On 8/19/2011 11:26 PM, jeff millar wrote:
> On 08/19/2011 07:05 PM, Roger (sub1) wrote:
>> I really don't see wide adoption of SS legal limit amps until the
>> devices can be made more rugged, can be cooled more easily, and become
>> available at much lower prices. Prices are based on tooling set up
>> costs, number of units produced per batch, amount of silicon in the die
>> compared to the base wafer size, and yield efficiency. At present they
>> are just too fragile and the protective circuitry (which is more
>> complicated than the amp) really pushes the prices far too high for wide
>> spread adoption.
> I used to think the same thing...but not any more.
>
> The Freescale MRFE6VP61 can withstand 60:1 VSWR for 10 msec, which is plenty
> of
Try as I might I've not been able to find a data sheet on the MRFE6VP61.
When it gets to 5 or 10 seconds then I'd be comfortable. I do not like
having to depend on fast acting protective circuits. You are still stuck
with the sensitivity to over voltage spikes which must be suppressed one
way or another. So the reliability of your device becomes the
reliability of the protective circuit.
> time for a simple controller to shut it down. That part costs $271 for about
> 1000W out...a lot cheaper than a 3cx800. Transistors will get cheaper. Tubes
> will get more expensive.
That part I agree with, but I'd be looking at at least 4 of them in PP,
parallel for a comfortable delta T at 1500 out and they still wouldn't
lend themselves well to running the legal limit in class A.
>The commonly used reference design for the amplifier uses a large block of
>copper to spread heat into a large Aluminum finned heat sink. But water
>cooling
>makes better sense.
Water cooling is almost as easy to implement (with the proper
precautions) with high powered tube amps.
With tubes and SS You also have to be careful about the coolant temp and
humidity. I've worked with water cooling from both perspectives.
Condensation can ruin a solid state amp almost as quickly as a voltage
spike. For the best efficiency the water needs to be a cold as possible
without causing condensation to maximize the delta T. The higher the
delta T the better the cooling. The same is true in computers. Actually
the liquid cooling systems used in computers work very well with solid
state amps.
> Mount the transistor to about a 1/2 inch square water pipe
> to minimizes the thermal path from transistor to water and use the water to
> spread the heat rather than a lot of heavy expensive machined metal.
You do need a certain amount of metal in there above the minimum
required for heat transfer, or thermal inertia to take care of cooling
failures. It really doesn't take much water to remove a KW of
dissipation which is capable of well more than the legal limit.
Unfortunately cooling systems can run the gauntlet from very simple to
elaborate. The simplest is to just throttle tap water through the heat
sink but from a conversation approach is not very efficient when it
comes to resources. It, at minimum needs a solenoid valve to turn the
water on and off to eliminate the "memory" factor.
These figures may sound like a lot but 1500 watts for an hour is
1,290,631 calories. To hold equilibrium that is 1,290,631 grams of
water per hour. One gallon of water = 3,785 grams so that is roughly 341
gallons per hour or 5 gallons per minute BUT it takes one calorie to
raise one gram of water 1 degree C. If we let it raise the water 5
degrees we've just dropped the water requirement to about 1 gallon per
minute. (and increased the cooling efficiency) If we are willing to
raise the water 10 degrees (which is a typical delta T) we are now down
to about a pint per minute (give or take a bit) and these figures are
for continuous power. For tubes you can average it out due to the mass
of the cooling elements. SS devices have only a tiny fraction of the
thermal mass which makes it desirable to have more copper than the
minimum necessary for the proper operation when the system is working.
IOW I'd want a copper block roughly a 1/4" thick by 2" on a side for
thermal mass. For the average system this would not add much if
anything to the over all physical dimensions but would certainly add to
the reliability.
BTW when you figure 20% duty cycle for SSB AND the typical 10 minutes
out of 30 transmit time the required water becomes miniscule.
The best "to me" are the refrigerated systems that cool a reservoir
down to just above the condensation point. These systems work well
summer or winter as they put the excess heat outside (or inside if you
wish). But being mechanical and requiring circulating pumps in addition
to the refrigeration system are more complicated and more prone to
failure, which means at least a method of shutting down the amp if the
cooling system fails. IOW coolant temp and flow.
The main thing I like about water cooling for high powered amps...IT's
SILENT! No blower noise. OTOH you still need a blower on the tube
seals where with SS cooling the drain cools everything.
> Transistors dump heat at ground potential while tube dump heat at high DC and
> RF
> potentials. That makes cooling a LOT easier.
Depends on what you are used to. If you have been working with several
hundred KW that was water cooled, then water cooling a legal limit amp
is not all that difficult. Also almost any external anode tube can be
easily converted to water cooling. OTOH it is true that water cooling a
tube with 3 or 4 KV on it takes a bit more plumbing than a grounded heat
sink and there are safety considerations.
OTOH you are running a far higher delta T which means a more efficient
cooling system. The downside for HV water cooling is the necessity for
using either distilled water or deionized water which are both highly
corrosive where metal and particularly copper and brass are concerned.
With the grounded heat sink you can use plain old tap water although
when recirculating it you do have to treat it as if it were a swimming
pool. IOW Algae and bacteria can be serious problems.
My choice of water/liquid cooling would be a 10 gallon PVC tank using
either direct flow to the amps or a heat exchanger in the tank made of
copper tubing. The water would be cooled by a modified air conditioner
negating the need for running any of the coolant outside which here in
the frozen North presents a number of hazards. Each amp would have flow
and temperature monitoring. Loss of pressure would shut down the
circulating pumps and close safety valves to the reservoir. The reason
for the heat exchange would be that little coolant would be lost even in
case of a piping failure.
> The cooling system becomes a transistor mounted on a short length of square
> pipe
> that has some hoses leading to a bucket for thermal mass or maybe a fan cooled
> heat exchanger.
My cooling systems would be some what different than that arrangement.
I want the heat dumped outside in the warmer weather and inside during
cold weather, but without affecting the humidity in the house or shop.
>I'm designing a transistor control board that uses a FET to switch the
50V drain supply. The micro-controller cuts off >the amp in about 2 msec
on any fault conditions. The FET doesn't need protection from drive
without supply >voltage. Homebrew amps can use a 3000W supply available
on ebay for $25. Commercial manufacturers will have >to pay real money
for a switcher, but you get power factor correction, light weight, and
huge volume discounts >because lots of supplies use the same components.
Agreed
>The trend from 12V to 24V to 50V supply has greatly simplified >output
network design because the device output >impedance goes up by 4 for
every doubling of supply voltage. >Total amplifier size will shrink and
become separate >subsystems, especially with water cooling.
Agreed, although the same concept can be used with tube amps where RF,
Control, and PS are separate modules.
I have one like that and am building several others following the same
approach. Unfortunately I've not found a switcher that would give me
around 3KV at roughly 6 or 7 KVA which is required for the legal limit
out when running class A.
As I stated previously the RF circuits are relatively simple in SS amps,
it's the control and protective circuitry as well as cooling that is
both complex and expensive.
> Amplifier pallet >smaller than a paperback. Power supply smaller than
a shoe box. Water cooling smaller than a >bucket. Control box >just
large enough for a meter and display. Multiple amplifiers can share the
subsystems.
Having worked in Industry with many types of controls, I do not like the
idea of amps sharing subsystems such as control and protective circuits.
The same is true for switching power supplies powering multiple amps at
the same time. You increase the odds of a total system failure when
taking this approach. I have no problem with the PS being switched
between amps and that works well with SS amps as there is virtually no
warm up time although the control circuitry may take a couple of seconds
to become fully operational depending on its configuration.
I would end up with an amp about twice the size of a paperback, a PS
about the size of two or 3 shoe boxes and a control board about 8 X 10
inches. Even with new components the control board should run less than
$50 USD so sharing them would save little and add to the likely hood of
failure.
That would all end up in one enclosure which would be about the same
size as today's SS legal limit amps.
>
>Parallel power supplies for >more capacity.
That is one of the main advantages of low voltage.
> Routing 50V around the shack makes more sense than routing 4 kV.
routing a high current buss around the shack scares me almost as much as
routing a HV buss around the shack and it's more of a fire hazard. It
can be done but takes a lot of precautions if properly done.
>Multiple amps can share the >same water cooling bucket.
> Multiple amps can share the same controller. jeff, wa1hco _
Even with one amp I'd want at least a 5 gallon reservoir with temp
control on it. I'd probably aim for 50F but even that could be too low
on a high humidity day. Yes, multiple amps could share the same
reservoir but I'd want separate circulating systems for each. A single
system could be used but then it'd need flow control and monitoring at
each amp. Flow monitoring is not as simple as pressure monitoring. You
measure a pressure differential across an orifice and the system will
shut down with either insufficient or excessive differential. Exhaust
temp needs to be measured right at the amp as well.
No way would I use a controller on more than one amp. They are cheap and
I'd prefer the reliability of one for each amp.
73
Roger (K8RI)
______________________________________________
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
|