On Saturday, February 16, 2013 7:00:30 PM UTC, harry wrote:

If you have ever had a system run purelyb y convection you will

realise what a puny force it is.

In a pumped system, it is neither here nor there.

I have.

It can be a significant factor, but then I could work it out if I felt inclined.

The differential pressure, provided by the pump, across the radiator is throttled down in the course of balancing until the flow is adequate.

Most radiators can work adequately on a one pipe system with no diverter tees (Harry will have to google for that), with convection alone driving nearly all of the flow through the radiator. That suggests to me that the differential pressures across the rad, whether provided by pump or convection, may be of a similar order of magnitude.

It's a hypothetical question since no-one would knowingly connect a radiator TBOE with the flow in the bottom. I don't see why it wouldn't work, is all.

It was a useful exercise, since my pondering of this gave me a bloody good idea about something else.