Sorting out expansion vessel and overtemperature valve

[jack]

4 hours ago, John Carroll said:

You said the EV was flat, what pressure was it recharged to?, it should also have been done at the very least with the cold mains to the cylinder shut off and a HW tap opened, if its located above the cylinder then that should be OK.

 

Originally 3 bar and it's located above the cylinder. I'll make sure I do what you say when I recharge it (didn't manage to do it over the weekend). I've turned off my immersion diverter for the moment, so the temperature in the tank won't exceed ~50° C until I've got all of this sorted out. 

 

4 hours ago, John Carroll said:

I would also suggest checking that there is a pressure reducing valve (PRV) installed (which there should be) on the cold water valve set and most/more important that its setting is checked.

 

There's a 3 bar pressure reducing valve on the incoming mains. I already have a pressure gauge on its way so I can test it.

 

4 hours ago, John Carroll said:

Even with your 22L EV and assuming that you will rarely if ever be reheating a full cylinder, more likely say 80% or 200L then a final pressure of 3.12bar at 60C & 3.25bar at 65C with a cyl prepressure of 2.3bar, 2.5bar (if available) will still give over 90% flowrate of a 3.0bar cylinder.

 

Okay, thanks. I'm going to have a think about the setup as a whole. Currently considering adding an additional expension vessel to the cold feed near the tank, as there's more space there than where the current one is near the ceiling.

 

4 hours ago, John Carroll said:

Is the blue PRV next to the (red) TPR also stamped as set to 4.5bar?

 

There are two valves: 3.5 bar pressure relief, and 90-95°/4.5 bar temperature/pressure relief.

Thanks again, I appreciate the feedback.

[John Carroll]

From the horse's (McDonald Water Storage) mouth, the MI's show, below, the expansion relief valve mounted on the cylinder as yours is and none on the valve set (again like yours), the PRV should be set to 2.1bar, likewise the EV charge pressure....set to 2.1bar  and that's why a 22L EV is perfectly acceptable under those conditions, a full reheat to 60C then gives a final pressure of 2.8bar, 65C gives 2.91bar & 70C gives 3.12bar. all comfortably below the expansion relief valve's 3.5bar setting.

 

 

 

 

 

ECOFlow-Installation-Manual.pdf

Edited August 18 by John Carroll

[jack]

13 hours ago, John Carroll said:

From the horse's (McDonald Water Storage) mouth ...

 

Thanks John. I found that document a couple of days ago. It was one of the triggers for my comment about rethinking the setup as a whole. It's frustrating the original plumber apparently didn't read the manual before installing a 3 bar PVR on the incoming mains 10 years ago! 

 

While it's been fine operating at 3 bar for 10 years, I'm going to lower the incoming PRV to the recommended pressure. I'll also install another expansion vessel to help keep the pressure low.

 

While static pressure is fine, it does drop noticeably if a couple of showers or taps are in use. I suspect the most likely culprit is the water softener, but I'll be doing some pressure tests on the weekend to see what I can learn.

 

Either way, I'd like to do what I can (within reason) to avoid making the crappy dynamic pressure any worse when I change to the lower pressure PRV on the incoming main.

[Nickfromwales]

43 minutes ago, jack said:

 

Thanks John. I found that document a couple of days ago. It was one of the triggers for my comment about rethinking the setup as a whole. It's frustrating the original plumber apparently didn't read the manual before installing a 3 bar PVR on the incoming mains 10 years ago! 

 

While it's been fine operating at 3 bar for 10 years, I'm going to lower the incoming PRV to the recommended pressure. I'll also install another expansion vessel to help keep the pressure low.

 

While static pressure is fine, it does drop noticeably if a couple of showers or taps are in use. I suspect the most likely culprit is the water softener, but I'll be doing some pressure tests on the weekend to see what I can learn.

 

Either way, I'd like to do what I can (within reason) to avoid making the crappy dynamic pressure any worse when I change to the lower pressure PRV on the incoming main.

If you can leave a 5bar PEV feeding the softener, and only drop to 2.1 at the cylinder group then that would help. 
 

If your incoming cold mains will never see 5 bar then just leave a raw feed to the softener. 
 

Can you install a 150 or 200L accumulator anywhere, as this would make a big difference. 

[Nickfromwales]

Ps kudos to @John Carroll as I am guilty of skim reading, and assumed stainless cylinder and 3 or 3.5bar at the control group.

 

Many heads always better than one.

[John Carroll]

If, (which it probably has been) the EV precharge pressure has/had been set to 3.0bar and even assuming that the cold water pressure after the PRV (set to 3.0bar) never went higher than 2.0bar before reheating then the expansion relief valve would still have lifted if ~ 150L or more of the UV cylinder required reheating to 60C.

 

My ancient spreadsheet, below, may be of interst and be a help in sizing calculations for either HW or CH.

EV CalcsJK Rev0.xlsx

[jack]

On 19/08/2025 at 11:19, Nickfromwales said:

If you can leave a 5bar PEV feeding the softener, and only drop to 2.1 at the cylinder group then that would help. 
 

If your incoming cold mains will never see 5 bar then just leave a raw feed to the softener. 
 

Can you install a 150 or 200L accumulator anywhere, as this would make a big difference. 

 

Just coming back to this after spending some time reading and thinking over the weekend.

 

First, I've installed the new 3.5 bar pressure relief valve, turned off the immersion diverter, and reduced the inlet PRV pressure to 2.2 bar while I sort all of this out. 

 

Second, @John Carroll, your spreadsheet was hugely helpful, so thanks again for sharing it. While I understood the general principle behind expansion vessels, seeing the actual numbers was eye-opening to say the least. A few percent expansion doesn't sound like a lot, but when you combine it with a limited maximum DHW tank pressure and an immersion diverter, the peak operating pressure very quickly gets problematic. 

 

As an example, with a starting temp of 10 °C and a finishing temp of 70 °C (not crazy numbers with the immersion diverter working on a sunny day with colder mains in the autumn), the inlet PRV set at 2.5 bar, and my current 22 L expansion vessel, the final guage pressure of the system would exceed 3.5 bar. Adding a 50 L expansion vessel would drop the final pressure to 2.8 bar. 

 

2.2 bar is proving to be unpleasantly low in practice, so I really want to bump that number up with a larger expansion vessel or accumulator.

 

Using the nominal operating parameters in the cylinder manual (2.1 bar inlet PRV, operating temp 65 °C), and assuming an inlet temp of 7 °C (doubt it ever gets much lower in my circumstances), gives a peak pressure of about 3.1 bar. This should be a reasonable guide to a maximum operating pressure.

 

By fitting an additional 80 L expansion vessel, I can run at close to 3 bar on the inlet PRV while keeping a similar peak pressure.

 

Even allowing for a final temperature of 80 °C, I'd still be 0.2 bar below the 3.5 bar set by the pressure relief valve. These are pretty extreme numbers, so I should rarely reach them.

 

@Nickfromwales, I know the difference in principle between an expansion vessel and an accumulator, but in practice, is there any benefit to choosing one over the other? Are they structurally different? I see diaphragm versus bladder - are diaphragms intended to stretch, so that the pressure buffering is partly provided by the elasticity of the membrane? In general, would you expect either type to last longer than the other?

 

The 22 L expansion vessel would stay in place near the outlet at the top of the tank. I assume the new vessel/accumulator would go after the water softener in this diagram?:

 

   

I also plan to install a new PRV between the softener and the new pressure vessel/accumulator, so that the 3 bar supply is set after the water softener.

[John Carroll]

The attached may be of some help re Accumulator.

For example assuming a 200L accumulator and a required UVC pressure of 3bar, then if 4.0bar mains is available to recharge the accumulator you end up with a net vol of 40L. If you install a pumped accumulator and assuming you can pressurize it to 6.0bar then the net vol is 86L

Accumulator CalcsJK Rev0.xlsx

[jack]

1 hour ago, John Carroll said:

The attached may be of some help re Accumulator.

For example assuming a 200L accumulator and a required UVC pressure of 3bar, then if 4.0bar mains is available to recharge the accumulator you end up with a net vol of 40L. If you install a pumped accumulator and assuming you can pressurize it to 6.0bar then the net vol is 86L

Accumulator CalcsJK Rev0.xlsx 15.46 kB · 3 downloads

 

Thanks John.

 

I think the numbers for using an accumulator as an accumulator (rather than as a glorified expansion vessel) don't work that well in my case. Our mains is only about 3.4 bar, so realistically we'd need to consider a pumped system to get useable volume out of it. A 200 L accumulator + pump + additional pipework/valves would add a lot of cost and complexity, plus we don't really have space for a 200 L accumulator.

 

Having done a bit of research, I don't think there's any benefit in using an accumulator instead of an expansion vessel in this application. They don't seem to be much cheaper.

[Nickfromwales]

The accumulator would need to be on the supply side, before any restriction were in place. Pointless (imo) putting it where you show as it's not able to "perform admirably" there tbh.

 

I'd also T the expansion vessel off the cold inlet instead of the hot outlet, as that's a far happier life for it, if possible. 

 

And yes, adding the much uplifted volume of expansion should put this to bed afaic, even if it seems overkill to most. At first glance it may get questioned, but when you have eagle-eyed folk on here spot the 2.1 bar inlet pressure then it would immediately make sense.

[John Carroll]

On 19/08/2025 at 12:49, John Carroll said:

If, (which it probably has been) the EV precharge pressure has/had been set to 3.0bar and even assuming that the cold water pressure after the PRV (set to 3.0bar) never went higher than 2.0bar before reheating then the expansion relief valve would still have lifted if ~ 150L or more of the UV cylinder required reheating to 60C.

 

My ancient spreadsheet, below, may be of interst and be a help in sizing calculations for either HW or CH.

EV CalcsJK Rev0.xlsx 17.79 kB · 12 downloads

 

Just added a few more calcs here assuming water at a very minimum of 5C, so if you increase the EV capacity by 80L to give a total of 102L, the final pressure after a full reheat from 5C to 80C will still be < 3.5bar at 3.31bar. (prepressure of 3.0bar)

EV CalcsJK Rev1.xlsx

Edited August 27 by John Carroll

[jack]

On 26/08/2025 at 22:38, Nickfromwales said:

The accumulator would need to be on the supply side, before any restriction were in place. Pointless (imo) putting it where you show as it's not able to "perform admirably" there tbh.

 

I'd also T the expansion vessel off the cold inlet instead of the hot outlet, as that's a far happier life for it, if possible. 

 

Thanks Nick.

I was talking about using an accumulator (i.e, a vessel sold as an accumulator) as an additional expansion vessel rather than as an accumulator per se. Neither here nor there now, as I'm going to go for an actual expansion vessel.

 

Does the best location change if we're talking about an expansion vessel? I figured it should be connected on the DHW tank side, without any other components (e.g., water softener, second PRV) between it and the DHW tank.

 

As for the location of the 22 L tank, that's where it's currently installed. I figured the simplest thing would be to add more capacity on the cold side, as you suggest. I suppose there's nothing stopping me moving the 22 L tank to the cold side as well.

 

On 27/08/2025 at 11:17, John Carroll said:

Just added a few more calcs here assuming water at a very minimum of 5C, so if you increase the EV capacity by 80L to give a total of 102L, the final pressure after a full reheat from 5C to 80C will still be < 3.5bar at 3.31bar. (prepressure of 3.0bar)

EV CalcsJK Rev1.xlsx 18.2 kB · 1 download

 

Thanks John. I modified your spreadsheet the other day to get the numbers I referred to above. My version lets you put in start and end temps (updated 29 Aug to include graph and comparison table):

 

System pressure calculator - updated for BH.xlsx

 

 

 

[John Carroll]

Nice one Jack, where you just input the start/end temperatures to do the calcs.

[John Carroll]

Can you have a look at the attachment sometime please.

Why do you add +1 to both C7 & C8 in the Indirect function?

I'm a 60 year experienced "steam man" so I use the Spirax Sarco Sub Saturated Steam Tables and the relative water densities.

My calc results are practically the same as yours but if I don't add that +1 to both C7 & C8 then it doesn't compute. I have just inserted a few values in column I.

I modified your calc slightly, it will work with either the +1 added on or not added to C7& & C8, gives slightly different values.

TEMPORARY Expansion System pressure calculator.xlsx

[Nickfromwales]

1 hour ago, John Carroll said:

I'm a 60 year experienced "steam man" so I use the Spirax Sarco Sub Saturated Steam Tables and the relative water densities.

There’s a crazy wealth of diversity on here.

[jack]

19 hours ago, John Carroll said:

Can you have a look at the attachment sometime please.

Why do you add +1 to both C7 & C8 in the Indirect function?

I'm a 60 year experienced "steam man" so I use the Spirax Sarco Sub Saturated Steam Tables and the relative water densities.

My calc results are practically the same as yours but if I don't add that +1 to both C7 & C8 then it doesn't compute. I have just inserted a few values in column I.

I modified your calc slightly, it will work with either the +1 added on or not added to C7& & C8, gives slightly different values.

TEMPORARY Expansion System pressure calculator.xlsx 76.19 kB · 0 downloads

 

I've put the answer to your "indirect" question in bold below, but in case anyone's interested in the way the whole thing works (which might give an insight into how my brain "works"):

 

As you know, the coefficient of expansion for water changes with temperature, and the actual equation for calculating expansion between two temps is quite complicated.

 

I made a table of temps from 4 - 95 °C, at 1 °C per cell.

 

I inserted the known cofficients for 4 °C, 10 °C, 20 °C, ..., 90 °C at the appropriate positions in the table.

 

I then did a linear interpolation for the values between the known values. For the values between, say, 10 and 20 °C, I took the average of the coefficients for 10 and 20 °C, then divided by 10 (call that value "x"). The value for 11 °C is the (known) value for 10 °C + x. The value for 12 °C is the value for 11 °C + x. It's not perfectly accurate, because the relationship isn't actually linear, but it's certainly close enough for this task.

 

Since the coefficient represents the expansion at that temperature for a 1 °C temperature change, all you need to do to find the total expansion is to add all of the values between the two temperatures.

 

The indirect function in Excel (which I only learned about when doing this spreadsheet) lets you construct a cell reference based on any other input. In this case, the row in which each temperature value sits in the table is one higher than value of the temperature itself. For example, the coefficient for 7 °C is in row 8. I therefore needed to add 1 to the values of the start and end temps to use them as references to the correct rows in column G.

 

I can't remember where I got the values for the expansion coefficients this is all based on, but the results are very close to the table you included in your original spreadsheet.

 

 

I assume the difference is due to the interpolation mentioned above, but either way, it's close enough for this purpose.

 

Edited to add: Updated version of my earlier spreadsheet includes a graph and the comparison table above:

 

 

System pressure calculator - updated for BH.xlsx

[John Carroll]

I finally got it to work without adding or subtracting from the D values after  reading the above but didn't fill in all the values, 

John Only Expansion Calc using Steam Tables.xlsx

[John Carroll]

On 29/08/2025 at 13:26, jack said:

 

I've put the answer to your "indirect" question in bold below, but in case anyone's interested in the way the whole thing works (which might give an insight into how my brain "works"):

 

As you know, the coefficient of expansion for water changes with temperature, and the actual equation for calculating expansion between two temps is quite complicated.

 

I made a table of temps from 4 - 95 °C, at 1 °C per cell.

 

I inserted the known cofficients for 4 °C, 10 °C, 20 °C, ..., 90 °C at the appropriate positions in the table.

 

I then did a linear interpolation for the values between the known values. For the values between, say, 10 and 20 °C, I took the average of the coefficients for 10 and 20 °C, then divided by 10 (call that value "x"). The value for 11 °C is the (known) value for 10 °C + x. The value for 12 °C is the value for 11 °C + x. It's not perfectly accurate, because the relationship isn't actually linear, but it's certainly close enough for this task.

 

Since the coefficient represents the expansion at that temperature for a 1 °C temperature change, all you need to do to find the total expansion is to add all of the values between the two temperatures.

 

The indirect function in Excel (which I only learned about when doing this spreadsheet) lets you construct a cell reference based on any other input. In this case, the row in which each temperature value sits in the table is one higher than value of the temperature itself. For example, the coefficient for 7 °C is in row 8. I therefore needed to add 1 to the values of the start and end temps to use them as references to the correct rows in column G.

 

I can't remember where I got the values for the expansion coefficients this is all based on, but the results are very close to the table you included in your original spreadsheet.

 

 

I assume the difference is due to the interpolation mentioned above, but either way, it's close enough for this purpose.

 

Edited to add: Updated version of my earlier spreadsheet includes a graph and the comparison table above:

 

 

System pressure calculator - updated for BH.xlsx 45.39 kB · 2 downloads

Have you installed the extra EV yet?.

 

[jack]

On 29/08/2025 at 23:13, John Carroll said:

I finally got it to work without adding or subtracting from the D values after  reading the above but didn't fill in all the values, 

John Only Expansion Calc using Steam Tables.xlsx 20.73 kB · 4 downloads

 

I don't know why you found the offset being there so offensive! 😆 I could have moved the table so the offset wasn't required, but it was easier to add the offset than to undo the cell merges, move the table, and re-merge the cells.

 

3 hours ago, John Carroll said:

Have you installed the extra EV yet?

 

Not even close. I've figured out which one to get but I need to figure out where to locate it in the plant room. This needs a bit of thinking due to the location of the water softener.

[John Carroll]

A extra 28/30L of EVs to give a total of 50L should satisfy your requirements but with precharge/filling pressures of 2.5bar/2.7bar with both teed into the cold feed if possible, the 22L looks as if its perched on top of the UVC so receiving hot expanded water which further increases final pressure, you may well find that the dynamic pressure at even reasonable flowrates is never higher than 2.5bar anyway, a total of 50L EVs with the above settings will then meet your "worst" conditions of 7C/80C.

 

 

 

 

Edited Wednesday at 07:24 by John Carroll

[jack]

25 minutes ago, John Carroll said:

A extra 28/30L of EVs to give a total of 50L should satisfy your requirements but with precharge/filling pressures of 2.5bar/2.7bar with both teed into the cold feed if possible, the 22L looks as if its perched on top of the UVC so receiving hot expanded water which further increases final pressure, you may well find that the dynamic pressure at even reasonable flowrates is never higher than 2.5bar anyway, a total of 50L EVs with the above settings will then meet your "worst" conditions of 7C/80C.

 

 

 

Thanks John. 

 

Yes, the existing EV is right above the cylinder. I'll take this opportunity to move it to the cold inlet. 

 

Given that the water softener has isolation valves, would it be best practice to connect the EVs after the softener, so they can never be disconnected from the cylinder?

 

Thanks.

[Ajm]

Our UVC has a NRV right at the cold inlet, part of the tank. If yours is like this fitting an EV on the cold won't work!

 

Also should there not be a NRV between the branch for the tank and the feed to the cold manifold to prevent expanded water from the tank feeding into the cold water supply?

 

[jack]

1 hour ago, Ajm said:

Our UVC has a NRV right at the cold inlet, part of the tank. If yours is like this fitting an EV on the cold won't work!

 

Our NRV is right at the mains inlet. There's no separate NRV after the branch to the cylinder from the cold feed.

 

1 hour ago, Ajm said:

Also should there not be a NRV between the branch for the tank and the feed to the cold manifold to prevent expanded water from the tank feeding into the cold water supply?

 

It's funny you mention that. Until I did the diagram above, I assumed we had one. Indeed, I thought it was a requirement (or at least a recommendation), but from a brief review I couldn't find any such requirement. No such valve is shown in the manufacturer's installation diagram, as shown in John's post above.

 

I was going to ask whether it's at least best practice, as this would be a good opportunity to add one. The cold feed to the cylinder is quite long and there's plenty of room to fit an NRV in the cold feed to the cylinder upstream of the expansion vessels.

[John Carroll]

1 hour ago, jack said:

 

Thanks John. 

 

Yes, the existing EV is right above the cylinder. I'll take this opportunity to move it to the cold inlet. 

 

Given that the water softener has isolation valves, would it be best practice to connect the EVs after the softener, so they can never be disconnected from the cylinder?

 

Thanks.

Yes, 

The bottom attachment shows the gist of it.

The combination valve set. below, has, from the right, cold mains inlet, then balanced cold out after the PRV, then the EV connection and out of the top of the expansion relief valve, the discharge to waste, but there is a NRV incorporated in (after) the PRV to stop any expansion back to the mains as suggested by @Ajm

 

 

 

 

 

 

Edited Wednesday at 09:38 by John Carroll

[JohnMo]

Sorry not read the thread - but an observation this is all G3 related stuff being discussed - should you have a G3 certificate to be playing with this?