You can’t really talk about radiator thickness without also including the variables of the core matrix; fin density, tube spacing, thicknesses, etc.  Fin density is one of the most important. 

 

The attached chart shows radiator cooling drag as a function of thickness for some different sizes (area).  Note that in this case the curves have a minimum point, an optimum; that occurs somewhere in the vicinity of 3” thick.  Also note that it is for staggered fins at a fin density of 8 fins/inch.  Typical auto radiators, and many used in racing, have densities of 14 – 16 fins/inch that are not staggered.  Higher fin density shifts these curves to the left, so for 14 fins/inch (not staggered) the optimum may be more like 2 1/2” thick.  For configuration reasons, NASCAR and other racers may find that using a thicker radiator with lower fin density is more advantageous.  This may also be true for aircraft use.  IIRC, the WW II liquid cooled airplanes had rads 6-10” thick with a rather open matrix.

 

The chart is also for 220 kts at 15,000 ft.; a little beyond where we generally go.  However; I believe slower trades off for lower (higher density) so the optimums MAY stay in about the same place.  Anyway, the idea is to give a feeling for how these things trade off. Putting this together with what seems to work well is that for the automotive type rad, somewhere around 2 ½” thickness is good.  Check the fin density of the evaporator cores.  I haven’t; but I think they have a more open core  making the roughly 3” thickness more suitable.

 

Al