X-Message-Number: 2079
Date: Thu, 8 Apr 93 12:00:10 CDT
From: Brian Wowk <>
Subject: CRYONICS Cold Reviews
Steve Harris:
> To Brian W: Why have 4 reservoirs when you can have one
> central one?
Four reservoirs make the system more redundant. If one of the
pipes or reservoirs ruptures the whole room isn't ruined. This is
critical when *all* of your patients are in it. There aren't many
scenarios that could cause such damage, but the slightest possibility
(earthquakes? terrorist bombs? bad workmanship? ) must be hedged
against. Pipe leaks cannot be fixed once this room is operating.
Also you will find that a single reservoir is not more space
efficient than four. This is because the insulation around a single
reservoir would have to be thicker to avoid sinking too much heat at
the room center. Remember the whole idea is to distribute your heat
sinks as widely as possible. That way temperature gradients are
minimized if the fans fail.
> I personally don't see the problem with LOx buildup...
Indeed. I had not considered speeding up boiling at the end
of a reservoir purge. That stratagem means that purging LOx from a
Cold Room is much easier than from a dewar (from which patients have
to be removed first).
> Your design with the LN2 in a lot of gravity fed horizontal
> floor pipes and upright pipes looks a lot more complicated than a
> simple metal plate, and moreover is likely to fill up with cold,
> and stratified air like those low open-top supermarket ice cream
> freezers, if you insist on getting "cold" in through the bottom
> and sides, and out through the top (i.e., having heat come
> through the top to be removed through bottom and sides of cell).
> Is this all worth it for top access? Might be better to cool
> everything through one thick top plate (with some -196 C re-bar
> running between the insulated sides of each cell, as necessary),
> and have each cell equipped with an individual square metal "roof
> access" plate...
My design description was insufficiently detailed. The
1 inch pipes pass >6 inches under the floor of the room in troughs
cut in the Trymer foam. Remaining space in the troughs is then filled
with perlite. In other words the pipe run under the room is heavily
insulated, with little heat transfer occurring down through the floor.
In fact most of the floor heat transfer is *upward* from the ground
below.
The insulation around each vertical pipe run is adjusted to
sink about 10 watts over its 3 meter rise. The 20 vertical pipes
around the outer room walls thus sink a total of about 200 watts,
which is the approximate amount expected to flow through the room
sides.
Finally, at the top of their vertical run, the pipes bend to
carry -196'C vapor through horizontal pipes running along the top
edges of cell walls. Vapor from the LN2 reservoirs is similarly
routed. The insulation around these pipes and the reservoir tanks
within the room will be designed to sink the expected heat flow coming
in through the ceiling. If anything, an excessively *warm floor* may
occur in this design.
"Excessively warm", by the way, means less than 1 degree
warmer. Let's say we use 3mm aluminum for all the walls and floor.
(While we're on the subject, this is 205 square meters, costing almost
exactly $10,000 from my local supplier.) No point in the room will be
more than half a meter away from a heat sink. It is then easy to show
from the thermal conductivity of aluminum and the expected heat flows
that a temperature difference greater than 1 degree cannot exist
between any two points in the aluminum walls, even with the fans off,
and maximum vapor stratification.
How does this compare to the LN2-on-top design? With a
central LN2 reservoir as the only heat sink, you would need a
continuous aluminum roof almost one meter thick (!) to achieve a similar
temperature uniformity. (To convince yourself of this, consider that a
5 meter outer circumference of the reservoir would have to draw in
500 watts gathered over the whole ceiling area. The thermal
conductivity of aluminum is about 200 watts/m/degC.) A vast improvement
could be made by routing LN2 from the above-ceiling reservoir through
pipes running along the cell wall tops and then vertically down the outer
walls. But then that is very similar to my design, and I believe that
the room is more secure and efficient with the reservoirs inside the
room rather than above it.
I also want to point out that vapor stratification is a non-
issue. Temperature gradients in the aluminum walls are the only issue.
This is because any open air spaces in cells holding patients should
be padded with fiber glass insulation. This should be done regardless
of the cooling technology we ultimately use.
> A refrigeration system that removes 400 watts at LN2 temps
> could be used only as a backup, in order to drastically reduce
> LN2 use in emergencies (we ought to see if they have a 40 or 50
> Watt one for our bigfoots).
If we install a Cold Room at the new facility, the bigfoots
may be offline in two years. No one has yet spoke up against my
suggestion of transferring existing LN2 patients to -130'C, so I
assume there is no problem with this. Of course some professional
cryobiological advice must be obtained before a final decision is
made.
> It might well be possible to get -135 with -110 freezers sitting
> around on the floor in a large room-sized freezer at 0 C., as
> CLarissa suggests, but that would be rather inefficient of space.
The proper way to do this is to have your 0'C freezer cool
only the heat sink of your nominal -110'C freezer. But then you have
built a multi-stage freezer, which is how most ultra-cold freezers
work anyway.
Thomas Donaldson:
> Furthermore, some of the mobility of current neuropreservation would
> be preserved by smaller modules: as the Dora Kent case showed,
> it may be useful or even essential to quickly move some or even all
> patients elsewhere.
The system we use to store patients need not be the same one
we move them around in. A Queue freezer with a small gasoline
generator could move selected patients off to parts unknown as
necessary. Alternatively, an in-house LN2-based system could
probably be built even more inexpensively.
The question of moving everybody at once is tougher. The
system we are now contemplating assumes Alcor will have 100 patients
in ten years. In 20 years (or less) we may have a 1000 patients.
Obviously at some point it becomes impractical to move everybody
quickly, and a stand has to be made. We may already be at that
point. Look at it this way: The money we save by implementing a low-
cost immobile storage system can hire more lawyers when we need them.
Steve Jackson:
> Your latest design still uses square cells. Is there a problem
> with hexagons? It's certainly the most efficient way to use the
> space.
I'm sorry, but it isn't clear to me how hexagons save space.
Our patients aren't shaped like hexagons, or sections thereof. If
anything, the profile of a human body looking from the top down is a
rectangle with 2/3 aspect ratio, which is probably why Hugh Hixon
thinks we can get 6 patients per square cell.
For the record, I don't have any particular obsession with
squares or cubes. The ideal cell will probably end up being a
rectangle of some yet-to-be-decided dimensions. I just don't see the
rationale for abandoning right angles.
Steve Harris:
> Before we toss the thermopile idea completely, I suppose I
> should point out that there is still one possible design to
> consider which makes use of them: If we can get a 50 C dif-
> ferential from them easily, as advertised, it might be possible
> to obtain -135 C temps in one thermoelectric stage, starting from
> a mechanically produced -79 C. The latter temp, which is dry ice
> temp, is the most commonly available temperature for ultracold
> freezers commercially, and such freezer systems are extremely
> common in research and should by now be maximally inexpensive for
> the temp gradient and power needed. A cold room, then, might run
> at -79, with each patient pod (and ethyl chloride) further cooled
> by Peltier devices to -135. Failure of the mechanical -79
> refrigerator then could always be backed up by easily and
> universally available dry ice, and electrical failure could (as
> in other designs) be backed up by a diesel generators.
This is a really neat idea. I'll investigate it further next week.
--- Brian Wowk
Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=2079