>Does anyone have any idea on the length of time that big (i.e. 1MW + )
>tubes take to heat up the cathodes, from the time filament voltage is
>applied ? A friend asked me, thinking I would know, but needless to say
>I have never looked at a data sheet for any of the really bid devices.
>Dave Kirkby G8WRB
Many of the large tubes are similar to the medium sized tubes as far as the
filament material.
Thoriated tungsten mesh, thoriated tungsten bars, thoriated tungsten wire.
As you know, the
3-500Z is thoriated tungsten. As far as electron emission goes, these
elements can be brought online in a few seconds. But there are other
factors that dominate the startup of bigger tubes (with handles and with
eye hooks).
Hairpin filaments are the quickest to start, as the filament wires are
vertically stretched. The upper support can allow for the expansion, by
taking up the slack. Like the 3-500Z.
The large filament 'baskets' called mesh filaments can take 20KW or so of
heater power. One must understand that these things may be mounted with
spring supports or pantagraphs (is that the correct word?). But the
filament wires may be wound spirally. The mechanical dimensions must change
as the heat is switched on. While the anode is fairly stable chunk of
material, the grids are fine wires in baskets, or pyrolytic graphite shells
in some of the big tubes. If the heater expansion rate is faster than the
adjacent grid, then there is risk of a filament to grid short. The spacing
is small there, for high mu. Some of the filament heat is radiated to the
grid and plate, and the grid changes also in dimension, at a differing
rate.
Care and Feeding does explain some of the filament basket designs and
concerns. Also
a booklet that is free from Econco in Woodland, CA (916) 662-7553, called
TUBE TOPICS. All of this leads to the requirement that a few minutes be
needed and that a ramp up of the filament voltage is required - to allow
the mechanical distortions time to relieve the changes and the geometry
change in the desired fashion. The pyrolytic graphite tube is a hard one.
The coefficient of thermal expansions of the filament is much different
from the graphite, which has it's own thermal expansions. PG does not have
symmetric thermal or electrical properties but instead has different
characteristics in the X and Y versus the Z directions.
I am giving 8 minutes of controlled ramp for a TH555A tetrode from France.
The Burle 7835 power triode gets 2 minutes. An Eimac 4CW250,000B gets about
3 minutes. Some of these are limited by the speed of a motorized variac on
the primary of the filament transformer. Others have electronic control.
More on that in a later posting perhaps. In particle accelerator
applications, one has the luxury of knowing when you want particles, and
the warmup can be planned for (except for frequent thunder storm outages
here in New Mexico!). Broadcast stations have a much tighter requirment to
return to the air. So transmitters may actually have a 'slam on' bypass to
the filament timer, which allows instant return to the air if power is only
interrupted for a few seconds. This is allowable if the tube manufacturer
agrees to it. Depends on the tube.
Oxide cathodes which are coated with special barium and calcium compounds
among other things have a required warm up time before one can switch on
the cathode current. These are the 2-15 minute tubes that are in some ham
gear, and in even high power tubes. The peak current for short pulses is
high in these monsters, so they are typically used in radar sites, such as
the monsters out on some Pacific Islands.
Something else I learned recently from Eimac engineers is that the big
tubes can have a sagging filament over time. This is especially prevalent
for tubes switched on and off daily. Eventually the electron optics are
disturbed enough that the cutoff conditions, the grid - cathode
capacitance, and the gain can change. And a short can develop. It's better
to leave them running, or reduce the filament to 'black' heat, than to
throw the switch off.
Finally, there is the getter. Some tubes actually have a getter that
appears to be reactivated or at least helped by the filament heat. These
don't work instantly. This leads to the requirement to wait 15 minutes or
so, before applying HV. This is what we learned for the 4CW250,000Bs used
in plate modulators here.
In this case it isn't the filament emission itself but another feature that
requires the filament.
The short answer for G8WRB is that each tube has it's own requirement and
that the manufacturer must supply that info, either on the data sheet or
upon your personal request. Otherwise you may destroy the tube very quickly
(or over a year or two).
I read a very interesting old article by RCA recently, where they explained
how vacuum tubes are really chemical reactors, that the emission of
electroncs from the cathode is an oxidation/reduction mechanism.
I never though of it that way.
John
K5PRO
--
FAQ on WWW: http://www.contesting.com/ampfaq.html
Submissions: amps@contesting.com
Administrative requests: amps-REQUEST@contesting.com
Problems: owner-amps@contesting.com
Search: http://www.contesting.com/km9p/search.htm
|