Surprise in lab supply - all electrolytics ok, even with cracked gaskets

[dom0]

So while cleaning up an older lab supply (fully discrete, hand-matched and glued together pairs for error amps - only datecode on power transistors, 1970!) I encountered a whole bunch of Roederstein electrolytics, all of which had somewhat swollen and cracked gaskets. These were Röderstein EB. Capacitance almost spot on, basically 5 % tolerance throughout (these are -10+50 % parts). Very good ESRs as well: 10 µF / 100 V with around 1.5 Ohms, 10 µF / 35 V with .8 Ohms, some 100 µF / 25 V showed .4 Ohms. The replacement parts were no better at all. I only checked the axial ones after replacing them, because I didn't expect caps with visually broken seals to be in perfect working order. I assume the seal is actually tight and the cracks are just superficial.

Also some Röderstein EY 2200 µF / 35 V capacitors. Both(!) had 2400 µF and <.01 Ohms ESR.

I've had some very good luck before with old electrolytics. But an instrument where all(!) ~50-odd year old capacitors were in perfect order? That's a first.

I guess the manufacturer made true on their advertisement: "Simple, proven design, high quality component for long life". Quality indeed. Only Röderstein electrolytics, Philips and TI transistors. Potted transformer with double shield winding; core is a premium cut, not the economy EI/UI stuff... not welded of course. Components Made in Germany/Englang/Netherlands as far as eye can see. Unthinkable today.

[Gyro]

Yep, capacitor quality varies hugely these days... However back in the day, most specs were conservative and they were built to last.

I encountered the same thing on one of my meters, '70s vintage axial Wima electrolytics (back when they actually made electrolytics!). I purchased some decent brand long life replacements, only to find that the ripple voltage was lower with the originals, so I swapped them back into service (easier to do with axials) no bulges or leaks after 50 years.

[rsjsouza]

Similar thing here... I couldn't find fault on the tank capacitor of my Power Designs 5015A, which was last calibrated 50 years ago. The cap is a GE Aluminium computer grade capacitor.

I think not only the parts were quite over spec'd at the time, but the materials and construction technology had reached a sufficient level of maturity to grant stable operation (and absence of leaks) over the lifetime of the part.

[dom0]

The shiny Sprague axials are also often very reliable. However, so far I've usually had at least one or two per device that had issues (high ESR or low capacitance... often both). The similarly looking radial Spragues ... I've seen many of these dead. In Tektronix 7xxx series supplies these were used en-masse and usually a large portion of those have failed by now.

The blue axial and radial Philips caps from the mid 70s (ca.) - mid 90s have a bad reputation with many it seems. In my observation, the blue axial ones are quite reliable. The radial ones, again, not so much. (A pattern emerges?)

--

This thread: https://www.eevblog.com/forum/projects/electrolytics-with-organic-(non-water)-electrolytes/msg2770316/#msg2770316 is almost a companion in a way to this one. Asking the question, in essence, what would you have to buy today to get the same longevity. The Röderstein EB axial capacitors from post #1 are 200k hours @ Ir @ 40 °C. That's 22 years of continuous operation. Probably more, because for most of these they are just blocking capacitors typically, so they never reallly see that kind of ripple and internal heating.

[rsjsouza]

The blue axial and radial Philips caps from the mid 70s (ca.) - mid 90s have a bad reputation with many it seems. In my observation, the blue axial ones are quite reliable. The radial ones, again, not so much. (A pattern emerges?)

Interesting your mention of these. In another thread I mentioned that I still have a huge number of blue radial 1990s Philips and orange axial Siemens made in Brazil that are still quite in spec (tan δ around 0.02) - sure, about one third of them showed signs of leakage, but after 30 years I could not say they are bad. I also still have a 1980 Siemens beast (75000µF/25V) that started to show signs of age only a couple of years ago (it can be seen on the picture at this post).

[wraper]

Yep, capacitor quality varies hugely these days... However back in the day, most specs were conservative and they were built to last.

Sky was bluer and grass was greener. In reality capacitor quality varied a lot, many did not last. Also in older equipment in general they were stressed way less.

[madires]

Also some Röderstein EY 2200 µF / 35 V capacitors. Both(!) had 2400 µF and <.01 Ohms ESR.

I have a bunch of old ROEs mainly used in power supplies for audio amps and can confirm that they are still in good condition.

[madires]

The blue axial and radial Philips caps from the mid 70s (ca.) - mid 90s have a bad reputation with many it seems. In my observation, the blue axial ones are quite reliable. The radial ones, again, not so much. (A pattern emerges?)

The blue axial electrolytics aren't any better. They don't leak but capacitance has decreased and ESR increased considerably. If you see any of the old blue Philips electrolytics just replace them.

[Gyro]

I suspect that axials tend to last longer because...

- Only one wire seal, which was visibly more substantial. The negative lead is spot welded the the case.
- Much lower heat stress on the seal during soldering - the long leads to the board would limit the heat relative to radial ones, where the seals are right up against the solder joint.
- Lower mechanical stress deformation of the seals, than radials both on insertion/soldering/board stresses vibration and accidental knocks to the top.
- Lower packing density than radials leading to better cooling?

Just guesses, but I suspect that they all play(ed) their part.

[SeanB]

Many of the older axial capacitors also have a secondary seal under the rubber, in the form of a SRBP disc that is sealed in to provide mechanical support for the main seal, so it also holds the electrolyte from leaking out. As well the electrolyte in general was a thickish paste, with a good reserve of it in there as well, simply because the volume required to not damage the bottom tab as the inner assembly was loaded into the case during construction, which requires this long lead to be folded without damage, so there is a dead volume filled with extra electrolyte. Yes an old electrolytic can still be close to spec ESR and capacitance wise, and, with careful reforming, also have similar performance leakage wise, if not being operated 24/7/365, which kept the oxide layer intact for the working voltage.

But you do not know just how much electrolyte is left to keep reforming defects as they occur, and thus how long the capacitor will carry on operating. You could have another decade, or it could go next day.