Failed grid leak resistors damage Leak Stereo 20

[GrimJosef]

I've just returned a Leak Stereo 20 amplifier to its owner who brought it to me with an exploded EL84 cathode bypass capacitor and a burnt out corresponding resistor. Looking underneath I could see another cathode resistor going distinctly brown. My first thought was that this was just EL84s going gassy late in life, as they are prone to do, and then going into thermal runaway with the resulting over-current doing the damage. In part this was true. The valves were gassy and the one connected to the blown components was running a very high over-current.

But having repaired the damage and fitted known good replacement valves I was checking the grid voltage when I suddenly saw it go through the roof and the new valve switch hard on. I powered off immediately and started checking the grid circuit components. Both the coupling capacitor and the grid leak resistor had been replaced at some point in the past. The caps were now Russian paper in oils (the four silver cylindrical ones you can see just below the centre of the attached 'before repair' pic). I know they can fail occasionally, but it isn't very common and testing of these ones indicated no problems at all.

The resistors however - the row of four little yellow ones which are above the cathode resistors, which are above the cathode capacitors, which are above the couplers - were a different matter. For some reason the previous repairer had fitted 510k ones instead of 470ks. And the one in the misbehaving channel was intermittently open circuit. Poking it caused continuity to come and go. I checked the solder joints and the tags, but they were fine. It was the actual resistor itself - as if an end cap was making and breaking contact with the track on the main body of the resistor. Sure enough the grid leak which corresponded to the browning cathode resistor was also iffy - intermittently reading either 510k or a couple of megohms.

Obviously much higher resistance grid leaks can lead to much higher grid voltages and thus to over-current and serious overheating. This can release gas inside the valve, so even if there wasn't a gas problem before, there very easily can be afterwards. I fear this is what has happened here.

I'm sorry the picture is so poor. I thought I'd taken a proper one before sending the amp back, but it seems I must have inadvertently deleted it as I can't find it now. All I have is the quick snap sent to me by my customer in advance of the amp arriving. I don't know what brand these iffy resistors are. They look at a glance like Iskras, bút more careful inspection shows they're not. They're a bit smaller, and darker in background colour and look a bit less well made. All I can say is, if you come across anything like them in some critical position then do take care. These ones have a lot to answer for.

Cheers,

GJ

[barrymagrec]

I had faulty grid leak problems on a number of Bias / erase oscillator units - 100K grid leak in a double triode tuned multivibrator circuit. Seemed to be a resistor batch problem but hard to be definite as it was all 10 year old kit. Also small carbon types.

Difficult to identify at first because as with your Leak it was intermittent - worked fine for ages then just went beresque for no apparent reason.

[kalee20]

Those four little yellow ones look very much like Iskra UPM Carbon film types, which I have generally found very reliable!

Though - difficult to see in the photo, but - no component likes being stretched between two tags. A bit of neutral slack (needn't be obvious) will avoid stressing the leadout joint.

[G6Tanuki]

I'm wondering if the little yellow resistors have suffered by being wired with their leads under slight tension, then subjected to repeat heat-cycling from the adjacent cathode-resistors?

(I remember that in some 1950s valve TVs certain resistors and power-diodes which were expected to run 'hot' had their leads formed into single-turn loops before the tails were soldered to tags/valve-bases; this confused me at first because I was imagining all sorts of cunning things to do with inductance)

[trickie_dickie]

Originally Posted by G6Tanuki;

(I remember that in some 1950s valve TVs certain resistors and power-diodes which were expected to run 'hot' had their leads formed into single-turn loops before the tails were soldered to tags/valve-bases; this confused me at first because I was imagining all sorts of cunning things to do with inductance)

I always did that. My first manager back in the 60s taught me to do that to give thermal relief to the resistor. Or put a 90 degree bend to the leads for the same result.

[GrimJosef]

I see quite a lot of equipment which has been constructed on tagboard and I have to say open-circuit component failure is rare. I can't remember the last time I came across it and finding two examples out of four resistors does tend to make me blame the components rather than the assembly technique.

I don't know much about the history of this amp. All the modern replacement components were put in by someone else before it came to me. If I had to guess I might say they were 1980s vintage or later (the Russian capacitors appeared in much larger numbers here after the disintegration of the USSR).

As for whether the yellow resistors might be Iskras, as I say I think they aren't. The kind I'm familiar with are like the one in the attached picture - a little longer and much fatter than the ones in the amp. The latter looked like they might be relatively recent but pretty 'cheap and cheerful'.

Cheers,

GJ

[vidjoman]

The resistors in the amp are 1/4 watt but the one you show from Anglia is 1/2 watt - both Iskra.

[GrimJosef]

I didn't realise Iskra made 1/4W in this range. I've looked at a few pictures of them on eBay and it's plausible that that's what the ones in the amp were. In which case I'm a bit more surprised that they failed. I've seen plenty of the 1/2W and larger Iskras in Radford amps, mostly mounted onto pcbs admittedly. They seem reliable there, even when they've ended up a bit scorched.

Cheers,

GJ

[Valvepower]

Hi,

In the dim and distant past when building the Roost Guitar amps back in the 1970's I remember Iskra 'alike resistors. They looked very much like Iskra but from memory they were not of European manufacture, possibly Indian manufacture?

I remember the wire coming out of the body of the resistor on the occasional resistor when pulling the leg gently through the long nose pliers to straighten the leg - not good! Also from memory the colour bands weren't as well defined as on the Iskra parts.

Ah the days where you could get Partridge output transformers, Mullard EL34's and GEC KT77's

Regards
Terry

[vidjoman]

Originally Posted by GrimJosef

I didn't realise Iskra made 1/4W in this range. I've looked at a few pictures of them on eBay and it's plausible that that's what the ones in the amp were. In which case I'm a bit more surprised that they failed. I've seen plenty of the 1/2W and larger Iskras in Radford amps, mostly mounted onto pcbs admittedly. They seem reliable there, even when they've ended up a bit scorched.

Cheers,

GJ

We used thousands during the 1960’s in transistor equipment. Still have quite a few and rarely find one out of tolerance. We also used the 1/2 watt versions in power amplifiers.

[GrimJosef]

Ah, transistors ... beyond my reach I'm afraid . I deal almost entirely with valve equipment, which might explain my unfamiliarity with them. My stock of larger ones do seem pretty good.

Cheers,

GJ

[Radio Wrangler]

Are we sure they're Iskra?

If someone's fitted Russian military PIO capacitors, have they gone boutique on resistors? Something designed to make people feel happy about their carbonaciousness, but without the care that makes for reliability.

David

[GrimJosef]

Given how many fairly ordinary blue metal films there are, I'd be surprised. The Russian PIOs are at the 'very reasonably priced' end of the PIO market. They don't appeal to the big spenders.

Cheers,

GJ

[bikerhifinut]

A couple of weeks ago, I had a similar problem with an output EL34 on my homebrew amp very loosely based on the Leak TL circuit.
A post mortem after isolating the cause and understanding the effects (Thanks Valvepower for the tutorial) led me to suspect my own enthusiasm for soldering grid stoppers as close to the input grid pin as possible, here the 10k grid stopper had failed mechanically at the end cap, but effectively created the same conditions re lack of referencing to ground for the grid bias. The replacement grid stopper (and the the other 3 were also checked/replaced as needed) was soldered with some extra "Play" in its leadout and not so close to the valvebase tag that I was stressing the actual endcap of the Resistor.
An informative and very useful post, Thanks GJ.
Incidentally I also have the Russian PIO in my Leak Stereo20, bought when they were initially available "en masse", never given a moments worry. They have what looks like a belt and braces gastight/liquid tight Glass seal, I understand they were designed for use in extreme conditions in Military gear.
These days i just use standard Polyesters like the BVWS shop jobs, the prices of the Russian ones are now getting a bit silly too.

Andy.

[Marconi_MPT4]

Reading this thread brought to mind a similar problem when matching 100k 1% 0.25W resistors for a project. Selecting two out of fifty from a tape bandolier, at least six were outside specification e.g. 762k which started to drift downwards! Another measured 113k with no drift while four others were 2 to 5% high. Supposed to be and look like MFR4 but I somehow doubt they are.

They has been in stock for some years and stored in a reasonably airtight box, but can not trust them now.

Rich