british-cars
[Top] [All Lists]

reply to questions on coolant

To: "S.O.L. Mailing List" <british-cars@autox.team.net>
Subject: reply to questions on coolant
From: Gregory T Fieldson <fieldson@jhunix.hcf.jhu.edu>
Date: Tue, 23 Aug 1994 11:02:44 -0400
OK, I clearly don't pick up a thread and run with it (not quickly, 
anyways). A week ago, a couple of questions appeared about my writing on 
engine coolants...


I. Greg Meboe asks:
     "One question  I have, I was considering lightly sandblasting the
core of  my rad  to get off the layers of black paint.  This because I
assume whatever thermoplastic is in the paint is a poorer conductor of
heat than  the metal  it covers.   I  see race  cars with  bare  metal
raiators,  and   I  assume   that  they're   doing  this  for  cooling
efficiency."

Response:
     Although I understand heat transfer better than electrochemistry,
I can't answer this question with as much precision as you might like.

Here is what I know:

1.According to  the symposium  papers, the  primary  loss  of  cooling
     efficiency in  a  brass  and  copper  radiator  occurs  when  the
     soldered joint  between the  tubes and the fins. Typically, brass
     and copper  radiators lose  about 5%  efficiency in  their  first
     couple of  years. Presumably  they stabilize after a this initial
     period of time.
2. Adding  a thin  layer of  paint will  certainly add a little bit of
     resistance  to  the  heat  transfer  path.  I  do  not  know  how
     significant this is. Obviously a thick layer will be worse than a
     thin layer.
3. I  found out  whether a  matte black finish has a higher convective
     heat transfer rate than bare metal. I suspect that it does have a
     better convective  heat transfer rate than polished metal, but is
     probably little different from brushed metal.
4. A  radiator exposed  to salt  air or  road  salts  will  definitely
     experience external  corrosion. This will irreversibly affect the
     solder joints.

Here is what I opine:

     Although I don't know anyone who races to consult, what I believe
is that  a racecar  requires greater  cooling efficiency than a street
car because  it is  running continuously  at high  rpms.  A  race  car
radiator probably  gets replaced every couple of seasons and generally
does not have sufficient exposure to salt water or salt air to corrode
significantly.

     Most road  cars have  excess  cooling  capacity  and  can  easily
accomodate a  corroded and normally painted radiator. I know that this
is certainly  true of  my TR7.  This  is  not  universally  true.  The
Spitfire, for  instance,  can  have  cooling  problems.  If  your  car
experiences chronic  overheating problems,  an oil  cooler is probably
the best  solution. Of  course,  it  is  possible  that  some  dreaded
previous owner has actually added more paint to the radiator, in which
case paint removal could make a significant difference.

     There is  the matter  of metal  hardness. I  have *no*  practical
knowledge of  sandblasting or  its effect  on copper.  I would suspect
that having  the radiator  steamed or  dipped would be a better way to
clean it.  If you  do clean  the radiator,  it is  possible to have it
anodized. This  will put  a thin protective metal oxide coating on the
radiator,  which  will  *probably*  enhance  its  durability  with  no
noticeable effect on cooling. I think that some radiator cores may
initially be anodized instead of painted.

     Finally, whatever  you do,  I doubt  that you can cause permanent
and irreversible  damage to  your car as a whole. As long as the tanks
are good,  a brass and copper radiator can  be easily re-cored. If you
can safely  blast the  radiator for free and have the resources to re-
core if things go awry, then go for it!

II. Joe Gorin writes:
     "My experience  of having to rod-out the radiator after 12 kmiles
implies that  time matters, not just mileage, in coolant lifetime. ...
But, let  me look  at it  from a theoretical angle. ... [1]The rate at
which the silicates are "used up" probably doubles every 10 degrees C.
Therefore, the  rate of  decay depends  on the  period of  time during
which the  coolant is  near its  maximum temperature.  A short trip to
the store and back, say 2 miles each way, might result in 6 minutes of
operation, but  30 minutes  of near-maximum coolant temperature. Thus,
the coolant  lifetime might be 5 times shorter in miles on a low usage
car than  on a  car driven  for 30  minutes at a time on freeways. ...
[2]Consider a  car that  is kept  in the  garage only  for 5 years, or
44,000 hours.   The coolant would have to be used up at a rate that is
lower by  a factor  of 73  (44,000 hours  vs. 600 hours) to be equally
used up  by sitting  in the  garage.  For this relative rate of "using
up," the  weighted average  temperature in  the garage  need  be  61.9
degrees  C  (111  degrees  F)  lower  than  the  stabilized  operating
temperature. ... I don't know if these models have any accuracy, but I
doubt that the common recommendation to replace coolant based on time-
or-mileage-whichever-comes-first  is   totally  bogus,  though  it  is
obviously not an accurate model."

     You have  certainly  caught  a  mistake,  or  as  I  prefer,  "an
oversight" in  my original  article. The  proper recommendation  for a
low-use car  would be  to keep  the coolant  no longer than its shelf-
life.   I have located no real information on coolant shelf-life but I 
would bet on 3 or 4 years, tops.  There were no manufacturer's numbers 
on the two containers I  have at  home. As soon as I get hard numbers,  
I will update the archived copy of the article.

     Addressing  your   arguments  about   temperature   and   coolant
consumption more directly, both are correct in principle. I think that
the first  [1] is  definitely a perceptive observation. I suppose that
one could  adjust the  estimate of  average speed, and hence hours per
mile, to  reflect the  type of  driving done.  The second [2] argument
moves into  chemistry. In simple approximation, the rate of loss of an
inhibitor would  depend exponentially  upon the  temperature. However,
this rate incorporates a significant constant that we don't, a priori,
know. We  could calculate  this constant from the shelf-life (assuming
some  temperature)   and  the   bench  test   results  (and  operating
temperature) by fitting with an Arrhenius type expression.

     As a  practical matter, rate increases of 100 to 10,000 times are
not unreasonable  for reactions  at 90C  (194F)  vs  28C  (82F).  This
temperature difference  is an  absolute temperature  increase  of  20%
(362K vs 297K).
     Not  knowing   the  shelf-life  of  coolants  but  applying  your
empirical evidence,  I would  suggest that  3 or  4 years is the outer
limit of coolant life in an aluminum engine.

     I do  appreciate your  reading and comments on the article. It is
nice to know that people are willing to think critically about it.
     

Gregory Fieldson                        fieldson@jhunix.hcf.jhu.edu         
Department of Chemical Engineering  --  The Johns Hopkins University


<Prev in Thread] Current Thread [Next in Thread>
  • reply to questions on coolant, Gregory T Fieldson <=