Pump-Over After Conversion to Pressurised Hot Water

[David Tong]

We just had the hot water cylinder in the bathroom replaced by a Heatrae Megaflo to get mains pressure hot water. Results are terrific (great shower performance, proper mixer taps) except that we're getting pump-over from the vent pipe into the expansion tank in the loft when the tank demands heat. I'd really appreciate any wise advice so I can discuss it sensibly with the plumbers.

It's a 1911-vintage three-story house with celler, attic, and loft over the attic. The boiler is in the cellar, the HW cylinder is in the upstairs bathroom, and the expansion tank is in the loft over the attic. The original HW system was conventional: gravity fed with flow and return pipes to the HW coil continuing up from the bathroom to the expansion tank in the loft (as overflow and down-feed respectively). Because the Megaflo is not suitable for 'gravity circulation primaries', the installers added a pump at the boiler in the pipe coming up to the coil. It's a Grundfos Selectric on minimum setting and controlled by the HW thermostat.

It looks to me as if the extra head from the pump must be what's forcing the water over the top of the overflow pipe. The data sheet shows it can support 2.1 metres of water at zero flow rate. This is the worst case, so I'm thinking that raising the overflow height by 2 metres should compensate for the pump. There's only room to add about 0.7 metres going vertically, but if the overflow pipe were to go diagonally from the expansion tank to the roof apex and back again we could get almost 2 metres of vertical height. Does this make sense, or would it violate good practice?

Otherwise maybe we should close off the existing overflow feed and replace it with a dedicated overflow pipe connected between the boiler and the new pump but it's a long and awkward pipe run so not an easy solution. (Eventually we'll want to fit a condensing boiler and pressurise the CH system, in which case we won't even need the expansion tank, but I don't think we're ready for this just yet).

Any other ideas? Thanks in advance for any help.

[Christian McArdle]

If the boiler supports sealed pressurised operation, then convert now. It only costs a few quid in parts, and is much quicker to install than fannying around fixing medieval vented plumbing.

Christian.

[David Tong]

Christian,

Thanks for the response. Not sure if it does or not. It's a 15-year old Potterton Kingfisher CF80. The manual says max working head is 2.7 bar but doesn't even mention sealed pressurised operation.

Hmm, maybe it's time to change to a new boiler. Condensing boilers seem to be all the rage but I heard they won't work up a long flue and I don't know how we'd feed the tundish pipe out of the cellar (below ground level). I'll start doing some homework...

-- David Tong

[Ed Sirett]

If you can't do that because the boiler does not support it. Then if the pump for the megaflow's primary is independant of the radiators then try the lowest pump speed setting and then try half closing the outlet coupling on the pump if needed.

[David Tong]

Ed,

Thanks for the tip. I'll give it a try.

Do you think changing the pump to an electronically controlled type such as Grundfoss Alpha would help? From the data sheet it looks as if you can turn it down to give a more or less constant pressure of only about 0.5 metre of water.

[MarkK]

The pump should be downstream, i.e. it should pump away from, the vent and fill pipes. If it pumps towards the vent, you'll get pumping over.

The pump must be a minimum of H/3 below the water level in the header tank, where H is the head provided by the pump. e.g. a 5m head pump should be at least 1.67m below. For reference, in my house I have a 5m head pump with the header tank ~1.9m above, with no pumping over.

Check the position of the vent in relation to the pump first, as this is the most likely explanation. If this is the problem, find yourself a competent plumber, as the one you've used is not!

Mark

[John]

Why spend needless money? As Ed suggested partly close the outlet valve from the pump If you really want to sort the problem reconfigure the pipes around the cylinder primary.

[David Tong]

John,

Thanks for the input. It's just that I'd read that the function of the vent pipe is to allow steam to escape in an emergency overheat situation. The boiler info says the vent pipe has to be at least 22mm dia suggesting that you need decent flow through it if the worst happens. I was thinking that having the pump in the pipe must have compromised this already, and that partly closing the valve would compromise it even more. So it would be less of a compromise if a controlled pump meant that I didn't have to close the valve so much.

Lack of practical experience with heating systems makes it hard to assess risk - I don't want to be paranoid but also don't want to risk something serious :-)

David.

[David Tong]

Mark,

Thanks for the input. Interesting about the figure H/3, though I can't see how it matters too much exactly where the pump is in a system. Isn't it the difference in height between the top of the vent pipe and the water in the tank that matters? The pump here is a full three floors below the tank :-)

As to finding a competent plumber - I thought they were, and they've completed all the other work very nicely (fitting a new bathroom in the attic and refurbishing an existing one at bedroom level). I'm pretty sure they are an honourable pair of guys and wouldn't like to write them off as incompetent over this. But they do seem to have made a mistake in the way they've treated the vent pipe when mating up the Magnaflo with the original system.

I'm just looking to tap into the collected wisdom of the group prior to discussing with them later in the week how to recover from this awkard situation .

David.

[Richard Conway]

In that case I would suspect them to come back and sort it for you FOC - even if it means re-plumbing the f&e pipes back to where they should be, i.e. on the feed after the boiler but before the pump

[Dave Plowman (News)]

It's main purpose is as an expansion pipe. Hot water expands, so has to have somewhere to go. Although this function is shared with the feed pipe.

Yes, but this shouldn't happen unless something else fails.

My system has a bypass of the same diameter as the pump pipes going between the feed and expansion pipes. I'm pretty certain this is to reduce the chances of pump over.

[Ed Sirett]

I had assumed that the pump was placed in the primary return from the cylinder. So try pump speed 1 and if that fails get the installer back put the pump in the return and/or rework the vent/feed pipe locations.

[MarkK]

No. Think about how a pump works - high pressure downstream, low pressure upstream. You want the vent & fill on the low pressure side. Incidentally, H/3 is from a text book.

My statement on their competence was conditional - there could be another problem causing this.

If you would care to check the pump location & direction in relation to the vent you can either confirm or eliminate this as the cause and hence establish their competence (or not).

Mark

[David Tong]

Thanks to everyone who's responded to my plea for help. It's got quite interesting (at least to me :-) and I've been spurred on to make some proper measurements and try a few sums. I may have made some daft assumptions - but what do you think? Am I on the right lines? If I am we ought to be able to restore the safety margin against pump-over to what it was before just by raising the vent pipe by another 20 inches. Since there?s

27 inches of vertical space available, this seems like a possible quick fix. (I know it's a bodge but somehow it seems safer than throttling the pump). Anyway here goes (sorry if it's a bit long):

Situation Three-storey house with boiler and pumps in cellar, HW tank at bedroom level, expansion tank in loft above attic. Height difference between boiler and water level in expansion tank is 10.6 metres (417 inches). Vent pipe rises to 18 inches above water level in tank.

Before Modifications Potterton Kingfisher CF80 Boiler, vintage about 1990, with two flow and two return ports. One pair for CH and the other for HW. Connected as a vented convection-fed HW system, with pumped CH. CH section has pump in flow side of boiler. On HW side, the flow port goes to input side of HW heating coil and then ascends to the vent pipe. (This peaks 0.45m [18 inches] above the water level in the expansion tank in the loft space). Pipe from return side of boiler goes to output side of HW coil and from there to bottom of expansion tank.

Question: How much pump-over safety margin in the expansion tank? Answer: 10 inches.

I worked this out as follows. The total pressure differential between the vent pipe and the expansion pipe will be about equal to the sum of the pressure drop across the boiler and the pressure drop caused by convection flow. From the boiler data sheet the drop across the boiler is 4.5 inches (with an assumed total flow of 30 litres/min). The convection pressure will be about equal to the expansion coefficient of water, times the height, times the temperature differential between flow and return. Assuming the temperature drop is 40 deg C this comes to 0.00021 x 417 x 40 = 3.5 inches. So the total pressure drop is 4.5 + 3.5 = 8 inches. This means that the water will rise in the vent pipe to 8 inches above the level in the expansion tank, so the original safety margin against pump-over was

18 ? 8, or 10 inches.

After Modifications The only significant difference is that there?s a Grundfos Selectric pump on minimum setting in series with the vent pipe and close to the flow port on the boiler. This will increase the 8-inch pressure differential calculated above by the head from the pump. At an estimated flow rate of 15 litres/min (0.25 litres/sec), the pump data sheet gives the head as 0.5 metres (19.6 inches). So the total head in the new scheme will be 8 + 19.6 = 27.6 inches. But the water can only rise 18 inches before it overflows, so now it can?t help but pump over when the HW tank demands heat!

David Tong.

[MarkK]

You're on the right lines if you just want to fix the symptoms. The cause can be addressed by telling your plumbers to move the pump from the cellar, where it was incorrectly located, to a position downstream of vent & fill pipes, presumably near your HW cylinder.

It really is that simple, and since you ask, I think you're making rather a meal of it by coming up with calculations of additional head needed to counter the pump pressure. Do you brake and accelerate your car at the same time? Do you calculate how much braking is required to keep down to a certain speed because your mechanic set your throttle to be half open all the time? You're doing the plumbing equivalent. If you don't trust the advice given you can verify it in any decent plumbing text book. I recommend "Plumbing - Heating & Gas Installations" 2nd Ed by R.D Treloar. ISBN

0-632-05332-1

Mark

[David Tong]

can verify it in any decent plumbing text book. I recommend

Mark,

Interesting about the book. I'll see if i can find a copy. Fifteen years ago when I was having to bleed attic radiators every couple of days I had a lot of trouble finding a plumber who understand the problem. I then looked at plumbing text books in Waterstones but the ones I could find seemed to be mostly a set of recipes and rules of thumb without much theory to back them up. In the end I did some sums that told me why the pressure in the attic radiator was negative and sure enough moving the pump from the feed to the return on the boiler solved it for ever after. But I had to really persuade the plumber (from a firm of 'heating engineers') that it was ok to do it. Taught me it's best to try to understand it yourself.

Anyway the 'symptom fix' worked this morning. Last night I bodged a temporary height increase of 70 cms onto the vent with garden hose. This morning after the system had settled with a fully heated tank, the vent pipe was cold just above the point where it leaves the return side of the tank. After a big heating demand from a shower the same pipe became hot but only to a height of 38 cms. So I reckon that represents the extra head from the pump.

I'm sure you're right about the 'proper' solution. Putting the pump in the tank cupboard would avoid an awkward pipe run from the cellar, but it'll be in the main bathroom and next door to a main bedroom. So it will have to be a really quiet pump. Any recommendations? Must say the Grundfoss seems silent now. Just wondering if they stay like that over the years?

Thanks for the help, David.