This answer [42] is a good treatise -> https://retrocomputing.stackexchange.com/ques … rst-time-in-yea

Thanks! Will read that.

Consult your electrician.
You should check what happens when a RCD triggers.
Perhaps you sit in the dark then.
To add, use a separate AC circuit with RCD/FI protection and perhaps a smaller and quicker FUSE like B6 or B10 (6 or 10 A) for the devices on your testing table.

In case you do a visual inspection on the power supply, watch out for metallized paper filter capacitors. Many of them are contained in plastic cases, and if that case breaks (which it does on its own as the plasticizer evaporates...), moisture might get inside them. Connecting moist capacitor to line voltage will usually make them fail violently. These capacitors are common in equipiment produced up to the late 1980s.

A common brand selling this type of capacitors back in the 80s was the Swedish company RIFA, which got their name associated with the issue of those caps failing. As these caps are usually used only for interference suppression, blowing a RIFA will most of the time just blow the fuse (internal device fuses and/or your mains circuit fuse), and create some soot you have to clean up. Serious damage to the electric circuit only occurs in rare circumstances. https://www.epanorama.net/newepa/2023/04/14/w … ifa-capacitors/ seems to contain a good overview of the problem with these capacitors. If you decide to replace RIFA (or similar) capacitors, make sure the replacement has the required safety rating. These caps are certified to not set your house on fire when they fail, and they are used in circuit postion that legally require capacitors certified that way. The important rating is the "Class X2" safety classification at your local line voltage + safety margin. Using 250VAC Class X2 safety caps in 230V countries is at the border of being acceptable, local regulations might require 275VAC certification as minimum. Class X1, Class Y2 and Class Y1 are even safer safety classes, and are permissable substitutes.

I heard the people servicing old audio equipment refer to them as "Knallfrosch" in Germany, which approximately means firecracker. I didn't experience a RIFA capacitor go boom myself yet (just tantalum DC filtering caps), but it's a common experience that RIFA caps blow up on retro computing conventions, for example in Apple II or Apple III computers.

Thank you all for contributing to the answers to my questions.

As with all old electronic equipment from all eras, it's always the capacitors.

One thing you can do is to unplug the psu from the computer, power it on, and test its output. It might require a load to work properly, you’d have to look into that, but powering on just the psu and measuring its output can tell you a lot. You might also try leaving the psu on for a while to see if it gives out, but again you might have to give it some kind of load.

Disruptor wrote on 2023-11-30, 00:51:

Consult your electrician.

I'm a professional electronic technician for over 40 years, but never worked with material that old.
I know my way around with (micro) processors, I designed (and still do for my hobby) micro electronic circuits, but this is a whole different ballgame...

Look into capacitor reforming. TechTangents on youtube has a good video about the subject.

I'm usually starting with a visual inspection inside the device, looking for swollen caps, leaking batteries, discolored/burnt PCB laminate, physical damage, foreign objects, etc, and general cleaning. If there are fans, I'm making sure they rotate freely.

Then it's time for the capacitor reforming mentioned above, focusing on large caps exposed to abuse in the power delivery chain, like smoothing caps in the PSU, voltage regulators, etc. Failing signal conditioning caps seldom or never lead to catastrophic failure. It's best done by desoldering the caps and reforming them individually out of circuit, but guess that would be quite impractical when dealing with a large collection. It might be more feasible to simplify it by doing it with a voltage adjustable AC supply instead, that easily could be designed and built around a triac, if there's nothing available commercially.

If possible, I disconnect and start the power supply unloaded, checking its output quality, verifying voltage(s), looking for excessive ripple etc.

Once that's done, I reconnect the PSU and power up normally, but keep an eye on temperatures etc for a while, making sure there's no runaway anywhere. A thermal imaging camera is immensely helpful. Nowadays there are Android phones with that function built in if you're on a budget. I'm also verifying the PSU's output quality loaded.

I also like the idea of using an additional smaller fuse earlier in the thread. I'd get an assortment of B and Z curve circuit breakers in small sizes, say 1, 2, 4, 6 and 10A and choose among them according to the device's rated current. It might limit the damage done if there's a problem.

Also, keep in mind that there are probably as many methods of doing this as there are electronics geeks, not trying to suggest that my method is perfect in any way 😀

Oh, and be ready for cracks, pops, smoke, arcing, and even fire. Always think it through and prepare with an action plan before you power anything up. You might also want to use some PPE, I've for instance had one of those through-hole mounted fuses from Littelfuse literally explode in my face. Good thing I had glasses on. The bang gave me ringing in the ears for hours afterwards.

keenmaster486 wrote on 2023-11-30, 15:01:

As with all old electronic equipment from all eras, it's always the capacitors.

Yes, tantals seem to be endangered when they taste 12 V again.

I’m hesitant in applying the Capacitor reforming using a variac. I know it’s frequently used with analog equippement, but with ttl chips? Maybe not a good idea because I’m afraid that ttl chips switching under low voltage conditions might act analog, not switching properly and thus draw an excessive amount of current.
Not to mention how the various psu’s might react on low voltage supply…

Adriaan wrote on 2023-12-01, 22:48:

I’m hesitant in applying the Capacitor reforming using a variac.

There is no need to reform electrolytic caps in modern equipment. It can be done but there is usually little point in doing so.

Any transformer based equipment will only have low-voltage caps. Now the really old and power-hungry stuff might be using those huge 10mF and bigger capacitors (often with screw terminals) - those are expensive and difficult to find these days, it might be worth taking them out of circuit to reform and test their health. Smaller, typical electrolytic caps are easily replaced should they fail. These machines had fuses for primary, often also secondary sides and also some form of overcurrent protection.
An exception to this rule is devices with old RAM chips that needed 3 voltages to run. The power-up and down sequencing for these voltages is critical and the RAMs will die if the PSU malfunctions. The PSU will have various protection circuits for that but too many faulty caps can make it inoperational. In these cases, if the state of the PSU is not known, a full service and test (out of the machine) is a very good idea before powering it.

Switching mode PSUs will have high voltage caps but only on input side, should have fuses though. And you can't really variac those in many cases, in fact design issues coupled with prolonged undervoltage might cause damage. If you can't easily take the PSU out and test it, just run it. In most cases the issue will be with degraded/dried out secondary caps, not the primary ones anyway. Any unformed but still good caps will just reform on their own with DC applied, so during normal operation. The only issue is the first second, then the first minute, then an hour. During that time the current through the cap drops a lot, but if the reforming is not fast enough (which usually means the cap needs replacing anyway) the heat can build up and cause a vent, short, or some other type of complete failure.

And then there are AC motors (possibly directly connected to line AC) and their starter/sustain capacitors, those might be present in cooling systems, old floppy or disk drives, etc. These caps can also reform but they usually degrade slowly by loosing capacity, which will prevent the motor from running. They can short though, due to spikes and other HF noise in the line. This rarely happens to switching mode PSUs because these tend to have current limiting thermistors and RF filters that reject a lot of that.

In general it's always a question of how expensive or important given equipment is, and how much time and money you are willing to put into it. Correctly sized fuses are your friend. You might attempt a fuse replacement without investigation into the cause if you suspect the problem was unformed caps - but never bypass any fuses. You can try to limit the inrush/reform current with a lightbulb in series but it's not always clear just how big of a bulb (or several) you'd need to both limit the current and not cause too much of undervoltage. As a rule of thumb the bulb wattage should be 2x to 3x that of the device being powered.

Adriaan wrote on 2023-11-30, 00:06:

I don’t know if this has been discussed before since the search isn’t working (on my iPad…) Since a few weeks I work as a volunt […]
Show full quote

I don’t know if this has been discussed before since the search isn’t working (on my iPad…)
Since a few weeks I work as a volunteer (I’m retired…) at the Historic Study Collection of the University of Twente, the Netherlands.
We have a huge collection of old electronic gear amongst which are a great number of old computers, like a PDP11, PET, PC’s, several types of Commodores etc. and some have been switched of for over 30 years!

My question: is there a scientific valid way to power up these old devices, or just switch them on and watch for magic smoke?
I would appreciate some guidance in the matter…

Thanks in advance and for those who would like a peek in our collection, here is our website (in Dutch)
https://studieverzameling.utwente.nl/
And for those who want to visit us, you are most welcome.

Have you considered contacting the nice folks at the https://www.homecomputermuseum.nl in Helmond?

Thanks @Deunan for your elaborate answer. Very informative and helpfull!

weedeewee wrote on 2023-12-02, 18:56:

Adriaan wrote on 2023-11-30, 00:06:

I don’t know if this has been discussed before since the search isn’t working (on my iPad…) Since a few weeks I work as a volunt […]
Show full quote

I don’t know if this has been discussed before since the search isn’t working (on my iPad…)
Since a few weeks I work as a volunteer (I’m retired…) at the Historic Study Collection of the University of Twente, the Netherlands.
We have a huge collection of old electronic gear amongst which are a great number of old computers, like a PDP11, PET, PC’s, several types of Commodores etc. and some have been switched of for over 30 years!

My question: is there a scientific valid way to power up these old devices, or just switch them on and watch for magic smoke?
I would appreciate some guidance in the matter…

Thanks in advance and for those who would like a peek in our collection, here is our website (in Dutch)
https://studieverzameling.utwente.nl/
And for those who want to visit us, you are most welcome.

Have you considered contacting the nice folks at the https://www.homecomputermuseum.nl in Helmond?

I'm sure I will in the near future, thanks!

If this is going to be some sort of museum or exhibition where these machines are going to be powered, even if only from time to time, it would be best to keep a logbook for each (at least for each group). Save yourself and others the trouble of figuring out what has been tested, and when, or repaired/replaced and how. Otherwise even with the best intentions you might cause some damage just by trying to open or disassemble these for testing, when it is not necessary. If you can then do a full maintenance of everything - doesn't need to be all at once. It's a good idea to look into devices that work as well, just to be sure you are not about to have some major malfunction.

I've not worked much on old stuff like PDP-11 and such but I do know a thing or two about older machines like these. They run hot. Forced air cooling is not optional. Depending on the configuration the fans might be replaced with slower and quieter ones (for AC motors there are options like step-down autotransformers) but this will require some testing. Again a rule of thumb is you want the inside temperature to not exceed 40C.
This might seem obvious but there can be some surprises. For example some really power-hungry PCBs full of chips might actually fry themselves even with cooling fan running, if the case is open and the airflow is not directed. Never assume it's safe to run anything for more than a few minutes with open case unless you can put some temperature probes and make sure nothing gets too hot. The same goes for PSUs.

Some of these machines have ferroresonant transformers - these run hot. Like, almost burning hot. This is normal, they work at saturation point. But it's critical the resonant capacitor is tested and good, and that the AC input voltage is not too low - this is one transformer you really don't want to variac unless you know how to quickly ramp it up past the self-regulation treshold.

And something I've never dealt with but I've seen on YT - some old 8" floppy drives have electro-mechanical ejection mechanism. You can't remove the floppy if the power is cut, or that mechanism malfunctions. You'll rip the heads off if you try. In these cases the drive needs to removed for service, usually there is a trick to free the floppy without taking it all apart but this depends on the machine, needs to be figured out.

Like I wrote in my starting post, I’m very new to old/vintage electronics and am just learning how to deal with equipment of 40+ years old.
The first thing that came to mind was Vogons.org where I suspected a wealth of knowledge about old equipment and behold: I wasn’t disappointed at all!
Thank you all contributing in answering my questions and I think this was a first post with a global survey, and I’m sure I’ll be back, asking for more input and probably with more specific questions.