Dave,

In getting to see the sketches, I am more convinced than ever that your problem was air in the system. Consider the following scenario.

 

1. you fill the system. air is bound to be trapped at the inlet of the radiator.

2. You make some low rpm runs. Assuming you had enough water to have the pump primed, you got some circulation, but at a low rpm, the flow is low enough that it cannot flush all the air out of the radiator inlet. The water will flow down some of the tubes, and others will have no flow, being blocked by the air bubble. (Uneven temperature distribution on the radiator)  Encouraged by the fact that it seems to be working you then go to:

3. Higher rpm. This increases (momentarily) the flow rate through the radiator, which then sweeps a slug of air up to the pump. The pump loses prime, and shuts off totally. You now overheat.

 

Until you get all the air out of the Radiator inlet, you cannot get the pump to keep pumping. Could you add a petcock vent on the high end of the rad -- pressurize the system with coolant and get all the air out?

 

Your proposed modification should work better, but you still could get a slug of air from the rad unless you get it hydraulically solid. The vent line will remove small amounts of air as the engine runs, but you still have to watch out for the 'big slug' of air when transitioning from idle to high rpm.

Bill Schertz
KIS Cruiser # 4045