Re: Are tube plate dissipation's peak or rms?

"Tim Wescott"
Phil Allison wrote:

Tubes are (or were), in fact, rated with much more margin than
semiconductor devices. Semiconductor specs are "never, ever exceed"
specs, where tube specs -- at least receiving tubes -- are "center
design" specs. The idea was that the specifications given were what
you designed to, assuming 117V (in NA) on the line, but if it went up
to 125 for a while, no tubes would be damaged.

I never experienced tube equipment much -- I'm too young. But tales of
putting 500V on the plates of 6L6's* and running the plates cherry
red** abound.

** The only real limitation on max plate voltage for most power tubes is
internal insulation breakdown. 6L6s are rated to stand continuous peaks
of up to 2.5 kV on the plate, at zero current.

The max rated DC supply voltage for 6L6GCs ( the only kind made for
decades) in a class B, push pull power stage is 800 volts - where
each plate swings up to almost double that figure on signal peaks. One
pair of tubes can deliver 100 watts of sine wave power this way.

* 6L6's are rated for something like 375V.

** All such figures need to be looked at in context.

A pair of 6L6GCs is rated for a 450 volt DC supply in class AB1 to get
55 watts at low THD.

The final limit is internal temperature of the mica supports and the
plastic base - which must not become so hot they begin to conduct.
Making the series grid resistance low really helps keep the issue at

I was going by me recollection of the plain ol' 6L6 data sheet -- which,
when I pull it out and look, says 360V.

** For a *very specific set* of other conditions you completely fail to

My point was more that you can take a tube -- like the 360V-rated 6L6 --
and put much more to it than it's rating (500V in transmitter usage) and
get away with it.

** You made no point at all.

It was all bollocks.

Which is basically what you say much more succinctly:
all such figures must be looked at in context. 500V on a "360V" tube, in
intermittent service, kinda works -- and if it doesn't, it's a socketed

** Did you bother to read anything I wrote ??

All of it went right over you pointy head - didn't it.

I wouldn't expect to be able to waltz into a design review and present a
design where's I'm putting 36V onto a 36V transistor, much less 50V onto
a 36V part -- my colleagues would insist that I use a 50V part for 36V
service, not the other way around.

** You are so far off with the fairies it is laughable.

Piss off.

..... Phil