Problem with underfloor heating

[MikeH]

I've recently had wet underfloor heating installed in my downstairs shower room. I specified that it needed to be *fully* independent of the main heating system so that I can set it to turn on and off earlier than the main heating and can specify a night set-back temperature to reduce the initial warm-up time in the morning.

My existing heating system is a pumped Y-plan layout with a conventional condensing boiler and a indirect H/W cylinder fed from a tank in the loft. The UFH extension has a separate pump, room thermostat with integral timer/controller and the usual two-port manifold, mixing valve (Danfoss VMV), etc to provide water to the underfloor pipewark at around 45 degrees C.

Initially I had a problem with the system in that because the 'normal' state of the original Y-plan 3-port valve is to open the port to the H/W cylinder, if the UFH was on it would also heat the water in the tank dangerously above the 55-60 degrees set on the cylinder stat. The plumber then put an additional two-port valve between the 'A' output of the three-port valve and the feed into the coil of the H/W tank. This ensured that the tank was only ever heated if the tank stat was calling. It also means that the original Y-plan 3-port valve now effectively has a 'fully closed' position when neither the H/W or C/H stats are calling for heat.

What I now find is that within a few minutes of being switched-on, the UFH causes lots of banging and whooshing and the water in the header tank gets very, very hot! The temperature on the output of the UFH mixing valve also gets much higher than 45 degrees - it once briefly got to 75 before I switched it off.

I think the plumber has fixed the overheating H/W tank problem, but has introduced another one in that if the ONLY load on the boiler is the UFH, the Danfoss mixing valve presumably provides such a significant flow restriction on the output side of the boiler that the system quickly kettles because the hot water output from the boiler has got 'nowhere to go'. If I correctly understand how the Danfoss VMV mixing valve works, it's a bit like trying to run the boiler with the flow side capped-off!

If I run the central heating at the same time as the UFH it all works fine, but I need to be able to run the systems independently.

How can this be resolved? How is a small UFH extension to a conventional heating system normally designed to avoid this problem?

Thanks for any suggestions.

Mike

[The Natural Philosopher]

1/. A bypass loop on the boiler with e.g. a ball valve in it to restrict flow to JUST enough to allow water to circulate is one option. If you put that across the danfoss valve, then its out of circuit EXCEPT when the UFH is on.. 2/. Another is to use a towel rail...

3/. Another one is to use a temp reducing valve that actually is never off - one that moves from bypass to UFH progressively. i think my polyplumb one is like that.

[MikeH]

Well I guess a towel rail is probably the favourite... As it happens the pipework is in the right place and SWMBO is going to want one as soon as it gets cold.

I'm pretty sure the Danfoss valve is progressive - I think it's just that the UFH presents a relatively small load on the boiler, so in order to achieve a 45 degree output temperature the valve has to be almost totally closed on its flow side. Without any other load (like the rest of the C/H radiators), the return temperature rises to almost equal that of the flow. The valve then closes the flow (and presumably opens the return side) more and more in an attempt to control the output temperature. That's my guess anyway, given the symptoms...

Presumably this should have been anticipated by the folk who designed the system?

Mike

[Tournifreak]

It doesn't sound like the plumber knows what he's doing. I hope he's not charging you extra for all the fixes... Sorry, I know it's not very constructive, but it has to be said.

Jon.

[Tournifreak]

Sorry to reply to my own post, but I thought of somethign constructive to say :-) I would have converted to an S-plan system with separate valves for HW, CH and UFH. With a bypass between flow and return.

Jon.

[Christian McArdle]

Whilst you might be able to fix this with a bypass, you may have trouble if your boiler can't modulate down enough.

Personally, I would have ripped out the Y-plan and replaced with S-Plan-Plus

2 port valves. Then ensure that there is an automatic bypass complying with the boiler manufacturer's instructions, should they require one.

Christian.

[MikeH]

Thanks for all the replies.

It's fixed-price - the plumber's actually working for the builder doing the job, so I'm not paying anything for the 'fixes'. If I were, I'd just do the job myself. Having said all that, I would have preferred if they'd got it right in the first place.

To be honest I was assuming that it would be converted to S-plan, but I didn't specify the system at that level of detail, so I don't have grounds for complaint.... As long as it complies with my requirement for operation independent of the main C/H and H/W system, that will do.

I'll bung the problem back at them and see what they come up with.

Thanks again.

Mike

[MikeH]

I've just had a look at the boiler manual and it appears that, given the changes that the plumber has made to the original Y-plan layout, there should now be a bypass valve installed with a 3 litre/min capacity.

I'll point that out to him and see what he says.

Mike

[Tony Bryer]

If he's kept the 3-port valve there will automatically be a bypass - when power is off the DHW port is open on the valves I've seen.

[MikeH]

port in order to prevent the UFH circuit heating the water in the cylinder. So instead of a conventional 3-port valve which, as you say, always has one port open, I now have a combination of a 3-port valve with an additional 2-port valve on the 'normally open' output. The result is the combined 3-port and 2-port valves present a 'totally closed' option when neither the C/H or H/W stats are calling for heat. In other words the built-in bypass that is normally a feature of the

3-port diverter valve no longer exists on my system.

Mike

[Aidan]

If it's a non-modulating boiler, then it's an on/off device with a fixed heat output.

The shower room UFH will have a smallheat requirement, so where does the rest of the heat go?

For example, say the shower room UFH requires 500W, boiler output is say 15kW, what happens the unwanted 14.5kW?

It will be used to raise the temperature of the primary circuit and the boiler, but this will have a small thermal capacity and the boiler will cycle on & off rapidly. It will run for 30 seconds every 15 minutes, but will probably fire and switch off 2 or 3 times in that 15 minutes. Not conducive to a long boiler life.

The only sensible approach would be a thermal store, i.e, make the thermal capacity of the primary circuit much larger to give the boiler a heat reservoir that will allow it to run for some time when the reservoir needs topping up.

I think the original spec could only have been achieved by spending a unanticipated large wad of cash; electric UFH might have been better here, but I loathe it.

[Aidan]

UFH in a concrete slab would have a similar effect. Is it a slab or a suspended timber floor?

[MikeH]

It's a slab, but even when the UFH is heating it from freezing cold it sounds as if it won't provide anything like sufficient load to the boiler on it's own. The boiler kettling and uncontrolled UFH flow temperatures that I've already seen certainly seems to bear that out.

The boiler does modulate, but only down to 30,000 BThU/hr (from it's

85,000 BThU/hr rated capacity)

Duff UFH design then? Presumably it's reasonable for me to claim that the people doing the work should have forseen this?

a bit, but not enough to prevent the boiler over-cycling during periods when the UFH is on but the C/H is not?

Perhaps I need to read-up on what's involved in installing a thermal store? :-/

Mike

[The Natural Philosopher]

As I said previously, the issue is whether the temp mixing valve acts as a 4 port true valve..i.e. allowing bypass when UFH is up to snuff..or whether it essentially stops the boiler pump dead in its tracks when its up to temp.

Faster speed on the UFH pump may help, BUT as I said, a bypass loop across the valve is probably the easiest thing.

[MikeH]

OK guys, let's see if I have this right...

There are two issues:-

First is the need for a bypass *if* the mixing valve doesn't provide a bypass mechanism itself. If necessary I'll call Danfoss next week to verify that. Either way, I plan to install a towel-rail otherwise SWMBO will start to complain once the weather cools a bit. Even if the Danfoss VMV doesn't provide a bypass, the provision of a towel rail should easily meet the boiler requirement for a 3 litre/min bypass route. Presumably this will stop all the boiler kettling, banging and whooshing noises, and will also stop it heating the water in the F&E tank.

The second issue is the relatively small load placed on the boiler by the UFH when there is no other heat demand from the H/W or C/H. Because the boiler only modulates down to 30,000 BThU/hr (sorry about the imperial units - haven't got used to kW for boilers yet) the UFH won't place enough load to prevent the boiler over-cycling. Again, the addition of a towel rail might help a bit, but not very much.

Is that right?

Mike (now off to Screwfix to look at towel rails...)

[The Natural Philosopher]

Should do.

Mmm. Won't the boiler simply switch off? If the exit temp is too high in down modulation? Not actually had any experience of a modulating boiler..

[MikeH]

Yes, first it will modulate down as far as it can then, if the flow temperature is still too high, it will switch off. At some point it will try to fire-up again - presumably when the flow temperature drops far enough. The other poster (I think it was Aidan) was saying that this could result in the boiler starting and stopping every few minutes (i.e. over-cycling) - something that might not be good for it in the long term.

Mike

[Aidan]

Yes. The point about it is that the rated heat output should only occur in design conditions, i.e., extreme cold. Most of the time it will only be taking a tiny amount of the primary flow from the boiler and mixing it with recirculated return from the UFH loop to maintain the UFH temperature. That's what it's meant to do.

Look up the installation instructions for your mixing valve on-line. It sounds like he may have arranged the 2-port zone valves so that the only flow is through the mixing valve, which will usually be very small. If incorrectly installed, it may explain why the valve isn't controlling the secondary UFH temperature.

Mixing valve loops are often installed with the secondary flow and return connections adjacent to each other on the same pipe; that way the operation of the primary pump doesn't induce or alter the flow in the secondary, modulated-temperature, loop.

A possible solution is to have a by-pass through a cross-over bridge, i.e., a connection between primary flow and return pipes with a regulating/lockshield valve to restrict the flow. That way you'll always get a minimum flow rate. The disadvantage is that condensing boilers don't like them, because the unused hot flow water enters the return, increases the return temperature and reduces the boiler efficiency slightly. It may even be verboten under building regs, although I think that applies to by-passes intended to prevent TRVs being over pressurized. .

The flow & return connections to the secondary loop are then close to each other on the cross-over bridge There's negligible pressure loss due to resistance between F&R tees, so primary flow doesn't create a pressure differential between F&R and doesn't affect secondary flow.

[MikeH]

Thanks for your detailed reply. I might have to read it a few more times before I can claim to really understand the points you make (especially the bits about pressure losses and differentials in the last paragraph). ;-)

I've had a look at the Danfoss VMV documentation - both on the Danfoss site and on the product paperwork that the plumber left with me. There isn't anything to say that it provides a bypass route from flow to return directly through the valve. On the contrary, there are references in the specs to values for the 'leakage' through closed ports, so it looks like there's no bypass function built into it. The symptoms bear this out as well, but I'll see if I can get a definitive answer from Danfoss next week.

The existing system design *certainly* results in the only flow being through the mixing valve if the existing C/H circuit and the H/W are not calling for heat, so I'm 99.9% sure that some form of additional bypass *is* required. I think the reason the valve isn't able to control the temperature of the UHF circuit flow side is because the entire primary circuit gets up to the boiler flow temperature (or near enough). In the absence of any other load on the boiler other than the UFH, the primary return side gets too hot and, in the absence of a 'cool' return with which to mix, the valve is unable to maintain the UFH flow temperature within spec.

I'll get in touch with the plumber again and see what he suggests.

Thanks again for your help.

Mike

[Aidan]

Hope it makes sense. I looked it up.

Data here;

confusion is in the diverting and mixing applications in the top left of page 2.

The diverting application is constant temperature but variable flow rate through the load. This isn't what you want. However, the primary pump's flow rate should always be the same regardless of what the control valve is doing. There should be a regulating valve in the by-pass, so the by-pass path resistance can be adjusted to equal the load path resistance; Manufacturers, eh?

The mixing valve application is variable temperature and constant flow rate through the load. The plumber has seen the flow and return arrows and connected it across the flow and return pipes, as he would a radiator. I can't blame him, but it won't work like that. The little electric actuator motor is being overwhelmed by the pressure difference produced by the pumps across the valve ports and it's being forced open when it is trying to close. Hence, loss of control.

The diagram only makes sense if the flow and return pipes are connected to the same pipe, giving negligible pressure difference between them. . The secondary pump will draw water from the passing primary as it needs it, but the primary needs a flow route to keep it going past.

The VMV application on the top RHS is a mixing application.

I think you need a cross-over bridge with a zone valve to isolate it, the limit switch on the zone valve enables primary & secondary pumps and boiler.