[Amps] Water Cooling

[Ian White, G3SEK]

N9RF wrote:
>Has anyone experience with water cooling (GS-35B in particular; any 
>tube in general)? I have been told it may not be feasible at 4KV. Also, 
>if the cooler is in the cavity (coaxial or stripline), there are other 
>issues. Any thoughts?

For many years there was a description of N6CA's water-cooled 1296MHz 
amp in the ARRL Handbook, and that shows how to do it safely. The main 
issue about HV is the leakage current due to impurity ions, introduced 
in the water and subsequently dissolved out of the metalwork of the amp 
itself.

Fortunately pure water is a very poor conductor. You have to allow a 
certain length of plastic tubing between the HV side and the point where 
the water is 'grounded out' by flowing through short sections of 
grounded pipe. This is a vital safety precaution.

In effect, the HV supply is shunted by a high leakage resistance in each 
of the two insulated tubes. One of the grounding sections is connected 
to chassis via a microammeter, so you can monitor the leakage current 
and change the water when appropriate.

Cavity amps are easier to water-cool when the outside surface of the 
anode is also outside of the cavity, so it's almost dead to RF. (There 
are many scattered references in the '432 and Above EME Newsletter' 
archives on W6/PA0ZN's website, but the technical contents are not 
indexed for searching.)

Water-cooling has also been done in stripline and coaxial-line cavities, 
where the whole tube live to RF. The best way to do this is to run metal 
pipes along the stripline until you reach a point of low RF potential, 
and then jump across in insulated tubing. (DUBUS magazine had a 
description of an amp using a water-cooled GS35B with a half-wave 
stripline for 144MHz, but I don't have the reference.)

On the other hand, 4kV is approaching the level where some GS35Bs will 
also become prone to internal arcing. This and the possibility of water 
leakage suggest the need for a good glitch resistor and an effective 
anode current trip.



-- 
73 from Ian G3SEK         'In Practice' columnist for RadCom (RSGB)

http://www.ifwtech.co.uk/g3sek