So how else do you explain the change of heat transfer between individual rads so suddenly .. where, in this case .. there is not even any head required?
Is the circulator deciding for itself which rad to heat at what time? This was the clue all along, wasn't it?
Needs a system study, doesn't it?