[Amps] New Elecraft Amps (GS35B sockets?)

[Ian White GM3SEK]

Tony King - W4ZT wrote:
>
>Although the GS-35B doesn't really need a socket as we know them, it is 
>best cooled using a mounting fixture that assures a good flow of air 
>around the filament, cathode and grid connections as well as the anode 
>cooler.  There is at least one commercially manufactured amp using the 
>GS-35B tube that has the tube directly fastened to the chassis. It uses 
>an air box above the chassis into which air is blown to cool the anode. 
>Unfortunately, there is really not a good flow of air around the 
>filament and cathode connections. There are several home made amps that 
>are mounted similarly and each has its own solution for cooling. 
>Unfortunately, those I have seen suffer from insufficient cooling of 
>the base of the tube or require an additional blower to cool the base.

The method of cooling by blowing the cold air in above the chassis can 
be made to work perfectly well - indeed, it is the *normal* method of 
cooling in many VHF/UHF amplifiers, both amateur and commercial.

The "blow from below" method and the "VHF" cooling method each have 
their own particular advantages and disadvantages.

The big advantage of the "blow from below" method is that there's 
absolutely no doubt about the cooling of the heater/cathode area. It 
sees exactly the same air flow as the anode - it even gets the coldest 
air, before the anode does. But the heater/cathode area of a GS-35B 
dissipates only about 35W, so it really doesn't *need* the same air flow 
as the anode, which has to dissipate anything up to 1500W.

The disadvantage of the "blow from below" method is that the air-flow 
resistances of the socket and the anode cooler are both in series. This 
creates a high back pressure, which greatly reduces the flow rate 
available from most blowers. Most centrifugal (squirrel-cage) blowers 
have a very nonlinear pressure/volume characteristic, so even a small 
increase in back-pressure can cause a disproportionately large decrease 
in volume flow.

Remember, what actually cools the tube is the flow rate (not back 
pressure as some mistakenly believe).

The big advantage of the "VHF" cooling method is that the flow 
resistances of the upward and downward air paths are in parallel. 
Compared with the "blow from below" method, this greatly reduces the 
back-pressure. Now the nonlinear characteristic of the blower works to 
your advantage, and produces a *large* increase in total air flow. The 
same blower delivers more than enough extra air to compensate for 
splitting the flow into two separate paths.

The disadvantage of the "VHF" cooling method is that you have to choose 
the size and location of the holes that bleed air downward past the 
heater/cathode seals. This is an area where mistakes can be made. People 
underestimate how large these holes need to be. They also underestimate 
how large these holes *can* be, without significantly affecting the 
anode cooling. Also the holes are often too far away from the tube, so 
the downward blasts of air go past the heater/cathode seals, missing 
them completely.

A recommended method for a GS-35B mounting is to drill four 3/4-in holes 
around the tube, and then fix four 3/4-in pipe elbows below the chassis 
to direct the air flow inward and onto the heater/cathode seals. (You 
can glue the elbows onto the chassis, using silicone or hot-melt glue.) 
This gives the heater/cathode all the cooling it actually needs.

Also a vent is needed to let this air escape from under the chassis. 
Another common mistake is to make this vent too small, so it chokes off 
all the downward air flow. Make the total open area of the vent much 
larger than the total area of the holes in the chassis.

With attention to these simple (but important) details, the "VHF" method 
of cooling works very well indeed.



Tony continued:
>That problem combined with keeping the stray inductance in the grid 
>circuit to a minimum is what drove me to build the mounting fixture 
>(socket) the way I did.

Just to be clear, the "stray inductance" problem that Tony mentions is 
the inductance that you'd get from using four metal pillars to support a 
mounting plate below the chassis. Both Tony's base and the simple "clamp 
direct to chassis" method avoid that problem.



-- 
73 from Ian GM3SEK         'In Practice' columnist for RadCom (RSGB)
http://www.ifwtech.co.uk/g3sek