From: Jim Garland [mailto:4cx250b@miamioh.edu]
Sent: 16 March 2013 13:43
To: 'Ian White'; amps@contesting.com
Subject: RE: [Amps] Coupling a blower to an air system socket
Another very effective method of cooling is to blow air directly into
a sealed anode compartment. Most of the air flows upward through the
anode cooler and is vented directly to the outside through a chimney
ABOVE the anode cooler. There is NO chimney between the base and the
anode cooler.
Meanwhile 25-30% of the air flow is allowed to bleed downward through
the tube socket to cool the base seals. This method reduces the back
pressure on the blower, and allows it to deliver much more air than
the conventional base-upward layout. It has been used very
successfully for decades in VHF and UHF amps - so much so, it is
regarded as "the normal method".
Very interesting concept, Ian. I'm wondering how the 25-30% downward
flow past
the filament pins is adjusted? Presumably one needs to size an outside
vent on
the underchassis to exhaust that air. Also, I don't quite understand
why the back
pressure on the blower is reduced. Seems like most of the air has to
flow up
through the anode cooler, which presumably is the largest flow
impedance. The
air flowing past the filament pins has to have a bottleneck where it
vents from the
enclosure to keep the flow down to 25-30%. . The combination of the two
vents
(the filament vent and anode cooler vent) presumably reduces the back
pressure
slightly, but I wouldn't think the effect would be very great.
On a related topic: In my experience, a problem with blowers is often
that the
motor rpm is too high, causing turbulence in the airflow. As noted by
somebodly
else, turbulent flow is less effective at cooling an anode than laminar
air flow. This
fact was known by the Collins engineers who designed the 30S-1 cooling
system.
They mounted a low speed blower directly under the tube socket, powered
to give
laminar airflow through the 4CX1000A. Some hams (misguidely, in my
opinion)
swap the orignal 4CX1000A for a 4CX1500B, in the hope that the 1500W
plate
dissipation of the latter tube willl provide a larger safety margin.
Unfortunately, the
reverse happens, because the fins in the 4CX1500B are much more densly
packed, which inserts additional flow impedance into the air path and
causes
turbulent flow. The actual net effect is to reduce the cooling and,
hence, decrease
the amplifier performance.
73,
Jim W8ZR
Sorry, Jim, but that is exactly backwards. Laminar flow is good for
aerodynamic design where the objective is to minimize drag, and
turbulence is your enemy. But in cooling applications the objective is
to maximize the heat transfer from the hot metal into the cool air...
and for that purpose, turbulence is your friend.
Laminar flow is slow, smooth and orderly. A defining feature of laminar
flow is that all of its streamlines (the lines that you'd see traced out
by thin streamers of smoke) are parallel. The highest velocity is in
the middle of the duct, tapering away to zero in the "boundary layer"
alongside the walls of the duct. Laminar flow with a static boundary
layer is great if your objective is to minimize drag; but laminar flow
is bad for air cooling because that stagnant boundary layer acts as an
insulating blanket.
Turbulent air is the exact opposite - quick, swirling and chaotic. The
turbulence breaks up the blanketing boundary layer and is far more
effective at transferring the heat away from the surface and into the
flowing air.
The air flow into a blower is generally quite laminar; if you trail a
streamer of smoke into the air intake, you can see that the streamlines
hold together and remain substantially smooth and straight. But once it
enters the blower, the air is stirred up violently by the high-speed
blades and comes out highly turbulent. This turbulent air at the blower
outlet is the most efficient means of cooling available, so ideally the
blower should always be just upstream of the tube.
The finned anode coolers of tubes like the 4CX1000 and 1500 are a form
of heat exchanger, and the fins are intended to increase the surface
area available for heat transfer. But this creates a large number of
very thin airways, which force the air to flow straight and parallel to
the fins - no matter what's happening outside of the anode cooler, the
air flow inside is *always laminar*. We'd like it to be turbulent, but
the spaces between the fins are simply too small to allow any whirlpools
and eddies to form.
The reason why it doesn't work to swap a 4CX1000 for a 4CX1500 is that
the 1500W dissipation rating requires more air to be forced through the
narrower gaps inside the cooler. If you don't change the blower as well
as the tube, that isn't going to happen.
Another major part of the problem is that small blowers are not very
good at generating the pressure that is needed to drive a sufficient
volume of air through the close-spaced fins of the anode cooler. A small
increase in back pressure can cause a disproportionately rapid reduction
in air throughout, which is known as "choking".
The traditional cooling method is to blow air into a sealed grid
compartment and then upward through the base, chimney, anode cooler and
exit chimney. The problem is that each of these items creates some
back-pressure and they are all connected in series so the back-pressures
add together. You are constantly fighting against the characteristics of
the blower and its tendency to choke.
The method of cooling by blowing air into a sealed anode compartment was
first popularized by the revolutionary K2RIW amplifier design for
432MHz. It was then exploited by Fred Merry, W2GN, whose amplifier
designs for 50 through 220MHz are detailed here:
<http://www.newsvhf.com/w2gn.html>
As I said, this method of cooling is completely normal in the world
above 50MHz. The advantage of this system is that it places the flow
resistances of the anode cooler and the base in parallel. Back-pressure
drops dramatically and the same blower can push a much larger flow rate
through the anode cooler. The blower characteristics are now working in
your favor.
Jim, you were quick to notice the need to regulate the fraction of the
total airflow that is directed downward to cool the lower part of the
tube, but this is surprisingly non-critical. If the tube is mounted in a
conventional base, that limits the downward air flow so all you need to
do is seal the grid compartment and provide a screened vent of a few
square inches. It's actually quite hard to get this wrong - if you make
the vent too large, the blower will compensate by delivering more air
without "robbing" the upward flow through the anode cooler.
73 from Ian GM3SEK
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps