Marshall 1959 SL100 (4 x EL34) Grid leak resistor size

[tubeswell]

I have an early 1959 reissue. I think it's "SLP" but not sure. ... It has 220k grid leaks.

Just a slightly related story I suppose.?.

Think this one is the 2005 UK-assembled RI 'handwired' - Gutshots attached. Note the unusual pin numbering on the (Chinese OEM supplied?) octal sockets LoL. (Maybe the originals also had misnumbered socket pin numbering? - Like, if you made a mistake in the past - why not do it again?)

[Chuck H]

That's funny/disturbing! But it's cool how they got the wire colors so close. The purple is too dark and the green is too bright, but not bad Mine is a PCB from '91. Whether that's a documented reissue or not I can't say, but that's what I have. A PCB 1959 reissue from '91. Long ago I checked the transformer numbers that are supposed to be "the ones" against what I have. That's the transformers I have. And as I mentioned, no great shakes IMO. But then I was playing the amp through British made 16 ohm Vintage 30's at that time. IMO they are not nearly as good sounding as the British 8 ohm version (or any of the China made speakers ) so that may have been part of the problem. But the amp is just languishing in the closet for now. I'm more interested in my own designs and whenever I'm strapped for $$$ and think about selling it I don't have the dough to rebuild it properly. Selling it "as is" would be a horrible waste of opportunity. So I continue to sit on it until "someday"

[tubeswell]

Damn, I just got a call from that customer - those new (Mike Matthews) 'Svetlana' EL34s blew up!

[pdf64]

Just to note that the g1 circuit resistance limit for real EL34 is 500k even in class B fixed bias; some info notes 700k in other modes http://www.mif.pg.gda.pl/homepages/f...010/e/EL34.pdf
So 220k grid leaks per pair is compliant with that.
Whereas the limit for 6L6GC in fixed bias is only 100k! http://www.mif.pg.gda.pl/homepages/f...93/6/6L6GC.pdf

It's difficult to give credence to info for CP tubes; the electro harmonix 500k seems dodgy whereas the tung sol 100k seems more realistic
http://www.dougstubes.com/el34b-tungsol.pdf
http://www.dougstubes.com/el34eh.pdf

It seems that g1 current can increase when a tube is dissipating heavily, such that if the g1 resistance is too high, the effective g1 bias voltage becomes reduced.
That can lead to things spiralling out of control (leading to red plating) and it may be that EL34 have always been somewhat prone to that, more so than other common power tube types.

[dstrat]

If you try and bias that amp at 70% plate dissipation , and plate voltage is much above 420vdc its going to redplate the tubes in most cases at full output, regardless of g1 resistance.

see this post for more from ampgarage, R Aiken explained the problem way better then I could.

superlead with master volume redplates - The Amp Garage

Hope that helps.

dave

[Chuck H]

Damn, I just got a call from that customer - those new (Mike Matthews) 'Svetlana' EL34s blew up!

We need a "don't like" button

It's remarkable how many repairs I read about that go back to the customer only to fail again. I'm sure you tested the amp. Probably burned it in for a few hours and probably had it up to clipping for a time too. So what is the customer doing to the amp? Have any of the previous tube failures happened at your bench or do they only happen when the amp is with the customer? Does the customer complain of any tonal anomalies (Sound cutting in and out, low power, etc.) or just bring it back when the tubes blow without mentioning anything else? How does he use the amp (Always dimed into an attenuator? Are we sure there's a proper load? etc.)?

[tubeswell]

For this unit, this was the first time I'd serviced it. Yes checked all the voltages and brought the bias up slowly over a 40 min period. Va = 507, Ik = 35.1mA. Then afterwards I tested the amp for 1/2 an hour on one of my cabs with correct load - no problems. Went around to his place yesterday and tested it for 2 more hours on another cab with correct load. No problems. Then last night he took it to a industrial warehouse location somewhere where I don't know how the unit was been run, or what the wall voltages are, where the problem occurred. Even so, you'd think new tubes would last longer than 2 days. Maybe 71% bias was too hot? Maybe the load resistance should be doubled?