X-Message-Number: 2062
Subject: CRYONICS: mechanical refrigeration
Date: Mon, 05 Apr 1993 12:42:50 -0400
From: "Perry E. Metzger" <>


> From: Brian Wowk <>
> 
> Perry Metzger:
>  
> > 1. a simple-minded question. I was under the impression that you could
> > buy commercial freezer units that operated at the temperatures in
> > question. Wouldn't it be simpler and cheaper either to buy such units,
> > or, if they are too small, to commission a manufacturer of such units
> > to build extra-large units?
>  
>         There are -130'C body-size freezers ("Queue" freezers) 
> commercially available.  I don't know the exact cost, but as I recall 
> the people who looked into them a few years ago concluded the per-person 
> costs would be very high.  
>  
>         To maximize economy you should store all your patients in a 
> single foam-insulated room, not a bunch of separate freezers.  The next 
> question is then cooling.  Thermoelectric cooling (no moving parts) is 
> still my first choice.

>From what I understand, Peltier effect cooling is far more expensive
both from a capital costs and from a cost of operation point of view,
which is why it isn't used in commercial refridgeration systems.  I
might be wrong on this, but I'd want to see figures.

> However if further investigation reveals it to 
> be uneconomical, then multi-stage mechanical freezer(s) operating 
> between the room and outside may be indicated.  (Of course internal 
> thermal ballast and external backup such as LN2 would be essential.)  
> This system should NOT be custom built.  It should be made from readily-
> available off-the-shelf cryogenic freezers with tons of spare parts 
> available.    

Well, yes, it should use standard cooling equipment for all the
mechanical components, but having the enclosure be custom built is not
much of a risk. I believe, however, that we are more or less saying
the same thing. My only point was that a commercial vendor would have
more experience on this than we do.

>         Bear in mind that every mechanical part added to this room 
> decreases robustness and increases maintenance requirements, quickly 
> eating up savings resulting from using electricity instead of LN2.  It 
> also increases dependence on the outside world for spare parts and 
> expertise.  The economics of mechanical refrigeration will have to be 
> looked at very carefully if this route is chosen.

Well, LN2 is made in -- guess what -- mechanical compressors powered
by electricity! Admittedly it is a byproduct of the production of
oxygen, but I would imagine that it can't be priced that much below
the cost of mechanical production even were it not a byproduct.
(Figures, anyone?)

Furthermore, although moving parts are a risk, thermoelectric systems
can quite readily break down as well -- electrical and electronic
systems are not failsafe and maintainance free. Assuming one built
one's units with dual mechanical cooling units, had backup power, and
had enough thermal balast to keep the units safely at temperature for
half a day during catastrophic failures, you should be fine -- you
need the dual systems, balast and backup power for a thermoelectric
system, too, so its not like you have gained that much by way of
simplification.

> > 2. Another issue I haven't seen addressed too much is this: precision
> > of temperature regulation. The systems that have all been mentioned
> > thus far sound like they would expose patients to relatively large
> > temperature swings compared to the "stay at the boiling point of LN2
> > forever plus or minus a few tenths" system that we have now.
>  
>         A system based on melting reserviors of ethyl chloride (as Steve 
> Harris suggests) would have essentially zero temperature fluctations.  
> It would have even less temperature variation than LN2 submersion 
> because freezing points vary less with pressure than boiling points.

Quite true. Any system -- even a mechanical freezer-room, ought to
have something like melting ethyl chloride as a thermal balast.

>         If a more active thermoelectric temperature control system is 
> used, I wouldn't expect temperatures to vary more than ~0.1'C (and very 
> slowly).  This is similar to what patients in LN2 are currently exposed 
> to because of atmospheric pressure variations. 

What is the rationale for believing that controlling the temperature
of thermoelectric systems would involve more precision than
controlling a mechanical system? Both would necessarily involve having
a temperature probe and a regulator of some sort. You can regulate the
power going into your thermoelectric system but you can also regulate
your compressors. Both depend on a feedback loop and both are thus
likely to produce a sawtooth temperature curve. It would seem
difficult to gain stability regardless of the cooling technology
without something like a melting solid as a thermal balast.

> > 3. Lastly, it seems that people have dismissed using a working fluid
> > other than liquid nitrogen out of hand. It would seem, however, that
> > other fluids do exist with an appropriate boiling point and that they
> > are not considered primarily because of expense or toxicity. However,
> > it is unnecessary to actually expose the patients or employees to the
> > working fluid or to consume it -- one could simply use the liquid in a
> > closed recirculating system to exploit its boiling as a way of very
> > stably maintaining patients at a desired temperature. Patients could
> > be surrounded by dry gas or by a liquid with a suitably low freezing
> > point. Is this impractical in some way that I haven't thought of?
>  
>         This is exactly what a mechanical refrigerator does.

Yes, of course, but a mechanical refrigeration system does not liquify
the working fluid -- it usually remains gaseous. It does not provide a
stable temperature. I was merely proposing that a system that
exploited something like (apparently) melting ethyl chloride to
provide temperature stabilization via the laws of physics instead of a
mechanical or electronic feedback mechanism.

Perry

Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=2062