• Please review our updated Terms and Rules here

Before I dare apply power...

JonB

JonB

Veteran Member
Joined
Jan 26, 2014
Messages
1,652
Location
South Herefordshire, UK
I have an S-100 based machine which I probably mentioned before - it's the one with the Telek FDD-1 CPU card.

I'm getting to the point where I am thinking of firing it up.

But - what should I check (after the PSU)? Recommissioning an old S-100 machine is something to be done carefully, I suspect. Especially given the lack of success I've had with other computer equipment from the same source.
 
If it's a linear power supply, you probably want to either bring it up slowly on a variac or remove the bulk filter capacitors and reform them on a bench supply with a current limiting resistor. I've had very good success doing this. There's guides elsewhere on the process, I believe several folks have written up their experiences here on the forums.

After that, the big thing is shorted tantalum capacitors. When they go, especially in a system with a big power supply, they can really go. I've gotten boards where it was clear that a tantalum cap on the board *next* to it in the card cage let go, and shot out a jet of flame which smoked/scorched the back of the board. If there's a hard enough short, you can blow traces off of the circuit board. I deal with this by bringing up S-100 boards on my current limited bench power supply. Watch for excessive current draw (which can be a little tricky to determine sometimes, since S-100 boards can draw a lot). Carefully feel tantalum caps to make sure they're not getting hot. Check the voltage after the regulators to make sure the regulators aren't dead.

If you don't want to deal with possible tantalum cap problems and/or don't have the means to test for them, you can of course just replace them all ahead of time.
 
Can I use the actual PSU to reform the capacitors? There are three of them, very large. I can imagine putting a resistor in series and measuring the voltage drop across it. The PSU comprises a fat transformer and a couple of fat diodes (1/2 wave rectification,?).

As I have no bench supply or variac, this is the only way I can do it.

See the diagram here: https://www.qsl.net/g3oou/reform.html
 
Yes you can use the supply but first make sure it doesn't have a shorted diode or bridge. I like to keep the forming current to less than 100ua although these are likely computer grade capacitors and designed for about 10ma leakage. Figure the voltage you are going to use as the source and calculate 100ua. Run that for a day or so and move up to 1ma if it looks good.
Check the capacitors for blown seals on the ones you remove. If you limit the current, you shouldn't need the variac.
Dwight
 
As for tantalums, I recommend putting one board in at a time. Used a 1/4 amp fuse on the AC side. Run them for at least an hour. shut the power off emove it and plug the next one in. Repat. Check the voltages before and after. That is about the best one can do.
Dwight
 
I built one of those 60w bulb - in - parallel devices to plug my machine into.

I have found the folk at google group S100Computers to be most helpful. Very active and responsive. They are manufacturing all sorts of S100 / Arduino hybrid boards to name but one example.
 
I had exactly the same concern when it came to powering my SOL-20 for the first time.

I decided (after doing the things below) it was better to power up the supply, with a power resistor dummy load, in isolation, not connected to the motherboard or any S-100 cards initially.

With regard to my linear supply it had some usual issues, with loose screws that retained power devices to the heat sinks and degraded white thermal coupling grease that had dried and broken up and turned to a patchy white dust. I checked the small electrolytics for value ESR and leakage, they were ok. The large main large filter electrolytics seen in these old computers often maintains its capacity unless any physical sign of leakage is evident. Probably because of its large volume.

Once I was happy with the physical condition of the supply and its components I wired up some ceramic power resistors to a plug fitting and plugged the supply onto that to test it. It only requires a modest load for testing around 1A for the 5V supply and 0.5A or less is ok for any higher voltage outputs. The resistors need to have an appropriate power rating and by placed on a heat resistant surface. If the supply was in fact ok on that light load, but not at full load, you would find out later and it wouldn't damage the computer boards if its output sagged down. The main thing is that the supply is checked not have have an output voltage that is too high, due to regulator/ pass transistor failure etc.

There would be nothing wrong with bringing the voltage for the supply slowly up with a variac either, but be aware, if you do that with the whole computer, some of the cards/circuits use an R-C coupling system to generate initial reset pulses, that circuit won't work if the supply voltage rises very slowly.

Once the supply is serviced, checked and working, I would go with the notion of plugging in boards one by one (with the power off initially of course) and checking for overheating/shorted Tant capacitors, and checking that the voltage regulators on each board have a normal output voltage. With my whole computer and testing 8 or more S-100 boars I only struck one shorted Tant and none on the mobo, I may have been lucky there.
 
....I could have added that although many vintage computer linear supplies have an SCR-crowbar circuit to help protect the scenario of the output going over-voltage, you can never be 100% sure it is working, or going to work (unless you deliberately modify a value in the regulator control circuit to get it to trip) so it is always worth checking the supply under test into a resistor dummy load first, just on the off chance the output voltage was too high and the crowbar circuit had failed. Better that, than applying high voltage to a computer pcb or logic board. Also, if you have the data on the supply's max rated output current on each of its outputs, you can set up dummy loads to run it at full output power for a while to soak test it and be sure its definitely ok in all respects prior to connecting it to the computer.

I don't believe in electrolytic capacitor re-forming. My view is if the capacitor requires that to recover its function, or any other form of conditioning therapy, due to age related changes, it is not fit for use, it should be replaced. Though, that is a whole other topic to argue over and I don't want to pick a fight with re-forming believers. None of the electrolytic caps in my SOL-20 had any re-forming or other therapy. If I had detected any anomaly with their uF value, ESR, or electrical or physical leakage, I would have replaced them Stat.
 
Last edited:
I agree that a cap that needs to be replaced, should be replaced. A slow bring-up or external reforming will tell you if it's any good though, and avoid popping the vent plug and spraying electrolyte everywhere (repaired a few machines like that!)

For dummy loads, I use automotive tail lamp bulbs. I've got a bunch of 6V bulbs for my old Willys Jeep, and modern 12V bulbs for everything else. A combination of the two will get you close enough to load down most supplies you might encounter. Especially important on switchers, which may not start at all with no load.
 
The interesting thing about electrolytic caps, is that they have a number of failure modes. When I replace electrolytics, I always open them up for a forensic inspection, I have been doing this for many years. I don't just throw them in the bin.The most common is drying out as water leaves the casing despite the seals over decades: have a look at the PM5519 that "ran out of water":

https://www.vintage-radio.net/forum/showthread.php?t=140044

Another mechanism is insulation failure and foil to foil shorts less common.

Another is corrosion of the internal links/connections , there is a photo of this on page 2 of this article.

http://worldphaco.com/uploads/ELECT..._IN_TEKTRONIX_TM_SERIES_POWER_MAIN_FRAMES.pdf

The real problem is there can be some fairly horrible things happening inside your vintage electrolytic capacitor, even with reasonable test results. So for old electro's now I adopt a policy that if any anomaly at all is detected on testing, or any physical leakage, or old surface mount types, they should be replaced, and if they cannot stand an immediate power up, without exploding their seals, they should also be replaced, not re-formed. That is just the way I have decided to do it after examining the internals of multiple failed electrolyics.
 
You're certainly entitled to your beliefs and choices and I don't want to start that argument either, but I think it's worth pointing out that the capacitor manufacturers themselves recommend re-forming after an electrolytic cap has been stored for a long time. Increased leakage and some thinning of the oxide layer over time are quite normal, especially at elevated temperatures, but this is a reversible process, not unlike recharging a rechargeable battery.

Considering how relatively scarce the large caps used in linear supplies have become these days, a newly purchased one may even also have been sitting on the shelf for a long time and be in the same condition as the one being replaced.

Suddenly applying full voltage to an electrolytic cap that's been stored for a long time may destroy a perfectly good capacitor or possibly even damage the equipment.

See P. 2-6:
http://www.nichicon.co.jp/english/products/pdf/aluminum.pdf

Also P.7:
http://www.cde.com/resources/catalogs/AEappGUIDE.pdf
 
MikeS said:
Considering how relatively scarce the large caps used in linear supplies have become these days, a newly purchased one may even also have been sitting on the shelf for a long time and be in the same condition as the one being replaced.

Suddenly applying full voltage to an electrolytic cap that's been stored for a long time may destroy a perfectly good capacitor or possibly even damage the equipment.
I agree with that in that those large linear power supply capacitors are getting rarer and should be preserved if possible. I agree they are special items. Low Z high current output supplies now, have much smaller value filter caps and the low Z output is effectively electronically synthesized and the capacitor recharge rate in the SMPS is much faster than the 8.3 mSec from full wave rectified stepped down line power.

Those big caps in computers and many appliances are a thing of the past and a valuable historical item. So it is worth it on those vintage linear computer supplies winding the supply up slowly if possible with a variac or lamp limiter at least for the first time initially if it has been in storage. Those capacitors are so large in value though and have a very low ESR, most supplies they are connected across have difficulty rapidly generating the full potential across their terminals and its more of a ramp at normal power up.

I regard common garden smaller electros as disposable items though, many are only rated by the manufacturer for a few thousand hours use. And with poor design practices, putting them close to heatsinks, their life is even shorter. So apart from winding up the power initially slowly over a some minutes for an old appliance, I don't do reforming. If any of the electros cannot handle that I replace them, if I have not already done that.
Click to expand...
Click to expand...
 
Hugo Holden said:
I regard common garden smaller electros as disposable items though, many are only rated by the manufacturer for a few thousand hours use. And with poor design practices, putting them close to heatsinks, their life is even shorter. So apart from winding up the power initially slowly over a some minutes for an old appliance, I don't do reforming. If any of the electros cannot handle that I replace them, if I have not already done that.
Click to expand...

Absolutely, the little guys are jelly bean parts even nowadays, and modern aluminum electrolytic caps are vastly superior to the old stuff.

The big bulk filter caps are what I was referring to w.r.t. reforming. I think we're all on the same page with that.
 
glitch said:
Absolutely, the little guys are jelly bean parts even nowadays, and modern aluminum electrolytic caps are vastly superior to the old stuff.

The big bulk filter caps are what I was referring to w.r.t. reforming. I think we're all on the same page with that.
Click to expand...

What capacitance value separates these so-called '"jelly bean parts" from the "big bulk" ones?
 
exidyboy said:
What capacitance value separates these so-called '"jelly bean parts" from the "big bulk" ones?
Click to expand...

Good question. The "big ones" we were talking about in vintage computer linear supplies are around 50,000uF with screw terminals. Small capacitors would generally be in the range of 4,700 uF or less with radial wire lead connections and pcb mounting. A 10,000 uF with screw connections might also be worth trying to save. Probably any large physical sized vintage electrolytic with a clamp and screw connections is worth evaluating/testing (perhaps re- forming if you are into that) and preserving if possible. Many of these also have interesting labels and markings and when new probably were a moderately expensive part.
 
I have attached a photo of the type of a big capacitor we are talking about, this is the one from my SOL-20, it is quite an impressive beast which is about 4 inches tall and about 2.5 inches diameter. It is quite the lovely part. The ones I have seen in other S-100 computer photos look similar. It has a great logo.

I bought a spare similar one for a good price, but it seems they sell well. These are not common parts these days and represent a part from a lost age of high current linear supplies for computers, that may all but have disappeared by the mid 1980's when switch-mode power supplies came to dominate the world.

https://www.ebay.com/itm/50-000uf-2...742169?hash=item41b7930e19:g:r7cAAOSwXeJYJiK-
 

Attachments

  • P6200006.jpg
    P6200006.jpg
    88.8 KB · Views: 1
One of the problems with the very old ones is that they simply dry out mostly from internal heating.. Reforming will get you nowhere.

I've had decent success with Sprague Computerlytics, but I've yet to find an unaffected CDE "Long Life" cap.
 
Chuck(G) said:
One of the problems with the very old ones is that they simply dry out mostly from internal heating.. Reforming will get you nowhere.

I've had decent success with Sprague Computerlytics, but I've yet to find an unaffected CDE "Long Life" cap.
Click to expand...

The Lambda "long life" capacitors in power supplies are all junk as well. Why would someone say long life and then have them fail left and right.
Dwight
 
You must log in or register to reply here.
Back
Top