Reliability of Mullard/Philips "Liquorice Allsort" (Tropical fish) Capacitors

[Leon Crampin]

The trouble is that you can't trim a film capacitor once it's wound. I tested hundreds of these devices and there was definitely a "valley" in the middle of the value distribution curve. They met spec. though - to about 99.9%. They also don't support combustion.... I rejected Rifa capacitors and earned my salary that month..

Well done Philips.

Leon.

[kalee20]

Originally Posted by G.Castle

As a general comment I have noticed that a lot of capacitors, of non electrolytic construction, do seem to be on the low side of their marked values often very close to the tolerance limit.

You've been a victim of the Great Capacitor Scam!

Many manufacturers managed to improve their production methods, and the raw plastic films became repeatable - such that capacitors could be routinely made with production 'spreads' of 2% rather than the 10% of yesteryear.

The market was happy with 10% in many cases, so what did the manufacturers do? They'd wind them short dials - if they were making "100nF capacitors" they'd set their dials at '92nF' knowing that everything which came out would be 90 - 94nF so still within the range of 100nF +/-10%.

And then they could get 8% more capacitors out of a roll of film/foil...

[G6Tanuki]

Pragmatically, I've never had any issues with these capacitors.

But I've always bought them from accredited supplies at full price as proper-quality parts, not Birkett's Bargain-Basement stuff.

If it's cheap there's probably a reason.

[stainless]

I knew them as Football Jerseys

[G.Castle]

Originally Posted by kalee20

You've been a victim of the Great Capacitor Scam!

Many manufacturers managed to improve their production methods, and the raw plastic films became repeatable - such that capacitors could be routinely made with production 'spreads' of 2% rather than the 10% of yesteryear.

The market was happy with 10% in many cases, so what did the manufacturers do? They'd wind them short dials - if they were making "100nF capacitors" they'd set their dials at '92nF' knowing that everything which came out would be 90 - 94nF so still within the range of 100nF +/-10%.

And then they could get 8% more capacitors out of a roll of film/foil...

That seems highly credible!

I have a learned habit of testing even new components before use and with capacitors that really looks like what's happening! Some other types of variation is as Leon says where you have a distribution near either extreme of tolerance so leading one to believe that the better ones from a batch were cherry picked for sale as tighter tolerance, with no doubt a premium, this is I've noted more with vintage components.

My comments about those tropical fish from Birketts were just as old John claimed, the colour coding on them was a bit wobbly on some and there were a few that had debris particles stuck in the dip coating, but functionally all fine. I tended to never bother with his "untested" transistors lines though.

To his credit, he never misrepresented any of his stock.

Greg.

[kalee20]

It was explained to me by a capacitor manufacturer's rep!

The close-tolerance capacitors would still be more expensive, but for a different reason.

In the old days, they'd be tested and cherry-picked from the Production output - which costs money (and which could give the 'notch' in the distribution of 10%-marked values that's been mentioned).

And now, no extra testing would be necessary, but the machine knobs would have to be turned up to use more film/foil, which still costs money.

The C280 capacitors... I like them! Though disappointingly, the covering did seem fragile. The orange dipped-epoxy radial capacitors which replaced them regrettably don't add the same colour to circuit boards, but they are more robust...

[G.Castle]

Thanks Peter that does make sense, I've found that motor start/run capacitors are annoyingly often actual value = stated value minus tolerance.

I suppose anyone with electrically good C280 capacitors could conceivably improve them mechanically by dipping them in clear epoxy if they really needed them to be used in period repairs.

[Leon Crampin]

Definitely not. The epoxy on Rifa capacitors is largely, but not entirely the cause of their dismal performance. Philips chose their encapsulation materials very carefully.

The orange, and earlier blue capacitors are not epoxy coated.

Leon.

[G.Castle]

Originally Posted by Leon Crampin

Definitely not. The epoxy on Rifa capacitors is largely, but not entirely the cause of their dismal performance. Philips chose their encapsulation materials very carefully.

The orange, and earlier blue capacitors are not epoxy coated.

Leon.

Thanks for that Leon, do you know what the encapsulant was on the orange drop ones? They do seem much more robust.

I just thought the Rifa caps were/are so dire because the paper dielectric would swell then split the case, (rather than being a problem with the epoxy itself), moisture would get in then bang!

Regards,
Greg

[kalee20]

Originally Posted by Leon Crampin

Definitely not. The epoxy on Rifa capacitors is largely, but not entirely the cause of their dismal performance. Philips chose their encapsulation materials very carefully.

The orange, and earlier blue capacitors are not epoxy coated.

Leon.

I'd agree with the first bit - transparent epoxy is indeed prone to cracking, because it has no filler material to stop crack propagation. Microcracks thus become macrocracks. However, with a plastic-based capacitor element rather than a paper one, it shouldn't be such a disaster.

But the second - I don't have my Mullard/Philips databook handy, but see the attached datasheet exerpt (Mullard changed name to Philips Components, then sold off the capacitor division as BC Capacitors, which was eventually bought by Vishay as part of their world-domination attempt on passive components)

It definitely says epoxy!

[G.Castle]

Thank you Peter. I had also seen the Vishay data sheet, but not the earlier Phillips one for the 'orange drops' so I can't say, though I will accept that the encapsulant perhaps contains some filler that makes it more pliable and have less tendancy to crack..

[kalee20]

Well, I had a look in my 1978 Mullard Quick Reference Guide, unfortunately it's too abridged to give the 'orange drop' encapsulation material!

I don't have the relevant full databook

Leon, any clues as to what you think it might have been? (My money is still on epoxy, but if a better material exists I'd be interested, whatever - it might be useful in my Day Job!)

I was also musing on this:

Originally Posted by Leon Crampin

Philips chose their encapsulation materials very carefully.

Philips might have been very careful, but they had several big encapsulation cock-ups all the same...

* The C280 capacitors as the subject of this thread for starters, as previous posts;

* The Lockfit transistors, the encapsulation material caused an early death;

* The TO-7 transistor encapsulation, as used on AF117's and several others... another metaphorical ticking time-bomb.

[Leon Crampin]

The C280 of encapsulant used an inert filler - I can't remember what it was. The binder was a soft precursor of epoxy, totally unlike the hard stress-inducing material that made the Rifas unusable for our applications.

I accept totally the existence of Philips cock-ups such as those mentioned. I clearly remember testing the lockfits and watching them fall off the board with broken tabs during my vibration tests. There was no strength of triangulation as all 3 tabs emerged at the encapsulation split line - idiotic. The icing on the cake was then their failure in my dishwasher tests for moisture ingress.

I dodged the tin whisker problems - Ge devices were no good to us at all due to their inability to work over the automotive temperature range.

It was an interesting job worth doing. If only other makers had attached the same importance to real world component testing.

Leon.

[Maarten]

The lockfit range was short lived, luckily, and quality might have varied somewhat between factories. The TO-1 and TO-7 "RoHS" construction might be the larger cockup by far of the three with a seemingly near 100% failure rate.

The C280 series was continued from the mid to late 1960's on as the 2222 342 and later 2222 352 series. There might have been differences in construction as well, though possibly only between the 342 and 352 series. I don't see them as particularly troublesome.

Fun fact: "Liquorice allsort" is called "English liquorice" (actually: Engelse drop) in Dutch, and so are the capacitors.

[joebog1]

Bit like Dropjes ( I think ) double salted licorice.
With respect Maarten

Joe

[Radio Wrangler]

Apart from the fragile leads, the other thing to watch out for with these things is that in many of them, the dielectric film is Mylar, which is PET film given a 2-dimensional stretch and then heat treatment to set the stretch in (~3:1 stretch, so not subtle) often referred to as PET to dodge brand names. It's a prime material for exhibiting 'dielectric absorption'

Charge a capacitor, discharge it completely, remove the short and then watch it partially re-charge itself out of nowhere. Presto Changeo Magic!

Be sure never to use these in ADCs, digital voltmeters etc as integrator capacitors... and voltage to frequency converters (analogue music synthesisers). You can run into some mystifyingly anomalous behaviour. Anomalous? Well Anglo-Saxon has some names for it.

They're perfectly fine for general applications.

David

[dougietamson]

Originally Posted by Maarten

The lockfit range was short lived, luckily, and quality might have varied somewhat between factories. The TO-1 and TO-7 "RoHS" construction might be the larger cockup by far of the three with a seemingly near 100% failure rate.

The C280 series was continued from the mid to late 1960's on as the 2222 342 and later 2222 352 series. There might have been differences in construction as well, though possibly only between the 342 and 352 series. I don't see them as particularly troublesome.

Fun fact: "Liquorice allsort" is called "English liquorice" (actually: Engelse drop) in Dutch, and so are the capacitors.

Would it be an idea to dip them a few times with drying between times in a modern clear (ish) varnish?

Liquorice Allsorts were mostly sugar, the worst 'liquorice' I've ever tasted was brought back by a Finnish colleague returning after a holiday with the aim of bringing back the most disgusting 'sweets' and watching us eat/spit them out!

[Maarten]

Dipping them might be problematic as some of them (may depend on which series) are already dipped in a very thin wax like coating. Varnish might not stick very well.

At least one of the ways they seem to constructed is first epoxy (or similar) dipped, then painted, then wax (or similar) dipped.

Metallised Mylar or PET capacitors are more commonly known as MKT capacitors. Just as other types there's a time and place to use them. The versions coated in an even mustard colour, are KT or wound Mylar capacitors with slightly better characteristics. For more critical applications, Philips recommended metallised paper (MP), polystyrene (KS) or metallised polycarbonate (MKC) capacitors. Later on, those were largely replaced by metallised polypropylene (MKP) or metallised/wound polypropylene (KP/MKP) capacitors. MKT still is the standard go to foil capacitor for non-critical applications.