Recommendations for 'refurbishing' antique electronics.

[BenKenobi]

I have a number of what could be described as antique computer hardware, pretty much none of it has been powered up in 20+ years, it was all working fine when stored.

I want to wake it all up but don't want to kill it in the process. I'll give it the visual once over looking for any old style caps etc, other than that exposing them to a gradually increasing AC voltage is the plan, not sure over what duration or what steps though - power up at low AC and watch the current I guess.

Any other recommendations ?

I'm sure there are some custom packages in there that cannot be replaced - or sourced so if they ain't dead I don't want to kill them, I have no schematics or service data for them. A couple of them use small CRT's. (Compaq Portable IIe), Compaq SLT286, one has a plasma screen (Toshiba 3200SX), I know they aren't worth much but that's not the reason for waking them up. I've used each and every one in a professional capacity - not ones like these but these very machines, they cost an absolute fortune when new.

[pelule]

Befor powering them up:
# if they have not stored at a dry environment, I recommend to keep them some days in a dry room, to ensure no humidity is left.
# do a visible check of damages in the electronics, mainly the electrolytic capacitors.
# if you have an adjustable AC source (variac), start up the gear with lowest specified operating voltage and observe the current. Keep it at that voltage for ~1 hour to warm up the gear before increasing the voltage.
BR
/PeLuLe

[lordvader88]

What era capacitors and types are in it ? Some electrolytic capacitors age even worse, when not used for periods of time, and 20yrs is a long time.

And as u probably know and to newbs, electro caps wear out on the order of years I guess, 5-10-20? depending on the tech and usage. So they can start acting like a big resistor passing current, or worse a wire and go zap

[edpalmer42]

The problem with using a Variac on this type of equipment is that switching supplies don't take kindly to low voltage.  Even starting at the lowest rated voltage is problematic because it causes highest currents in the primary circuits.

The first thing I would suggest is, if possible, disconnect the power supply outputs from the main board(s) and attach resistive loads with a value of maybe 30% of the rated output to the supply.  Older switching supplies sometimes needed a load to work right.

The second thing I'd do is use an incandescent bulb in series with the unit under test to limit the mains current.  Size the bulb so that the bulb's rated current is somewhat more than the expected drain.  It might be difficult to set up bulbs that meet this criteria, so just do the best you can.  There's nothing to say that you can't use multiple bulbs in parallel.

If the main caps use screw terminals, forget all the above and just remove the caps and test and/or reform them seperately.

Of course, this only tests the power supplies.  The rest of the circuits can't be seperated, so about all you can do there is cross your fingers.  If you have an ESR meter, you can give that a try, but in-circuit measurements may or may not be reliable depending on your meter.

Ed

[bitseeker]

Great suggestions, Ed. In addition, for bringing up the motherboard, if you have some lab power supplies with sufficient output capability, you can make use of their current limiters to detect potential problems without letting all the smoke out. I'd power up the motherboard alone, and if OK, test with different expansions. Of course, this all depends on how much power your supplies can deliver.

If the only way you have to power up the computer is with its own power supply, after verifying that the supply works fine based on Ed's suggestions, then I'd also start with just powering the motherboard before adding the load of lots of other components/peripherals, etc.

Note that it's not unusual for old dipped tantalum caps to spontaneously pop.

[BenKenobi]

I'm trying to source a decent cap tester, don't do this for a living so didn't want to spend mega bucks - but I also need a decent through hole desoldering tool - can caps etc really be verified in circuit successfully - i.e. leakage - I currently desolder them then use a multimeter and bench PSU but only have a manual pump and braid so it takes forever - most of the stuff in these computers is older through hole type.

Am I correct in the belief that capacitors degrade if they don't see power for extended periods or is this an unwarranted fear - aka power on and be damned.

I wish I knew what load to add - I have no schematics or supporting docs so it would all be guesswork based on component ratings - I guess I could inject 5 / 12 / 16 VDC or whatever is expected assuming that I can work it out - my biggest fear is the PSU dying and taking something out on the way really so disconnecting the boards is a good idea.

I have enough bench PSU capacity I think - 10A at 42V DC and I can parallel / track those.

The Caps range from 1982 on, the newest is 20 years old ...

[bitseeker]

I'm trying to source a decent cap tester, don't do this for a living so didn't want to spend mega bucks -

It depends what aspects of the caps you're wanting to test. For general pass/fail ESR verification, the so-called $20 transistor tester (some models are much cheaper nowadays) works fine.

For leakage testing, your current method with a bench power supply and DMM is fine, provided you can achieve the working voltages of the caps being tested.

but I also need a decent through hole desoldering tool

Lots of threads about these. I picked up a used, portable Pace Microportable MP1 some time ago for less than $100. Works great. Availability of particular brands/models in your area could be significantly different, of course.

- can caps etc really be verified in circuit successfully - i.e. leakage

It depends on the circuit. For example, if you don't notice that the cap under test is in parallel with several others, you won't know which one is leaking when it's fully in-circuit.

- I currently desolder them then use a multimeter and bench PSU but only have a manual pump and braid so it takes forever - most of the stuff in these computers is older through hole type.

Just desolder one side and test. No need to completely remove them. Also, although not a substitute for leakage testing, a preliminary check of the cap's ESR will give an indication if it's suspect.

Am I correct in the belief that capacitors degrade if they don't see power for extended periods or is this an unwarranted fear - aka power on and be damned.

It depends on the type of capacitor and what's in it. Aluminum electrolytics rely on a thin oxidized layer inside. Over time, this can degrade and may be able to be restored by slowly reforming the capacitor. However, given how inexpensive they are these days, it's better to change them if they leak DC. Also, electronics from the era of the Capacitor Plague will be at risk of damage from leaking electrolytic caps, especially if they've just been sitting without a charge.

I wish I knew what load to add - I have no schematics or supporting docs so it would all be guesswork based on component ratings - I guess I could inject 5 / 12 / 16 VDC or whatever is expected assuming that I can work it out - my biggest fear is the PSU dying and taking something out on the way really so disconnecting the boards is a good idea.

You can start with small loads. If there isn't enough for the power supply to start, it'll just sit there. Note that many computer power supplies expect to see a power good signal (i.e., a particular voltage level on that pin) before they'll start.

For the voltages, the power supply should have a label as to what they output. Once you've successfully gotten the supply running, you can measure what voltage comes out of which pin and replicate them with your lab power supplies.

The Caps range from 1982 on, the newest is 20 years old ...

Really old caps can be fine for many decades, believe it or not. I have some power supplies from the 60's that are just fine. It really depends what kind they are, when they were made, and how they were used. For example, electrolytic capacitors in switching power supplies, beside heatsinks, and in other high-temperature or high-ripple current environments have much shorter lifespans.

[BenKenobi]

 

Sounds like a plan.

There are numerous desoldering stations but very few that seem to be ESD safe, I've been looking for the one that Dave tested but haven't found it in the UK yet - I've found ones that 'look' the same but E-Bay can be a bit of a lottery.

[edpalmer42]

Great suggestions, Ed. In addition, for bringing up the motherboard, if you have some lab power supplies with sufficient output capability, you can make use of their current limiters to detect potential problems without letting all the smoke out. I'd power up the motherboard alone, and if OK, test with different expansions. Of course, this all depends on how much power your supplies can deliver.

The thing I worry about with external power supplies is whether or not the circuit will be sensitive to power supply sequencing.  I don't know how often you might run across this.

Note that it's not unusual for old dipped tantalum caps to spontaneously pop.

Oh, yes!  I bought a piece of equipment cheap because the magic smoke escaped.  I saw a charred tantalum cap on the +15V supply and replaced it.  The tantalum on the -15V supply promptly smoked.  Changed that one and the unit was fine.

Ed

[BenKenobi]

I'm hoping that the power supplies aren't that smart.

Not even started on them yet just setting out the stall for some winter projects.

[bitseeker]

Yeah, hopefully the old stuff will be fairly straightforward. Oh, and regarding the tantalum caps, wear safety glasses as I've seen the outer shell fly quite high into the air when they pop.

Whenever you dive in, I'll be looking forward to seeing the old gear and hearing about your progress.