mgs
[Top] [All Lists]

Re: speed of water in radiator

To: triumphs@autox.team.net (Triumph Mailing List),
Subject: Re: speed of water in radiator
From: "Randell Jesup" <jesup@mailhost.scala.com>
Date: Tue, 1 Oct 96 16:36:27 EDT
Barry Schwartz writes:
>Your overheating problem sounds like your not getting proper  air/water flow
>at speed.  I personally suspect its water flow, and might actually be
>flowing through too fast, not staying in the radiator long enough to
>transfer the heat away from the engine before returning.  Could also be the
>reverse (had the same problem in another car found the water pump impeller
>had disintegrated causing the very same symptoms you describe)

        There seems to be a common (mis)perception that water moving "too
fast" through the radiator will not lose enough heat and will cause
overheating.

        While I can think of some wierd possible designs where something
like this is, in theory, possible, I don't believe that this "problem" is
possible in any reasonable car engine cooling system.

        Think of it this way:
1. Heat transfer per unit time is directly proportional to temperature
   differential.
2. For given amount of water, moving faster through the radiator it will
   spend less time in the radiator, and thus (from 1) be cooled less.
3. However, for any given unit of time, more water will move through the
   radiator.  If you work out the figures, this more than compensates for
   the reduced cooling of any specific cc of water on one trip.  Another
   way to think about it is that while that cc of water gets cooled less
   per trip, it makes more trips.
4. In fact, cooling will be better, since with faster moving water the inlet
   temperature will be closer to the outlet.  This means that the temperature
   differential between water and air will be higher in the lower part of
   the radiator (compared to lower-flow), and so more heat per unit time will
   be transferred to the air in the lower part of the radiator.
5. Note that too little flow causes the bottom part of the radiator to
   transfer little heat (little temperature differential), and thus total
   heat transfer is lowered.
6. A thermostat reduces flow (and thus heat transfer) when the radiator is
   transferring too much heat, keeping the engine temperature around optimum.

Conclusion: heat transfer to the air will be increased by increased flow;
therefore the average coolant temperature will go down (and the heat
distribution will be, in theory, smoother).

Caveats: if you try to speed up the coolant flow, you can cause caviation of
the impellor (which is bad and reduces flow), increased turbulence and
increased power spent pumping water around.  Also, in theory an engine could
be designed to want cooler water coming from the outlet of the radiator for
some purpose, though I've never seen such an engine.  Also, increasing
pumping may not evenly distribute the extra flow around the engine, so some
cylinders/etc may not show as much change as expected, or in theory the flow
pattern could change and rear cylinders could get hotter.

        Trying to increase flow by removing a thermostat can cause reduced
flow in back cylinders (such as the infamous TR3(?) design), depending on how
it normally interacts.  Usually the thermostat reduces flow to the radiator.
Most water pump designs cause flow blocked by the thermostat to be used to
circulate water around the engine, reducing hot-spots by keeping the
temperatures even.  This (removing the thermostat causing hot-spots) is
probably part of why the myth about flow vs heat-transfer came into being.

        Still, within reasonable limits, increased flow transfers more heat,
all else being equal.  Even if it isn't better in the circumstances, the
reason isn't what was mentioned above (lack of loss of heat due to flow).
Just don't remove the thermostat to try to increase flow, unless you
know exactly what you're doing.

p.s. I'm still recommending voting against rec.autos.makers.mg/triumph/jaguar,
at least until there's consensus here in the mailing lists that switching
to newsgroups is a good idea - and so far there's never been such a consensus.
Vote ASAP, voting closes soon.  See news.announce.newgroups for voting forms
(the CFV's).
-- 
Randell Jesup, Scala US R&D, Ex-Commodore-Amiga Engineer class of '94
Randell.Jesup@scala.com
Exon food: <offensive words no longer censored - thank you ACLU, EFF, etc>

<Prev in Thread] Current Thread [Next in Thread>