Gravity siphoning of the hot tank - suggestions?

b318isp

A few years ago we had our attic converted into two bedrooms. The fitted radiators were tapped off the hot tank coil send and returns.

When the system is off, the cold water in the radiators falls and displaces the hot water in the coil, eventually cooling down the hot water in the tank. So, by 10am the next day, the water is only luke warm. We have managed this my a simple manual isolation valve.

I tried fitting a non return valve, however the pump can only overcome this at its highest setting (and because it is tapped off the coil, the pressure differential is low).

I then installed a cold side trap on one radiator, by looping the return line higher than the radiator. The thinking being that the cold water in the rad would not rise, hence trapping itself. This didn't work, I assume due to some siphoning effect.

I don't have an easy way of getting power from the central heating controls to the hot tank or the radiators, so solenoid valves or a separate lift pump are out.

I also considered changing the return line so it goes to a common header, or even putting a balance valve on the coil. Both of these are tank out jobs - and the latter would introduce a risk of a dead headed pump.

I was thinking of an unsprung non return valve (like the type used in solar heating circuits) but this may not prevent a reverse flow siphoning effect.

All I'm left with is to change to timed electrical heaters...

Any other ideas on overcoming this, a fit and forget solution?

K.Flyer

My first thought from your post is that you are loosing hot water through a leak somewhere on the downstairs pipework.
Switch off all of your taps and make sure no water is being used anywhere, then after about 30 mins check you main attic tank and see if the ball cock is running slightly.
This has to be ruled out first before looking at anything else.
Otherwise, if it is as you describe, a plate to plate heat exchanger plumbed in correctly should solve the problem.

b318isp

K.Flyer wrote: »

My first thought from your post is that you are loosing hot water through a leak somewhere on the downstairs pipework.
Switch off all of your taps and make sure no water is being used anywhere, then after about 30 mins check you main attic tank and see if the ball cock is running slightly.
This has to be ruled out first before looking at anything else.
Otherwise, if it is as you describe, a plate to plate heat exchanger plumbed in correctly should solve the problem.

Thanks. We've outruled a leak, as the cooling effect does not occur when the radiator circuit is isolated (and any leak in that would be very obvious!).

Would a secondary side pump not be needed for a PHX?

K.Flyer

b318isp wrote: »

Thanks. We've outruled a leak, as the cooling effect does not occur when the radiator circuit is isolated (and any leak in that would be very obvious!).

Would a secondary side pump not be needed for a PHX?

Yes, it would be needed, but set up correctly should eliminate the problem.
Have you checked the central heating pump to make sure that it is not over-running for any longer than it needs to?

b318isp

K.Flyer wrote: »

Yes, it would be needed, but set up correctly should eliminate the problem.
Have you checked the central heating pump to make sure that it is not over-running for any longer than it needs to?

Yes, it's tied to the boiler running.

K.Flyer

Did this problem start as soon as the attic radiators were fitted or did it develop sometime later?
When the boiler is switched off, does the circulation pump switch off as well or does it keep running? If it keeps running, it should go off after a couple of minutes.

b318isp

Did this problem start as soon as the attic radiators were fitted or did it develop sometime later? - Yep


When the boiler is switched off, does the circulation pump switch off as well or does it keep running? - Goes off with the boiler, both simultaneously controlled by the timer.

RJF

When you fitted the non return valve was it on the flow or return of the rad circuit? Should be on the flow. Also the location and size of the rad circuit connections will be important - flow above the return, rad circuit 1/2 inch- I presume the coil circuit is 3/4. There should be a balancing valve on the coil return - this will need to be balanced. Possibly a high head pump needed. Can you give a fuller description of rad and pipe sizes and heights above boiler and cylinder. I presume a factory insulated cylinder on a sealed system? It is possible for this system to work if set up properly but more info needed. If you could post pictures that would help.

b318isp

When you fitted the non return valve was it on the flow or return of the rad circuit? Should be on the flow.

- Flow side.

Also the location and size of the rad circuit connections will be important - flow above the return, rad circuit 1/2 inch

- Flow and return on the same level. 1/2" pipes.

I presume the coil circuit is 3/4. There should be a balancing valve on the coil return - this will need to be balanced.

- Yes, 3/4". Balance valve on coil, however rad stab-ins taken within this, so closing the valves, closes the rads too.

Possibly a high head pump needed.

- 6m pump fitted

Can you give a fuller description of rad and pipe sizes and heights above boiler and cylinder.

- 3 story, boiler & pump at ground floor, HW tank in middle floor, problem on top floor. All rads piped with 1/2".

- Boiler set at normal head height, say 1.8m; pump above that, say 2m, HW tank at say 2.5m, top rads at maybe 5m.

I presume a factory insulated cylinder on a sealed system?

- Yes. Closed system, mains pressurised.

It is possible for this system to work if set up properly but more info needed. If you could post pictures that would help.

- Thanks for your reply. LOds of idea, but looking for a simple mechanical solution first.

K.Flyer

Its an odd thing to happen, but from your first post I understand why its happening. I have never had the problem occur to any of my installations, but recently did have to rectify a similar problem for someone.
The solution was to fit a plate to plate with a small circulation pump. Its easily wired up with a stat on the cylinder and one on the main circuit.

RJF

The rad return connection needs to be on the far side (after) of the balancing valve. Can you describe precisely where you made the connections for the rads? Just to be clear when you said stab ins did you mean screw in clamp connectors or did you use regular "tee" fittings?

b318isp

RJF wrote: »

The rad return connection needs to be on the far side (after) of the balancing valve. Can you describe precisely where you made the connections for the rads? Just to be clear when you said stab ins did you mean screw in clamp connectors or did you use regular "tee" fittings?

This was installed for us. Basically the plumber tapped into the coil supply and return lines for the coil at the hot tank.

I agree that the balance valve is not in the right place, is the tap in for the rad return is before the balance valve. However, even if I move this it wouldn't prevent the siphoning, but it would improve the pressure drop.

Fittings are all tees.

K.Flyer

Is it possible to take a picture of the pipework and show us how it has been left. Might be easier to establish where things have gone wrong.

RJF

I agree with KFlyer that a picture would increase chance of finding a solution. You said that the connections are on the same level. How is this? The flow connection should be somewhere near the top pipe going into the cylinder and the return should be after the balancing valve at the bottom of the cylinder. The balancing valve restricts the flow out of the coil and if the connection is before it, this can possibly provide an easier flow route for convected water through the rad circuit. Can you describe it any further?

Calahonda52

So the rads heat when the HW cicuit is on.
What about a flow switch on the HW feed to open a normally closed motorised valve on flow and return on rad circuits
I know they are not duck tight but ... perhaps

b318isp

@KLF@ I'll take a picture in the next day or two.

@ RJF: It is as you describe. The balancing valve is downstream of the coil AND rad return. Ridiculous, as it means both are restricted.

@ calahonda52: I'm trying to avoid an electrical solution

RJF

Sounds like an easy fix then, just move the return to the far side of the valve and make sure the flow is above the return connection if possible, and then balance the system. If you post the picture should be able to confirm and tell you where to make connections.