pumping over - sort of

I've just finished installing an oil fired boiler in the kitchen to replace my old coal fired roomheater CH. It is now a fairly standard Y plan. Boiler is wall mounted in the kitchen. Pipes go up to floor space, horizontal for 2m then up into airing cupboard. Feed pipe then rises to vent into loft header tank and 4 inches along the cold comes in on the underside of the feed pipe.

Pipe then goes to pump, mounted to pump vertically. Next is 3 port valve arranged so it look like the letter T. One side to CH, other to top of HWC coil. CH return has non-return valve, then joins HWC return, then back to boiler.

So what happens is when CH switches off it seems that pump stops dead but water doesn't so I get a pressure surge on the pump inlet, where the vent pipe is. Water goes up the vent pipe and say a mug full goes into header tank.

To try to fix this I lowered the water level in the header tank - no good. Then I raised the vent pipe but the roof is low pitch so I could not raise this very far. About 21in is all I could get between water level and top of bend. Still not enough. I should also add that as I had a bit of 28mm spare and some fittings the section I inserted in the vent was 28mm which I thought would give the water more volume to take up before it overflowed.

Finally I tried a non-return valve. I already had a pipe connecting the CH feed to vent pipe via a gate valve (open gate valve to bleed air from pipework). I changed this for a no return valve which would open when vent pipe pressure was above CH feed, i.e. pressure across pump. When pump is on the pressure closes the valve. Better, perhaps, but I still get a trickle.

Suggestions?

Can you turn the pump speed down?

If not, try giving the pump some additional resistance by partially closing one of the pump IV's - I'm not suggesting this as a permanant solution but it could identify if your pump is generating excess head. I know that your cold feed and vent connection give a neutral point but odd things happen when pumps start and stop which move neutral points.

I should have said I've balanced the radiators - sort of. I've a IR thermometer but this had trouble reading the valves and pipes. Much better reading the flat radiator surface so I set the radiator LSV to give same centre temperature on the basis that should be the equivalnet as setting the same drop.

The radiators are now about balanced but the system has been added to over the years and extensions so the pipe sizes are not ideal. There is still a few C between rads but some of the LSV are down to <1/4 turn and some are fully open. Since I'm now tweaking <1/8 turn.... Anyway the radiators mostly have TRVs so that will do the fine adjust as long as they heat up all together.

In the process I set the pump which is on 2 which gave 10C drop. If I set it to 1 I get 20C drop and this seems to cut out the boiler (flow >80C) before the room thermostat cuts out when the CH first starts. Also, I find that when the heat is on the thermostat cuts before the boiler hits its limit of 80C. That would mean with a 20C drop the return water would be below the 55C limit to avoid condensation (and hence rust). So I can't really slow down the pump.

I guess I could look at unbalancing the system, a bit, to get better flow, which might let me set the pump slower. BTW I did try lower pump speed and that did reduce the effect, as you would expect.

Could it be an air lock somewhere that is acting like a spring (air is of course compressible, water is not)

What about running the vent up to the maximum height of the loft and back down again? It doesn't HAVE to go only just above the tank.

Now that's something I had not thought of. I'll see if I can bleed anything (pump rads etc.) tomorrow

Well I bleed the radiators and got some air. I'd forgot to bleed them after inserting the check valve but then I only dropped the water level a bit to get the valve dry as it was at the highest point.

Bleeding helped. With the pump on #2 there was no pump over whereas there was before (and all should be the same). On #3 it still pumped over so I took out the check valve, dismantled it and cut the return spring down until the valve only just closed. Replaced and retried and no pump over on #3 . The check valve must now be opening and so I have plenty in hand with the pump on #2.

Why did you need the valve?

I already had a gate valve in pipework linking from the CH port of the 3 way to the vent pipe. Open this valve to bleed air from pipework highest point. It was also used when I was running the old coal fire in parallel with the new boiler to finish off the coal (new boiler off of course).

Since I also had a 22mm check valve I reckoned that if I replaced the gate valve with the check valve then the over pressure driving the water up the vent would open the check valve and thus stop the overflow. It would also make this point self bleeding everytime the pump turned off.

As to why I was getting the overflow I don't know. My guess is pump stops but water doesn't. Or it could be pump pressure compresses some air which then kicks back when pump stops. Bleeding rads helped but did not cure it so air isn't the whole problem, unless its somewhere else that's not working its way out. All my pipe work slopes up and I would have thought the odd bit of trapped air unlikely to have such a dramatic effect.

oilman said:

What about running the vent up to the maximum height of the loft and back down again? It doesn't HAVE to go only just above the tank.

Click to expand...

all the **** i was told at college i was told about 150mm +40mmpmh.

Connect the cold feed between the expansion pipe and the pump, (within 150mm) .

Fit an auto by-pass valve after the pump and into the common return, ditch the non return valve.

problem should be solved.

He did! The feed pipe comes 4 inches after the vent! (the feed pipe IS the expansion pipe).
A manual bypass can be used to raise the temperature of the boiler to be clear of condensation, though of course it reduces the available head.

Old systems with corrosion, so gas around the system, are quite prone to a spurt-over. Bleeding everything usually fixes it until more gas is created from corrosion! I daresay you have inhibitor in your system, perhaps it needs topping up?

There is always the final solutuion - connect the vent pipe to the f&e and remove the feed pipe so you have a combined feed and vent!

Not how I read it Chris.

anyways, it should be boiler flow, open vent, cold feed, pump, auto by-pass connection, three port valve.

is that what we have ?

Perhaps I misunderstood, what did you mean exactly Malc?:

Pipes go up to floor space, horizontal for 2m then up into airing cupboard. Feed pipe then rises to vent into loft header tank and 4 inches along the cold comes in on the underside of the feed pipe.

Click to expand...

So what happens is when CH switches off it seems that pump stops dead

Click to expand...

Don't you need any pump overrun with your boiler???
Maybe if the pump ran on and things cooled a bit the 'pump over' would stop too.