Modern EL34 reliability?

[pdf64]

...The amp I was popping el34's in uses a 6.6k primary and a 470r resistor between the plate and screen supply. Chances are that the OT HF impedance is sufficient that spikes during clipping were finding their way through the screens rather than the plates....

I'm a bit sceptical of the 'spike' hypothesis
Rather, consider that as the load impedance increases above the max power zone, plate dissipation will tend to reduce but screen grid dissipation will tend to increase, so for pentodes that can be a problem.
And because guitar speaker impedance is somewhat above nominal for much of the bandwidth of interest, things are already a bit pushed.
It's perhaps why the EL34 loaded Marshalls have a rep for being fussy about load impedance.

[nsubulysses]

IMO, sort of. Lemme splain.

If you're making classic designs then you are using output transformers that have a lowish primary impedance and, since you're using good iron (no one said Mercury wasn't good, just that it's overpriced to cover their marketing) the choke is likely a key performer here also. IME it's high frequency spikes during overdrive that over dissipate the tender screens of modern el34's. A lowish plate impedance and a choke who's coil has sufficiently high impedance at high frequency helps prevent high frequency spikes from finding a path through the screens rather than the plates. The amp I was popping el34's in uses a 6.6k primary and a 470r resistor between the plate and screen supply. Chances are that the OT HF impedance is sufficient that spikes during clipping were finding their way through the screens rather than the plates. This, combined with the newer tubes tender screens was almost certainly my trouble. Increasing screen circuit impedance and lowering plate circuit impedance is the solution. Lesson learned. In your case I would bet it's the design, rather than any sort of brand/quality for the iron that is working in your favor since it has the aforementioned properties.

Can you explain further what you mean by increase screen impedance reduce plate impedance. Do you mean use a choke rather than 470R resistor, or increase 470R resistor?

So say for instance you use a choke with 65R DC resistance instead of 470R resistor,and placing it in circuit gives you the high impedance to AC as you desire, BUT then the screen voltage is now too high in overdrive conditions because it doesn't sag much because of the 65R choke resistance. Can you add a 400R resistor in series with the screens and get the best of both, high impedance to AC, and voltage drop to reduce screen dissipation at max volume? Or would this be kind of a dundering arrangement?

[bob p]

...no one said Mercury wasn't good ...

I did.

http://music-electronics-forum.com/t44662/

Actually, when the iron is good it's great. My experience was that their QA was not as high as their prices when I tried to purchase in bulk.

I noticed Sowhat was having good success with CE-D as a supplier. I've had success with them too. Maybe there is something to the idea that not all suppliers are the same?

On the other hand, I noticed that he's using high B+ and master volumes. Amps that have master volumes are kinder to the EL34 than amps that don't have master volumes and are run hard into a loadbox.

[Ted]

Hey Leo

Something I've experimented with is using a FET with a zener and a couple of resistors as a shunt regulator on the Screen Grids.

You need to select a suitably large wattage dropper resistor, but it worked OK dropping the 500V rail down to 250 for my home made GU50 amp

[nsubulysses]

Can you explain further what you mean by increase screen impedance reduce plate impedance. Do you mean use a choke rather than 470R resistor, or increase 470R resistor?

So say for instance you use a choke with 65R DC resistance instead of 470R resistor,and placing it in circuit gives you the high impedance to AC as you desire, BUT then the screen voltage is now too high in overdrive conditions because it doesn't sag much because of the 65R choke resistance. Can you add a 400R resistor in series with the screens and get the best of both, high impedance to AC, and voltage drop to reduce screen dissipation at max volume? Or would this be kind of a dundering arrangement?

Chuck what's up ?

[pdf64]

Yes, as a protective measure I think it's a good idea to add a series resistor to a low resistance choke, so as to get an adequate degree of screen sag at high signal levels / overdrive.
But for a given inductance, higher resistance chokes are cheaper

[Chuck H]

Ok, sorry for the late response.

My inference was that the reason some are building classic Marshall circuits and not having any trouble with modern el34 tubes could be because the choke in those designs presents a high enough impedance to HF spikes so that the screen circuit isn't a sufficiently lower load than the plate circuit at those frequencies. Marshall chokes aren't typically small or cheap. IME HF spikes DURING CLIPPING can be extreme. If an OT's inductance is such that there is a higher impedance to these spikes than there is through any screen grid circuit, then the screen grid will serve bear the "load" so to speak. It's appropriate to note here that Marshall amps typically use a fairly low primary impedance. Making it even less likely that the screens circuit will present a lower load. Reduced screen voltage can only prevent static over dissipation. If the tube is clipping hard and there are HF spikes they can still find their way through the screen circuit rather than the plate circuit if the screen circuit impedance is lower than the plate circuit impedance at those frequencies. With the typically big, fat Marshall choke I think it's entirely possible that the screen circuit impedance to these HF spikes is greater than 470R. So even though the static screen voltage is higher with the choke the screens themselves may be better protected because the HF spike voltage through them is lower. I haven't tested for this, YMMV, but it's as good an explanation I can reason for why JCM800 clones, aren't eating tubes while other amps are. Considering that Marshalls often have high plate and screen volts and use a lower primary impedance it would be easy to think that such amps would eat more tubes. But the devil is in the details.

And as Pete indicated, if you have a low resistance choke then adding a resistor is a good idea. A low resistance choke is almost sure to be a smaller choke with fewer turns. Lower inductance. And it's probable that such a choke wouldn't present as high impedance to HF spiking. In some cases I think adding resistance in series with the choke might even be a good idea with the burly Marshall chokes. I'm considering amps with over 500Vp. Where with a choke fed screen you could have nearly the same voltage on your screens.

[cjenrick]

back in the 70's you could run those EL34's with the plates cherry red and they did not care,

nowadays, one little splotch of red and they go into meltdown mode,

Vacuum Tube Valley EL34 issue>

http://www.pearl-hifi.com/06_Lit_Arc...ation_Date.pdf

[Thoriated Tungsten]

back in the 70's you could run those EL34's with the plates cherry red and they did not care,

Yes, they did care. They just didn't have a voice to scream out their pain. Tubes are supposed to last many thousands of hours. Most guitar and Hi-Fi amps run the tubes into the ground with zero regard for tube longevity. Only output power for the least amount of manufacturing cost matter. Leo Fender, I blame you.

Look at licensed radio amateurs. They frequently get decades of life out of tubes for their transmitters. They have to, because by now they can *only* rely on NOS for replacements.

Though to be fair, then the current manufacturers may be limited by other factors than their own competence. There are many cases of tube manufacturing requiring specialty metal alloys or compounds either due to function or practicality of manufacture. One example off the top of my head would be that we can probably no longer make tubes like Philips' "Special Quality" long life types, as the required cathode sleeve Nickel alloy with virtually zero Silicon content is no longer available.

I do wonder if the real reason why CED/ex-Svetlana stopped making glass tubes was actually due to them running out of 'new old stock' of the materials required for making reliable tubes. Thus they were faced with a choice of either dramatically reducing their tube reliability or drastically increase prices as they had to pay for having small batches of unusual materials made.

[cjenrick]

the reliability of the Sevetlana 6L6 might have had something to do with their decision to get out of the guitar amp tube biz,

it was a great sounding tube but supposedly had problems, probably the screen grid getting a bit warm,

[Mick Bailey]

Svetlana still makes glass tubes, just not guitar/audio tubes - they make large transmitting and industrial X-ray tubes. I think it was Stan who reported that the guitar amp market is so small that the economics didn't stack up for them. As a market shrinks, then the price goes up and/or the quality comes down. Like making tyres for a Ford Model T.

We're hanging on to a sentimental obsolete technology; a world of Ge transistors, linear power supplies, tubes and wooden cabinets. The rest of the world has moved on and it's difficult to sustain manufacturing on a world-wide scale to support what amounts to a minuscule market.

[cjenrick]

NOS prices ain't gonna go down, that's for sure, I got an email from Eric Barbour one day, he said that Svetlana was getting out of the audio tube game, come on over to my office in Woodside CA for some free tubes, like an idiot i never went, still kickin myself,

have you seen this Mullard factory tour>

https://www.youtube.com/watch?v=-GgWIlvyEL8

[Jazz P Bass]

Really awesome video.

I was dumbstruck by the equipment used.

The R & D to engineer the machines must have cost a bundle.

Pity the techs that had to maintain the stuff.

[Justin Thomas]

I <LOVE> the rhythmic clikety-clack of the machines when they would do the closeups. Seems valves are able to make great music at <EVERY> stage of life...

Justin

[Thoriated Tungsten]

Svetlana still makes glass tubes, just not guitar/audio tubes - they make large transmitting and industrial X-ray tubes. I think it was Stan who reported that the guitar amp market is so small that the economics didn't stack up for them. As a market shrinks, then the price goes up and/or the quality comes down. Like making tyres for a Ford Model T.

I should probably have stated, that Svetlana stopped making small all-glass receiver type tubes.

(Snip.)
Pity the techs that had to maintain the stuff.

You might be surprised. In modern factories employing similar mechanical machinery the weak link is usually the control electronics. With the exception of any deliberate wear parts, like welding electrodes, and at the speed machines like this operates at, the mechanical parts last a very, very long time with just a tiny bit of preventive maintenance.

A friend of mine owns a company, which makes similar machinery though usually performing less complex operations. With modern metal manufacturing techniques the mechanical parts pretty much lasts forever, unless you take an angle grinder to them. For instance he considers modern, general purpose robotics to be hilariously unreliable by comparison, to the point that he literally refuses to use robots in his projects.