Intermittent hot water air lock problem

Situation is a flat with roughly the following configuration. The hot water flow in all the flats is poor due to relatively low head of water - even with the old style of taps. The pipework also has to dip under the floor to get across the hall where the front door is.



So, the hot water is boosted at the tank with a Grundfoss Home Booster pump which has a flow switch (and the low pressure cold to the thermostatic bath mixer is also boosted with the same type of pump as it has a shower attachment on it).

This setup works really well and has allowed modern taps to be fitted. The only tweak needed was to fit a high flow aerator to the bathroom mixer tap as the flow with the standard one was too low to encourage the flow switch on the booster pump to trip in.

The problem is this setup occasionally seems to air lock and I can't understand why given that nothing is above the level of the cold water tank and its a relatively simple pipe run. We're talking maybe once every 3-6 months (AND after every occasion where the hot water stop valve is turned off at the input to the tank and the pipes to the taps drained down for any reason).

The symptoms of the intermittent air lock are a gradual reduction in flow from the taps and the pump not kicking in. If you notice and catch it early enough, simply turning on all the taps can sometimes sort it - where the pump then kicks in and presumably helps shift the lock judging by all the gurgling noises.

But if that isn't done, the water flow will eventually slow to a trickle and it doesn't matter what you do with the taps at that point - it won't shift without help. Even switching the booster pump to manual operation which I would have expected would blast any air lock through, doesn't work if it gets to this point.

So, what I usually do is draw off a bit of water at the drain point under the kitchen tap and after a few seconds, the pump will kick in and the problem will sort itself out with loads of gurgling. With it being a dual flow mixer in the kitchen, I would have expected opening that tap up would have had the same effect - but it clearly doesn't. The only thing I haven't tried is taking the aerator/filter off the kitchen mixer when this situation occurs to see if that sorts it.

Any ideas? Thanks.

Nice CAD drawing.
How large is the cold cistern...is it running dry or too small?
Your cold mains inlet to the cistern could be aerating the water...move it away from the outlets.
Is your cylinder vent pipe dipping down at any point.
Some of these cylinder flanges can block with scale and prevent proper air venting.
Have you complied with the minimum head for the pumps...I'd have thought they would be better down at floor level.
Peripheral (regenerative) pumps would be better than centrifugal pumps...not sure what the Grundfos is.

Didn't fancy explaining it! Comments added below. Thanks!

Gasguru said:

Nice CAD drawing.
How large is the cold cistern...is it running dry or too small?

>Tank is 50 gallon. The home booster is only a light boost (0.5-0.75 bar) to bring it within the operating range of most modern low pressure taps + cold water pressure is very good, and far greater than the pump can shift - so I'm pretty certain the cold isn't emptying. Only has one elderly person living in it so it won't be the case that more than one tap is in use at any time either.

Your cold mains inlet to the cistern could be aerating the water...move it away from the outlets.

>Inlet is on the opposite side of the tank from the outlet so that shouldn't be a problem.

Is your cylinder vent pipe dipping down at any point.

>No

Some of these cylinder flanges can block with scale and prevent proper air venting.

>It's all been re-done over the last couple of years, it's not a hard water area and this has been a problem since the changes were first done. The Warix is definitely connected the right way too. So pretty certain there is no vent issue.

Have you complied with the minimum head for the pumps...I'd have thought they would be better down at floor level.
Peripheral (regenerative) pumps would be better than centrifugal pumps...not sure what the Grundfos is.

>The cold water booster is actually further down towards ground level than I've shown it, and I'm pretty certain it complies with the 1.5M head recommended for the pump. The HOT booster is where I've shown it - horizontally from the Warix flange on the top of the hot tank so that will be less than 1.5M - probably 1M'ish. It would be difficult (i.e. hot tank out job) but not impossible for me to change the pipework so the pump is down at floor level and still pumping horizontally (it shouldn't be installed pumping directly downwards). If I was certain that was causing it, I'd make the change - but I'm just struggling to understand how.

Click to expand...

How is the cylinder heated?
Electric immersion heater thermostats are notorious for being inaccurate (even brand new)...it should cut-out at 60 degrees but I've had many at 70 or even 80 degrees.
Ideally they need checking before installation.
If the cylinder's heated with a boiler is the cylinder thermostat in contact with the wall and set to 60 degrees.

Gasguru said:

How is the cylinder heated?
Electric immersion heater thermostats are notorious for being inaccurate (even brand new)...it should cut-out at 60 degrees but I've had many at 70 or even 80 degrees.
Ideally they need checking before installation.
If the cylinder's heated with a boiler is the cylinder thermostat in contact with the wall and set to 60 degrees.

Click to expand...

Two immersion heaters - one on E7, and the other as a boost if required (hardly ever used). New branded immersion heaters (Tesla Incoloy, Screwfix item 53389) were fitted at the time rest of the work was done. These include a dual-safety stat so if there was any over-heat event, that would have actually tripped the stat, requiring a physical reset.

Doesn't look like you've got an anti grav loop on the HW, if the pumps above the tank? Air's obviously getting into the pump/pipework from somewhere.

Madrab said:

Doesn't look like you've got an anti grav loop on the HW, if the pumps above the tank? Air's obviously getting into the pump/pipework from somewhere.

Click to expand...

Thanks Rob. I've seen reference to the anti-grav loop for fitting shower pumps when the pump is above the level of the tank. In this case, it is literally just above - i.e. on a Warix, the pump is the upward facing spigot and that is turned immediately at 90 degrees and there is just around about a 3 inch bit of pipe to the pump. So I wasn't expecting there would be much chance of an air lock in and around the pump.

But thinking about it, I remember seeing that the bore of the pump is smaller than the 22mm pipework so when the system is filled with the pump switched off as it typically would be, the flow through will be restricted. The output from the pump runs for another 1.5ft horizontally, and then straight down into the floor. I guess that could result in air pockets in the top of the pipework where the pump is that might not shift easily even when the pump itself is bled.

What puzzles me though is where a shower pump is shown below the height of the hot tank in diagrams, it would still seem possible for an air lock to occur in the pipework coming out of a top flange in roughly the same place, before it goes down to the pump.

If we're thinking the air lock is in the pipework around the pump (still not clear why it causes such an infrequent problem....), could a way to test this be to add an automatic air valve each side of the pump? The 90 degree out the Warix is a compression, as is the 90 degree down to the floor (i.e. you can just undo the nuts and lift the pump and horizontal pipework off in one piece). If the bends are replaced by tees, the auto air vents can easily be added.

Don't think of it as an air lock per sé, rather air bubbles from the hot cylinder getting into the pump impeller when the pump isn't running, once that happens you get cavitation & the impeller blades won't grab the water properly and slowly but surely the flow dies, then the switch turns off the pump, opening other taps increases the flow through the impeller and pushes the air out.

Flanges remove a lot of the air bubbles but not all and air bubbles will naturally rise to the highest point and get into the pump. The loop stops them getting into the pump as they can't get around the loop.

I'd assumed that any formation of bubbles in the hot tank (e.g. around the immersion heater element when its heating overnight) would have floated up and vented before the taps are used. But I guess its still possible that there are enough bubbles still around to be drawn into the side holes of the tube of the Warix when the pump is running. It would explain the fact it takes so long for the problem to show up too.

Decision time - if I was going to add an anti-grav loop, I might just as well relocate the pump down towards the floor and dispense with the need for the loop altogether. The pump isn't in the most accessible place as it is + it will then have the correct head of water above it too.

Thanks Rob.

It's a process of elimination. You remove out the chances that air bubbles can get to the pump and if it still has issues then that's one area ruled out and time to look at other areas.

Just to close this one down Rob, I relocated the pump down to floor level a few weeks ago, and I also decided to put a manual air vent at the highest point where the hot comes out the warix. No problem filling the system back up afterwards (that was a first!), and it has behaved fine since. I've just checked the vent this evening and it had trapped a tiny bit of air over the weeks which looks like it ties in with the fact it can go months before the problem occurs. So I'll just check it from time to time. Thanks again!

No Bother @YorkshireMidge ... good stuff that seems to have settled it down .... things always go so much easier when they're installed as per the norm ... anything else and it just gets complicated