Since the oil specific heat is only
about 1/2 that of water, it requires approx twice the flow rate (based on that
difference along) of coolant to bring the same amount of heat to the
cooler. Then if the airflow through the oil cooler is impeded in
any way (or you are pumping hot air into it) the situation becomes worst.
There's not really much
we can do about oil flow rate, short of not restricting it. Does the oil
typically flow at double the rate of the coolant in the stock
configuration? I've never really thought about that, but I can't see how
it could.
If the oil doesn't flow
that much faster than water, and it has half the heat rejection of water, then
do we need to size the oil cooler the same as the radiator? I've been
amazed at how little radiator I need, compared to how hard it is to cool the
oil, so this is starting to make sense.
Don’t forget
that the amount of heat rejected to the oil is about 40% of what goes into the
water. The specific heat of oil is actually about 70% that of a 50/50 mix of
water/glycol (.6 vs .84 in metric) My dyno test showed the flow rate of water just
a bit more than double that of the oil if you have a thermostat in the coolant circuit
(triple without the thermostat). Now you put in the factor of typically allowing
the oil about 20 F higher the coolant, and you have the air flow right, it
works out that the heat transfer area (core volume, if you wish) of the oil
cooler can be just a bit less than half the radiator. And Mazda had things
figured correctly.
This, of course,
assumes that you have the same incoming air temp to both. Putting the oil
cooler behind the radiator throws a big wrench into optimizing the cooling
systems; and is a bad idea from the git-go. Heat rejected is pretty much proportional
to the temp difference between the air and the coolant/or oil. But if that’s
your only configuration option, and the air temp increase through the rad is,
say, 35F; and you want to climb out on a 90F day, you can expect to need an oil
cooler core volume of roughly 1 ½ times normal; or about ¾ the size of the rad.
This is rough
estimating and assumes equal effectiveness of rad and cooler; but you get the
idea.
Al