dia said:
"Pump in the worse possible position sucking out from the least line of resistance, the top of the boiler, "
but it isn't, it's pumping towards the boiler.
This is a problem I've seen half a dozen times, surprised nobody else has come across it. (One result of it is that you get reversed flow in combi boiler's radiator circuits when they're providing hot water, cos the diverter valves let-by a tiny bit.)
Felix is right but does he know why!
It isn't a venturi, as we look at the pipes. You only get a venturi, and the suction effect, where there's a big reduction in the diameter of the pipe. We have no reduction?? No reduction means no suction. There is a reduction in the sense that the size of the pipe 22mm? is smaller than the size of the boiler, which is where the vent goes off. If you do your Bernoulli sums you'll find the difference. I doubt it will be huge because the pipe is fairly large compared to the boiler - I'll work it out when I have some time. There will be a difference, though the set-up is common enough and we would see it more often if it were large.
Except that there is a venturi. A magnet will confirm that there's a build-up inside the pipe, of scale containing magnetite, at the point where the feed pipe joins the return. THAT is what has caused the reduction in diameter and resultant venturi effect. Berni's figures wouldn't be accurate here because of
flow separation, which is to do with the real situation of the flow at the side of a pipe being much slower than in the middle, and the more important dramatic effects caused by
turbulence where you dont get
laminar flow. The effects due to turbulence can outweigh the effect of pipe diameter by a large factor.
It's made worse by the fact that the vent pipe rises very little above the level of water in the f/e tank. It would be better in this case if the vent pipe were larger in diameter in the length where the water level is because it would then rise less.
The scale will have been building rapidly once pump-over started, as that leads to introduction of air therefore oxygen into the water.
Solution may well be to drain down, cut the pipes and have a butcher's. It's common to find that replacing the adjacent pipes is easier than cleaning them. When you put them back it would be good to make the feed pipe come in from underneath, ie with bends, so that you don't get convection currents mixing the hot and cold (airy) water much. Of course you could always put a fat bit of pipe in there to tee off!
http://home.earthlink.net/~mmc1919/venturi.html