
Such supplies scare me silly. I've had good results running the high voltage side at just a few tens of volts provided you power the control logic externally. This allows you to follow the switching signals right up to and including the power stage without the risk of imminent death! You can exercise pretty much every component in the supply this way.
Another strategy is to power the supply through a resistor, which limits the current and so avoids explosions. This doesn't do much for the shock risk, however. Once you are happy it is essentially working (albeit at very low power), swap the resistor for a filament light bulb. At low currents, this has low resistance. If the current goes up unexpectedly, all that happens is that the light comes on! It's an old idea, but still handy in cases like this.
Old transistors are a pain. I have a few 50va 125 volt 400hz aerospace inverters that I managed to blow up the transistors on. I got lucky and found the schematic that told me what type of transistor they were and what they needed to be rated at. Took a while but I found 10 of them for about $30 from China. Got them and swapped them in and my inverter lived again, surely not to fly again but good as new if not better. I think they needed to be 60 volts at something like an amp or two but the replacements I got were rated for up to 10 amps. I examined an original circuit and the circuit with new transistors with my oscilloscope and they seemed to be electrically identical.
A variac is very handy for restoring power supplies that are having overcurrent problems. That way you can monitor for excessive current through the circuit as power is applied gradually.
Might help a little bit to reforms caps in a linear power supply. But if there is a hard short in a linear supply it’s useless too. It won’t help you with a switching power supply - the transistors are used as temporary shorts. And if it’s not temporary, they blow themselves up. Nature of the beast. Might even hurt: you could get a good switcher in a stuck undervoltage state and destroy it.
You missed the opportunity for a "walk in the PARC" joke! Good job though, I'm curious what the cost of replacing the PSUs with similar, modern ones would be!
It's not a question of cost obviously. You don't want to destroy the historical authenticity of the machine by blindly sticking new generic parts in it, even though it would be way easier. Plus these are are quite noble, early switching power supplies. Priceless? It's a fine balance sometimes trying to preserve the functional as well as the historical - we try very hard to do both. That said, I was pretty close to giving up on that one...
A dedication to authenticity I can support. If they fail again with something you can't replace, would you consider modern (probably higher efficiency) supplies?
For a machine as rare and historically significant as the Alto, and belonging to a museum, unlikely, unless there is no other practical solution. For a more widely available machine, that I own, and with a generic consumer grade power supply, possibly. I'd still try to repair the original first!
At the end of the 1980s I had a hard time with Solitron devices darlington transistors that were connected in bridge in servo controllers a customer brought me in order to repair. I consulted my suppliers and the price quote was absurd. The solution was replace by power mosfets.
@02:05 How to find shorted tantalum capacitors: "Turns out -- it's still bad . . . because the output's shorted. So . . . and then we get the classic shorted tantalum, which we found by basically just powering it with several amps, and we looked for smoke."
I worked for a guy in the mid 70's that started his own company, designing and constructing microcontroller-based systems. He was especially proud of the switching power supply module that he sweated and slaved over for long hours to get designed properly. Last I checked he was still in business.
The fuse problem is the worst... Many times I have rolled across the shop to steal fuses from other equipment to diagnose what's on the bench. If only someone made a resetable ckt breaker for those push and turn holders...
Because it's the one I have with a big screen that shows well on camera. And boy was it cheap. A poor excuse obviously. Let's use the 8-digit HP DMM next time. Or alternatively I could do I Chinesium DMM repair video ;-)
I would do a Chinesium DMM destroy video instead... It would probably help dealing with the frustrations and pressure of fixing a stubborn PSU, by destroying something else for a change ! lol ! Anyway, you have waaaay more patience than me trying to guess your way in reverse engineering old technology. Great job !
NO WAY! As he mentioned a few times, at least 400 V capable transistors are needed. 3055 is rated 60 V. Moreover, a 3055 is SLOW. For the switching, you need FAST or even faster! No deal...
Personally, I find the MCL videos to be somewhat drawn out... If he measures a resistor and finds that it's drifted high, the following 30 mins of video explain the replacement of that resistor. That's fine if you're working for the government, but in the REAL world, it'd be replaced in 30 seconds rather than 30 minutes
It is not "restoration project", if you throw all innards away and emulate them with Rasberry Pi and LCD-screen. From user point of view the result would the same (or better), but CuriousMarc wants to suffer.
I think the point was more of "These are just off the shelf regulated PSUs so why not grab some new ones". They aren't Alto specific and even for testing reasons I'd keep a newer set of things on hand, just in case one like this decides to go again while trying to get the rest of the system up and running. It's not entirely out of the picture that a PSU with this many failures so far will have more down the line.
+Laurent For the challenge, and for preserving the integrity of the historical original of course. But Carl and Ken were questioning my sanity just like you!
600$ for 2 transistors -> #mademyday! God safe this kind of customers! I know, trying to keep it historical correct, but spending 600$ for stupid transistors properly nobody else see's the next 100 years is really mind "blowing". ;-P
They're only $300 nowadays because they're no longer manufactured. So, technically, he didn't blow a pair of $300 transistors - but that's the price for replacement original parts. I think so long as the replacement parts perform the same as, or better than (as is often the case), the originals and they are a good physical match then it's completely acceptable to use them. After all, many of the modern replacement capacitors are smaller than the original ones. Surely the main aim is to have a fully functional piece of kit, rather than a faulty, albeit totally original one!
I would, for safety reasons, replace all them PSU's with a regular ATX power supply, everything is there except for the 15v lines in which I would either use a another supply or a stepup.
Hint: Use the old trick of putting an incandescent bulb in series with the line input. Saves fuses and components on initial startup!
They had a variac which gives more flexibility. But that doesn't really limit current on a short either, so a light bulb is often the weapon of choice around here too.
That's correct.
I do that every time 😂 a 50W works perfectly on crt devices
i know I'm kinda randomly asking but does anyone know a good site to stream new movies online ?
@Timothy Leland Flixportal
Ahh, the joys of SMPS repair. Most painful I have ever repaired was from my own Sony Profeel KX-27PS1 TV Monitor circa 1981. It was a strange self oscillating power supply with a epoxy coated hybrid control board that would control the supply's output by injecting a signal into a control winding on the isolation transformer/inductor. It would randomly blow the chopper on startup once every few days to every few weeks. After replacing half a dozen expensive sony transistors I modified the soft start circuit which cured the problem. Soft start was very odd too consisting of a huge current limiting resistor and relay that would switch the resistor out of circuit as the secondary voltage would rise. Cheers.
After watching this I decided to learn how SMPS work and tonight I managed to fix my PDP-11 PSU :D
Congrats!
This really gives you the chills like a proper horror movie..."noooooo please don't go in there sweet Fairchild transistor, it's not safe...aahhh and in you went".
Normally the transistor is there to protect the fuse :D
Exactly. The design worked perfectly!
LMAO! 😂😂🤣😂😂
Power supplies have become too cheap to justify repairing them for our customers. Replacement is generally the better economic bet. But when we did repair switchmode supplies back in the bad old days, an isolation transformer was absolutely mandatory.
It breaks the galvanic path from the AC mains input to the SMPS, and permits you to choose any arbitrary point in the primary circuit as your reference ground. Makes a 'scope more useful. Also prevents the "exploding ground clip" hazard.
Kinda takes me back. Great work!
Loving this series Marc! Being able to journey with you through the testing, diagnosis, repair and final testing is great and I’m learning a lot. You guys have some excellent test gear too, which comes in handy. Good luck on the next power supplies, hopefully they won’t be so painful to repair.
Another Restoration adventure! I will be watching from the edge of my couch the whole way! keep up the good work guys. :)
Excellent. I really enjoyed your repair video.
Whenever I do PSU repairs for industrial customers, I replace ALL caps, re-seat each and every IC, and check every resistor for proper values, in prior to any measurements. C20 is very likely to fail soon, even though, I can see you have checked it already.
Some of the greatest videos on all of RUclips here. Covering so many realms.
Thank you is all I can offer. Set up a Patreon and I’d be happy to chip in.
I had a chance to visit Digibarn about 5 years ago and was sad that the Alto was non-operational. Thank you for your excellent restoration work!
that was interesting.
my suggestion would be:
a) during testing put really low rated quick fuses, this may save transistors
b) for regular operation, put sand filled fuses. it's a lot more secure if they blow. Or alternativly put silicone caps on them
c) get a transformer for propper earth separation plus an easy accessible switch. looking at your "extention cord operation" really makes me shiver. this is not safe.
I put a 300w tungsten lamp insted of a fuse
Very good. I worked on switching power supplies for the AUTOVON system in the US Air Force. The power transistors would blow even when you did everything right. Tough to diagnose and repair!
That feeling of satisfaction is why we do what we do!
Thank you for your perseverance.
I love to see this. Always a joy to see you work on this really old machines. It really nice to see you bring them back from the dead. They are finally undead again ;-)
Great job of repairing a PSU without any supporting documentation. I don't envy your task with the other PSU's - let alone the rest of the Alto! I'll be following your endeavours at every stage. Good luck!
Nice work....
Typically when I do this, the thing explodes, the house burns down, and the neighborhood catches fire. lol
Very interesting vid, as always.
Only a handful of repair vids that go to this detail.
My first thought about the multivibrator is that it has a feedback path, so will probably inhibit the chopper signal if the output is missing.
Anyway - still watching. lol
Thanks! And the depth or detail requires more effort on the part of the viewer. That's why the retr0bright restoration vids gets 500M subscribers and mine get 8k. Nothing wrong about that though :-)
CuriousMarc
lol. True.
RUclips seems to be flooded with Nintendo Switch vids atm, too. Easy ingestion for the masses.
Magic smoke was from a thermistor, no?
(for in-rush current?)
only a handful of repair vids ever go into this much detail, is right. FFCossag is another I know of that will indeed get into this much detail, and the occasional MaxxArcade, when he wasn''t swamped with work and put out plenty of audio amp repair videos some years ago. I think FFCossag gets more special mention since he's one of the no-frills electronics guys on youtube that not only repairs stuff in alarming detail, but also will bend stuff to his will and "improve" the design of commercial products to make them less prone to inevitable failure otherwise.
Off that, I'm very pleased the replacement caps were Cornell-Dubilier. I don't think any restoration job on something such as this deserves anything less.
+Jordan I agree, 8-bit is very entertaining! His videos are very well made, well researched and instructive. And he works on vintage hardware that is quite accessible to most folks.
I like 8-bit guy / 8-bit keys as well, but he's not so experienced in electronics (self-confessed).
Don't get me wrong, David makes some great vids.
I just like to watch the more in-depth stuff, where possible.
The Alto stuff is right up my street, as I spent many years helping to restore a classic Quantel Paintbox from the BBC. ;)
(which, for various reasons still isn't booting.)
Actually +CuriousMarc - do you know of any SMD HDD readers / writers / emulators that don't cost a fortune, or any FPGA stuff that your colleagues may want to open-source?
We're still trying to recover the (possible) data from a Fujitsu SMD HDD with 14" platters.
I've done most of the PCB layout for the differential transceivers, and will need to write the FPGA logic soon for the MFM encoding / decoding.
Great! Really enjoyed it.
Perhaps difficult to repair, but makes it even more interesting to watch!
From what I can remember, it was considered 'common practice' to set the 5V output voltage somewhat above the nominal 5V to help account for cable resistance.
eg: If you set it to 5.1 volts (+2%), then there's a decent chance it will still be within tolerance (normally +/-5%) even when it's fully loaded down.
Mesmerizing video! It's always a great pleasure to see a master at work.
Wow!!! So much of components for a 5v regulated power supply.good to see repaired
Outstanding work. I love to watch the gear I was raised on come back to life.
I was very scared when you started to test the PSU. Especially when you couldn’t find the source of the buzzing. I’ve had multiple caps blow up in my face and so maybe now I’m a bit traumatized. This turned into a thriller video for me haha
One trick I use is to use and incandescent lamp instead of a fuse during debugging. The lamp will limit the current to a safe value in case too much current is drawn. Also as the resistance of the lamp when cold is about 1/10 of working resistance the voltage drop is not so big at low currents. This has saved me many blown power transistors.
Here in April 2020, that OEM transistor is $40 used on Ebay. Multiple vendors selling NOS have quantities under 100 - probably none have sold since you made this video. You sure do have perseverance!
Excellent video, thank you very much to make this stuff, it's pure gold i learned a lot!. Keep the good work. Ohhh man that scope... really cool gear (and brains of course) to restore this piece of history! it totally deserves it.
I was fixing a microcontroller circuit whilst this was on and at the exact moment you said "ooh! 5 volts"... I said exactly the same thing.
Love switching power supplies in a sick kind of way. Currently in the back of a telly and it's perhaps the most challenging thing I have ever tried to fix
Blowing fuses in spectacular fashion. Love it! :)
You've never seen a power supply with that many problems? Welcome to my world. 😭
Good job on that fix. Must have been satisfying adjusting it to 5v
I work in aviation qualification. Unfortunately, for the qualification of equipment, the exact spec of the configuration is essential. Any old components from the zenzyg market could result in unsafe performance. That makes for $+++ and there is profit to be made, sure, but don't underestimate the work in qualifying aerospace equipment. Just look through DO-160! Snakes on a plane are the least of your worries.
That is very true.
Use a an underrated voltage isolation transfomer. Probably 50VA to 100Va will be enough to kick the PSU in. Two good points. First, it protects your scope. Second, you CANT blow a fuse nor explosions, nor sparks since the isolator drops very gently (when underrated). I repairing ATX PSUs, I always use a 50VA isolator, never got a serious event, its is not powerfull enough to blow the fuse. I put a led on the output so I can tell when the isolator is shorted or overloaded (to avoid toasting it). The isolator started most of the PSUs. Strugled once with a POE PSU from a CISCO switch, because it has active PFC and it need to start working even with in stby and takes a nice amount of current. But eventually kicked in.
Great job!
It's never a good idea to open the control loop of a SMPS; you learned that the hard way ;)
I've modifed some small 20 to 50 watt SMPS to turn them into constant current power supplies for LEDs, and I've always double checked that the control loop is closed before powering it on.
SMPS using the flyback converter design (common for up to like 100 watts) will blow up instantly when starting them with an open control loop.
This Power Supply appears to be an early switch-mode power supply,
that resembles some aspects of older Linear Designs.
Given that these power supplies are ~35-40 years old ... leaking Electrolytic capacitors are common.
Blown diodes suggest an external surge voltage for their failure.
I’d be very nervous leaving the IC in a socket - should the socket develop a bad connection over time (thermal cycling etc) and the regulation go haywire, I shudder to think what damage would happen to the the Alto. Well done on getting it up and running though. Switch mode power supplies can get quite entertaining when they blow up their switching transistors...
Had a desktop power supply let loose the magic smoke once. Was one hell of a bang then loads of smoke. Took out a cd rom drive as well as a hard drive. Never want to experience that again
If you used at least a dim bulb tester, I think you would know immediately if there was a short without popping the fuses. Of course the lamp(s) would start out at a minimum wattage threshold way below the power supply's normal consumption then work your way up in steps.
The Antique Radio Forum helped me find standard cross reference for transistors with proprietary 1960's 3M part numbers stamped on them. After a couple months someone came forward with access to a database which showed the cross reference (for which it was a very common ) transistor.
If interested, the thread is here:
antiqueradios.com/forums/viewtopic.php?f=13&t=319927
I suspect these may have been manufactured by someone else and a proprietary p/n was placed on them for the power supply manufacturer, in order to keep someone from easily repairing it, therefore keeping the manufacturer and/or their service techs employed.
Seems a LH research domain was created for Pioneer Magnetics which builds replacement power supplies. Probably a lot less expensive than $600 for those two transistors.
www.lhresearch.com/
Is anyone else screaming "You have a Variac, I've seen you use a Variac, plug it into the damn variac"?
A variac can't really tell you if there is a short circuit or not. A dim bulb tester definitely does without letting the magic smoke out first.
Excellent suggestion. I always use a variac and lamp limiter when powering up any bit of kit that hasn't been used for some time, saves fuses and tears.
Looks like season 2 of Alto Restoration is going to be the best.
Congrats on this one ! Btw I know you are carefull, but please, be even more carefull, working on switching power supplies is damn dangerous.
Welcome aboard Sir! :)
That large watch is making me very nervous....
Such supplies scare me silly. I've had good results running the high voltage side at just a few tens of volts provided you power the control logic externally. This allows you to follow the switching signals right up to and including the power stage without the risk of imminent death! You can exercise pretty much every component in the supply this way.
Another strategy is to power the supply through a resistor, which limits the current and so avoids explosions. This doesn't do much for the shock risk, however. Once you are happy it is essentially working (albeit at very low power), swap the resistor for a filament light bulb. At low currents, this has low resistance. If the current goes up unexpectedly, all that happens is that the light comes on! It's an old idea, but still handy in cases like this.
Old transistors are a pain. I have a few 50va 125 volt 400hz aerospace inverters that I managed to blow up the transistors on. I got lucky and found the schematic that told me what type of transistor they were and what they needed to be rated at. Took a while but I found 10 of them for about $30 from China. Got them and swapped them in and my inverter lived again, surely not to fly again but good as new if not better. I think they needed to be 60 volts at something like an amp or two but the replacements I got were rated for up to 10 amps. I examined an original circuit and the circuit with new transistors with my oscilloscope and they seemed to be electrically identical.
A variac is very handy for restoring power supplies that are having overcurrent problems. That way you can monitor for excessive current through the circuit as power is applied gradually.
Might help a little bit to reforms caps in a linear power supply. But if there is a hard short in a linear supply it’s useless too. It won’t help you with a switching power supply - the transistors are used as temporary shorts. And if it’s not temporary, they blow themselves up. Nature of the beast. Might even hurt: you could get a good switcher in a stuck undervoltage state and destroy it.
Lol, watching this makes me feel so much better about all the components I've blown during fault finding.
My theory on the odd buzzing is the transistors have lost their high frequency response. Hence the switching isn't as clean.
You missed the opportunity for a "walk in the PARC" joke! Good job though, I'm curious what the cost of replacing the PSUs with similar, modern ones would be!
It's not a question of cost obviously. You don't want to destroy the historical authenticity of the machine by blindly sticking new generic parts in it, even though it would be way easier. Plus these are are quite noble, early switching power supplies. Priceless? It's a fine balance sometimes trying to preserve the functional as well as the historical - we try very hard to do both. That said, I was pretty close to giving up on that one...
A dedication to authenticity I can support. If they fail again with something you can't replace, would you consider modern (probably higher efficiency) supplies?
For a machine as rare and historically significant as the Alto, and belonging to a museum, unlikely, unless there is no other practical solution.
For a more widely available machine, that I own, and with a generic consumer grade power supply, possibly. I'd still try to repair the original first!
here's a man who needs to be introduced to the dim bulb current limiter!!!
how do you just find these ancient computers like this????? on the last video i saw a comment mentioning you having a cray supercomputer???????
Strangely you can still get transistors in these cases including the slides for insulating from heatsinks.
At the end of the 1980s I had a hard time with Solitron devices darlington transistors that were connected in bridge in servo controllers a customer brought me in order to repair. I consulted my suppliers and the price quote was absurd. The solution was replace by power mosfets.
Very well done repairing that.
@02:05 How to find shorted tantalum capacitors: "Turns out -- it's still bad . . . because the output's shorted. So . . . and then we get the classic shorted tantalum, which we found by basically just powering it with several amps, and we looked for smoke."
Very nice repair. :-) Thanks for posting there videos.
No variac with a dim bulb ti put in front of this stuff before you power it on?
I think it has been run with very bad caps for a long time. All kind of horrible stuff might happen to PSUs when their ESR is too high.
I wasn't expecting switching supplies in a computer that old...
I worked for a guy in the mid 70's that started his own company, designing and constructing microcontroller-based systems. He was especially proud of the switching power supply module that he sweated and slaved over for long hours to get designed properly. Last I checked he was still in business.
@@donmoore7785 what’s the name of his company
The fuse problem is the worst... Many times I have rolled across the shop to steal fuses from other equipment to diagnose what's on the bench. If only someone made a resetable ckt breaker for those push and turn holders...
Why use a crappy Chinesium DMM when working with precious vintage parts (@18:00)? Why don't you use a Fluke and be sure of the readings ?
Because it's the one I have with a big screen that shows well on camera. And boy was it cheap. A poor excuse obviously. Let's use the 8-digit HP DMM next time. Or alternatively I could do I Chinesium DMM repair video ;-)
I would do a Chinesium DMM destroy video instead... It would probably help dealing with the frustrations and pressure of fixing a stubborn PSU, by destroying something else for a change ! lol !
Anyway, you have waaaay more patience than me trying to guess your way in reverse engineering old technology. Great job !
Your videos should be called "Battles with thermal migration"
I love all the AVE References!
23:47 Are those disk capacitors? They look like MOVs to me.
Great video! Thanks for posting.
Thermistors, for limiting inrush current.
What's that tool you used to trace the noise on the board?
It’s just a condenser lavaliere microphone hooked up to my camera.
THat's one hell of a part graveyard.
Yikes! I wish I had this much patience haha
Ask Mr Jones at the EEVblog...
The board is awesome
Can't you use 2N3055 or MJE3055? general purpose power transistors, can you use a power supply out of an IBM compatible?
NO WAY! As he mentioned a few times, at least 400 V capable transistors are needed. 3055 is rated 60 V. Moreover, a 3055 is SLOW. For the switching, you need FAST or even faster! No deal...
BUX80 would likely work, and is easily obtainable.
Looks "fun" :) but maybe next time send it over to Mr. Carlson's Lab. He does a lot of PSU work and sensitive restorations.
Personally, I find the MCL videos to be somewhat drawn out...
If he measures a resistor and finds that it's drifted high, the following 30 mins of video explain the replacement of that resistor.
That's fine if you're working for the government, but in the REAL world, it'd be replaced in 30 seconds rather than 30 minutes
A way to improve the directionality of a microphone (or your ear) is to put a long tube in front of it.
wow. i could Never stay so calm 😣😂
What is the name of the tool you use at 14:00 to find the noise source?
Nm, looks like your mic lol
unbelievable!! very very cool
Fascinating!
Stupid question: why spend so much time to fix what looks like a set of stock power supplies?
It is not "restoration project", if you throw all innards away and emulate them with Rasberry Pi and LCD-screen. From user point of view the result would the same (or better), but CuriousMarc wants to suffer.
And we all want to suffer along with Marc on this. It's so much more fun that way! :D
I think the point was more of "These are just off the shelf regulated PSUs so why not grab some new ones".
They aren't Alto specific and even for testing reasons I'd keep a newer set of things on hand, just in case one like this decides to go again while trying to get the rest of the system up and running. It's not entirely out of the picture that a PSU with this many failures so far will have more down the line.
See now that's the Fun part. Can't spell defunct & disfunctional without fun. :D
+Laurent For the challenge, and for preserving the integrity of the historical original of course. But Carl and Ken were questioning my sanity just like you!
600$ for 2 transistors -> #mademyday! God safe this kind of customers!
I know, trying to keep it historical correct, but spending 600$ for stupid transistors properly nobody else see's the next 100 years is really mind "blowing". ;-P
Tune in next time for more Alto Boys!
i think the fan spins backwards
-----> 2:43
Marc! Don't touch it like this! Can we somehow donate towards your thermal camera budget or something?
I like to pat my transistors :-)
23:26 hits the high note. :)
Wow, you blew a pair of $300 transistors???
Oops.
They're only $300 nowadays because they're no longer manufactured. So, technically, he didn't blow a pair of $300 transistors - but that's the price for replacement original parts. I think so long as the replacement parts perform the same as, or better than (as is often the case), the originals and they are a good physical match then it's completely acceptable to use them. After all, many of the modern replacement capacitors are smaller than the original ones. Surely the main aim is to have a fully functional piece of kit, rather than a faulty, albeit totally original one!
AFAIK, if his second set of substitutes didn't work, he'd probably be forced to pay the current-going-price, which just over $300/pop.
No way I'd ever pay that much to these crooks! I'd put in an equivalent. Which is what I just did.
I would, for safety reasons, replace all them PSU's with a regular ATX power supply, everything is there except for the 15v lines in which I would either use a another supply or a stepup.
Shoutout to AVE?
Isn't more practical to switch the entire power supply for something not from hell?
Yes, but that would be a repair, not a restoration...
Jesus 60amps
Stupid question: why spend so much time to fix what looks like a set of stock power supplies?
Because they can >:D