Cossor 364 - the dreaded brand new faulty component

[stevehertz]

In short, when restoring a Cossor 364 I fitted quite a few new components and one of the new capacitors was faulty. But it's a long story, read on if you're interested.

Back in 2012 my vintage audio/video activities were in a lull; no wireless, no hifi, no video. However, I was using a 1935 Cossor 364 as my domestic radio and I decided that the old cotton covered rubber mains lead needed changing for safety reasons. It was while dong this that my interest got fired up again, and seeing lots of 'iffy' components in the 364 I decided that I would start right there with that radio. I exchanged lots of capacitors, checked all the resistors for value, tested the valves, and got stuck in with switch cleaner wherever it seemed the thing to do. The 364 is quite awkward to dismantle and this entails disconnecting three wires from a separate speaker sub chassis to the main chassis. After completing the work I put it all back together and to my surprise and dismay, it didn't work, not a jot. I spent several sessions looking for a mistake in my restoration work and came to the conclusion that I had mixed up the three wires from the main chassis to the speaker. But as there are five different options of how these three wires could be connected (that's 15 wired permutations) and being fed up with it, I simply put the set back together and replaced it with my Cossor 375 to use every day.

Fast forwards nine years and I decided that it was time to revisit the 364 and get it working. Still thinking that I had muddled up the three speaker wires I spent some time identifying exactly what was what by measuring resistance values at the speaker transformer and inductor, identifying the three long wires from the main chassis, and then connecting them as appropriate. It still didn't work. Again I checked my under chassis wiring, looked for loose earth connections, valve base issues etc etc, all to no avail so time to start testing properly and take some voltage readings. It wasn't long before I found that there was no voltage on the screens of V1 and V2. This either meant that the HT feed resistor to the screens was o/c or the bypass cap was short circuit. Being an old component I checked the R first, it was spot on at 50k. This now meant that the new cap fitted nine years ago was the likely culprit - !. Measuring it on the ohms scale of my AVO, it was short circuit. It's rare to fit a new, replacement component and it be faulty, but clearly it happens. Anyway, a second replacement was fitted and the Cossor now works very well. The cap was a 0.1uF 400V lime green ERO. I've used lots of these in the past as replacements (I got them 'free' from my place of work) and they've always been very reliable, but not this one! Moral to the story, don't assume that the new component fitted is a good one!

[Boulevardier]

Thank you, Steve. A very "educative" story. I think we are all strongly biased to assume that a new component is healthy. Particularly with capacitors, I always try to remember to test and measure the value of a new part before I fit it (the manufacturers' part/value codes for capacitors can be - well, just impenetrable), but it's easy to neglect this check if in a hurry.

Mike

[mister valve]

I had a similar thing happen with an Electrolytic Capacitor (a 10uF @ 450V). Brand new and unused, I fitted it as an H.T. reservoir, to replace the aging original.

It worked for about 10 minutes then there was a few loud 'crack' sounds then nothing, the set went mute. Fortunately I had done one of my usual modifications when restoring the set, and fitted an H.T. Fuse. This had blown, so I checked the resistance between the H.T. line and chassis.... It was a dead short !
This brand new, modern, axial 10uF @ 450V had gone dead short after a grand total of about 10 minutes !!!! The sets operating voltage even before the the rest of the valves has 'pulled' the H.T. down, was well below the 450V rating, and the max ripple current is within the manufacturers specifications. I replaced this faulty capacitor with another new 10uF @ 450V of exactly the same type and make, and I've had no trouble at all since.

I'm so glad I fit H.T. fuses as standard practice when restoring a set; as I would have been needing a new rectifier valve, or possibly a new mains transformer - instead of just a new fuse costing pennies !

Getting back to the point.... It's quite true - you can't just assume a component will be good, just because it is new !

[joebog1]

Good story, BUT its a bit frightening. I have literally thousands of the said components.
My pic is just one box of five !!.
When the Aus govt got out of valves, I bought up all their "NOS" bits for a song. Capacitors, resistors and electrolytics. I have yet to have a failure!! ( touch wood ). It might teach me to be a bit more careful in the future.

Your pic shows a cap where the outer sleeve doesnt fit "properly". I am not accusing you of overheating it, but is there a possibility it HAS been overheated?. I have "toasted" a few with a heatgun while testing faulty equipment, and the caps assume the shape that yours has. They did not fail however!!.

Joe

[snowman_al]

I have experience of the green ERO caps, the MKT 1813 types mostly. They do leak and are on my check on sight list. Though I've never seen a short circuit one.
Audio Innovations used them and I always replace them if I get an AI500 or similar to service.

[stevehertz]

Originally Posted by snowman_al

I have experience of the green ERO caps, the MKT 1813 types mostly. They do leak and are on my check on sight list. Though I've never seen a short circuit one.
Audio Innovations used them and I always replace them if I get an AI500 or similar to service.

Thanks for the tip. As can be seen from the photo, mine are marked KT1801.

[Leon Crampin]

I spent some time years ago in my job designing automotive electronic assemblies assessing components. I think I earned my month's salary by not approving Rifa capacitors for my employer's applications.

As a quick summary, I would not approve any capacitor encapsulated or backfilled with hard epoxy. The stresses caused by thermal cycling are almost certain to cause failure. Although the Philips C280 and other derived capacitors have a somewhat friable encapsulant which can chip off at the lead entry points, it's soft enough to allow thermal flexibility and hence resist damage from thermal cycling, and wets the element sufficiently well to avoid moisture ingress. It's also flame retardant.

Leon.

[stevehertz]

Is that one reason why X and Y rated caps fail as they're often constructed using hard epoxy?

[stevehertz]

Originally Posted by joebog1

Good story, BUT its a bit frightening. I have literally thousands of the said components.
My pic is just one box of five !!.
When the Aus govt got out of valves, I bought up all their "NOS" bits for a song. Capacitors, resistors and electrolytics. I have yet to have a failure!! ( touch wood ). It might teach me to be a bit more careful in the future.

Your pic shows a cap where the outer sleeve doesnt fit "properly". I am not accusing you of overheating it, but is there a possibility it HAS been overheated?. I have "toasted" a few with a heatgun while testing faulty equipment, and the caps assume the shape that yours has. They did not fail however!!.

Joe

No, definitely not overheated! I've installed a few thousand components in my time. And in fact, each wire end was only 'laid on' and soldered quite quickly, no wrapping to fill up with solder. I too have used dozens of these in the past without problem, hence my disbelief at what I was finding. I have dozens more and won't hesitate to use them, but for the sake of 30 seconds, I will drop them across the AVO for a quick leakage check.

[G6Tanuki]

A salutary story which we can all learn from.

I had a similar sort of 'new part fail' when replacing a Lockfit transistor in my Roberts R707 - the new TO18 transistor I took off a bandolier supplied by a reputable company turned out to be mislabelled and was PNP not NPN.

cue much cussing and a snotty email to the supplier.

(all the others on my few inches of bandolier were equally wrong too. I stopped buying stuff from that particular suppliers on the basis that I no longer had confidence in their quality control process.)

[Uncle Bulgaria]

Even with the best intentions it's easy to forget to test a new component. Certainly I get caught up in wanting to see the result, so I am not as methodical as I ought to be. This is a good story to encourage me to be more so. Just think how long it's been sitting there as an irritation and it wasn't down to your workmanship at all!

[joebog1]

I always test components, but with the Rodestein stuff I have I origionally tested a "random sample bunch" without any leakage or failures. All my stock was made in Germany and in general I have always believed they made very high quality components. All my caps are 1813 types, which as I remember was their highest spec devices. I do know that the tolerance is excellent even though they are branded 5%, they are usuallly within about 2%. All the resistors are 2% branded and I have thousands of .6 watt and 1.2 watt devices. Another quirk for Rodestein, they are the only people that made .6 watt and 1.2 watt.

Steve I wasnt assuming YOU had overheated the cap, I was thinking more was it too close to a valve

Joe

[stevehertz]

Originally Posted by joebog1

I always test components, but with the Rodestein stuff I have I origionally tested a "random sample bunch" without any leakage or failures. All my stock was made in Germany and in general I have always believed they made very high quality components. All my caps are 1813 types, which as I remember was their highest spec devices. I do know that the tolerance is excellent even though they are branded 5%, they are usuallly within about 2%. All the resistors are 2% branded and I have thousands of .6 watt and 1.2 watt devices. Another quirk for Rodestein, they are the only people that made .6 watt and 1.2 watt.

Steve I wasnt assuming YOU had overheated the cap, I was thinking more was it too close to a valve

Joe

The cap is underchassis, 'miles' away from heat, the valves are on top of the chassis.

[stevehertz]

The set is now back up and running in a corner of my dining/office room and the 375 has been 'retired' again. I always loved this 364 when I saw it in a friend's collection, and many years later (over 30) we came to a swap arrangement and now it's in mine and I've no plans to let it go! I love mid 30s Cossors, they're so Art deco and even though they've never been considered the Rolls Royce of sets, they work very well!

[mister valve]

Your Cossor 364 is one lovely set and in such a nice condition too. One thing I do like about Cossor's of that era is the wave change switches, easy to clean and retention the contacts if necessary also very precise and positive action.
The Cossor build quality, in that era, was good, and the 'point to point' wiring is neat and direct, and not the dreadful "birds-nest" that some other manufacturers seemed to produce.

[stevehertz]

Thanks, and yes I agree about the wiring neatness.