water cooling

There are 1200 volt pulses on the drains of the mosfets, which is where the heat comes out. What I have now is an AlN insulator and then the cold plate, which could be optimized, but I'd really like something more radical. Water flowing directly on the fet case would be ideal, but probably not practical. Water flowing over the bottom side of the AlN insulator might be practical, with some interesting machining.

In my bench setup, water is flowing through those hoses into a commercial cold plate. The assembly aluminum baseplate is bolted to that. The fets are insulated from the baseplate with the AlN ceramic slabs. Lots of thermal interfaces: Drain:AlN:baseplate:cold plate:water. All but the last with silicone grease.

The baseplate is milled, so has some surface roughness. I really need a near-optical surface finish between the AlN and the baseplate/cold plate. The AlN is lapped to a very fine finish, and the mosfets are very flat. Thermal silicone grease will squash down below 100 micro-inches (my measurement resolution) IF the mating surfaces are flat.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics

?

With properly de-ionised water it would be perfectly practical.

When I was in the Nijmegen science faculty workshop, one of the graduate st udents set up 40kV water-cooled X-ray source, with the 40kV water-cooled me tal anode insulated only by the - very carefully de-ionised - cooling water .

It worked. There would have been enough CO2 diffusion into the circulating water to make it slightly acid and slightly more conductive than pure water (which is 10^-7 molar for both hydrogen and hydroxyl ions) but it was good enough.

Your 1200V pulses should not be a problem.

Since steam is totally non-conductive, a boiling pool of water sitting dire ctly on the FET, renewed by a steady stream of droplets of condensed water falling back through the steam, could be even better insulated.

A ceramic wick in a rather freaky non-conducting heat pipe - or a least a h eat-pipe containing insulating segments in the right places - could work to o.

Unleash your ingenuity! Though this wouldn't seem to require a full brain-s torm to fill in the details.

o

ow.

f

. 1

urity Re-

google.

It doesn't need to be. Properly de-ionised water is a good enough insulator .

--
Bill Sloman, Sydney

You have two types of corrosion to worry about. One is galvanic corrosion where dissimilar metals are in contact, such as aluminum and steel. You can minimize this by using the same metal for the heat sink and the fittings that attach to the tubes. In this case, a copper heat sink with copper fittings would be ideal.

The other type of corrosion is through electrolysis. The original copper telephone wires were biased with a positive voltage. This soon resulted in the copper eaten away by stray currents to ground. The solution was to use a negative voltage, which is why the fast-disappearing copper telephone wires use -48V bias.

In your case simply grounding the copper heat sink with a ground strap should work well.

Water flowing on the fet case would have little effect due to the small surface area. The anode wires would be quickly eaten away through electrolysis.

Instead of pure water, you should use a recognized coolant such as Prestone

50/50, available most everywhere:

coolant-0293023p.html

Since the heat sink is grounded, you don't care about the dielectric constant of Prestone.

The last concern is the water pump. The case should be grounded to the same ground as the heat sink.

This arrangement should give a very long life to your apparatus.

Not if you don't expect the system to be exposed to below zero temperatures.

Corrosion inhibitors for mixed metal central heating systems do all the corrosion inhibition required. The one's I used were added as 1% of the water in the system. The latest version seems to recommend 4%.

Very true.

Nowhere near as long as a heat pipe. the only moving parts in a heat sink are the vapour moving to the cold end, and the liquid wicking back.

Nothing gets eroded.

--
Bill Sloman, Sydney

On Oct 21, 2018, snipped-for-privacy@ieee.org wrote (in article):

Are you sure that these are mild steel? The sheet metal radiators I?ve seen in Europe are made of some kind of stainless steel, as are the tubing and fittings connecting the radiator. I never saw rust or discoloration on any exposed metal.

I would think that it would be easy to arrange a permeable barrier that would alsow the Hydrogen to diffuse out to the atmosphere while keeping water et al in place.

Hmm. Circulating hot water systems in the US are equipped with air-release valves, located at the highest point in the system, and hydrogen can escape via these valves as well. (The air enters dissolved in the makeup water.)

The air-release valves of my experience are based on a compressed stack of vulcanized fiber rings. When wet, the fiber expands and seals, just like oakum in the boat hulls of yore. When dry, the fiber shrinks, allowing gasess to seep through. Eventually, the fiber wears out and the valve seeps water all the time, and the valve must be replaced. This happens every ten or fifteen years.

Most of the electronics cooling systems of my experiuence use an antifreeze of some kind, with corrosion inhibiters. Some systems use pure deionized water, but this tends to be fussy to maintain. The reason to want pure water is that it?s less viscous, and so flows more freely.

Joe Gwinn

Not so complicated at the first place I worked -- use copper pipe and brass/bronze fittings, and use silver solder (braze) when arbitrary joints are needed. They had far more problems with those push-on plastic fittings rocketing off, than corrosion in pipes. Aluminum was a strict no-no for them.

Typical customer water was somewhere between tap water and automotive antifreeze, AFAIK. It's expected to scale shut at some point, hence the flow meters and thermometers on various water circuits we designed into the systems.

We supposed that anodized aluminum would probably be fine, as long as it's not touching anything else -- avoiding galvanic couples -- but that it would only last as long as the anodize, which would be pretty good in a clean, closed system, but impossible to assure in an arbitrary system with grit in the line and who knows what else.

Tim

--
Seven Transistor Labs, LLC 
Electrical Engineering Consultation and Design 
Website: https://www.seventransistorlabs.com/

rote:

m

or

ink

e

cubic

on

ong.

hange

ermal

mum;

ld

sed

y t e g

be

g

er

We didn't buy any of them ourselves. Bits of exposed metal (when the paint had been nicked) did show some signs of rust, so they probably weren't stai nless steel.

y

The house was built in 1936, with cast iron radiators. We bought it in 1994 , and were more modern radiators had been added over the intervening period .

or

ould

t al

Sure. You can buy replacement "vent valves" that seem to work that way.

You have to put them at the highest points in the system. We had two - one in my study in the attic and another on another pipe loop on the second flo or below it.

I didn't bother getting one. I liked to know how much hydrogen the system w as generating.

e

pe

European systems - or at least the one we had - are fully sealed, with an e xpansion tank to take up any volume variations. I had to pressure it with m y bike air-pump every few years.

f

sess

r

a

as

s a

tal

e

Not in a fully sealed system.

It's also cheaper.

--
Bill Sloman, Sydney

On Oct 22, 2018, snipped-for-privacy@ieee.org wrote (in article):

Ahh. I never looked at anything that old.

.

US systems are also fully sealed to the degree possible without true hermetic seals everywhere, and also have those expansion tanks. But water always seeps out and air always enters through the various moving seals - it?s a matter of rate.

Was it really 600 liters? I recall draining my system into a couple of buckets many years back.

And therein lies the problem: the circulator pump seals are not perfect, especially while the pump runs. Nor are the valve stem seals.

.

In those applications, coolant cost doesn?t much matter. Long term stability and low maintenance costs are more important. Not to mention heat transfer rate.

Joe Gwinn

Plastic tubes and a $16 aluminum cold plate: I don't have a lot to lose.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics

Good point. You might want to send to the owner of the "Atomic Rockets" web pile: The site specializes in trying to bring some reality into space flight, space travel, and spaced out science fiction authors. Maybe add it to the "misconceptions" page:

See the index at the bottom of the home page.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558

from

eld or

at sink

spite

es.

afer

n a cubic

lution

e wrong.

for

to change

t thermal

optimum;

would

uncased

Every

that

that

l the

5 deg

ght

uid

int

and

n any

1994,

od.

op

p

t it

at

ng

one

em

lease

eup

an

ith

etic

eeps

a matter

ck of

ike

g

seeps

y ten

ter a

dded

that

tem

g up

This was a very old system, the cast iron radiators held a lot more water t han their modern equivalents, and some of the connecting pipes had a larger diameter than you'd find in a modern system. The house was pretty big as w ell.

reeze

zed

Negligible in a 600 litre system.

m

at

True, but nobody ought to spend more than they need to.

--
Bill Sloman, Sydney

I've never seen stainless steel radiators in the UK. All the (domestic) radiators I've seen are either pressed steel (plus rust) or cast iron. The latter are found in architectural salvage yards.

Driving a Beetle in slippery weather was like trying throw a mallet with handle first.

I survived one Nordic winter driving a Beetle (model 1951).

--

-TV

I'm not the first one to notice it, at least one of the producers who worked on some of those shows (name escapes me) noticed the same thing and went on to play a larger role in the later Battlestar Galactica series re-make. wWere more "realistically", or at least as realistic as you can make a show about technology that doesn't exist, the Galactica was a battered, ruined hulk on the verge of structural collapse at the end of several years in space...

The several direct hits from nuclear bombs probably didn't help much either despite the super-duper duranium magic alloy the hull was made from or whatever

We don't get ice but before I was born, my uncle was crossing a flooded creek in his beetle and it began to float. Luckily it was just the bridge which was under water and it had side railings so he managed to get across instead of floating away!

Then again, it was part of the pedigree:

--
Cheers, 

Chris.