Suspected Leak in Hot Water system vs Cylinder behaviour?

[stormyuk]

Good morning,

I have a Kingspan Tribune XE (TXN210) indirect unvented cylinder in my property (installed in 2020, in a new build property) and I am trying to understand what I originally suspected was a leak from the HW system.

My cylinder is heated twice a day on 2 x 30 minute cycles, this generally gives us enough hot water for washing and showing for us, boost is used in circumstances where extra hot water is required.

I was taking a water meter reading some weeks ago and I noticed that the small meter dial x0.0001 (in millilitres, one revolution being one litre) was moving while I was taking the reading, indicating water usage when no one was actually using the water. So I started to monitor this usage suspecting I have a leak somewhere on the property.

After making sure there was no visible leaks I could see, checking all water outlets and toilets etc. I then checked that there was no water usage when the stopcock to the property was closed. This proved the meter was not faulty as when the stopcock is closed no water is used.

I then turned off the cold water feed to the XE Cylinder and this also resulted in no water being used. When the feed to the cylinder was re-enabled I had the unexplained water loss again. So this proves the loss of water is in the HW system somewhere. I then double checked I had no drips from taps, showers etc and actually did fix a small drip on the kitchen tap (although it was a very slight drip, not explaining the figures below)

Below is a typical example of the water loss, which is not consistent and not linear. It seems to follow this pattern each day with slight variation in the figures but the "loss" always gets less from the early reading, to the later reading as time progresses.

The tank is heated between 6.30 and 7.00 in the morning and then 2 or 3 showers are taken between 7 and 8.30. We then purposfully didn't use anymore hot water during the daytime (times below) over a series of a few days to monitor.

Water Meter reading; |00308|844| (960) (start meter read, in cubic meters, followed by litres in red 844, and millitres (960ml in red also, shown on the small red dial).

Monitoring over the next few hours, shows the following change, where the number in red is last number on the meter, reading in litres. The numbers in (brackets) are the "water loss" over the previous time period.

4.960 08.30am
5.750 09.30am (0.79l)
5.900 10.35am (0.15l)
6.080 11.33am (0.18l)
6.270 12.37pm (0.18l)
6.380 13.37pm (0.11l)
6.510 14.37pm (0.13l)
6.590 15.37pm (0.08l)
6.600 17.16pm (0.01l)

Total. 1.64 litres between 8.30am and 17.16pm.

Could this be explained by the hot water tank topping up as the cylinder cools during the day, due to contraction of the volume in the tank leading to the diaphragm moving, leading to a top up of the system?

The house is a new build and I cannot see any water marks anywhere on any walls or ceilings. The HW pipes are all in the walls and ceilings, and again there are no water marks anywhere that I can see.

I am a bit baffled by this.

Thanks,

 

[Ric2013]

0.0001l is 1/10 of a ml, not 1 ml.

 

[stormyuk]

Yeh it's a typo. Basically this dial is 1000ml's isn't it? So a full rotation is equal to 1litre. So each numbered marker 1,2,3 etc etc is 100ml?

That's how it seems to work each rotation of the small dial increases the last red number on the meter by 1litre.

I can't edit the original post. I think the figures in the list are still valid.

 

[John.g]

If the unvented cylinder has been installed properly, NR valve etc, then with no usage but cold feed still open (or shut) then a 180 litre cylinder with 165 litres of water and 15 litre expansion space pre pressurized to say 2 bar, same as mains pressure, will absorb 2.67 litres of water if heated from 15C to 60C, obviously a bit less if heating from a different base, the cylinder pressure will increase from 2.0 bar to 2.65 bar but in cooling down again this 2.67 litres will again be taken up by the system and the pressure will/should return to the system. So, IMO, the flow meter should not register anything.
The mains pressure would have to increase by 0.42bar to force 1.84 litres into the expansion space and might happen if the PRV is passing and isn't "droptight"

 

[Ric2013]

If the unvented cylinder has been installed properly, NR valve etc, then with no usage but cold feed still open (or shut) then a 180 litre cylinder with 165 litres of water and 15 litre expansion space pre pressurized to say 2 bar, same as mains pressure, will absorb 2.67 litres of water if heated from 15C to 60C, obviously a bit less if heating from a different base, the cylinder pressure will increase from 2.0 bar to 2.65 bar but in cooling down again this 2.67 litres will again be taken up by the system and the pressure will/should return to the system. So, IMO, the flow meter should not register anything.
The mains pressure would have to increase by 0.42bar to force 1.84 litres into the expansion space and might happen if the PRV is passing and isn't "droptight"

Interesting thought is what if the mains pressure varies?

 

[John.g]

A PRV is essential with a unvented cylinder because if its as high as 4/4.5 bar bar then the expansion valve will lift at 6 bar when the cylinder heats up of if the PRV is passing, it only needs ~ 3.7 litres which when added with the expansion will rise to 6 bar.

 

[Ben-gee]

Do you have a balanced cold take off after the isolation to the hot cylinder?

 

[John.g]

Even though this one has a external "air bubble" they are all pretty similar.

 

[stormyuk]

If the unvented cylinder has been installed properly, NR valve etc, then with no usage but cold feed still open (or shut) then a 180 litre cylinder with 165 litres of water and 15 litre expansion space pre pressurized to say 2 bar, same as mains pressure, will absorb 2.67 litres of water if heated from 15C to 60C, obviously a bit less if heating from a different base, the cylinder pressure will increase from 2.0 bar to 2.65 bar but in cooling down again this 2.67 litres will again be taken up by the system and the pressure will/should return to the system. So, IMO, the flow meter should not register anything.
The mains pressure would have to increase by 0.42bar to force 1.84 litres into the expansion space and might happen if the PRV is passing and isn't "droptight"

Interesting, so it looks like I potentially do have a small leak somewhere, just no idea where and why the leak changes leak rate throughout the day, I have no idea. This is what confuses me. I would expect a leak to be pretty constant (or get worse over time), this is why I wondered if it could be something with the tank and the fact its cooling.

Could a faulty PRV cause water to be taken into the tank, but at a varying rate? The first hour after water usage is always the worst and then it slows down during the day as my figures show. I will do some more monitoring next week to see if its still similar rates.

There is no dripping from the tundish or overflow of water elsewhere.

Unless of course, its not installed correctly but, interestingly, the original tank had "popping" noise issues and Kingspan themselves replaced the tank under warranty and never mentioned it being installed incorrectly as potentially they could have used that as an argument not to honour the warranty, although see the next quote.

Do you have a balanced cold take off after the isolation to the hot cylinder?

Interestingly no, non of the mixers use a balanced cold feed which is probably wrongly installed. The tank balanced cold is just capped off at the PRV assembly, but Kingspan engineer themselves never queired this when they replaced my tank under warranty and its the pre-plumbed version, so they would have known about this on inspection and replacement.

Interestingly the show homes never had the balanced cold installed either (as I have photo's of the airing cupboard installs) so I am guessing this builder doesn't install the balanced colds which is probably wrong, but cheaper. Not sure that would cause this issue though? The performance of the showers etc is very good and so not sure if and how much back pressure there is to the tank because of this. How would that manifest? Dripping at the tundish due to increased pressure in the tank? I have recharged the pressure gap on the tank and the tundish did have a weep for the following day, but since then has been dry and has been dry for the rest of the time.

The house is a 2019 new build, tank was replaced in 2020.

 

[Chuck]

A leak of a litre or more a day into the fabric of a building should be quite noticeable if it continues for a few days.

Based on your observations, I suspect that the PRV is most likely to be responsible. You could confirm this by connecting a pressure gauge to the DHW and see how it correlates with your 'leak'. The mains inlet pressure should also be checked. PRVs need to have a reasonable drop across them to work correctly. If the incoming pressure is not high enough, e.g. because all your neighbours are having their baths and showers, then the cylinder will be left at too low a pressure and will top up once the inlet pressure rises again.

You do sometimes get leaks that are temperature sensitive, e.g. open after hot water has been run due to the increased temperature and close as the pipework cools, but they are not common. And, as I said above, a litre a day from your DHW is probably going to be fairly obvious once you start looking.

 

[John.g]

Unless you can shut off the hot water outlet at the cylinder then you don't really know if the hot water is leaking back through the mains via a mixer, when the stopcock is shut then obviously can't/wouldn't be able to leak back. With no usage, the hot will be higher than the mains and higher still with cold water usage but no hot water usage.

 

[stormyuk]

A leak of a litre or more a day into the fabric of a building should be quite noticeable if it continues for a few days.

Based on your observations, I suspect that the PRV is most likely to be responsible. You could confirm this by connecting a pressure gauge to the DHW and see how it correlates with your 'leak'. The mains inlet pressure should also be checked. PRVs need to have a reasonable drop across them to work correctly. If the incoming pressure is not high enough, e.g. because all your neighbours are having their baths and showers, then the cylinder will be left at too low a pressure and will top up once the inlet pressure rises again.

You do sometimes get leaks that are temperature sensitive, e.g. open after hot water has been run due to the increased temperature and close as the pipework cools, but they are not common. And, as I said above, a litre a day from your DHW is probably going to be fairly obvious once you start looking.

Believe me, I have looked for the leak and its probably been doing it a long time now (since the house was built or potentially when the PRV was changed with the old faulty tank), it just I only just noticed it when I happened to be looking at the meter one day.

For now I think I will just keep an eye on it, as getting a plumber to do all that work checking or changing the PRV (in case) is probably going to cost more than the water I am "losing".

Annoying for sure although if what John suggests below is whats happening I might have bigger issues....

Unless you can shut off the hot water outlet at the cylinder then you don't really know if the hot water is leaking back through the mains via a mixer, when the stopcock is shut then obviously can't/wouldn't be able to leak back. With no usage, the hot will be higher than the mains and higher still with cold water usage but no hot water usage.

Unfortunately there seems no way to do this, the hot water feed off the tank is just piped with no stop valve etc.

If it is what you suggest I am not sure there is an easy fix as its gonna be neigh on impossible to retro fit a balanced cold feed to every mixer without some serious intervention (2 showers, one bath, and 5 other mixer taps upstairs and down).

Thanks for the responses.

 

[Ben-gee]

The reason I asked about a balanced cold after the hot shut off is that I suspected the leak would be in a toilet - as 90% on ‘unnoticed’ leaks are, usually flush valve letting by and then fill valve making up.
But sounds like nit in your case if isolatingonly the cylinder stopped the leak.

 

[Timmy D]

Just FYI, unvented cylinders shouldn’t have a hot outlet isolation valve (just in case you’re thinking of adding one).

 

[stormyuk]

Unless you can shut off the hot water outlet at the cylinder then you don't really know if the hot water is leaking back through the mains via a mixer, when the stopcock is shut then obviously can't/wouldn't be able to leak back. With no usage, the hot will be higher than the mains and higher still with cold water usage but no hot water usage.

Do water meters not run backwards?

Would the pressure really be enough to push water back past the meter?

Just FYI, unvented cylinders shouldn’t have a hot outlet isolation valve (just in case you’re thinking of adding one).

To be honest I wouldn't be messing about like that to test a theory.

Even doing all the pressure measurements suggested is beyond me. I am an electrical engineer not a heating and plumbing engineer. I'd get someone in. Just not sure it's worth it for the amount of water I am losing. It would be nice if I could get a full 24 hour measurement of how much water I am actually losing. Asking two girls to not use water for 8 to 9 hours is a stretch!

I am just trying to get down to the bottom of the system behaviour and if there is something I can do about it without ripping the fabric of the house apart.

I am pretty sure I don't have a physical leak somewhere considering I should be able to find some evidence of it. But, still it's a worry in the back of my mind.

Of course wasting water is never good either for the environment or my wallet.

If the problem is actually the lack of a balanced cold or something else. Why I asked the original question.

The builder who built this house seems to do this on all their homes so I am surprised more people have had issues..

Thanks

 

[John.g]

It can certainly flow backwards against the mains if the pressure is higher but whether it would turn the meter in reverse or not, don't know.

Have you got a pressure gauge mounted anywhere after the NRV?, say on the cylinder itself?, if you have, then as the cylinder is cooling down and with no usage the pressure should start falling, if it doesn't then it points to the PRV increasing its downstream pressure.
Wthe temperature falls from 60C to 50C then the pressure would have to rise by ~ 0.4bar to increase the water volume by 1.85 litres (per meter reading).

 

[moonlight]

Is the pressure relief valve dripping into the tundish when heating the water?

 

[king of pipes]

If the water meter stops when the isolation valve on the combination valve is closed and no water is being used in the property? then this points to a leak you have tempreture and pressure relief valves fitted to the cylinder I would be monitoring these especially at heat up times do you have any idea what your incoming
water pressure is ? water pressure can certainly rise over night and a faulty combination valve is quite possible, your cylinder is still under warranty but you must determine what and where the problem is and where you losing water . Kop

 

[John.g]

IMO, a pressure gauge is the first requirement to track down the problem.

What we havn't mentioned is a holed heating coil but if the heating system is sealed then eventually if the cylinder pressure is consistently higher than the coil pressure then the PRV will eventually lift on the boiler, however if the two pressures are very close and at times the cylinder pressure may be lower than the coil pressure, in this case water will be shunted back into the cylinder thus keeping a balance.
If the boiler& system is open vented then eventually the F&E tank should start overflowing.

 

[stormyuk]

Is the pressure relief valve dripping into the tundish when heating the water?

No the tundish is dry it dripped a little bit for a day after I recharged the air gap but, since then it's been fine.

 

[John.g]

Keep a eye on the boiler pressure when off, it should fall slowly and should not rise under any circumstances if the filling valve is shut and if there is no cylinder coil leak.

 

[stormyuk]

If the water meter stops when the isolation valve on the combination valve is closed and no water is being used in the property? then this points to a leak you have tempreture and pressure relief valves fitted to the cylinder I would be monitoring these especially at heat up times do you have any idea what your incoming
water pressure is ? water pressure can certainly rise over night and a faulty combination valve is quite possible, your cylinder is still under warranty but you must determine what and where the problem is and where you losing water . Kop

My only issue with I understanding if it is the combination PRV/NRV on the input to the tank, where is this water going? What goes in should come out I suppose..

IMO, a pressure gauge is the first requirement to track down the problem.

What we havn't mentioned is a holed heating coil but if the heating system is sealed then eventually if the cylinder pressure is consistently higher than the coil pressure then the PRV will eventually lift on the boiler, however if the two pressures are very close and at times the cylinder pressure may be lower than the coil pressure, in this case water will be shunted back into the cylinder thus keeping a balance.
If the boiler& system is open vented then eventually the F&E tank should start overflowing.

Yeh it's a sealed system.

To track it down it looks like I am going to have to get an expert in to check the pressures. It's finding said expert...

I am not convinced there is a leak into the heating circuit to be honest. The boiler is reading around 1.2bar on it's gauge and the gauge on the heating pipework upstairs reads slightly lower which is what I would expect. Maybe 1.1bar but then who knows how accurate they are probably not very.

If my tank is ~3bar the boiler would have been chucking water out by now I would think. Also when I had the system serviced in March / April I also had the water condition checked and it appeared ok with correct inhibitor level and such. This would probably have all been flushed out. Its conceivable the fault has just occurred but, my gut feeling is it's always been doing it and I've just never noticed.

 

[stormyuk]

Keep a eye on the boiler pressure when off, it should fall slowly and should not rise under any circumstances if the filling valve is shut and if there is no cylinder coil leak.

Yeh boiler pressure only appears to rise when there is heating or hot water demand and it's running. It does rise to around 2bar which I always thought was quite high from a standing start of 1.2bar (1.8bar if hot water demand and 2bar for the radiator circuits).

Although it's always done this and two different plumbers who have serviced the boiler two different years have said it's fine...

 

[John.g]

If its returning to 1.2 bar when cold then no coil leak, any pressure gauge on the cylinder??.

 

[stormyuk]

The only gauge at the cylinder is for the heating circuit. There is nothing for the tank itself unfortunately.

I also still can't get my head around the declining 'leak rate' throughout the day.

4.960 08.30am
5.750 09.30am (0.79l)
5.900 10.35am (0.15l)
6.080 11.33am (0.18l)
6.270 12.37pm (0.18l)
6.380 13.37pm (0.11l)
6.510 14.37pm (0.13l)
6.590 15.37pm (0.08l)
6.600 17.16pm (0.01l)

I am doing more readings today to see how it compares to the other weeks, it will send me around the bend I think.

 

[John.g]

Without any more ado I would install a pressure gauge on the cylinder circuit, there may be a blanked tapping on the PRV in which you can install a PG, wonder how was the system was set up originally?.

 

[Chuck]

I also still can't get my head around the declining 'leak rate' throughout the day.

You said you're an electrical engineer so perhaps the following analogy might help.

The PRVs used in these systems are fairly simple. As the pressure increases at the outlet port it pushes a piston against a spring that closes the valve. The top of the spring can be moved up and down to set the nominal pressure. In electrical terms, as the pressure (voltage) at the valve outlet approaches the set point the valve impedance (resistance, R) rises. The water tank is essentially a large capacitor (C) containing a mass (charge, Q) of water. So, if you consider the system to be a capacitor being charged from a constant voltage (supply pressure) via an increasing resistance (the closing PRV) you'd expect the time constant (RC) for the process to increase as it approaches equilibrium. The settling time to 99% of steady state is minutes, but the settling time to 99.9% seems to be hours.

 

[John.g]

I just think of a PRV as a pressure reducer with a sort of proportional control where the pressure loss depends on the flow rate.
The example given below tells me that this particular (1/2") valve will maintain a downstream/cylinder pressure of 2.3 bar with a upstream pressure of 8bar and a flow rate of 20LPM but the pressure will not be higher than 3 bar with no demand.
So the PRV in question, whatever its sizing/set pressure should also not rise above its set pressure with no demand except that its leaking past the seat or it may increase the downstream pressure with no flow if mains pressure rises??. Edit: no shouldn't increase with increase in upstream pressure at no flow conditions.

 

[stormyuk]

Thanks again for the responses guys.

Here is today's meter data;

Apologies for the poor photos, it's quite difficult to get decent shots.

First photo is 8.30am and the other photo is 5.32pm. The small dial and last digit don't actually line up properly (why am I not surprised).

So it's moved just over a litre today.

0.00 8.30am (dodgy line up)
0.40 9.30am (0.41l or 0.40l)
0.54 10.32am (0.14l)
0.65 11.32am (0.11l)
0.88 12.32 (0.23l)
0.95 13.59 (0.07l)
0.99 14.30 (0.04l)
1.04 15.29 (0.05l)
1.04 16.31 (0.00l)
1.04 17.32 (0.00l)

Between around 15.30 and 17.32 the meter didn't move at all.

Figures in brackets are the loss over the period I also missed a reading on the hour at 13.30.

I am tempted to just to get the PRV/NRV assembly replaced but, it seems an expensive that might not really be warranted with the volume of water if it is the PRV not seating properly. Whatever is 'leaking' has leaked a little less today.

 

[Deleted member 123817]

Yes, you could say "nothing to see here, move on" but is intriguing if nothing else but very few would notice that small change in usage anyway.
Would that period of no usage coincide with heavier cold water usage I wonder.

 

[stormyuk]

Yes, you could say "nothing to see here, move on" but is intriguing if nothing else but very few would notice that small change in usage anyway.
Would that period of no usage coincide with heavier cold water usage I wonder.

No water usage at all, I've purposefully not used any water at all for that time period to get a read on the behaviour.

So it's probably likely in normal usage scenarios it's probably letting more water into the tank but, this is an almost fake scenario with no water usage over that time frame.

It's definitely something on the HW side considering if the cold supply to the tank is turned off water usage is exactly zero over the same time period.

 

[John.g]

Maybe your neighbors were using more water during this period resulting in lower mains pressure at your stopcock.

 

[IND_Nick]

Do you know what the main incoming pressure is at your property. It could simply be that in the morning as more people are using water the pressure in the mains is lower, during the day the pressure increases in the main and this pushes through the combination valve and into the expansion vessel capacity, especially if the mains pressure, in the morning, is below what the PRV would shut off on.
If you had a leak of 1 ltr a day it would normally be very apparent quite quickly.
Just a thought.

 

[John.g]

Something like that but if a PRV is working properly then it should only open if there is some demand for water, however little, and any mains pressure rise shouldn't open it as its balanced unlike a PRV (safety valve) on a boiler which isn't balanced. When used as a Auto Fill valve for system make up (and left open) they must also be drop tight.

 

[stormyuk]

Do you know what the main incoming pressure is at your property. It could simply be that in the morning as more people are using water the pressure in the mains is lower, during the day the pressure increases in the main and this pushes through the combination valve and into the expansion vessel capacity, especially if the mains pressure, in the morning, is below what the PRV would shut off on.
If you had a leak of 1 ltr a day it would normally be very apparent quite quickly.
Just a thought.

No not sure on my pressure, think I am going to see if I can buy a cheap gauge as that's something I can probably check easily with say my outside tap which is pretty much at the stopcock.

Speaking to my neighbour again with the same Tribune XE his system exhibits similar behaviour although he's not logged it like me..

He is often away and after he has been away for a week (he turns his stopcock off while away) after he returns and puts the water back on his tank 'fills' for a while.