This definitely qualifies as a brain fart, because I don't even know if the stuff to make it work exists.  What if you put a bi-metal strip maybe 1/4" wide and a couple of inches long in the radiator outlets of parallel radiators.  Calibrate them so that when cool they lay flat against the side of the rad outlet, and when hot bend up into the coolant flow.  If you had asymmetric flow, the branch getting more flow (for whatever reason) would have hotter water exiting the rad.  Such a device would bend up into the flow of the warmer outlet more than in the cooler side and restrict the flow ever so slightly.  Since you are shooting for symmetrical geometry everywhere in the system, the asymmetric flow is caused by a very slight underpressure in one branch or the other, so a very slight correction should even things out.  Of course the first step toward a solution is to quantify the problem, so a temperature reading at each outlet would be in order.  If you found you had a substantial enough discrepancy to warrant corrective action, and it was always in the same branch, you would only need to [conditionally] restrict the cooler (highest flow) branch to even things out.

But first measure both branches and see if there's actually a problem ..... Jim S.
 

Russell Duffy wrote:

Neil(Giving Rusty something else to think about)Thanks, that's just what I need :-) Actually, I'm not committed to series or parallel at the moment, so I'll be interested to hear any comments about the benefits of parallel cores.  I'm sure parallel "can" be better, if the water behaves, but if it doesn't, you probably lose all the extra cooling margin.  I suspect that it can be better than your figures indicated.  With less restriction, there will be a greater flow rate.  Also, efficiency is improved in parallel, because the hottest water contacts a larger area of radiator. Series is certainly easier, and cheaper if you're using AN hoses. Cheers,Rusty (probably series) 

--
Jim Sower
Crossville, TN; Chapter 5
Long-EZ N83RT, Velocity N4095T