Re-forming electrolytic capacitors.

[Bazz4CQJ]

GJ, I've been ham-radioing since the mid-60's and I won't claim total recal on every occasion I've got involved in bringing something back in to use. I know that you work professionally in restoration, and feel sure you have dealt with far more instances of this than most hobbyists like myself. At the end of the day, we all have to set our own criteria for restoring/re-using/replacing old electrolytics allowing for the various trade offs (originality, reliability, etc) involved. That said, I'm curious about the extent to which reforming is being done based purely on an empirical approach as distinct having a good technical understanding of the why's and wherefores.

B

[G6Tanuki]

Has any electrolytic worth its name produced since the 1950s ever needed 'reforming'. Can anyone point me at any manufacturer-supplied documentation showing a need for 'reforming' of consumer-grade electrolytics??

If I even vaguely suspect a component I replace it. To my way of thinking, workshop-time taken 'reforming' caps (with no 100% guarantee of success) is better spent on replacing with current-standard components anything that may be suspect.

Components are cheap. In-the-field equipment failure (and the reputational-damage involved) can be expensive.
[This *really* bites hard in the case of 1970s-era wet Tantalum capacitors: a competitor nearly went bust because they were reissuing 25-year-old 'serviced and warranted' gear to the MoD with these things in them]

[GrimJosef]

That is a good question Bazz.

I'm a physicist by education and an engineer by experience. The engineer in me has asked quite a few other people how they go about reforming and has read more than a few threads about it on here. I've come away with the clear view that, within reason, the more gently the reforming is carried out, the greater the 'success fraction'. Since I'm working in a field where retaining an original component can add certainly tens and sometimes a hundred or two pounds to the value of an amplifier, the extra care and time it takes to reform slowly seems a price worth paying. This is the empirical approach.

Despite what we sometimes say, physicists tend to be 'big picture, let's not worry too much about real-world gritty details' sort of people. So some back of the envelope 'physicist' figures might be

Voltage on the capacitor: 200V (sort of mid-range for an HT cap)
Reforming current: 10mA (this feels like too much for the 16uF 450VW caps I often work with)
Power therefore being dissipated in the capacitor: 2W (200V x 10mA)

The big questions are then "What's the effective volume that that 2W is being dissipated in ?" and "What's the average heat capacity of the material within that volume ?". Taking the second one first, electrolytic capacitors contain aluminium foil, paper separator and some moisture with chemicals in it. The volumetric heat capacity of aluminium is about 2.4 J/cc/K. That of paper is about 1.1 J/cc/K. The moisture is probably negligible, but for completeness it's 4.2 J/cc/K. Assuming a 50:50 volume ratio of foil and paper and adding a bit for the water let's estimate that the average heat capacity of the stuff in the capacitor is 1.8J/cc/K.

Turning to the first question, the 'best case' answer is that the power is being dissipated more-or-less uniformly through the whole capacitor, mainly because heat conduction within it is so good that the heat will flow very quickly indeed away from any local hot-spot. My 16 uF capacitor occupies half the volume of a 5cm long by 2.5cm diameter can. So its total volume is about 12cc. Its total heat capacity is therefore 12cc x 1.8 J/cc/K i.e. about 22 J/K. So with 2W being dissipated in it it will take more than 10 seconds to heat up by 1K. If the last part of the reforming only takes, say, 3 minutes then the total temperature rise in the core of the capacitor will be less than 20K or so and that won't be enough to do any harm. However a 'worse case' answer is that the heat is not actually dissipated unfirmly through the whole capacitor. Instead some smaller volume takes most of the heat and can't conduct it away fast enough to keep the temperature down. If the effective heated volume were just a quarter of the capacitor, rather than the whole thing, then in the hot part the temperature would rise by 80K and if the ambient temp in the workshop was 290K to start with then the hot part would end up very close to the boiling point of the water. Boiling the water even in a very small part of the capacitor just doesn't seem a very clever thing to do.

People might say "Oh you don't need to worry about heat, I've felt the outside of the can and it doesn't get hot at all." They're missing the obvious point that the outside is quite a thick-walled metal tin often sitting in free air which can cool it. Of course it will stay cool. It's the capacitor itself, particularly right in the middle, that's vulnerable to being cooked.

I realise that this is a hand-waving calculation and therefore not likely to be very accurate. I don't think it will be much more than half an order of magnitude wrong either way though. Unfortunately I think that to do significantly better would require a rather detailed understanding of conductive heat transport in mixed media and, probably, a decent 2D thermal modelling package running on my PC, neither of which I have, I'm afraid .

Cheers,

GJ

[TonyDuell]

One thing to remember is that reforming is not a 'bodge'. It is how the capacitor was made in the first place (OK, called 'forming' then). The dielectric oxide layer was formed on the positive 'plate' by exactly this process I believe.

So IMHO if a capacitor has simply lost the oxide dielectric (by having it re-disolve in the electrolyte) and is otherwise as new (not dried up, no damage to seals, etc -- and that is a big 'if') then reforming should make it as good as new.

The other thing is that there have been a number of 'dodgy' new capacitors in recent years. There were well-known stories of them turning up on PC motherboards. Those in particular had the wrong formulation of electrolyte, so reforming them was not a cure. I have no idea if any HV ones were made that could be used in valve equipment. Probably not, actually, but... What I am saying is that a new component may not be any more reliable than an old reformed one.

[Bazz4CQJ]

Originally Posted by GrimJosef

Since I'm working in a field where retaining an original component can add certainly tens and sometimes a hundred or two pounds to the value of an amplifier, the extra care and time it takes to reform slowly seems a price worth paying. This is the empirical approach.

Yes..and that is a very effective approach in those circumstances .

[GMB]

Can anyone point me at any manufacturer-supplied documentation showing a need for 'reforming' of consumer-grade electrolytics??

I have a 2007 document from Vishay which gives the age for reforming as 3 years for non-solid types and 20 years for solid types, and gives a way to calculate maximum current and the calculation of expected life.

Some manufacturers do publish very detailed info on their products if you browse around.

The US Dept of defense gave guidance on reforming in 2004, where 5mA was given as maximum current to use for a maximum of 1 hour, before checking leakage after allowing it to cool down.

[G8HQP Dave]

Originally Posted by Bazz4CQJ

I vaguely recall reading articles where there was talk about reforming the oxide layer on aluminium, but cannot recall if that was pure speculation or whether it had been observed via some experimental study?

I would not call something based on how electrolytics actually work 'pure speculation'. Reforming is precisely that: reforming the oxide layer. This was originally formed at the factory, then maintained by electrolysis from the leakage current. Left unused the oxide slowly degrades so leakage current is needed to reform it. Too much leakage current causes heating and gassing and failure. So you need a small leakage current, so it takes a long time. That is why they are called 'electrolytics': the oxide layer (which is the capacitor dielectric) is maintained by electrolysis and needs to be reformed by electrolysis. This is also why you can't have an electrolytic with zero leakage current.

If it is quick then it never needed doing in the first place - hence those who claim that they never do any reforming but just turn it on can get away with this except for the times that they can't. I suppose they conclude that an exploding electrolytic was beyond reforming anyway - but it might not have been if they had done things properly.

PS Like many of you, I maintain a stock of components. If I have had a new electrolytic sitting in a drawer for more than a few years I will reform it before using it, unless it is low value or low voltage and being used in a non-critical circuit position where it can reform safely in situ. Manufacturers datasheets often specify shelf life, after which reforming is needed.

[David Simpson]

Most folk consider reforming in relation to acquisitions of radios & equipment which are suspect due to age or unknown condition. e.g. Junk sale purchases, finds in old lofts, SK sales, etc. Owned items which are regularly used or just switched on several times a year, unless it is known that something contains a suspect capacitor, don't need to require re-forming of electrolytics.
However, due to their relationship to the high DC voltages from the rectifier circuitry, particularly in 'scopes, valve amps, & Tx's, then it always best to be safe than sorry.
There is a vast amount of gen on capacitor construction on the internet. But basically, if one strips out the innards of an e.c., you get two long strips of metal foils(which can be several feet long) with a paste of polarise-able molecules of electrolyte sandwiched thinly between them(a long thin chemical butty). The re-forming current of 1 or 2 or 4mA, whatever, is used to pass -ve charged electrons through this paste over the whole surface area. For any particular sq.cm - just a few uA will be passing. Taking time of several hours ensures that an even distribution of electrons takes place. Not just in those areas where the electrolyte might be a smidgen thinner. Therefore ensuring that the resulting +ve/-ve charge can be established over the whole interacting area of two strips of metal.

Regards, David

[PaulR]

This is a very interesting thread.

I used always to replace electrolytics in radios I acquired until I got a Griffin 0 - 300v power supply (ex-school I think as someone had stuck a label on it to say it was not suitable for use by children!). I now use this to crank up the voltage slowly until I reach working voltage or the maximum the unit will supply. I have to admit that in my ignorance I restricted the current to 10mA rather than 5mA but I have only had one failure which was in a set that had the smoothing filter resistor directly under and almost touching the can. In this the can quickly got hot and the current went up rather than down over time.

It is sad to think of the presumably good caps I have gutted to put modern ones inside.

[GMB]

Griffin 0 - 300v power supply (ex-school I think as someone had stuck a label on it to say it was not suitable for use by children!

Quite right too.
300V is barely enough to entertain a child. 30kV would be more suitable.

I used to take unbelievably dangerous PSUs into school for various projects.
No-one understood them enough to worry.

[Hampus]

As I don't have a suitable power supply I just use half-rectified mains. A variac, a resistor and a diode. No isolation here Am I the only one doing this?

[Refugee]

I have a Variac on an isolating transformer and I just clip a diode and resistor into a safe block and connect up with croc clips and that is it.

[G8HQP Dave]

Originally Posted by Hampus

Am I the only one doing this?

I hope so! To monitor the reforming you need a voltmeter attached, so you have to be careful if a lash-up circuit is connected to live mains. OK if you are really careful, and nobody else has access.

In many cases the equipment itself can be temporarily converted into a capacitor reformer.

[Hampus]

Ah, I forget about the meters! I use two, one for voltage and one for current.

[Bazz4CQJ]

Searching the internet for more information on reforming, I stumbled upon the operating manual for a British services instrument called “Reforming Units, Electrolytic, No.1 Equipment”. This appears to be a reformer produced for the RAF(?) in ~1957. It’s on the web at http://home.btconnect.com/gmb/ReformerNo1.pdf. It's about 3MB so would not upload to the forum. I then realised that it's actually GMB's website!

The page appears to be the only page on the entire web which refers to the unit.
I wonder if any of the amateur designs in circulation show “great similarity” to that instrument, perhaps dating back to the time when the manual was probably still classified?

B

[David Simpson]

This is an interesting piece of equipment. Many thanks to GMB & Bazz4CQJ for its info availability.
I couldn't see a CT number, perhaps it never had one. However, the rear main enclosure has the exact dimensions & front lid securing holes as my old Marconi CT44 AF Wattmeter. From my old RAF Test Equipment repair days back in the 60's, I certainly don't remember ever seeing one. But I think that that particular 12" x 8" size of airtight enclosure was used for several items. Maybe a troll through VMARS online catalogue might reveal more. On various other stations I also don't remember seeing one in the radio bays. When a duff or suspect electrolytic was found it was immediately replaced with a new item. As was the case with most components. We had a vast abundance of store's items available.
I wouldn't mind getting my hands on one. A very interesting gem, I reckon.

Regards, David

[Mike Phelan]

I've actually got one of these and used it a couple of times, successfully.

[GMB]

As I don't have a suitable power supply I just use half-rectified mains. A variac, a resistor and a diode. No isolation here Am I the only one doing this?

I pretty much do this but often without the variac.

But I use a capacitor to limit the current, and usually use a doubler type configuration so the driving voltage is quite high. Then just watch the voltage rise.

If you want a lot more control and to fiddle about then look on my web site for this rather good device: http://home.btconnect.com/gmb/TE46.htm although it did need a tweak to avoid meter damage. I use this when assessing leakage resistances as you can set any voltage and see the current - but limited to 10mA. It can measure the capacitance too and gives an idea of the ESR a bit.

Oh, and I nearly forgot this: http://home.btconnect.com/gmb/C25.htm - the vintage answer to ESR meters!

[Skywave]

Originally Posted by G6Tanuki

Remember: "Original" does not automatically mean "Still Good". I'll always go for reliability and performance over appearance or originality. [Others may choose to differ].

Generally speaking, I take the same path as in that penultimate sentence.
Why? Because for me, a repair is usually a necessary 'evil': that is to say that I don't relish fault-finding. Once I've done all the necessary dismantling, diagnosis, effected a repair, re-assembly and then a 'soak test', I want to be quite assured that I won't have to go down that same path again - certainly not in the 'short-term'. By replacing 'old' with 'new' (providing that the electrical and mechanical suitability is O.K., of course), by fitting 'new', I feel that I have minimised the risk of such a repeat repair, plus any consequential faults / damage that may arise.

But as per the above quote - others here have different viewpoints - and I respect that.

Al.

[GrimJosef]

Originally Posted by Skywave

[... By replacing 'old' with 'new' ... I feel that I have minimised the risk of such a repeat repair...

But as per the above quote - others here have different viewpoints - and I respect that.

Yes. Our approach to this determines where we sit on the spectrum between 'repair' and 'restoration'. It's stating the obvious, given the name of the forum, but we have members here who cover the whole of that spectrum. There are even one or two who would leave a set untouched, and not working, to preserve its historical integrity.

Cheers,

GJ