Heating system for an ICF house with UFH

[Nelliekins]

Just now, PeterW said:

 

Flow temperature has to be below 40c with a chipboard based floor and also depends on the finish above - some wood flooring and laminates have a maximum floor temperature. 

 

There is also a comfort factor, and the issue with overshooting the  target temperature as the higher the flow temperature, the easier it is to get a continuous increase of room temperature once the flow is switched off. 

 

The spreadsheet from @JSHarris said a floor temp of just over 21C, so we want a flow well under 30C I would think... 

[scottishjohn]

I did this zone areas thing with my current ufh i fitted years ago ,but failure of zone valves and people not closing doors  actually made it a waste of time 

and if going for mvhr ,then with gaps under the doors you need for air flow it would be a pointless complication 

thats my view anyway ,

I got fed up of replacing valves ,so just removed the heads as they failed and opened up the actual  valves --seems to work just as well

[scottishjohn]

I must be a very tender person --cos i want room temp  when sitting at 23c  or i get chilled

Yes i know every 1c you up temp by is 7%  extra costs approx on old style house --but that the idea of very well insulated house--so it don,t change 

Edited December 29, 2018 by scottishjohn

[PeterW]

7 minutes ago, Nelliekins said:

 

The spreadsheet from @JSHarris said a floor temp of just over 21C, so we want a flow well under 30C I would think... 

 

You have to know the delta loss and the heat transfer of the floor build up. 33c would be fine in concrete but need longer running, 37c under wood and shorter run time.  

[PeterW]

6 minutes ago, scottishjohn said:

I did this zone areas thing with my current ufh i fitted years ago ,but failure of zone valves and people not closing doors  actually made it a waste of time 

and if going for mvhr ,then with gaps under the doors you need for air flow it would be a pointless complication 

thats my view anyway ,

I got fed up of replacing valves ,so just removed the heads as they failed and opened up the actual  valves --seems to work just as well

 

Zoning rooms gets challenging, but zoning different floors and especially with different floor composition is essential. 

[scottishjohn]

9 minutes ago, PeterW said:

 

Flow temperature has to be below 40c with a chipboard based floor and also depends on the finish above - some wood flooring and laminates have a maximum floor temperature. 

 

There is also a comfort factor, and the issue with overshooting the  target temperature as the higher the flow temperature, the easier it is to get a continuous increase of room temperature once the flow is switched off. 

I am not saying you are not correct ,but the current UFH ,which idid by routing the chipboard floors and fitting pipeshas run at 45c water temp at mixerbefore  manifolds for last 17 years and no downside as far as i can tell

[PeterW]

If that’s surface based pipe using the old 10mm stuff then it is about right - that was designed to work the same as a radiator in reality and just warm the air above. The newer systems use the heat capacity of the concrete to reduce the thermal “swing” and lessen the variations. @TerryE did some really nice modeling of his slab to show how it works. 

[scottishjohn]

I was getting far too technical i fitted the UFH to both floors ,  800meters of tubing routed into floors +covered with 6mm ply to put back some strength in floor,not sure you would get away with that with more modern wood frame house as joists are alot thinner

I did notice that the upper was not on as much --due to convection from lower I presume 

but in new house i won,t be heating up stairs directly anyway ,possibly,

may even go for a concrete upper floor as well --

all depends on what system i finally decide on for house build 

[scottishjohn]

3 minutes ago, PeterW said:

If that’s surface based pipe using the old 10mm stuff then it is about right - that was designed to work the same as a radiator in reality and just warm the air above. The newer systems use the heat capacity of the concrete to reduce the thermal “swing” and lessen the variations. @TerryE did some really nice modeling of his slab to show how it works. 

I have to say it works very well -- its KEE triple tube system  and is very responsive to ambient temp changes --floor can go from cold to heating room in under 1 hour,which is good in some ways as the climate here is so changeable day to day .

maybe i shouldn,t be thinking of a mega thick slab -but maybe thinner  and get quicker temp change -too many variables ,GRRR

that is a common complaint of people around here ,takes long time for floor to  change then  wife opens up widows to cool the house down --

that just does not work for me --heating it up too much to then let it out--maybe be better to on other side of the curve

Edited December 29, 2018 by scottishjohn

[scottishjohn]

 

[scottishjohn]

this solvis max system  was what i wanted when i first did UFH --but no british agent --so it never happened-

love the way you can change the heat input from gas to  oil burner at top .or any other heat input as well

[Nelliekins]

44 minutes ago, PeterW said:

 

Zoning rooms gets challenging, but zoning different floors and especially with different floor composition is essential. 

 

I think in my mind I was going to address the differences in floor requirements / performance by just changing flow rates across the manifolds as a whole. If our EPC consultant is right, then we won't even need UFH upstairs, but it was put in "just in case". Kim wasn't convinced the house would be warm enough without, and she's a lizard (needs to bask because she is always cold). 

 

There is a fairly large stairwell, so a large volume of hot air is free to circulate between all 3 floors of the house. I am guessing that will help the upstairs temp as well?

 

If the coil can deliver 6kW of heat energy, and our peak heat requirement from the UFH is 2.5kW to maintain 20C when it is -5C outside, that should be adequate for heating the UFH provided the flow rates are adequate (through the coil to the 3 manifold pumps).  Failing that, we will have to get more cats... 

 

Even better, we can use the home automation system to control whether the immersion is used to put heat into the cylinder or the boiler or both, so some of the time it'll be free to heat... Provided I can get accurate temps into the automation system in the first place! ?

[PeterW]

 

1 minute ago, Nelliekins said:

Even better, we can use the home automation system to control whether the immersion is used to put heat into the cylinder or the boiler or both, so some of the time it'll be free to heat... 

 

I wouldn’t bother relying on Loxone or anything similar to control the actual flow temperatures - they are too long a time constant to use with something that can do 0.01c changes. Just get it to do the on/off and then let some decent thermostats do the room temperature control. 

 

And trusting an EPC consultant to design your heating system is like asking @Nickfromwales to do your interior design .....

[Nelliekins]

14 minutes ago, PeterW said:

 

 

I wouldn’t bother relying on Loxone or anything similar to control the actual flow temperatures - they are too long a time constant to use with something that can do 0.01c changes. Just get it to do the on/off and then let some decent thermostats do the room temperature control. 

 

Wow that's a good guess, I was looking at Loxone! 

 

Any recommendations on the thermostats? I have seen the one @JSHarris used in his thread but the best I have seen are 0.1C increments - can we improve on that? 

 

Quote

And trusting an EPC consultant to design your heating system is like asking @Nickfromwales to do your interior design .....

 

PMSL! Maybe it's just as well that Kim is a lizard then! 

 

FWIW I designed the system in the original post based upon the DIY Heat Bank elsewhere on the tinterweb. And I am a self confessed cheapskate, bodger, curmudgeon. My rainwater system will use IBCs to save money, and I melted my own channels out of EPS70 to save about 500 quid on the upstairs insulation... The wood crossing between the walls has a series of loops of nichrome wire underneath, and the walls acted as a jig to guide the panels. Took about 5 minutes per 8' sheet of EPS, with 44 sheets to groove (and that's the feet of my eldest nephew who acted as a labourer for part of the project). 

 

[PeterW]

IBCs for rainwater has been done before on here - think it is @Bitpipe who did it. 

 

And I’m all for value engineering ...!! I did an MVHR from electrical duct and pipe fittings. 

[Onoff]

29 minutes ago, PeterW said:

IBCs for rainwater has been done before on here - think it is @Bitpipe who did it. 

 

And I’m all for value engineering ...!! I did an MVHR from electrical duct and pipe fittings. 

 

@Bitpipe's IBC thread:

 

 

[Nickfromwales]

1 hour ago, Nelliekins said:

OK. So the buffer tank affords us the luxury of no mixing valves, provided we control the cylinder temp accurately enough. Then we tee off the coil flow and return for each of the 3 manifolds, and have a pump on each circuit? Presumably we put a bypass and PRedV on the boiler flow/return, and a pump from the boiler?  

 

I will have to mount a couple of stats on the cylinder I guess, to give us tighter control over the temperature of the flow to the manifolds?

 

Then there is the UFH temp... We have temp gauges on every manifold, but since most of the house will be automated, I'd like to improve on that... Can we just put DS18B20s into brass tee fittings, and connect them over 1wire to the automation system? I guess a simple way to know if the floor is up to temp is to track the differential between flow and return across a manifold? Would that work? 

There is absolutely no way you should scrimp on having no mixing valves on the manifolds, especially as you'll have different floor constructions. You will need to have a very fine tuneability for the basement and the ground floor, and trying to achieve that with one global flow temp will be impossible. Do not cut corners here. Do not use your potable water TMV for heating, you'll need a Heating type Reliance valve for that, two totally different beasts.

 

For flow control to the manifolds you use the reliance valve and a pump to circulate between the cylinder and the manifolds. You should be fitting a 2-port zone valve to each leg to stave off unwanted convection heat flow and to be able to isolate each leg according to whether or not that particular manifold is being used. If you've already run the 22mm as a single pair then you simply put the 2-ports at each manifold. The cables you require to control these should be found in the manifold wiring centres, if you've fitted them. If not, get cables in now.

 

The UFH temp is controlled by the mixing valves on the manifolds, ( 20^oC to 60^oC range so very accurate in a PH ), and the flow and return temps are shown on the manifold temp gauges normally. If you don't have them then cheapo Chinese digitals are available or chuck in the thermistors and reference off the HA system.

 

Ok, much to @PeterW's entertainment I am going to retract my statement re deriving anything other than cold mains uplift from the cylinder. I looked at the link and its just a domestic hot water cylinder not a TS so is completely unfit for gleaning various services / inputs to / from. Too many good films on yesterday is my excuse, with 'Where Eagles Dare' being the highlight. With the mention of TS I dropped into default and assumed the coil was a DHW top mounted. 

 

Hydraulically separating the UFH from the main body of water is a step in the wrong direction, as you'll not be able to use the F&E tank to service that circuit. You'll then need to fit a separate fill loop, expansion vessel, pressure relief valve and pressure gauge to fill up and maintain pressure in that circuit. Is the F&E tank higher than the upper floor of UFH? 

 

Adding a SA seems the easiest way of dealing with DHW as they're compact and don't need an overflow ( discharge ) arrangement like an UVC does. A 6 will suffice if you're heating it from the gas boiler, but you'll probably be better off with a 9 ( £300 more then the 6 so not a great uplift when you look at the incremental price increase ) so you have somewhere to dump PV in the summer ( when you don't want to heat the buffer ).

 

 

[Nickfromwales]

6 hours ago, Nelliekins said:

Any recommendations on the thermostats? I have seen the one @JSHarris used in his thread but the best I have seen are 0.1C increments - can we improve on that? 

If you want your pumps and valves operating 60 times a minute then yes. 

Best get to the NASA boot sale this coming Sunday   

[Nelliekins]

1 hour ago, Nickfromwales said:

If you want your pumps and valves operating 60 times a minute then yes. 

Best get to the NASA boot sale this coming Sunday   

 

Doesn't @JSHarris have his pump running all the time at a low level? (Actually IIRC it's all day rather than 24/7)

 

Surely (a small amount of) heat injected sporadically into the buffer tank (let's call my 140L cylinder that if it is being used for UFH) and then circulate the warm (26C?) water around the UFH. Sooner or later the return temp will equalise to the flow temp, at which point the UFH has reached equilibrium and will not get any warmer with that flow temp. 

 

Which then means we control the floor temp by choosing the correct water flow temp, and keep the buffer tank at that temp.

 

5 hours ago, Nickfromwales said:

There is absolutely no way you should scrimp on having no mixing valves on the manifolds, especially as you'll have different floor constructions. You will need to have a very fine tuneability for the basement and the ground floor, and trying to achieve that with one global flow temp will be impossible. Do not cut corners here.

 

OK this I don't understand. If you run the heating pump constantly, and want the floor temp to settle everywhere at around 22.3C (taken from the @JSHarris spreadsheet) what is there to tune? We have individual flow control on each UFH circuit, and with a separate pump per manifold, can't we dictate the proportion of heat energy going to each manifold anyway? 

 

Quote

 

Do not use your potable water TMV for heating, you'll need a Heating type Reliance valve for that, two totally different beasts.

 

OK. 

 

Quote

For flow control to the manifolds you use the reliance valve and a pump to circulate between the cylinder and the manifolds. You should be fitting a 2-port zone valve to each leg to stave off unwanted convection heat flow and to be able to isolate each leg according to whether or not that particular manifold is being used. If you've already run the 22mm as a single pair then you simply put the 2-ports at each manifold. The cables you require to control these should be found in the manifold wiring centres, if you've fitted them. If not, get cables in now.

 

Not fitted the 22mm pipes yet. No manifold wiring centres though - we aren't zoning the floor for individual control. 

 

Quote

 

The UFH temp is controlled by the mixing valves on the manifolds, ( 20^oC to 60^oC range so very accurate in a PH ), and the flow and return temps are shown on the manifold temp gauges normally. If you don't have them then cheapo Chinese digitals are available or chuck in the thermistors and reference off the HA system.

 

Ok, much to @PeterW's entertainment I am going to retract my statement re deriving anything other than cold mains uplift from the cylinder. I looked at the link and its just a domestic hot water cylinder not a TS so is completely unfit for gleaning various services / inputs to / from. 

 

Indeed it is a hot water cylinder, with a single coil in the bottom 1/2. But for holding a body of water at a given (low) temp, does that matter? The tank is either for DHW or (more likely now) a buffer tank for the UFH... 

 

Quote

Hydraulically separating the UFH from the main body of water is a step in the wrong direction, as you'll not be able to use the F&E tank to service that circuit. You'll then need to fit a separate fill loop, expansion vessel, pressure relief valve and pressure gauge to fill up and maintain pressure in that circuit.

 

Noted, so the coil is either unused (probably) or used to preheat incoming mains for the DHW which is then heated fully by some other system. Make sense? 

 

Quote

Is the F&E tank higher than the upper floor of UFH? 

 

Yes. 

 

Quote

Adding a SA seems the easiest way of dealing with DHW as they're compact and don't need an overflow ( discharge ) arrangement like an UVC does. A 6 will suffice if you're heating it from the gas boiler, but you'll probably be better off with a 9 ( £300 more then the 6 so not a great uplift when you look at the incremental price increase ) so you have somewhere to dump PV in the summer ( when you don't want to heat the buffer ).

 

Completely agree, but can't really afford it before the VAT reclaim...

 

[Jeremy Harris]

I played around with several ways of controlling the temperature of our house, and settled on a pretty simple system, with a +/- 0.1 deg C hysteresis thermostat that turns the heat pump on and opens the UFH.  The UFH pump runs all day, whenever the timer turns the system on, and does work to even out the temperature in the slab well. 

 

I personally don't see the need for anything more complex than this, as it seems to control the house temperature very well.

[Nickfromwales]

2 hours ago, JSHarris said:

I played around with several ways of controlling the temperature of our house, and settled on a pretty simple system, with a +/- 0.1 deg C hysteresis thermostat that turns the heat pump on and opens the UFH.  The UFH pump runs all day, whenever the timer turns the system on, and does work to even out the temperature in the slab well. 

 

I personally don't see the need for anything more complex than this, as it seems to control the house temperature very well.

Yes, noted, but you don't have a mix of emitters over a number of floors. To have one flow temp to all the manifolds with them just having pumps is not a good idea, IMO, given I've done such installs and found it essential to be able to choose a slightly higher flow rate for the UFH over the timber floors, even with aluminium spreader plates. Bear in mind the UFH here is router'd into insulation with no aluminium plates to conduct and dissipate the heat effectively.

[Nelliekins]

57 minutes ago, Nickfromwales said:

Yes, noted, but you don't have a mix of emitters over a number of floors. To have one flow temp to all the manifolds with them just having pumps is not a good idea, IMO, given I've done such installs and found it essential to be able to choose a slightly higher flow rate for the UFH over the timber floors, even with aluminium spreader plates. Bear in mind the UFH here is router'd into insulation with no aluminium plates to conduct and dissipate the heat effectively.

 

One question then - can't we compensate, and achieve the variation in flow rate, by simply tweaking the flow meter settings on each circuit of the upstairs manifold? Whilst the UFH is router'd into the EPS (with no spreader plates), sooner or later the heat in the floor achieves equilibrium no matter what. The EPS and chipboard would have to be perfect insulators to prevent it, and neither come close to being a perfect insulator!

 

Whilst there are legitimate issues that you have identified in getting up to temperature initially (e.g. the lack of conductor to distribute the heat evenly across the floorboard), that's hopefully a one-time problem (or at least, few enough times during the lifespan of the heating system to not be worth additional expense or contemplation). Is that fair to say? Remember also that the target floor temperature is only 22.6C, and to achieve that temp, we are likely to have a flow temp of 27C or less.

[Nickfromwales]

3 hours ago, Nelliekins said:

Surely (a small amount of) heat injected sporadically into the buffer tank (let's call my 140L cylinder that if it is being used for UFH) and then circulate the warm (26C?) water around the UFH. Sooner or later the return temp will equalise to the flow temp, at which point the UFH has reached equilibrium and will not get any warmer with that flow temp. 

 

Which then means we control the floor temp by choosing the correct water flow temp, and keep the buffer tank at that temp.

The whole point of the buffer is to allow your big gas boiler to chunk heat into the buffer at the optimum condensing range flow temp. The boiler should not be constantly idling to keep the buffer temp low, nor will it be able to as that model has ( IIRC ) 7kW minimum burner output at maximum modulation. The way you propose to run it will see the boiler constantly cycling, fan constantly spinning, gas valve constantly active etc, so will reduce the longevity of the boiler too.  

 

3 hours ago, Nelliekins said:

We have individual flow control on each UFH circuit, and with a separate pump per manifold, can't we dictate the proportion of heat energy going to each manifold anyway? 

Please elaborate?

 

3 hours ago, Nelliekins said:

Not fitted the 22mm pipes yet. No manifold wiring centres though - we aren't zoning the floor for individual control. 

 

You'll still need 2-port zone valves on each run to stop flow when not required. If the pipes aren't in yet then just throw a cable in as you go.

 

3 hours ago, Nelliekins said:

Noted, so the coil is either unused (probably) or used to preheat incoming mains for the DHW which is then heated fully by some other system. Make sense? 

The kick in the nuts is that you cannot reduce the size of the DHW device if you only have seasonal uplift ( eg the heating is on so you get preheat as a side effect ). Not so problematic if there is enough excess PV to heat the buffer an the SA / UVC but then its just additional summer heat gain, and often unwanted.

 

31 minutes ago, Nelliekins said:

One question then - can't we compensate, and achieve the variation in flow rate, by simply tweaking the flow meter settings on each circuit of the upstairs manifold?

You'll be there for years trying, and it would be a nightmare to balance and keep balanced. 

[Nelliekins]

9 hours ago, Nickfromwales said:

The whole point of the buffer is to allow your big gas boiler to chunk heat into the buffer at the optimum condensing range flow temp. The boiler should not be constantly idling to keep the buffer temp low, nor will it be able to as that model has ( IIRC ) 7kW minimum burner output at maximum modulation. The way you propose to run it will see the boiler constantly cycling, fan constantly spinning, gas valve constantly active etc, so will reduce the longevity of the boiler too.  

 

Which is why on the drawing in the OP, the UFH was leeching heat from the DHW tank via the coil. Thus the DHW cylinder was being kept very warm (75C), and the UFH cylinder would draw some of that off when needed, using the coil. The upshot is that the gas boiler gets nice long burns (maybe 5-10 minutes?) to reheat the DHW cylinder to avoid short cycling and to not have to modulate much (or at all) when it does so (although that depends on what the minimum and maximum temps for the DHW cylinder are to still deliver the right water temp from the PHE output. 

 

Quote

Please elaborate?

 

Each circuit has a flow meter on the top, and a zone port that can take a valve at the bottom. Adjusting either top or bottom controls the flow rate through that circuit, so adjusting all 6/12 flow meters (depending on which manifold we are talking about) therefore dictates overall flow rate through that manifold's outputs. 

 

Quote

You'll still need 2-port zone valves on each run to stop flow when not required. If the pipes aren't in yet then just throw a cable in as you go.

 

I can do that, no problem. However the differential between flow and floor temps is only a few degrees C, and the buffer tank water is to be held at ideal temp for distribution, so under what conditions do we need to stop the flow, beyond what would be achieved by simply turning off the pump (which we don't want to do much anyway)? 

 

Quote

The kick in the nuts is that you cannot reduce the size of the DHW device if you only have seasonal uplift ( eg the heating is on so you get preheat as a side effect ). Not so problematic if there is enough excess PV to heat the buffer an the SA / UVC but then its just additional summer heat gain, and often unwanted.

 

That is fine. If we revert to the design in the OP, we have :

 

- A 140L cylinder, which you suggested yourself would be adequate for our peak DHW usage scenario of 2 showers, provided the boiler can replenish it quickly enough

 

- a 30kW boiler dumping heat into the 140L cylinder directly (so maximum efficiency, apart from possibly too high a cylinder temp?) 

 

- the 140L cylinder receiving a boost from a PV-powered immersion during daylight hours, so no gas burn unless the cylinder is drained faster than the PV immersion can replenish, and the cylinder allows us to extend the PV regen time because it will still work within a temperature range (ie if the heat is taken out by a tap running for 1 minute, the immersion can take 20 minutes to replenish that energy, because the cylinder might have dropped 2C but it operates properly in a possible 15C range before needing additional heat anyway) 

 

- the 140L cylinder supplying heat from the top outlet to a PHE to give DHW via a TMV to a hot water manifold (which has 22mm outlets for bath and showers, and 10/15mm outlets for sinks) 

 

- separate UFH buffer tank heated by PV-powered immersion and the bottom-mounted coil from the DHW cylinder (at a flow rate TBD to ensure the low temp for the UFH is maintained in the UFH buffer tank). Here the pump for the DHW coil - > UFH buffer can run at whichever flow rate is necessary to control the heat energy being delivered to the UFH buffer (and the home automation can control the on/off side simply enough provided it knows the flow and return temps and the cylinder temp(s) as well) 

 

The DHW cylinder takes maybe 10-15 minutes to fully replenish from the gas boiler (except it'll never get fully cold, and the boiler will hopefully be able to inject heat as fast as it is taken out if 2 showers are running so they never run cold). That sort of lag on the UFH side of things is inconsequential anyway - I doubt we could even measure the temp drop that occurs in 15m if the floor is at equilibrium. 

 

As I say, that (probably not full) replenishment will only take place when there is a big drain on the DHW, eg running showers or baths, or a number of smaller drains (several lots of sink usage). Running a sink tap should only result in a small topup being needed, which will either come immediately from the PV immersion or get held back until the cylinder needs a bigger boost (eg at night) at which point the gas boiler does a single longer burn again. 

 

Quote

You'll be there for years trying, and it would be a nightmare to balance and keep balanced. 

 

So maybe we don't bother trying. If we follow and extend @JSHarris' design, the 3 manifolds receive water from a UFH buffer tank held at the desired flow temp to eventually achieve the desired floor temp. It doesn't matter if the lag on the first floor (where the pipes are only 20mm below the surface) is less or more than the concrete floors below (where the pipes are 75mm below) - at some point all floors will achieve equilibrium and stop extracting heat from the pumped water.

 

There may be a case that we don't want the upstairs floor to get as warm, if the EPC guy was right. But that's a whole different ballgame, and even then if we know the temp in the upstairs floor, we can simply shut off the 2-port valve supplying the manifold that you have recommended we install, to avoid an overshoot, yes?

 

Which leaves us with what seems to me to be a potentially workable solution that doesn't stress the gas boiler unduly (although i agree it could be improved considerably by the use of a Sunamp to allow the boiler temp to drop a bit) and has more than desirable heat losses because of the F&E tank (which also would be addressed by the use of a Sunamp) 

Edited December 30, 2018 by Nelliekins
Auto correct corrections

[PeterW]

13 hours ago, Nelliekins said:

Sooner or later the return temp will equalise to the flow temp, at which point the UFH has reached equilibrium and will not get any warmer with that flow temp. 

 

Ok so that will never happen until the air temperature and fabric temperature all equalize and given you have continuous losses through both MVHR and fabric you will be chasing rainbows ...

 

The “simple” approach would be to run a pump and zone valve on each manifold with a bypass across the flow and return inside the zone valve so the pump can circulate the water when the zone valve is off but when heat is needed then it allows the flow in. A single blending valve at the tank “may” work, but if you are going this route you definitely need to isolate each manifold with a 2 port valve.

 

As a complete random thought, the other trick could to be to use a 3 port diverter valve on each manifold, when it was activated it would open the flow to the buffer, closed would be a “closed loop” so the pump can recirculate and even out the slab etc.... Hmm... ?.. that removes the need to try and balance a bypass valve, and these do have a habit of squealing under low pressure.