Ed Huntress wrote:
On Thu, 15 Mar 2012 14:40:40 -0400, Ed Huntress
wrote:
On Thu, 15 Mar 2012 11:05:08 -0700,
wrote:
Hey Ed.
You were going to tell me about the time you created a
Peltier-based test jig destined for Japan's Ministry
of International Trade and Industry to be used in Ocean
Thermal Energy Conversion research back in the 80's.
You used a Bridgy, a SB lathe and an ancient Walker
Turner drill press. Is that right?
g Oh, yeah, I remember that...
Okay, a long tail...
I left AM to pursue a juicy contract with MITI (Japan's Ministry of
International Trade and Industry), to produce reports on new materials
applications. One of their big energy projects was OTEC (Ocean-Thermal
Energy Conversion). I wrote a paper on corrosion-resistant materials
for that project and then, in discussions with them, they expressed an
interest in the new semiconductor thermoelectric cells, as possible
candidates for electricity generation from
low-temperature-differential, high-volume seawater. They actually use
the Seebeck effect, but Peltier and Seebeck are like yin and yang.
Anyway, they said they wanted some of the new cells from a company in
Trenton, NJ, to test and evaluate. I drove down and bought a few dozen
for them. When I delivered them they said they wanted to get started
quickly because they had contracted with a company in Japan to build a
refrigerator-size test module and it would take a couple of months to
get it, and then another month to run tests.
I had a design in my head for a four-cell test unit that I could make
in one weekend. They said it wasn't enough cells. So I explained that
you could change cells in it in less than five minutes, and they could
easily test 20 cells in a day. They they could run some statistics,
figure the variance and standard deviations from a few day's worth of
testing, and have the results they wanted. So they threw some money at
me and said go ahead.
It was really simple. I took two 6-inch squares of 5/8" aluminum
tooling plate (2024 -- it was all I had) and cut a serpentine groove
in one side of each plate with a 1/2" end mill (mill -- I had to use a
friend's Bridgeport). Then I turned some custom barb fittings from
brass (turn). I drilled holes to clamp the two plates together,
leaving room for the Seebeck cells, and then drilled and tapped holes
for the barb fittings (drill). There were two barb fittings on each
plate, at opposite ends of the serpentine groove. Through one plate
you ran cold water. Through the other, warm water.
I placed four cells between the plates, smeared them with conductive
silicone grease, and clamped them together.
I guess your clamping pattern assured that every cell
had an equal share of hot and cold sink coupling.
That apparently is really difficult to do.
One of our own disassembled a shorted PWM control for an RC application
and found that say 30% of his SMT MOSFETs barely made contact
with the heat sink. (You could really see the problem by inspecting
the 'sil pad' that was wedged between the heatsink and devices.
Some FETS made quite a deep impression and some were not apparent
at all.)
Then I hooked aquarium
hose to the barb fittings. One pair of leads from each cell projected
out of the space between the plates. Ta-dah. About five hours of work,
IIRC.
You could measure the volume of water and the in-and-out temperatures
from each plate, and the electrical output of each cell. You can see
the implications -- a simple way to measure energy in/energy out at
different absolute temperatures and temperature differentials (there
was lots of insulation on the whole thing, in use) and to test the
output variance among cells. Very, very simple.
Sounds like Science, too. One could bolt on thermocouples for instance.
That's the story. They loved it. And they decided not to invest in
OTEC.g
Ouch! That's like deciding not to buy a car because all
you could test were Yugos.
I love the TE concept but despair over their inefficiency.
I suspect that Stirling motors would work the best in
this application, no? ('Sounds like a great purpose
for 'flare gas' at last.)
Oh, I forgot the Plexiglass covers. The grooves were on the outside of
the sandwich, and I covered them with 1/8" Plexi. That way you could
see if a bubble was trapped in there that would screw up the output
readings from one cell.
Very nifty!
That'd be a way to create a solar heat collector, too.
I expect the plexi would have to be replaced with another
aluminum sheet, though.
--Winston