Pp el 34 ht waaay too high!

[unparalleled]

I am working on a NZ made Gunn 50 amp made in the 1970s.
The mains supply here is still usually around 230V as it was back then.
The standard PS is a voltage doubler identical to the attached Holden Wasp schem. (another NZ made 1970s amp.)

With the standard set up I get 880v plate, which is enough to cause repeated valve failures.
Is there a trick to lose about 100 Volts?
By my calculations @ 800V the amp would produce 100W, not 50.
At present I have it rewired to a non doubling full wave ct via S/by switch. using a diode on each leg of the PT.
Voltages with this set up:
AC 317-0-317
HT 423v
Screens 367v ( added dropper)
Bias -35v
The OPT imp. is 11K!
Should be able to get 50 watts with this set up but actual OP is 37W, not enough for the punk band owner!
I have experience servicing but have not done any valve work since 1979 and have forgotten a lot of tricks.
Hope someone can help?

[tubeswell]

Hi there unparalleled

What is the HT winding VAC output with nothing connected?

I guess it could be a number of things (so I'll guess away):

a) no load (disconnected cathodes or plates) causing the B+ to be at maximum potential

b) running EL34s in an amp designed for KT88s?

c) Was there some kind of voltage doubler in the power supply?

d) was your meter set to DC voltage? - does it have good batteries?

[Enzo]

Are you sure it was voltage killing your tubes? it is usually dissipation rather than votage that does that. EL34s can safely handle a ton of voltage. You mention 35v of bias, but no mention of current through the tube.

[unparalleled]

Hi there unparalleled

What is the HT winding VAC output with nothing connected?

I guess it could be a number of things (so I'll guess away):

a) no load (disconnected cathodes or plates) causing the B+ to be at maximum potential

b) running EL34s in an amp designed for KT88s?

c) Was there some kind of voltage doubler in the power supply?

d) was your meter set to DC voltage? - does it have good batteries?

All guesses wrong, Amp was as originally built/sold.
Gunn and Holden amps are known for high HT, currently available tubes don't seem to cope as well as valves made decades ago.

A: load connected via OP trans.

B: correct tubes, screen and bias volts.

C:Yes PS same as attached schematic.

D: I started using my first multimeter in 1965 age 12, I use one every day, I think I have that covered.
I am looking for the lose 100v trick, not a fault.
Cheers.

[unparalleled]

Are you sure it was voltage killing your tubes? it is usually dissipation rather than votage that does that. EL34s can safely handle a ton of voltage. You mention 35v of bias, but no mention of current through the tube.

I put a matched pair of Electro Harmonix bottles in this amp a few weeks ago along with a full cap job/ pots/ jacks cleaned etc.
Unless the owner has run it with a dodgy load the high HT is what I suppose would cause a short from screen to plate (120 ohms or so across valve pins out of unit)
I am now using Duncan's amp tools so the currents should be close to optimum specs.
The only thing out of specs is the extra 100 Volts.

[tubeswell]

So how much plate current?

[unparalleled]

So how much plate current?

Ia 60-240 but that is with Z out 3.4K and this amp has 11K xformer.

[tubeswell]

The primary impedance is surely way too high. Marshall 100W usually runs 4 x EL34s with a primary of 1k7 or so. Why is it so high? (Is that usual for these amps?)

[unparalleled]

The primary impedance is surely way too high. Marshall 100W usually runs 4 x EL34s with a primary of 1k7 or so. Why is it so high? (Is that usual for these amps?)

Yep, I agree it's unusually high impedance but Clarry Schollum told me what the impedance was from the original drawing and 11K is what is stamped on the Beacon made iron.
He also has schematics should I need them.

[Enzo]

Your output transformer impedance won;t affect the idle current. The lowest current reading you mention is 60ma. Is that your idle current???

If your screens are shorted to your plates, it is most probable that the screen wires melted and fell against the plate. And that would be from excess screen current, not high voltage.

At 880v plate, 60ma would represent about 52 watts. That seems a bit steep to me for the 25 watt EL34. Dropping that to 780 and 60ma would still be a 47 watt dissipation. Even down to 700v on the plate, 60ma still represents 42 watts of dissipation.

Your schematic shows screen resistors of only 270 ohms. They would limit screen current a lot less than the more common 1k resistors. Screen grids are the weak element in the EL34.

[Steve Conner]

I agree with Enzo. It most likely died due to overdriving the screen grids.

With 880V on the plates and an 11k OPT, it doesn't take much anode current to make 50 watts. So the tubes hardly need any screen current either. Assuming you were running the HT at 880 and the screens off the 440V node of the doubler, you'd probably be wanting screen resistors of 2.2 or even 3.3k. Basically cut and try to find the biggest screen resistor that allows the amp to make 50W on the bench.

With 440V, it's the 11k OPT that causes the problem. It's the wrong load for the tubes at the lower HT voltage, and the mismatch causes the screen grids to draw lots of current. So the amp can't make its rated power, and burns out its screen grids trying. If you want to stick with 440V, you need to get that load impedance down, perhaps by mismatching the speaker one step so the tubes see 5.5k.

KT88s or 6550s could probably take the 5.5k load at the full 880V HT. The amp would then make nearer 100W than 50, at least until the transformers burnt out... But if they look reasonably hefty they might hold up fine.

If you can't mismatch the speaker (say because it's 16 ohms and the OPT has no 32 ohm tap) then maybe you can replace it. A more sensitive one would help make the amp louder still (Vintage 30 anyone?)

[MWJB]

"B: correct tubes, screen and bias volts." Negative bias voltage is only a means to an end, it is only "correct" when it results in acceptable plate current at the power tubes. If, as Enzo also points out, your idle current is 60mA per tube (or even 60mA for both tubes) your plate dissipation is an issue.

Why not just switch to KT88? 6550 will also need the screen voltage issue addressed (a with EL84).

[Steve Conner]

Why not just switch to KT88? 6550 will also need the screen voltage issue addressed (a with EL84).

Well, I used to think this too, but Merlinb convinced me in a recent thread that KT88s have just the same screen voltage issues as 6550s. Both tube types may well need the 2.2 or 3.3k screen resistors to stop them drawing lots of screen current under clipping conditions. I'll dig up the thread if anyone's interested, or maybe Merlin could comment on this thread?

[tedmich]

D: I started using my first multimeter in 1965 age 12, I use one every day, I think I have that covered.
I am looking for the lose 100v trick, not a fault.
Cheers.

Ouch! The CT 100v zener trick (R.G. Keen's MOSFET + 95v zener, see http://www.geofex.com/Article_Folder...osfetfolly.htm) would work to lower your HT, but the issues raised may need addressing in the long run (after the amp goes home!)