[Amps] Coupling a blower to an air system socket

[John Lyles]

I learned this in 1981 when 4CX20,000A/8990 tetrodes in Broadcast
Electronics FM30 transmitters would have a darkened center post for 
filament as the silver tarnished from heat. We applied a thermocouple 
and learned how hot it was. Eimac and most tube companies recommend to 
never operate at 200 degrees C on a ceramic metal tube envelope. Our 
mechanical engineer figured out that the air could be applied
directly to this post, from either direction. As long as the air wasn't 
preheated, it could exhaust from the center area as well. So all of the 
transmitters of later production had a small tube sticking out from 
under the cavity, where warm air exhausted after cooling the filament 
stem. No connection required to the blower as it pressurized the box 
already. Some air came up through the socket to cool the screen grid 
connector but the bulk of the anode cooling came up through four 
RF-screened ports on the deck where they then escaped from the cavity up 
through the anode fins to the inside diameter of inner conductor in the 
resonator.

Point is, one doesn't have to 'hose' forced air to the filament stem 
area, if a measurable flow of air coming out of that can exit the box 
through a port below.

In a dual tetrode HF 100 kW amplifier I used two dedicated Rotron 
"Spiral" high pressure blowers with hoses, just for filament cooling. 
The anode was water cooled and only needed residual air through the 
circuit. Some air was needed for screen grid contact, that being 
provided by the same filament air leaking through the fingers of the 
socket. Being HF, the circulating current to the screen grid was not 
exorbitant.

In the currently produced high power 200 MHz cavity amplifiers, I use a 
dedicated 125 CFM blower (Cincinnati Fan with 10 inch cast aluminum 
impeller) and the back pressure is 10 inches of water pressure. It comes 
up through the lower filament connection (a 2 inch water pipe) and a 
UDEL plastic cylinder forces it around the center post, and through the 
lower fingers of the socket to escape back down the annulus between the 
pipe and a second larger diameter pipe. Being cathode driven, the pipes 
are resonant elements in the input circuit with high RF voltages. That 
air is very warm, but the tube remains 150 deg C under all conditions.

A second 325 CFM blower (14 inch impeller!) forces filtered air into the 
output cavity itself, where it blows across the screen and plate 
connector rings through the finger contacts to escape into the room 
through RF-tight penetrations. Being VHF, circulating current is high. 
With this air cooling, I am still using 1 GPM of water to cool the 
screen grid contact rings (the tube has small fittings for hoses) and 90 
GPM to cool the plate. Nothing easy about cooling a tube of this size. I 
spent at least a month of the design and prototype time on cooling 
issues like these. The proof is in the testing, as Ian alluded to using 
temperature paints in the 1970s. I still use Tempilaq paints very 
frequently, not only for tube testing but also checking the operating 
temperature of various amplifier components. Then adjustments to air 
distribution are done to put the air flow where needed.

Now using 3D printed plastic parts to make air distribution components 
like 4 way hose splitter. Much easier than current and welding metal 
tubing and boxes. My air splitter alone is costing about as much as a kW 
HF amplifier.
73
John
K5PRO

> Message: 1
> Date: Sun, 17 Mar 2013 07:12:57 -0700
> From: "Jim Thomson" <jim.thom at telus.net>
> To: <amps at contesting.com>
> Subject: [Amps] Coupling a blower to an air system socket

> ## Brilliant, never thought about cooling tank components that way.
> For tubes like 3CX-3000A7 or 3CX-6000A7, they both use a
> coaxial filament stem.   If your cooling scheme is used,  with air
> pumped into the above chassis area, and chimney from  anode to
> top lid, like the k2riw arrl method,  it will work,  but with these coaxial
> fil tubes like the 3x3 and 3x6,  Eimac sez to shoot 5 cfm  BETWEEN
> the coaxial fil stem assy.   A simple method of doing this is to use a
> small diam, flexible  plastic tube, and a fitting to chassis... and use some of the
> compressed air from top half of amp compartment, above the chassis, and direct
> it down the flex tube... then direct the tube between the fil stems.