Amps
[Top] [All Lists]

[Amps] water cooled tube or transistor PAs

To: amps@contesting.com
Subject: [Amps] water cooled tube or transistor PAs
From: "John T. M. Lyles" <jtml@lanl.gov>
Date: Fri, 11 Mar 2005 10:42:05 -0700
List-post: <mailto:amps@contesting.com>
Look up Thermacore in Lancaster PA, they are 
making some innovative heat pipe cooling loops 
for integrated circuits and other applications.

As for adding sodium or other ions to DI water: 
this will increase the conductivity greatly and 
cause current flow in the hoses. Eventually leads 
to migration of the copper and brass fittings, as 
deposits along the inner walls of the hoses. 
Also, this will erode away the fitting where the 
hoses connect to the tube from the mA flowing. In 
our water cooler RF amplifiers, I try and keep 
current under 250 uA in the hoses if possible. 
When we have large OD hoses like 2 inches or more 
(for big toobes) this is a problem, and it 
requires length to get enough resistance between 
the tube and ground, in each hose. In a hose such 
as 1 inch PEX (which is sold at Home Despot and 
'Lows' for home plumbing) about 60 inches are 
required with 5 Megohm-cm water conductivity 
(very clean). We trip HV interlocks on many of 
our systems if the Foxboro conductivity meter 
indicates below 5 Megohm-Cm. This indicates time 
to change the Culligan bottles. Call Culligan 
man. For smaller hoses like 1/2 and 3/8 inch 
(more likely at the 5 kW and under level) lenght 
is shorter. Just treat the hose full of water as 
a big resistor in your ohms law analysis.

For occasional use, be sure that the water 
quality is still good. There must be test kits 
available? If you open the system to air, 
dissolved oxygen is a bad thing. We use 
deoxigenator bottles as well as RO bottles. We 
keep it below 5 ppb as recommended by CPI/Eimac.

Thales tube company sells expendable electrolyic 
anodes which are placed inside the water fittings 
at the connections to their tubes, so that you 
sacrifice them, and run higher conductivity 
water. It can be as low as 100 kohm-cm using 
these devices. But they need to be checked 
annually in 24/7 operation, as the tips will 
erode away to below the edge of the brass 
fitting, then you expose the fitting to 
erosion/corrosion.

One more point, do not use regular steel in the 
plumbing for a DI cooled loop. It must be copper, 
brass, or stainless steel.

I agree that water cooling is a proper subject 
for some QRO systems. Probably has been beat to 
death here this past two weeks though! Try it, 
you'll like how quiet it is.  BUT, as you all 
know, HV and water don't like to mix.... However, 
for solid state, its relatively painless:

I helped design and install two 2.8 kW CW water 
cooled amplifiers at 805 MHz, at work here in 
1994-97. They replaced a room full of RCA 
Cermelox Tetrodes in cavities, which had been 
made obsolete when Burle Industries quit making 
the tubes. Our complete 40 dB amplifier with 
power supply fitted in one tall 19 inch rack. We 
used a slug of MRF899 bipolar transisitors, as 
LDMOS wasn't the rage then. You can read about 
the amplifier and see photos in the 1998 RF Expo 
Proceedings from San Jose, or send me a note and 
I can forward a copy.

73
John
K5PRO


>Will writes:
>>
>>  Personally, I dont see anything wrong with the posts as they
>>  pertain directly to water cooled tubes. If some addative can
>>  be used to cool a tube better, this is an amp forum, and it
>>  should be discussed. This especially if the addative can
>>  prolong the life of a very expensive tube.
>
>Neither do I see a problem.  Fact is, by this time next year
>I expect to see a water cooled solid state amplifier available
>at Dayton.  There should be some 120 - 200 volt transistors
>available at a reasonable price per watt by then the problem
>will be keeping them cool with a typical heatsink/fan.  A
>water cooled substrate (copper block) with a small closed loop
>system would make for a very nice package.?
>
>I hope we see some 1500 Watt (CW) output 100% duty cycle solid
>state amps (4 transistors) at reasonable prices (competitive
>with the Ameritron, TenTec, Commander, etc.) within the next
>two or three years.
>
>73,
>
>    ... Joe, K4IK
>
>
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps

<Prev in Thread] Current Thread [Next in Thread>