Search the forum,

Discuss System sucking in air - suggested fix is convert to closed - looking for more info in the Central Heating Forum area at PlumbersForums.net

Messages
4
Hi everyone,

I'm a newbie here, a layperson homeowner, stopping by in desperation to try and get a better understanding of my options regarding a diagnosis.

I have a conventional gravity heating system (cold water tank).

British Gas put in a new boiler and replaced most of the other bits (pump, condenser, flue) back in 2011.

The current boiler is a Worcester Greenstar 18Ri.

Over the last few years I have had nothing but problems. First the pump became really noisy. Then the boiler kept getting a flashing light and turning off.

I have had the same heating engineers working on it for at least three years and have had done everything they said needed doing. I must've spent at least £2000 so far.

I've had the boiler serviced every year, the pump replaced twice, a full power flush, new 3 way valve, several new rads, a magnaclean, new programmer and finally a new condenser.

And yet, still the blasted boiler keeps turning off with a flashing blue light.

Now the engineer is saying that the issue is that the system is sucking in air because the original pump was installed with "not enough head room". Apparently the newer pump is more powerful and is now causing the sucking in of air.

Every time it has happened I have to wait for them to come out and get rid of the air at £90 plus VAT per time. Plus we are without heating for the whole time which is a major issue as my mother is 72 and a cold will put her in hospital.

We were desperate the last time and Coronavirus had just started so we got another guy out cause he could come quicker. He said the same thing. He bled the system, turned the pump down, showed us how to reset the boiler and told us to see how that went. He said if having the pump set lower didn't solve the issue then the best choice was to convert the system to a closed system. At a cost of at least £500 not including VAT.

And then of course Coronavirus hit properly. It's happened again twice in two days and now I'm stuck.

I've googled converting an open system to a closed system but I can't find anything about it except for the more modern Megaflo type setups.

I'm hoping someone here can maybe explain a bit more about it...?

What does not enough head room even mean?

Is converting the system really the only thing that can be done?

What's involved in a conversion?

What are the implications of converting it with regards to maintenance and servicing etc?

Neither of the current engineers are able/willing to explain any of this in laymans terms.

I'm sick of throwing money at this problem. Then next time I do I want to be sure it's going to be the last time for a good long while.

Can anyone help me understand? Any help would be greatly appreciated. I've got a bunch of photos of the setup.

Thanks for reading my post.
 

Attachments

  • BG quote.jpg
    BG quote.jpg
    124.4 KB · Views: 22
  • boiler.jpg
    boiler.jpg
    67.6 KB · Views: 23
  • condenser.jpg
    condenser.jpg
    66.3 KB · Views: 19
  • hotwatertank.jpg
    hotwatertank.jpg
    122.3 KB · Views: 21
  • pump.jpg
    pump.jpg
    81.5 KB · Views: 24
  • magnaclean.jpg
    magnaclean.jpg
    100.9 KB · Views: 23
  • valve.jpg
    valve.jpg
    97 KB · Views: 24
A closed heating system just very basically speaking is the wee feed/expansion tank that is normally in loft to be removed and pipes to it capped off.
The system, (boiler type permitting), then requires a safety valve (pressure relief valve of usually 3bar) and an expansion vessel and a filling loop from mains water pipe connected to your heating system to top up pressure manually.
It shouldn’t be expensive to do, but your system must be sound, with no leaks or slight weeps including rad valves.
Remember a sealed heating system will be filled to and operate at a higher pressure to what it was, perhaps 3 or 4 times the pressure, so always a risk of dodgy work leaking, although hopefully all will be okay.
Do not be attempting this yourself as you could cause damage or leave system dangerous.
Your copper cylinder is a different water to your heating system and can remain untouched.
 
"Not enough head room" he means the static head above pump. This should be a minimum of maximum pressure created by pump divided by three. If you dont meet this requirement you can get what's called cavitation, this is one possibility to gases forming where the low pressure water boils at lower temperatures and releases gas.
Another possibility is the orientation of pump, cold feed and vent. It's hard to tell from the pictures but it should be oriented vent and cold feed within 150mm of each other on the flow and the pump afterwards. If not installed correctly you could be drawing air in.
When bleeding the system is it neat air or is it hydrogen? If it's the latter then theres possibly a few other things going on as well.
 
"Not enough head room" he means the static head above pump. This should be a minimum of maximum pressure created by pump divided by three.
Where did that formula come from?

Acording to Grundfos the required static head depends on the water temperature, not the max pump pressure: e.g Alpha3 (4m to 8m head) 0.5m @ 75°C, 2.8m @ 90°C, 10.8m @ 110°C.
 
Where did that formula come from?

Acording to Grundfos the required static head depends on the water temperature, not the max pump pressure: e.g Alpha3 (4m to 8m head) 0.5m @ 75°C, 2.8m @ 90°C, 10.8m @ 110°C.

That's from my old college text books, admittedly that was from 15 years ago and a bit before modulating pumps took off but a reasonable standard to work to
 
A closed heating system just very basically speaking is the wee feed/expansion tank that is normally in loft to be removed and pipes to it capped off.
The system, (boiler type permitting), then requires a safety valve (pressure relief valve of usually 3bar) and an expansion vessel and a filling loop from mains water pipe connected to your heating system to top up pressure manually.
It shouldn’t be expensive to do, but your system must be sound, with no leaks or slight weeps including rad valves.
Remember a sealed heating system will be filled to and operate at a higher pressure to what it was, perhaps 3 or 4 times the pressure, so always a risk of dodgy work leaking, although hopefully all will be okay.
Do not be attempting this yourself as you could cause damage or leave system dangerous.
Your copper cylinder is a different water to your heating system and can remain untouched.

It may be worth combining the cold feed and the vent as this will stop any pump over/pump back and is very easy to implement.
OR, a semi sealed system could be considered as its far more forgiving than the fully sealed system, they were fitted to literally dozens of houses around me where people changed solid fuel to gas.
You just blank the vent and install a swing check valve in a horizontal (or make a short piece horizontal) part of the cold feed, install a 12 litre E.vessel as close to the pump suction as possible with a pre charge pressure of 0.5 bar, the final pressure, after expansion will only be 0.7/0.8 bar and should solve the air problem.
 
That's from my old college text books, admittedly that was from 15 years ago and a bit before modulating pumps took off but a reasonable standard to work to
A 1997 Selectric document says 1.3m @ 75C and 1.4m @ 85C; and that's for a non-modulating 6m pump. That's nearer the 2m which your formula gives for a 6m pump. In any case, too much is better than too little.
 
Thats my old R.D Treloar plumbing and heating guide from college as I said. A rough guide and like you said better to have too much than not enough. I've had a quick look at a couple modern pumps specs and as you say the requirement is stated in temperatures. At the end of the day as long as one understands the importance of avoiding cavitation that's all that matters
 
A 1997 Selectric document says 1.3m @ 75C and 1.4m @ 85C; and that's for a non-modulating 6m pump. That's nearer the 2m which your formula gives for a 6m pump. In any case, too much is better than too little.
Where did that formula come from?

Acording to Grundfos the required static head depends on the water temperature, not the max pump pressure: e.g Alpha3 (4m to 8m head) 0.5m @ 75°C, 2.8m @ 90°C, 10.8m @ 110°C.

For interest, have any of you links to the NPSH for these pumps/temps. I calculated that the NPSH based on the three readings above are 6.45M,5.53M&6.07M. The flow rate, as well as temperature has a influence on the head required.
 
For interest, have any of you links to the NPSH for these pumps/temps. I calculated that the NPSH based on the three readings above are 6.45M,5.53M&6.07M. The flow rate, as well as temperature has a influence on the head required.

I can't provide you with a link to the "Net Positive Suction Head" but there are many calculators online (engineering toolbox i use a lot) to demonstrate the relationship between temperature and saturation pressure, although i think you have a grasp on this? In order to avoid cavitation the NPSHA must be equal to or greater than NPSHR.
[automerge]1586210476[/automerge]
Update.

I've dug my old college books out to double check what I thought I'd remembered and I was wrong.
The maximum delivered head of pump divided by three is in relation to prevent undue movement in header tank when pump is circulating.
I was prompted to dig it out when my calculations just weren't adding up, I feel like an idiot and I apologise for the mistake.
For an average flow temperature of say 75°c the minimum pressure required on suction side of the pump equates to 0.386 bar. 0.386 bar is NPSHA required to avoid cavitation, NPSHA (Net Positive Suction Head Actual) must me equal to or greater than NPSHR (Net Positive Suction Head Required).
So to recap as long as the pump has a NPSHA equal to or greater than 0.386, or just under 4 metres of head then cavitation will be avoided.
[automerge]1586213428[/automerge]
Nope thats not it either
[automerge]1586213789[/automerge]
I can't provide you with a link to the "Net Positive Suction Head" but there are many calculators online (engineering toolbox i use a lot) to demonstrate the relationship between temperature and saturation pressure, although i think you have a grasp on this? In order to avoid cavitation the NPSHA must be equal to or greater than NPSHR.
[automerge]1586210476[/automerge]
[automerge]1586214503[/automerge]
I've dug my old college books out to double check what I thought I'd remembered and I was wrong.
The maximum delivered head of pump divided by three is in relation to prevent undue movement in header tank when pump is circulating.

The available NPSH is absolute pressure head minus saturation pressure.
 
Last edited:
I used to do these calcs many moons ago in imperial units so see what you make of these.
I have always used Spirax Sarco "Steam Tables" to obtain the vapour pressures at different temperatures.


The vapour pressure of water at 75c is 3.94M (0.384bar) but better to keep to meters as you will see below. So, as above, the NPSHR is 3.94M
(atmospheric pressure= 10.39M absolute)

NPSHA = Static Head (or lift)(M) - suction line losses(M) + atmos.press (M. absolute) - vapour pressure (M.absolute). but friction losses can be ignored if the cold feed is just in front of the pump suction so the sum becomes...…??
NPSHA = Static Head(M) + atmos.press (M. absolute) - vapour pressure (M.absolute)

Now Grundfos stated that "1.3m @ 75C and 1.4m @ 85C" is the static head required so lets see.
NPSHA (at 75C)= 1.3+10.39-3.94 or 7.75M which is fine as it is 3.81M > NPSHR of 3.94
The vapour pressure at 85C is 5.9M so.....
NPSHA (at 85C)= 1.4+10.39-5.9 or 5.89M which is not so fine as it is actually less than the NPSHR of 5.9M?? so if the same 3.81M differential as @75C is desirable then the static head at 85C should be ~ 5.2M?
 
Last edited:
Sorry for the confusion last night, I'd had a few beers and was getting it partly wrong. I tried editing it but to no avail.
The saturation pressure of water at 85°C is 0.5787 bar. So in quick terms the NPSHA is absolute pressure minus the saturation pressure. So say we have a header tank 2m above pump inlet and a saturation pressure of 0.5787 bar, the absolute pressure is roughly 1 bar (1atm) plus static head equalling roughly 1.02 bar, now minus saturation pressure and you get 0.4413 NPSHA, now compare that to NPSHR for the Grandfos Alpha at 90°C and you see that at that temperature a head of 0.28 bar or 2.8 metres head is required to avoid cavitation and our NPSHA is 0.4414 bar. In fact if I'm getting my calculations right then that pump could be installed 6 inches below header tank and not cavitate?
The other problem with pump being mounted close to header tank fill line is it can create a area of low pressure in the vent actually drawing the water line down and causing air ingress that way?
 
Sorry for the confusion last night, I'd had a few beers and was getting it partly wrong. I tried editing it but to no avail.
The saturation pressure of water at 85°C is 0.5787 bar. So in quick terms the NPSHA is absolute pressure minus the saturation pressure. So say we have a header tank 2m above pump inlet and a saturation pressure of 0.5787 bar, the absolute pressure is roughly 1 bar (1atm) plus static head equalling roughly 1.02 bar, now minus saturation pressure and you get 0.4413 NPSHA, now compare that to NPSHR for the Grandfos Alpha at 90°C and you see that at that temperature a head of 0.28 bar or 2.8 metres head is required to avoid cavitation and our NPSHA is 0.4414 bar. In fact if I'm getting my calculations right then that pump could be installed 6 inches below header tank and not cavitate?
The other problem with pump being mounted close to header tank fill line is it can create a area of low pressure in the vent actually drawing the water line down and causing air ingress that way?

Ok, my method is always using head in M
vapour pressure at 85c is 5.9M
NPSHA = Static Head(M) + atmos.press (M. absolute) - vapour pressure (M.absolute)
Let H = Minimum static head required then:
5.9 = H+10.39-5.9 so:
H (minimum static head required) is 1.41M

vapour pressure at 90C is 7.16M
NPSHA = Static Head(M) + atmos.press (M. absolute) - vapour pressure (M.absolute)
Let H = Minimum static head required then:
7.16 = H+10.39-7.16 so:
H (minimum static head required) is 3.93M so if my calcs are correct then a static head of 2.8M will not satisfy the minimum requirements?.

Re: your calc above, you said "the absolute pressure is roughly 1 bar (1atm) plus static head equalling roughly 1.02 bar" but the static head is 2M or 0.2bar so 1.02 bar should be 1.2bar, this then gives you a NPSHA of 0.6213 bar?.
 
Yes the 1.02 was a mistype. So 1 atm (the pressure exerted on top of water in tank) plus static head would be 1.2 bar in my example, then you would minus the saturation pressure to find NSHPA. NSHPA and NSHPR are always actual pressures, not gauge pressure.
So water saturation pressure at 90°C is 0.7018 bar,
1.2 bar minus 0.7018 = 0.4982 NSHPA which is lower than saturation pressure and cavitation will occur. If in any time the NSHPA is less than required then you would either relocate pump and alter the feed and vent in an open vented system or increase the height of header tank.
I'm sure I've got my calculations right, I'm a bit more with it now lol but bare in mind to find NSHPA I'm simply taking actual pressure and minusing the saturation pressure (which I'm using an online calculator to find), it doesn't take into account the head losses due to friction in the suction line, which should be minimal anyway.

Edited due to yet another stupid error 😒
 
Last edited:
I may be wrong but I don't see it that way, the vapour pressure at 90C is 0.7018bar so this is the minimum NPSH required??, the NPSHA is (1.0 - 0.7018) or 0.298bar so the required static head is 0.7018-0.298 or 0.4048bar, say 4M.
So now the NPSHA= 0.4048+1-0.7018 = 0.703 or nearly the vapour pressure
 
My online calculator says saturation pressure of water at 90°C is 0.7018 bar. So to find NSHPA you take actual pressure 1.2 bar and minus saturation pressure giving a result of 0.4982. This is below saturation pressure and cavitation occurs. As long as NSHPA is equal to or greater than NSHPR then its not a problem?
 
Its quite a complex formula, the method i stated above is basic but gives a reasonably accurate NSHPA result. I have saved a couple calculators which are a bit more in depth. Perhaps I'm getting it wrong. Bottom line is the NSHPA must be equal to or greater than NSHPR or saturation pressure.
 
Yes, you are correct there so static head required in my calcs is 3.23M and not 3.99M, I think that it is common practice for the pump manufacturers to add +0.5M to the NSHPR. so the 2.8M grundfos number does seem a little low. I would love to see their figures for the newer types as a lot of people seem to be getting cavitation/noisy operation, I hope their NPSHR hasn't increased on these pumps, for whatever reason
 
I was going to point out the figures of that pump weren't adding up for me either. As long as the NSHPA is greater than saturation pressure then cavitation is avoided. Surely this is standard science and the figures are what they are, how could a pump manufacturer change its required minimum pressure?
 
They can't make the minimum lower than the theoretical, but poor design can certainly raise it.
[automerge]1586279368[/automerge]
I think I am back on track now.....I was actually correct in post #11 but lost my way after that.
I found a reference to the static head required for a Wilo Yonis Pico (like my own) and it requires 3M @ 95C.
Vapour pressure @ 95C = 8.63M Atmos pressure = 10.34M
NPSHA = Static Head(M) + atmos.press (M. absolute) - vapour pressure (M.absolute)
NPSHA = 3+10.34-8.63 or 4.71M
Even if the static head were 0 then the NPSHA would still be 1.71M.

Going back to the old selectric which requires 1.4M @ 85C
Vapour pressure @ 85C = 5.9M Atmos pressure = 10.34M
NPSHA = Static Head(M) + atmos.press (M. absolute) - vapour pressure (M.absolute)
NPSHA = 1.4+10.34-5.9 or 5.84M
Even if the static head were 0 then the NPSHA would still be 4.44M.
 
Last edited:
Grundfos states the following for the current Alpha1 and Alpha3 Pumps

Liquid temperature75 °C90 °C110 °C
Inlet pressure0.5 m head2.8 m head10.8 m head
0.005 MPa0.028 MPa0.108 MPa
0.05 bar0.28 bar1.08 bar

For the UPS3 the numbers are:

Liquid termperature75 °C95 °C
Inlet pressure0.005 MPa0.05 MPa
0.05 bar0.5 bar
 
3m static head plus atm is roughly 1.3 bar minus saturation pressure of.0.8461 bar for a temperature of 95°C gives NSPHA of 0.45 bar, which is below the requirement?
 
A static head of 3.3M is required IF 0.5 bar is the NPSR?

Vapour pressure @ 95C = 8.63M Atmos pressure = 10.34M
NPSHA = Static Head(M) + atmos.press (M. absolute) - vapour pressure (M.absolute)
NPSHA = 3.3+10.34-8.63 or 5.01M (0.5bar)
 
I think we're getting a bit muddled between us. I will attach an image if not for your benefit but others.
My understanding is simple. The NPSHA must be equal to or greater than the saturation pressure. If I'm reading you right from your example above the 3.3m static head is not enough for a flow temperature of 95°C to avoid cavitation. By my calculations you would need a static head of roughly 8m above centre line of pump, whether you raise header tank or lower position of pump and of course altitude will have an effect as well on atmospheric pressure.
Screenshot_20200406-203703_Samsung Internet.jpg
 
I'm happy enough with my interpretation of it, At 95C the vapour (saturation) pressure is 8.63M BUT the atmospheric pressure alone is 10.34M and any pressure > the vapour pressure of 8.63M will stop the vapour forming, Even with zero static head the atmospheric pressure is giving a NPSA of (10.34-8.63) or 1.71M but Grundfos have a NPSR of 5M so a tank with its water level 3.3M above the pump CL must supply the remainder, (static head of 3.3M.)
so the NPSHA=NPSR?
 
NPSHR is minimum pressure required to avoid cavitation, so in essence equal to saturation pressure.
 
The Americans generally explain things in simple terms, try this very old explanation that I have had for years, It talks about NPSH, NPSHA and NPSHR.
The first few paragraphs explains it well, I think.
 

Attachments

  • NPSH1.pdf
    757.9 KB · Views: 14
Last edited:
My first port of call would be to ensure there is enough height on the vent pipe for the header tank. I can't see the pipework properly in your pictures however it could be pumping over the vent.
This can cause all sorts of strange problems including air being dragged round the system and finding itself in the boiler causing lockout.
Also are the Flow and Return pipework in the correct orientation?
 
450mm is commonly given as the minimum height from the top of the vent to the F&E tank level.
There is a formula that uses (static height) the distance from the F&E tank level to the lowest point in the system, normally the boiler return. The calc is this height (in meters) X 40) + 150. If this static height is like my 2 storey house, 5 meters, then the calc is 5x40 + 150 or 350mm.
 

Attachments

  • Vent Pipe Height.jpg
    Vent Pipe Height.jpg
    16.9 KB · Views: 16
Last edited:
450mm is commonly given as the minimum height from the top of the vent to the F&E tank level.
There is a formula that uses (static height) the distance from the F&E tank level to the lowest point in the system, normally the boiler return. The calc is this height (in meters) X 40) + 150. If this static height is like my 2 storey house, 5 meters, then the calc is 5x40 + 150 or 350mm.
I have often wondered how the 450mm was calculated; but I don't see how the distance from tank level to the lowest point is relevant.
 
Yes, I've wondered a few times about it myself, the calculation part is also used I think for DHW cylinders except the measurement is taken from the bottom of the HWC but in either case the levels in the header tanks will only rise a few inches taking the expansion and in the event of cold water supply blockage will rise up the vent (open safety valve, or OSV) and overflow into the tank. You would need the vent sticking up through the roof to accommodate the expansion in either case without overflow. I can sort of understand the vent rising up 450MM or whatever in the vented boiler case as it possibly prevents the water overflowing when the circ pump starts/stops causing surging. The "X40" calculation might have some basis in the fact that water expands 4% at 100C?.
 

Reply to System sucking in air - suggested fix is convert to closed - looking for more info in the Central Heating Forum area at PlumbersForums.net

Creating content since 2001. Untold Media.

OFFICIAL SPONSORS

OFFICIAL SPONSORS

Back
Top
AdBlock Detected

We get it, advertisements are annoying!

Sure, ad-blocking software does a great job at blocking ads, but it also blocks useful features of our website. For the best site experience please disable your AdBlocker.

I've Disabled AdBlock