6GB5 and Variants as Screen Driven Amps

[wrenchone]

I've been semi-obsessively collecting 18GB5/LL500 horizontal output tubes, as they are cheap and have a nice tall, skinny form factor. Has anyone out there tortured them in a screen driven amp? They have nice, thick glass envelopes, and the only disadvantage that I see so far is that they need a magnoval socket. I suspect that they are fairly common in Europe, as even the domestic versions are largely made in Holland, probably by Philips. They are rated for 17.5W plate dissipation, just like a lot of their squatty duodecar siblings, but I suspect the plates can take more, as they look pretty substantial for all manufacturers I've seen so far.

[tubelab.com]

I got some 27GB5's somewhere. I now have the ability to melt screen grids with drive, so I will add them to my list of suspects.

[whitelabrat]

I'm looking forward to seeing some of those 6GV5's getting melted.

[tubelab.com]

I blew one of the 98 cent 6BQ6GTB's that I got in the AES sale in a most spectacular fireworks display last night, but of course the camera wasn't ready. I succeded in extracting 125 watts from a pair of the skinny 6BQ6's right before the fireworks started.

It seems that I have found a new way to blow up a tube. In extreme screen driven situations (125 watts from 6BQ6's IS extreme) the screen grid is going highly positive while the plate is approaching zero. This makes the screen grid glow. When it glows is can emit electrons which travel to the positively charged plate. The PowerDrive circuit uses a resistor to return the output (screen or control grid) to the negative supply. As the screen grid starts to draw a lot of current it is no longer under mosfet control and the screen voltage will runaway. The screen voltage rose quickly causing the tube current to shoot toward 1/2 amp. I was observing this effect and managed to catch it 3 times, but on the 4th time the fireworks erupted inside the tube and it doesn't work so good any more.

I will do some more experiments tonight!

[kenpeter]

Perhaps your screen powerdrive needs an active pulldown to
survive those glow situations. Maybe something like a White
Source Follower?....

Or turn the powerdrive upside down, so that pullup is resistive,
and the MOSFET is what yanks it back toward ground...

[tubelab.com]

Perhaps your screen powerdrive needs an active pulldown to survive those glow situations.

No, I really should try to avoid extracting 120 watts from a pair of tiny tubes, but I just can't help myself. I thought about building a push pull PowerDrive with active pull down, but then I realized that once the grid is glowing the tube itself is already in trouble.

I may do some experimenting along these lines anyway, but I believe this mechanism is an inherent limitation in screen drive. The screen really doesn't like being 200 volts more positive than the plate while being fed from a very low impedance source. It is probably OK for a normal music source, but continuous sine wave power testing melts grids!

I noticed the glowing screen syndrome last year, but the runaway never happened. I now realize that my wimpy power supply is what saved me. The plate and screen were fed from the same (already tapped out) source. Last night I had the plate connected to its own supply which was set on 600 volts and can source 1.5 amps.

Or turn the powerdrive upside down, so that pullup is resistive, and the MOSFET is what yanks it back toward ground...

That is what I am trying to avoid with the PowerDrive. I want to be able to slew the grid in the positive direction quickly. It is necessary for A2 operation in normal control grid drive. The pull down is only needed to discharge Mr. Miller.

Now, it is time so stop typing and hook up some fresh tubes!

[wrenchone]

I was just looking at my stash of 18GB5/LL500s last night. I like their tall, slim form factor. What I don't like is their 275V plate voltage rating. Since these are, after all, horizontal output tubes, I assume that this plate rating can be played with somewhat, especially in a screen drive context. Still, this does tend to limit the output power as compared to some of the tubes with less challenged plate voltage ratings. I may just decide to just let it rip and run a pair of these at 400V B+ to see what I can squeeze out of them in a screen drive P-P amp the peak cathode current is in line with the rest of its squatty little 17.5W brethern, so I can reasonably expect as high a peak power. Am I getting a little jaded, or does 45-60W/channel seem to be insufficient motivation to pursue a screen driven P-P design using these tubes?

[Michael Koster]

Originally Posted by kenpeter

Perhaps your screen powerdrive needs an active pulldown to
survive those glow situations. Maybe something like a White
Source Follower?....

Or turn the powerdrive upside down, so that pullup is resistive,
and the MOSFET is what yanks it back toward ground...

Or turn the follower upside-down and use an active pullup - allows a depletion mode FET to be used for the pullup in a mu-follower configuration. Low impedance drive in both directions.

[wrenchone]

I already use active pulldown on all the cathode followers I've done so far - depletion mode devices are no mystery - wish there were more of them.

[kegger]

Did you ever try these tubes out?

I'm running an SE,UL amp with EL500/6GB5 at roughly 300v plate to cathode an 60ma.
I'm using a 2.5k output with no feedback, these things sound really really nice as well.
(it's been running on an off for about a year, very stable at these parameters anyway)

I was running 6SL7 direct coupled to 6SN7 CF, I thought with the bias voltage they'd
be a bit more difficult to drive or least need some voltage gain but the amp was "way"
to sensitive, turned out just the mu of a 6SN7 was enough gain, I also had some plate
to grid feedback in there, turned out to be better with no feedback. Quite nice tubes!
(if running conventional I'd say an mu of around 30 is a good solid voltage gain stage)