Power Supply Recap, electrolytics vs polymer

[Laval]

Hello everyone.

I'm about to order the capacitors to recap the power supply of a Tektronix 2792 spectrum Analyzer and I was wondering: would it be a good idea, whenever possible (values and form factor) to replace electrolytics by organic polymer capacitors ? This is a switch mode power supply and the secondary part has 23 electrolytic capacitors.

[adinsen]

I had to google, but according to this resource, it will be a good idea: https://www.digikey.com/en/maker/blogs/aluminum-organic-polymer-capacitors

I guess it depends on your needs and budget. Often its possible to go up in voltage and/or temperature rating on ordinary electrolytic capacitors. I find that the typical 5000 hours of guaranteed lifetime of most caps is plenty enough when I think about what kind of use I'll be giving whatever unit I'm working on. I have just worked on a 46 years old PSU, and I haven't had to replace a single cap, not even the big filter caps 

[fmashockie]

I personally think the whole replacing every electrolytic cap on every piece of electronics you come in contact with is completely overblown and overrated.  I think people got scared during the 'capacitor plague' (here's the wiki link https://en.wikipedia.org/wiki/Capacitor_plague) where bad caps were being manufacturered accidentally between 1997-2007.  These caps would fail prematurely and were even from brands known for good quality. 

Unless you are working on an old radio that's pushing 75-100 yrs old (it this case you might not have a choice because the caps in there were made of wax and paper and likely have already gone bad), I don't understand why people stress replacing every single cap.  It is a waste. 

[Laval]

That has been on my mind for some time now. When I recapped my Agilent 53131A I measured all the replaced capacitor and only one had a high ESR that justifies the replacement. Of course, it's not a 30 years old device and to really measure them, you need to remove them. I think I'll only change the one that is bulged and the Rifas because they started clacking. The cracks don't go through the epoxy yet but they probably won't last very long.

[fmashockie]

There are some exceptions of course - like the RIFAs.  Those I do replace regardless 

[wraper]

I see no advantage over using high quality aluminum electrolytic capacitors is a device like this. Also there are potential problems to be had.

[Laval]

I see no advantage over using high quality aluminum electrolytic capacitors is a device like this. Also there are potential problems to be had.

I'm curious to know what potential problems could there be. They have better ESR as well as higher temperature rating. Capcitty is the same and voltage can be the same or higher. I'm just curious.

[SeanB]

Plus will add the scopes normally used high voltages, much more than polymer is rated for. Just use good brand aluminium electrolytics, rated for 10k hours at 105C, and when buying them get same height or similar, same diameter or slightly smaller, and higher voltage than old ones. Generally higher voltage is bigger can, which works well, and in most cases you will only get radial types, and then sleeve the one lead, and extend it around the can, to make it roughly axial, with a short wire soldered to it as extension. Stick down to the board after cleaning, using mirror silicone, as that is a very cheap non acetoxy silicone, and thus does not harm copper, but is good at holding to PCB and plastic parts.

The originals normally are, to a great extent, 85C rated, regular ESR, so low ESR ones might run rectifier diodes a little hotter, and make the ripple current in the transformer windings higher. But really not a worry, and if critical, say high current applications, where the original diodes ran hot enough to discolour the board, the diodes might need to be upgraded as well, to more robust modern ones.But that is really on a case by case basis, most of the time the board traces, and the transformer windings, provide enough resistance anyway.

[wraper]

Lower ESR is not necessarily better and may actually cause malfunction if circuit was not designed accordingly. They have higher leakage current. Low resistance to humidity, may cause short circuit failure eventually.

[Laval]

I see, thank you for replies, I'll just use high quality Electrolytic caps with better temp/voltage rating.

[p.larner]

change nothing unless they test bad or are leaking,unless you want to keep fixing it until its broken!.

[inse]

Lower ESR is not necessarily better and may actually cause malfunction if circuit was not designed accordingly. They have higher leakage current. Low resistance to humidity, may cause short circuit failure eventually.

An example from my own experience with low ESR: I put a ceramic capacitor (uF range) on the output of a linear regulator (not 7805 type) and had to notice that the output voltage was way off.
Upon further investigation I found out that the regulator was oscillating, the datasheet specified a minimal ESR for the output filter cap.
So I swapped it for a small electrolytic and everything was fine.

[T3sl4co1l]

Older SMPS often depended on ESR for a compensation zero in the control loop, and using polymers likely results in singing, instability, incorrect voltages (peak or average), or outright destruction.

Likewise linear regs, the ESR matters for the control loop.  Most older ones are more tolerant (emitter follower types ala LM317, 7805, etc.), but negative and LDO types (notice LM337 and 79xx are open-collector output) were designed around what was available -- medium to high ESR electrolytics, and tantalum.

Usually when they needed a low-ish stable ESR value they specified tantalum.  Nowadays, polymer can be selected for such (while most are low ESR, there is a range of values available), or a resistor added explicitly (especially for ceramic).

Tim