UFH mixing valve bypassing hot water directly to cold return

[Andy Lewis UK]

I have a UFH manifold from Wundatrade with a Grundfos pump. The pump feed has a mixing valve which has a hot inlet from the thermal store, and a cold inlet from the return manifold. In-between the mixing valve cold inlet and the return manifold, is a tee which returns the cold water to the store. This cold water return to the store has a higher temperature than what is returned to the cold manifold, suggesting hot water is flowing into the mixing valve hot inlet and out of the cold inlet (i.e. reverse flow). The mixing valve inlets are marked blue and red, suggesting the installation is correct. The phenomena occurs no matter the mixing valve setting. Any ideas?

 

[ShaunCorbs]

Tbh sounds normal as your only needing 35-40 dc for the ufh and your store is at 80 normally so half the flow will bypass roughly

 

[John.g]

Is the "cold" water return to the store at the mixing temperature and what mode/setting is the grundfos pump?

 

[Ric2013]

Hmm. I'd need to see a photo as I'm struggling to visualise.

If this is running off a pumped flow and return with another pump elsewhere, you may need a bypass to prevent pumps 'fighting'.

 

[John.g]

Do you mean it might or should have a low loss header?.

 

[John.g]

A small bit of topic but is correct to assume that a UFH TMV will be calling for full HW (boiler) flow when started up from cold, and if so is the manifold flowrate higher until the TMV reaches its setpoint.?.

 

[Ric2013]

A small bit of topic but is correct to assume that a UFH TMV will be calling for full HW (boiler) flow when started up from cold,

From experience, mine seems to, though I wonder if, like tap TMVs, it also allows a small amount of cold at all times?

and if so is the manifold flowrate higher until the TMV reaches its setpoint.?.

Yes.

 

[John.g]

From experience, mine seems to, though I wonder if, like tap TMVs, it also allows a small amount of cold at all times?

Yes.

This might indicate that some TMVs have some adjustment via a allen key, see post #8.

UFH blending valve reduces flow

Hi guys, I'm having trouble with a newly installed UFH system. There are 2 manifolds, one with 5 loops, the other with 9. Both are made by Herz, model no 1732007. Pump is a Wilo Yonic Para something something. Boiler is Vaillant Elo Block 14. My trouble is that setting the TMV to its minimum...

www.plumbersforums.net

 

[Ric2013]

This might indicate that some TMVs have some adjustment via a allen key, see post #8.

UFH blending valve reduces flow

Hi guys, I'm having trouble with a newly installed UFH system. There are 2 manifolds, one with 5 loops, the other with 9. Both are made by Herz, model no 1732007. Pump is a Wilo Yonic Para something something. Boiler is Vaillant Elo Block 14. My trouble is that setting the TMV to its minimum...

www.plumbersforums.net

Ah. I think we're talking at cross purposes. You meant the UFH manifold and I was thinking about a hypothetical LLH manifold from which the UFH manifold may be plumbed, so I've been using the wrong terminology and mislead you.

Start again.

My observations are based on my own UFH at home which I have ample access to.

My blending valve does not have a remote sensor on the manifold. You can turn the valve to a fully shut position and close the hot path off altogether (full recirculation), but I doubt this situation ever occurs in normal use as it would imply a DT of 0°C across the UFH circuit. Whether it ever goes into 100% hot mode (closes the recirculation entirely) or stays slightly open I do not know. There is no evidence that it is mixing in a small percentage of return water at all times, but I have not tested this theory. The only adjustment the manufacturer suggests is to adjust to the correct flow manifold temperature and this is all I have needed to do.

I have observed that when my UFH manifold is cold, the blending valve on the UFH pump inlet seems to take more flow from the boiler primary than when it is hot. Being a thermostatic blending valve, that is exactly what I would expect.

My UFH manifold is plumbed from a non-condensing heat only boiler and so needs to blend the flow temperature down from 70°C to 40°C. The UFH manifold is hydraulically separated by using CCTs on the boiler primaries with the maximum flow rate through the CCTs being regulated by a gate valve (I'm aware ideally it should be a globe valve, but there was not one in my scrap box). The flow through the CCTs is set to be just a little more than required by the UFH mixer valve. Ideally, it would be set to be exactly the flow required by the blending valve, but, being a fixed setting in a world with multiple variables, it needs to be a little more to allow some leeway as I would rather increase the return temperature to my non-condensing boiler than reduce it. Thus, in cold condition, the CCT flow acts as a recirculation path for the return water from the UFH manifold to be fed back to the UFH pump and manifold circuit to maintain whatever flow rate is set; in hot condition, the CCT flow direction is reversed and acts as a very slight system bypass from the boiler flow to the boiler return.

The flow through the UFH manifold pump is, as far as I can tell, a constant.

 

[John.g]

Daughter has a mixed Rad+UFH system which works quite well with no LLH or CCT or any bypass from a SE 20 year old Firebird (oil) Boiler, will probably be changing it sometime (if it ever breaks down) to a HE Grant where the return temp cannot be allowed below 40/45C.
I was thinking of just installing a (auto) bypass in the form of a tapstat where the sensor is attached at the boiler return to control the bypass valve, it might give very tight control of the boiler return temp even though I've never seen it suggested anywhere?.
Two systems compared below, I'm assuming a 5kw UFH system with a dT of 7C.

 

[Ric2013]

I think I see where you're coming from. I originally configured my UFH and radiator system without hydraulic separation of any kind. The question of starving the radiators is a bit academic when the UFH return warms up in only 7 minutes and I never noticed a problem with the smart pumps conflicting. It was Y plan, no bypass was needed, and it worked happily.

I did, briefly, experiment with a setup similar to yours. I set the flow across the bypass to a fixed rate to bring the DT across the flow and return, with the UFH circuit in a heated state, to around the same as the design DT for the radiators. I can see that a tapstat could be used to maintain a set return temperature no matter whether the UFH circuit is hot or cold. I have seen solid fuel (wood) boiler with what must be an essentially similar setup, although I think the one I saw had electronic rather than mechanical thermostatic control. The flow is short-circuited until the boiler is at 60° and only then does flow to the emitter gradually start.

 

[Ric2013]

Shall we try to get back to the OP?

Andy, I've read what you state above again several times. I get a bit confused and it sounds like what you have is normal and then I read it again and realise it isn't.

The return to the store should be at exactly the same temperature as the return manifold. In your case the pipe that should be the return to the boiler is at a higher temperature than the UFH return manifold.

That said, I have come across something a bit like this. At the time, I wasn't as knowledgeable about how UFH should connect to a system and wasn't really able to make sense of it, but it was along these lines... In what is almost a mansion, a system had been set up (by others) with UFH and radiators and the UFH was a large part of the large overall system. No attempt had been made to balance the system properly. But more critically, there was no attempt at hydraulic separation and much of the pipework was hidden under floors and expensive new carpets. If I remember correctly, the UFH took its flow from one pipe connected to the boiler and put its return to another pipe connected to the boiler. The system hadn't really been planned with any level of understanding and the UFH had literally been connected to the flow and return wherever seemed most convenient. I know I did what I could, which was balance the radiators and set the system pump to a more appropriate setting and this seemed to help, but I never did quite work out how to correct this, though now I think I have an affordable solution which might be a close coupled tee system similar to my own. Unfortunately I've lost the customer in the meantime, so I'm unable to look at the system again.

If you aren't in a mansion, it may be you have a similar problem, but I'm sure your pipework will be much easier to understand!

 

[John.g]

Can you post a photo of your manifold if not like this one Andy..
Is the temperature on either side of the TMV return definitely different, just wondering if you have a low loss header or another heat exchanger, (unlikely with a H.Store).

 

[Ric2013]

on either side of the TMV return

Can you redefine that please? I'm not sure what you mean, so the OP (if he's still lurking) probably won't either.

 

[John.g]

What I intended to mean is to check the (manifold) return temperature to the right of the temperature gauge (bottom) and check it 6 ins or so below the temperature gauge all three should be the same if no leakage past the TMV?.

 

[John.g]

Ah. I think we're talking at cross purposes. You meant the UFH manifold and I was thinking about a hypothetical LLH manifold from which the UFH manifold may be plumbed, so I've been using the wrong terminology and mislead you.

Start again.

My observations are based on my own UFH at home which I have ample access to.

My blending valve does not have a remote sensor on the manifold. You can turn the valve to a fully shut position and close the hot path off altogether (full recirculation), but I doubt this situation ever occurs in normal use as it would imply a DT of 0°C across the UFH circuit. Whether it ever goes into 100% hot mode (closes the recirculation entirely) or stays slightly open I do not know. There is no evidence that it is mixing in a small percentage of return water at all times, but I have not tested this theory. The only adjustment the manufacturer suggests is to adjust to the correct flow manifold temperature and this is all I have needed to do.

I have observed that when my UFH manifold is cold, the blending valve on the UFH pump inlet seems to take more flow from the boiler primary than when it is hot. Being a thermostatic blending valve, that is exactly what I would expect.

My UFH manifold is plumbed from a non-condensing heat only boiler and so needs to blend the flow temperature down from 70°C to 40°C. The UFH manifold is hydraulically separated by using CCTs on the boiler primaries with the maximum flow rate through the CCTs being regulated by a gate valve (I'm aware ideally it should be a globe valve, but there was not one in my scrap box). The flow through the CCTs is set to be just a little more than required by the UFH mixer valve. Ideally, it would be set to be exactly the flow required by the blending valve, but, being a fixed setting in a world with multiple variables, it needs to be a little more to allow some leeway as I would rather increase the return temperature to my non-condensing boiler than reduce it. Thus, in cold condition, the CCT flow acts as a recirculation path for the return water from the UFH manifold to be fed back to the UFH pump and manifold circuit to maintain whatever flow rate is set; in hot condition, the CCT flow direction is reversed and acts as a very slight system bypass from the boiler flow to the boiler return.

The flow through the UFH manifold pump is, as far as I can tell, a constant.

Could you please post a simple schematic of your systems please including the CCTs, have you 3 circulating pumps in all?.

 

[Ric2013]

Could you please post a simple schematic of your systems please including the CCTs, have you 3 circulating pumps in all?.

2 pumps, 3 circulating valves.

The CCT is just below the UFH blending valve. It would arguably be better to have the motorised valve for the DHW coil as a 3-port, with the zone closed position being through the bypass as, if there weren't an additional feature (not shown), this system might potentially reverse circulate as it does, in effect, break the 3 tees 'rule'.

Unfortunately, the cast-iron heat exchanger in the boiler needs a certain amount of heat dissipation, so a longer bypass run is preferable. Unfortunately, this is one of those systems where you just just can't meet the requirements of everything due to the MIs of the old components being along different lines of thinking from what is now 'best practice'.

 

[John.g]

The main difference I see between the systems that give trouble ie the UFH robbing the Rads and your system is the CCT? (which looks more like a bypass?) Is the UFH return throttled in?, what are the makes/models and modes/settings of the primary and manifold circ pumps.

 

[Ric2013]

The main difference I see between the systems that give trouble ie the UFH robbing the Rads and your system is the CCT? (which looks more like a bypass?) Is the UFH return throttled in?

It is throttled. That is the gate valve used to limit maximum flow through the CCTs and adjusted to just more than the flow required by the warmed-up UFH as previously described.

In fairness, I previously ran this system without the CCTs and there was no evidence of it causing a problem, so not sure if this discussion is relevant to the OP's problem (which wasn't directly related to radiators being robbed) or not. What I would say is that it is possible my previous system did rob the radiators at start-up, but since the UFH circuits content only takes around 7 minutes to change, it was a non-issue. In a sense, the UFH was previously able to make use of the (oversized) 12kW output of the boiler, whereas it's now limited to the 1.4kW which the gate valve is allowing to pass, so the revised system is not an improvement on all fronts.

, what are the makes/models and modes/settings of the primary and manifold circ pumps.

UPS2 15 50/60 on the primaries, set to PP3

UPM3 AUTO 25/70 on the UFH manifold, set to CP1

 

[John.g]

Very interesting.

Is the modified system below the same as your system?, if so, what, in your opinion,do the CCTs achieve.

 

[Ric2013]

Very interesting.

Is the modified system below the same as your system?, if so, what, in your opinion,do the CCTs achieve.

Essentially yes.

What do they achieve? On my system, not a great deal. It was an idea Heat Geek had about how having two 'smart' pumps essentially in series would confuse/upset the pumps as the effect of one pump responding to feedback from its circuitry would affect the other pump and reduce their lifespan and I wanted to try the CCTs to see what happened. I hadn't originally put CCTs in as I was of the opinion that the UFH would essentially only take what it needed from the primaries, and so it seemed to. Whether the pumps will last longer is impossible to say as there is no control experiment.

I can see they might be needed if the maximum flow through the UFH exceeded what the system pump or boiler could allow, but as my boiler is running at 12kW output (IIRC) and my UFH load is around a tenth of that (between 1 and 2 lpm), I suspect the question is academic. If they do anything, I suspect they extend the UFH warm-up time but reduce the radiator warm-up time when both zones call for heat in the morning.