All this is very interesting.... There's a bunch of things I find
confusing.

1) Radiation vs conduction
On one hand it's obvious that a layer of dull, IR-black paint has a
thermal insulation effect. But the same may be said of ANYTHING
surrounding a tube. Even one of the "good" IERC shields must have an
insulating effect of sorts. Moreover, the copper "fingers" are metal-
shiny, obviously designed to sink heat by conduction, not by absorbing
radiation.

2) Terminal filament temperature
The issue here isn't avoiding failures as much as increasing tube
life. The heat issue in low power devices was much disregarded till
close to the end of the tube era, due to engineering, commercial, and
I believe psychological effects. In the end, microtubes used in the
NORAD systems reached 500k h MTBF, and even before that, repeater amps
in transatlantic underwater phone cables had already made major
advances in reliability. There is a DoD or Collins study on the
matter, claiming that IERC shields improved MTBF over unshielded
tubes, if I remember correctly. As in all stable thermal systems, once
equilibrium is reached the terminal temperature of cathode and
filament can't be independent of what happens at the tube surface. I
know how to scrounge up a Wiener bridge or a shunt and measure small
deltas in heater current - but where do I look up some ideas on how
current is related to temperature, whence what deltas to expect, at
least in order of magnitude? Without a theory, even a rough one, as
Popper pointed out, there's not much to test.... :-(

3) Cooling effect of going black-body in the IR range
Granted, if 90% of the exchange surface eg in an aircooled engine
faces . . . itself (think deep cooling fins facing each other) most
cooling must come from conduction to a moving ambient medium (aka
air). So blackening should not make much difference there. But what
about situation where there IS open space around a hot device?

Right now, I remember that I know a guy who makes heating systems,
including a successful line of radiation heaters:
http://www.sabiana.it/download_pubblici/catgen_en.pdf, see the 1st
product, called Duck-Strip. The name's an inside joke: they were
designed by a Mr. Anatrella - Italian for "cute duckling". The things
run on hot water. At a Volkswagen plant those radiators heat people
from a vertical distance of over 20m. From an economic standpoint I
am not sure it's totally in the company interest to maximize per-
surface-unit radiation, but I believe they took a look at paints. They
also used to have a gas-fired radiation heater running at a much
higher temperature than the water type.

I'll ask him... stay tuned. Moreover, some things I found by
googling words that came up in this thread:


overclocking & paints (uh...)
http://www.overclockers.com/tips684/

irrelevant but funny
http://members.optusnet.com.au/mcdjim/100_4062s.jpg
http://forums.bit-tech.net/showthrea...2a61& t=53048

read what Dampney writes on improving IR absorbtion with "visible and
IR" black paint
http://thurmalox.com/Upload/Products/Products28.pdf

teacher's guide to experiment :-)
http://www.ed.psu.edu/ci/Papers/STS/gac-3/in05.htm
(no word on measuring temperatures in unreachable recesses)

On Thu, 17 Jan 2008 04:27:06 -0800, spamhog wrote:

Dull, black, heat resistant paints
have been used to help cool engines for ages.

It would be cool (literally) if one could spray and heat-cure unshielded
tubes
and improve their heat-shedding

Is there any indication that such paints, or some vacuum-tube specific
types,
would help keeping tubes cool
by improving heat radiation?

I'd love some factual info, if it exists, or educated guesses, rather
than uninformed blind guesses, as I am awfully good at doing uninformed
blind guesses already! :-)

What I know is that the glass will pass a proportion of the IR energy
being generated by the outside surfaces of the plate. Depending on just
how great a percentage, you may get more heating of the glass from the
paint capturing the radiation from inside than you get cooling from the
paint re-radiating it to the outside.

What I don't know is what will actually be the case.

--
Tim Wescott
Control systems and communications consulting
http://www.wescottdesign.com

Need to learn how to apply control theory in your embedded system?
"Applied Control Theory for Embedded Systems" by Tim Wescott
Elsevier/Newnes, http://www.wescottdesign.com/actfes/actfes.html

spamhog wrote:
On Jan 17, 12:48=A0pm, Chuck Harris
No, but don't let that stop you from trying it anyway.

-Chuck

:-) What keeps me is the fact that I have no idea how I could measure
the core temperature in a tube!

With an infrared thermometer or optical pyrometer. Auto parts stores
should have a model in the $20 range.

Is heater current temperature-dependent enough as to provide a proxy?

No, because the plate temperature is very different than the cathode
temperature, and the plate temperature is what you worry about.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

Scott Dorsey wrote:
spamhog wrote:
On Jan 17, 12:48=A0pm, Chuck Harris
No, but don't let that stop you from trying it anyway.

-Chuck
:-) What keeps me is the fact that I have no idea how I could measure
the core temperature in a tube!

With an infrared thermometer or optical pyrometer. Auto parts stores
should have a model in the $20 range.

The hand held infrared thermometers will end up measuring the
temperature of the glass, not the elements inside.

An optical pyrometer would be the way to measure the filament's
temperature (quite crudely), but they are expensive.

Is heater current temperature-dependent enough as to provide a proxy?

No, because the plate temperature is very different than the cathode
temperature, and the plate temperature is what you worry about.

The heater temperature should be somewhat dependent on the plate temperature,
half of the plate's radiation goes into the center of the tube, which is
where the heater is. The heater temperature should vary slightly with
plate temperature. A sensitive bridge might be able to measure it, but
it would be down in the noise.

The plate of a tube doesn't wear out, so it's temperature isn't
by itself important. When the tube is evacuated, the plate is
induction heated to a nice red/yellow temperature. This is done
to remove any absorbed gases and other contaminants. As long as
the plate is never heated above this temperature, it won't release
more gas.

The big reason to worry about the plate getting too hot is the
heat it radiates will heat the glass envelope, and may cause
it to crack, or melt.

-Chuck

On Sat, 19 Jan 2008, spamhog wrote:

All this is very interesting.... There's a bunch of things I find
confusing.

1) Radiation vs conduction
On one hand it's obvious that a layer of dull, IR-black paint has a
thermal insulation effect.

But at very thin layers, it is negligible. Look at house insulation. R-30
is 1-1/2 feet thick plus. A single pane window (single strength) is more
like R-1. Glass is a great insulator (compared to, say, copper) but at 1/8
inch thickness its almost not there. Vacuum tube glass is even thinner.

But the same may be said of ANYTHING
surrounding a tube. Even one of the "good" IERC shields must have an
insulating effect of sorts. Moreover, the copper "fingers" are metal-
shiny, obviously designed to sink heat by conduction, not by absorbing
radiation.

You have to compare conduction, convection, and all the mechanisms.

2) Terminal filament temperature
The issue here isn't avoiding failures as much as increasing tube
life. The heat issue in low power devices was much disregarded till
close to the end of the tube era, due to engineering, commercial, and
I believe psychological effects. In the end, microtubes used in the
NORAD systems reached 500k h MTBF, and even before that, repeater amps
in transatlantic underwater phone cables had already made major
advances in reliability. There is a DoD or Collins study on the
matter, claiming that IERC shields improved MTBF over unshielded
tubes, if I remember correctly. As in all stable thermal systems, once
equilibrium is reached the terminal temperature of cathode and
filament can't be independent of what happens at the tube surface. I
know how to scrounge up a Wiener bridge or a shunt and measure small
deltas in heater current - but where do I look up some ideas on how
current is related to temperature, whence what deltas to expect, at
least in order of magnitude? Without a theory, even a rough one, as
Popper pointed out, there's not much to test.... :-(

There may be some small effects, and it might be more tied to how many
times a tube is warmed up from cold, cooled off to cold than actual
temperature (in many applications, folks would turn them on and leave
everything running [eg. computer monitors, even today]).

3) Cooling effect of going black-body in the IR range
Granted, if 90% of the exchange surface eg in an aircooled engine
faces . . . itself (think deep cooling fins facing each other) most
cooling must come from conduction to a moving ambient medium (aka
air). So blackening should not make much difference there. But what
about situation where there IS open space around a hot device?

That black-body radiation works in both directions, not just absorption.
And, it may also be spectrum-dependent so that would have to be measured
with instruments, not our (human) eyes.

Right now, I remember that I know a guy who makes heating systems,
including a successful line of radiation heaters:
http://www.sabiana.it/download_pubblici/catgen_en.pdf, see the 1st
product, called Duck-Strip. The name's an inside joke: they were
designed by a Mr. Anatrella - Italian for "cute duckling". The things
run on hot water. At a Volkswagen plant those radiators heat people
from a vertical distance of over 20m. From an economic standpoint I
am not sure it's totally in the company interest to maximize per-
surface-unit radiation, but I believe they took a look at paints. They
also used to have a gas-fired radiation heater running at a much
higher temperature than the water type.

I'll ask him... stay tuned. Moreover, some things I found by
googling words that came up in this thread:


Its good that you did some google searching, but on the whole I think you
are worrying too much about cooling. And, if you do manage to cool the
cathode, then emmission would surely suffer. As the ultimate wacky
suggestion, you could immerse the tubes in liquid air/nitrogen and really
keep them cool (cost a lot of money), but then I'll bet you couldn't
"light up the tubes" (with filament voltage) at all.

===== no change to below, included for reference and context =====

overclocking & paints (uh...)
http://www.overclockers.com/tips684/

irrelevant but funny
http://members.optusnet.com.au/mcdjim/100_4062s.jpg
http://forums.bit-tech.net/showthrea...2a61& t=53048

read what Dampney writes on improving IR absorbtion with "visible and
IR" black paint
http://thurmalox.com/Upload/Products/Products28.pdf

teacher's guide to experiment :-)
http://www.ed.psu.edu/ci/Papers/STS/gac-3/in05.htm
(no word on measuring temperatures in unreachable recesses)