You *must* have a clear route from the boiler to the expansion pipe with
no valves in the way. You must therefore put the valve in the short bit of pipe between the T-piece and the connection to the top of the cylinder coil.
This would mean that the 2port *would have* to be fitted in the main loft on the cylinder which would be a pain in the a**. I was already going to use a RF tank stat to get from the boiler to the main loft.
It would be very much harder to go for a Y-plan because this would require
> the 3-port valve to be installed in your downstairs loft - where the CH & HW
> circuits split - and you would have to re-pipe your overflow all the way
> from the boiler side of this junction.
Whilst not strictly correct I reckon that it you add a 2 port valve on the HW return it *should* work ok. Others seem to have done it. Do a search on google and have a look at:
formatting link
have a similar system (gravity HW and pumped heating without controls) - but in my case I have a strange expansion tank setup. In my system the Gravity HW pipes are also feed/vent, except that my vent pipe extends
*into* the water in the expansion pipe. If the water in the expansion tank drops so that the vent pipe is not underwater then I get no hot water (after having a good look) the vent pipe has been extended at some time by 3-4 inches to keep it below the water line. It therefore looks to me that what looks like the vent pipe (top of cylinder coil to vent) is actually also a feed pipe and both pipes could act as either a vent or a feed as required. If a valve was added into either pipe - then the system could still vent though either the HW flow or HW return (which are also feed/vent pipes). If the valve is closed in the feed - then the system could feed water from the vent pipe (which is below water level) if required.
Has anyone else seen this before?