[R-390] Re: Tube shields? We don’t need no steenkin' tube shields!

[2002tii]

Barry wrote:

>Ideally, these would provide maximum thermal physical contact 
>between the two, while, at the same time, allow for maximum 
>convection cooling -- vertical "chimney" effect -- from the base of 
>the tube and shield out the top of the shield.
>
>If combined with a shield with a large top flange AND the insert 
>located high up, might well block the airflow, resulting in heat buildup.
>
>Bottom line though -- if electrical/signal shielding isn't needed 
>(to avoid spurious effects, whatever), the best bet is no shields -- 
>with some form of forced air cooling -- like well-positioned muffin fans.

I would have thought that most boatanchor afficianados would know all 
about IERC shields by now.  There is a military study that was pretty 
widely available on the net a few years ago, and probably still 
is.  It details the temperature of tube elements with dozens of types 
of shields and other variables.  The study shows convincingly that 
IERC tube shields with proper bases (see below) are the best way to 
reduce tube element temperature.

Most of the heat loss from tube elements is by radiation.  This is 
true with or without tube shields.  Without them, the radiative sinks 
are the walls of the chassis, the transformers, etc.  With them, the 
radiative sinks are the insides of the shields.  So, IERC tube 
shields do not work by convection, but by conduction and 
radiation.  The liner in an IERC shield does two things -- cools the 
tube envelope (glass) by conduction and, much more importantly, 
absorbs heat from the tube plate(s) by radiation.  (Note that the 
liners are flat black.)  All of this heat is passed to the outer 
shield can, then conducted into the chassis through the shield 
mounting base.  Thus, you need the appropriate, rather high shield 
mounting bases for them to work correctly.  Without the right bases, 
IERC shields are worse than nothing (i.e., the tubes run hotter than 
they run bare).  With them, IERC shields are way better than anything 
else (including nothing).  Most other tube shields are worse than nothing.

Airflow through the liner/shield is not a significant source of 
cooling, as shown by the net rise in tube element temperature without 
the proper bases (though undoubtedly some marginal convection cooling happens).

Remember, it is the temperature of the tube elements (grids and 
plate) that is important, NOT the temperature of the glass.  Sure, 
the glass can get too hot, but by the time it does, the plate and 
grids are long past fried.  And cooling the glass won't do much to 
cool the elements -- there is a vacuum between them -- so fans don't 
really do much good directly.  By reducing the temperature of the 
radiative heat sink (or, in the case of IERC shields, reducing the 
chassis temperature), fans can have a secondary effect.

Best regards,

Don