First, I share your views that this question should be addressed from the
efficiency aspect.
Heat that is not generated does not need to be cooled away.
But, if this is not possible,my vote is for using deionised water with
corrosion inhibitors added,such as anti-freeze compounds.
As long as the water is allowed to stand for long periods in the cooling
system, it would be possible to use ordinary tap water, taking account for
scales formation which could easily block the cooling channels. Tap water would
have sufficient resistivity for all reasonable supply voltages for LDMOS
amplifiers.
My own professional experience from liquid-cooled amplifiers comes from the
ITT-Standard Radio SSA1000 and PA204 HF and UHF 1 kW+ CCS output solid-state
amplifiers. They had a very "tough" specification from their military (Air
Force) users, that they should stand operating and standby ambient temperatures
of between -40 and +60 C and be able to "cold-start" immediately from even the
coldest conditions. Also,there should be no risks for contamination of the
cooling system by very long stand-by or storage periods (years). This ruled out
water-based coolants, and the designers chose the 3M Fluorinert fluid which is
completely chemical inert,non-freezing and non-corrosive. The down-side of
being very expensive and its low viscosity made it easily escape from even
minute leaks.
SSA1000's and PA204 amplifiers were for these reasons quite low in demand, as
"filling up" with Fluorinert was beyond the allowable price range of most
amateurs and some professional users. They preferred Collins HF-80's and
similar air-cooled PA's.
However, chance made it that got one of the design engineers for the SSA1000 as
a colleague in 2002.
We came to discuss the cooling system. He told me that the very "soft"
Stockholm tap water had been used in all development work and during early
performance verifications except for the full ambient temperature range tests.
There were no ill effects noticed from the use of this "simple" coolant. Also,
the mess of cleaning up the lab floor after being flooded due to leaks became
considerably easier...
This coincided with a "gift" of two SSA1000 systems from a defunct diplomatic
HF network. The Fluorinert had since long escaped, so for getting them running,
anti-freeze and deionised water in a 20/80 mixture was used. This worked well,
and the amplifier ran considerably cooler using water than with Fluorinert. A
"soak test" at 1300 W carrier for several hours during a summer day netted a
coolant temperature rise of about 30 C with the cooling fan running all the
time.
I parted with these amplifiers after losing interest in HF QRO some years ago,
but as far as I know, their new owner still run them with water coolant.
73/
Karl-Arne
SM0AOM
----
Ursprungligt meddelande----
Från : manfred@ludens.cl
Datum : 2017-03-06 - 15:04 (UTC)
Till : amps@contesting.com
Ämne : Re: [Amps] SS amps watercooling - was PowerGenius XL
Okay, folks, can we then perhaps reach some consensus, or recommendation
about what exact liquid to use in watercooled electronics?
It has to be water-based, because there is simply no other liquid that
matches water in terms of specific heat, availability and cost.
We don't want buildup of anything in out heat exchangers. So we must
make sure that the liquid is pretty much free from calcium, magnesium,
carbonate, silicate, sulfate and other such ions that tend to
precipitate on hot metal surfaces and form scales. This rules out tap
water (unless it's very soft).
We don't want algae growing either. This pretty much requires some sort
of alguicide added to the liquid. Otherwise we WILL get them, in the
long run. What to use? Chlorine ions? Hypochlorite? Some alcohol?
We don't want corrosion. But a tiny amount of corrosion should be no
problem. If we load the amplifier with distilled water, and that water
rips some copper from our tubing and heat exchangers, until reaching a
balance, is that so bad? I think we can live with it. If not, then we
would need some sort of corrosion inhibitor if we start from distilled
water.
Conductivity may or may not be an issue. With a typical LDMOSFET amp,
that has the source (and cooling block) at ground potential, and every
other metal part of the cooling system at ground potential too,
conductivity is a non-issue, as long as there are no metals with very
different electrochemical potential there (because then one of them
would corrode the other, by electrochemical action). It should be easy
enough to make the whole system just from copper and plastic, so that
there would be only one metal, and no electrolytic corrosion. Then we
simply don't need to worry about conductivity.
But if I ever build the amp I planned years ago, that uses cheap
switching MOSFETs, I would have two cooling blocks at drain potential.
It would be easy to keep the entire water circuit at the supply voltage,
so there would be no DC between any parts of the water system, but there
would be RF voltage between the two cooling blocks. Typical water
solutions have a resistivity that's plenty high enough to cause only
negligible loss, but would the tiny RF current between the two cooling
blocks cause harmful electrochemical action? If yes, then an insulating
liquid would be needed.
Cars typically use distilled water mixed with concentrated "antifreeze",
which is usually a mix of glycol (the antifreeze proper) with corrosion
inhibitors, and possibly alguicides, etc. We don't need antifreeze
action on our amps, used at room temperature. And in countries like
mine, where in most places the weather never gets below freezing, car
parts stores sell a refrigerant for cars that doesn't contain
antifreeze, but just corrosion inhibitors and a lubricant for the water
pump. Would perhaps this stuff, mixed with distilled water, be a good
candidate? In cars it's typically used at a concentration of just 1 to 3%.
A note on rainwater: Even in areas free from industrial pollution (if
such a place exists...), rainwater is still moderately acidic, because
the rain drops absorb carbon dioxide from the air, forming carbonic acid
(the same stuff that in much higher concentration makes soft drinks
fizzy). I live in a pretty clean, non-industrial forest area, and some
months ago I measured the acidity of freshly collected rainwater. Its pH
turned out to be 5.8.
And another note, on tap water: It doesn't seem to be terribly
corrosive, at least not to copper, brass and solder. Many houses here
have copper water pipes and brass fittings, soldered with either
tin-lead solder, or the more modern ones with pure tin, or with alloys
having mostly tin and a small amount of some other metal, and these
installations run reliably for many years, with hot and cold tap water.
Only in areas that have extremely hard water, scaling can become a
problem after several decades, mainly when scales shed off and
accumulate at the lowest spot of the piping, causing a clog there. But
corrosion doesn't seem to happen.
Instead when zinc-plated steel hot water tanks are connected to copper
piping, this forms a shorted galvanic battery! Given slightly acidic
tap water, it eats the zinc in a short time, and then starts eating the
steel below, and we get rusty hot water! Instead with slightly alkaline
tap water there is no big problem.
Anyway I can't avoid the feeling that it would be better to develop good
high efficiency linear amps, so that cooling requirements become very
low, and nobody would have any need to consider watercooling.
Manfred
========================
Visit my hobby homepage!
http://ludens.cl
========================
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
|