Combining radiators with UFH - some surprising conclusions

[JohnMo]

16 minutes ago, sharpener said:

Sounds ideal, have you got a link?

https://www.heatingcontrolsonline.co.uk/ivar-thermostatic-mixing-valve-pump-set-p-1073.html

 

Available with or without pump

 

Read the Balancing and regulations section

UNIMIX-C.pdf

Edited January 26 by JohnMo

[sharpener]

Thanks.

 

The primary valve (a) in Fig 3 looks to be the one to control the overall flow from the HP. I think my existing manifold has an equivalent so I will try it.

 

There is also a motorised valve in the return pipe (was easier for them to fit there as all rather cramped), and a manual 22mm iso valve in the flow. If there is room downstream to get in a 22mm flow setter I might do that so I can see flow rate. If no room I could alternatively replace the iso valve with a ball valve/flow regulator like this with a 4 lpm insert.

 

The mixing valve is suspect anyway and unfortunately I have not been able to trace the original part or find a replacement with the same porting, so it will need some complicated pipe-fitting in a restricted space, a job for summertime I think.

 

If I do need to replace it then I could instead buy the whole Invar caboodle. After checking it will fit my manifold! £158.82 is not a lot considering what you get for it.

[JohnMo]

Will let @Nickfromwales comment on the merits of Ivar mixers

[sharpener]

Well it gets more complicated by the day!

 

The flow limit of 4 l/min to the UFH is required while the radiators are operating as described above.

 

But when they are not, I realise I need to be able to supply the UFH at a higher flow rate and lower temperature, say 35 deg, in order to achieve a correspondingly higher CoP.

 

Perhaps when there is no flow to the rad circuit the internal pump in the outdoor unit will be powerful enough to get sufficient flow through a fixed restriction like a partially closed valve as upthread - I will certainly check this out first.

 

Otherwise the flow restrictor will need to be bypassed when there is no demand signal for the rads. Fortunately I have a spare motorised valve which can be supplied via the n/c contacts on the rad circuit zone valve which should achieve this nicely.

[Beelbeebub]

Just a thought. 

 

UFH tends to have a fairly high thermal inertia compared to rads (especially if it's in slab). 

 

Could you have 2 zones UFH and Rad. 

 

Each zone is set for the ideal flow and temp and you just "multiplex" between them say 30min on UFH, 30min on rads. 

 

Alternatively, you could flow to the rads all the time and, as needed, drop the flow temp and open up the UFH loops, so your rads will cool a bit (but still put out some heat) for a while whilst your UFH slab heats up. 

 

 

[sharpener]

3 hours ago, Beelbeebub said:

Could you have 2 zones UFH and Rad. 

 

Each zone is set for the ideal flow and temp and you just "multiplex" between them say 30min on UFH, 30min on rads. 

 

Alternatively, you could flow to the rads all the time and, as needed, drop the flow temp and open up the UFH loops, so your rads will cool a bit (but still put out some heat) for a while whilst your UFH slab heats up. 

 

 

I think the heat losses are too much for "either or" to work unfortunately. In the recent cold snap the HP needed to work at max o/p much of the time. The hall and the end of the kitchen furthest from the AGA got very cold without the UFH, and I had not at the time worked out why the UFH crippled the output from the rads. Living room has both so it matters less there.

 

We do have the system divided into the two zones but the flow temp from the HP is determined by the heat curve for the higher of the two emitter circuits, so if there is any demand from the rads side it will run up automatically to the higher temp. Not sure how I could prevent that. Perhaps use the microswitch contacts on the UFH zone valve to do something cunning like open the rads zone valve without the HP controls knowing!

 

I still think I would need to modulate the flow rate to the UFH though.

 

 

Edited January 27 by sharpener

[JohnMo]

Think I would try to simplify first before adding more complexity.

8 hours ago, sharpener said:

flow temp from the HP is determined by the heat curve for the higher of the two emitter circuits, so if there is any demand from the rads side it will run up automatically to the higher temp

Could you balance the system, so there is always a demand from the radiator circuit? Do this by lowering the flow temperature on the higher curve, so it never exceeds thermostat set point.

 

Think you need to get closer to running as a single zone, so the heat pump has lots of time to settle out the flow temps, and find it's best running point.

[Beelbeebub]

Are your rads on separate circuit to your ufh? Ie can you run just rads and just ufh?

 

If so, and your controller has the facility for separate wc curves per zone you should be able to run each set of emitters at the optimum temp.

 

It might be worth working out which emitter is your "worst" ie needs the highest flow temp.

 

It might be that it is just one or two rads are markedly worse. If those were upgraded it might allow a significant drop to a lower temp.

[sharpener]

14 hours ago, Beelbeebub said:

Are your rads on separate circuit to your ufh? Ie can you run just rads and just ufh?

 

If so, and your controller has the facility for separate wc curves per zone you should be able to run each set of emitters at the optimum temp.

 

Yes they are physically separate with their own zone valves. But as upthread the HP controller will cause the flow temp to suit whichever is the higher. So when the rads are in use I only want to allow a small amount of flow to the UFH. At other times a larger flow at a lower temp and better CoP.

 

14 hours ago, Beelbeebub said:

It might be worth working out which emitter is your "worst" ie needs the highest flow temp.

 

It might be that it is just one or two rads are markedly worse. If those were upgraded it might allow a significant drop to a lower temp.

 

No, we upgraded 6 out of the 8 when the HP was installed, they should all be quite well matched now.

 

Current plan is to restrict the flow to the UFH by partly closing the isolating valve, and then see if the pump is sufficiently powerful to push more flow through it to the UFH when the rads are not on. The Wilo internal pump in the Vaillant HP generates a high head and will try and do 34 l/min under most circs.

 

If this does not work well enough I will just set the times so they do not overlap as you suggested earlier, until the heating season is over and I can cut into the pipework to install the bypass valve round the flow restrictor. Am still thinking how to do this without running more cable from utility room to the UFH manifold. Yes I have a spare Shelly plug so might do it wirelessly.

[JohnMo]

7 hours ago, sharpener said:

Yes I have a spare Shelly plug so might do it wirelessly

Then you need internet to get your heating to work - not good 

 

Something like this, which an RF relay not WiFi.

 

https://mainsswitch.co.uk/

 

You can get them for £43 on eBay 

[JohnMo]

Just occurred to me, you just paid an arm and leg to get this system installed, get the installer back to rectify. You have an MCS performance guarantee. 

[JoeBano]

My living room wasn’t hitting 20c I wanted it to be, tried increasing the flow temperature that didn’t work. In the end I drop my flow temperature down back to original temperature and dialled my flow rate on the hallway Ufh and now it’s achieving 20c, hallway is still hitting 20c aswell. 
you might just play with it’s the balancing, how loops do you have? 

[Beelbeebub]

13 hours ago, sharpener said:

 

Yes they are physically separate with their own zone valves. But as upthread the HP controller will cause the flow temp to suit whichever is the higher. So when the rads are in use I only want to allow a small amount of flow to the UFH. At other times a larger flow at a lower temp and better CoP.

 

 

No, we upgraded 6 out of the 8 when the HP was installed, they should all be quite well matched now.

 

Current plan is to restrict the flow to the UFH by partly closing the isolating valve, and then see if the pump is sufficiently powerful to push more flow through it to the UFH when the rads are not on. The Wilo internal pump in the Vaillant HP generates a high head and will try and do 34 l/min under most circs.

 

If this does not work well enough I will just set the times so they do not overlap as you suggested earlier, until the heating season is over and I can cut into the pipework to install the bypass valve round the flow restrictor. Am still thinking how to do this without running more cable from utility room to the UFH manifold. Yes I have a spare Shelly plug so might do it wirelessly.

S I can get this straight in my head.

 

You want to run in 3 conditions

 

1) Just rads at WC curve suited to rads

 

2) Just UFH at WC curve suited to UFH

 

3) Both rads & UFH with HP outputting WC curve for rads

 

You can currently run 1 & 2 no problem, the issue is with 3 where the UFH circuit drops the return temp so much the HP cannot achive the desired rad temp?

 

Is that correct?

 

Assuming it is,

 

I imagine you essentially have 2 choices.

 

During the overlap time you can either run with Rads WC, and (as you suggest) have a lower flow rate through the UFH such that the return flow/temp doesn't drag the bulk flow down and cause the HP problems.

 

The other option is the inverse, run both at the UFH WC temp. The UFH will be fine (obviously) the rads will be a bit cooler and output a bit less but the return temp shouldn't be an issue for the HP.

 

The downside of this is the rad zones will be a little underpowered but this could be compensated for by switching to rads only for periods to "blip" a bit more heat into the zones.

 

The upside is there shouldn't be any plumbing changes.

 

 

[sharpener]

5 hours ago, JohnMo said:

Then you need internet to get your heating to work - not good 

 

Not necessarily, you can programme it over the LAN using a local IP address. Then it will run without reference to the web. Only mentioned it as is surplus to another project, not worth the hassle to get a £17 refund, I am sure it will come in handy for something.

 

5 hours ago, JohnMo said:

Just occurred to me, you just paid an arm and leg to get this system installed, get the installer back to rectify. You have an MCS performance guarantee. 

 

They installed to my spec and did a good job within various constraints including re-using the existing cyl and UFH and the wiring betweeen these and the utility room. Nor can either of us know what exactly is under the floor from 1995.

 

Anyway, before asking them back I want to understand how the combo of rads and UFH works in theory, and what might be possible in practice. Which is why I started this thread before these or any installers were on the scene. I hope I have shown the much-discussed use of rads and UFH in the same HP system is more complicated to optimise than many ppl think.

 

Maybe motorised mixers like Esbe are the answer but I am not convinced there is any thermodynamic advantage and the time, expense and disruption to install is unappealing.

 

4 hours ago, JoeBano said:

you might just play with it’s the balancing, how loops do you have?

 

There are four circuits each with their own wall-mounted thermostat

• Living Rm E, ~ 25 sq m, important as it is our main sitting area. There is also a rad rated at 1150W @ 47.5C
• Living Rm W, ~ 25 sq m, not used as there is a raised dais limiting heat flow, and a piano. There is also a rad rated at 1150W @ 47.5C

Installers' estimates for the whole room: heat loss 3653W, output from UFH 5321W @ 40C flow. I have no idea where they got that from, @JohnMo's nomogram upthread gives 46W/sq m for a MWT of 37.5C which implies more like 2300W, so would need half of that plus both rads to meet the heat loss of the room.

• Hall, ~ 15 sq m, needed to take chill off hall at least. Manifold is here in cupboard under stairs. Front door and side panels are d/g and well draught-proofed (apart from the keyhole!) but still make it cold.
• Kitchen E, ~ 15 sq m, needed as the AGA is at the opposite end.

[sharpener]

50 minutes ago, Beelbeebub said:

You can currently run 1 & 2 no problem, the issue is with 3 where the UFH circuit drops the return temp so much the HP cannot achive the desired rad temp?

 

Is that correct?

 

Yes

 

50 minutes ago, Beelbeebub said:

During the overlap time you can either run with Rads WC, and (as you suggest) have a lower flow rate through the UFH such that the return flow/temp doesn't drag the bulk flow down and cause the HP problems.

 

The other option is the inverse, run both at the UFH WC temp. The UFH will be fine (obviously) the rads will be a bit cooler and output a bit less but the return temp shouldn't be an issue for the HP.

 

Problem here is that reducing the dT between MWT and room from 27.5 to 12.5 means the rad output is more than halved bc of the 1.3 exponent in output vs dT, so is reduced by a factor of 2.79 or to 36%. This is a big drop and wouldn't meet the heat loss of the room as per my last post.

 

So as stated I think it is worth experimenting with the existing valve setup first before making any permanent changes to the plumbing.

 

 

 

Edited January 29 by sharpener

[Dillsue]

14 hours ago, sharpener said:

No, we upgraded 6 out of the 8 when the HP was installed, they should all be quite well matched now.

Did you do heat loss calcs for each room to size the rads and factor in the the massive reduction in output for low temp flow?? I'm working my way through all our rooms/rads just now to resize everything for 35 degree flow and intrigued to know the approach you took??

[Beelbeebub]

48 minutes ago, sharpener said:

Problem here is that reducing the dT between MWT and room from 27.5 to 12.5 means the rad output is more than halved bc of the 1.3 exponent in output vs dT, so is reduced by a factor of 2.79 or to 36%. This is a big drop and wouldn't meet the heat loss of the room as per my last post.

Understood, but is the root problem that there are some rooms that cannot be heated without using rads at 45C and UFH at 40C at the same time? 

 

What is the.max temp the UFH can stand? 

 

Edited January 29 by Beelbeebub

[sharpener]

Yes and no. I used the free heatpunk s/w to make my own estimates of the whole house heat loss to confirm HP size.

 

The room-by-room estimates were then done in detail by the three potential installers as part of their MCS approved quotations. All used Heat Engineer (IIRC you can buy a one-project licence for a reasonable fee but the learning curve looked too steep).

 

They all quoted figures at a range of temperatures. Here is one extract:

 

 

This doesn't quite cover your case, but without reverse engineering it all I think you would get a close result by using an exponent of 1.3 to work from published 50C dT rad outputs. Certainly it would work for the small correction from the figures tabulated above for 35 to get to your 32.5C MWT.

 

All assuming that the scale factors are independent of actual physical rad sizes but I think that is reasonable.

 

I think you would struggle to get a working rad flow temp as low as 35 though without ending up with a lot of K3s.

 

Our oil boiler had been set to 65C, and at 50C flow we only just avoided using any, the average rad size upgrade factor was 1.9. The installer wanted to fit one in B2 but I persuaded him the 5% increase in output from plumbing a K22 as TBOE was sufficient. Even so there was a subsequent iteration to get the sizes down by increasing the MWT to 47.5.

 

Of course all these figures are for worst case OAT (-0.2C, Plymouth), using the WC we expect them to be much lower most days.

[sharpener]

11 minutes ago, Beelbeebub said:

Understood, but is the root problem that there are some rooms that cannot be heated without using rads at 45C and UFH at 40C at the same time? 

 

What is the.max temp the UFH can stand? 

 

It's right on the margin. Normally we get away with using just the two rads in the living room, as neither of us likes the sensation of a warm floor. When we tried it with the oil boiler I think the mixed down flow temp was set at 45 but we certainly wouldn't like it that hot on a regular basis.

 

It's further complicated by ToU tariff as Cosy is cheap from 1 pm to 4, then it goes to peak rate until 7. We are elsewhere earlier in the day so start to heat the room from 3, and after 4 continue to run the rads from the thermal store (which runs down from a 50C starting point) as we want the room up to temp by 5 o/c.

 

But in the recent cold snap the rads were not enough and adding the UFH as currently configured did not help as it pulled down the flow temp. So the aim now is to avoid that and at least have enough heat from the UFH to prevent heat loss through the floor, I think that will do it. You might think it would be sufficient to run one or the other, but the UFH thermostats click on and off every 30 mins or so in the equilibrium state so the time constants involved are surprisingly short.

[marshian]

50 minutes ago, sharpener said:

Yes and no. I used the free heatpunk s/w to make my own estimates of the whole house heat loss to confirm HP size.

 

The room-by-room estimates were then done in detail by the three potential installers as part of their MCS approved quotations. All used Heat Engineer (IIRC you can buy a one-project licence for a reasonable fee but the learning curve looked too steep).

 

They all quoted figures at a range of temperatures. Here is one extract:

 

 

This doesn't quite cover your case, but without reverse engineering it all I think you would get a close result by using an exponent of 1.3 to work from published 50C dT rad outputs. Certainly it would work for the small correction from the figures tabulated above for 35 to get to your 32.5C MWT.

 

All assuming that the scale factors are independent of actual physical rad sizes but I think that is reasonable.

 

I think you would struggle to get a working rad flow temp as low as 35 though without ending up with a lot of K3s.

 

Our oil boiler had been set to 65C, and at 50C flow we only just avoided using any, the average rad size upgrade factor was 1.9. The installer wanted to fit one in B2 but I persuaded him the 5% increase in output from plumbing a K22 as TBOE was sufficient. Even so there was a subsequent iteration to get the sizes down by increasing the MWT to 47.5.

 

Of course all these figures are for worst case OAT (-0.2C, Plymouth), using the WC we expect them to be much lower most days.

 

Those heat losses seem really high unless they are massive rooms

 

I used the Heat engineer software (before I was aware of the spreadsheet on here) and it didn't seem like a steep learning curve to me

 

 

I'm using flow temps of between 28 and 35 with T22 Rads and bedroom temps are set 2 deg higher than the std 18 for bedrooms

I think I'd be trying to run the house with a single boiler temp for both UFH and Rads

[Beelbeebub]

The lowest flow temp for a given power output is going to be with the ufh and rads both on together (and balanced) because that is the largest emmiter area. 

 

So you could run open loop WC, and just balance the UFH and rad flows to give you the best results. 

 

What about the salus autobalancing trv/actuators? They monitor the flow and return and adjust the flow (they sit on standard trv valves or even ufh loop actuators) so the dT is 7c (or maybe 5c, I forget) 

[JohnMo]

1 minute ago, sharpener said:

ToU tariff as Cosy

Think that's half your issues, running more like a boiler than a low and slow heat pump. You have a mad blast to heat in short bursts, 

 

29 minutes ago, sharpener said:

cheap from 1 pm to 4, then it goes to peak rate until 7.

Trying to stuff loads of hours of heat in 3 hrs just doesn't work, may do with a boiler but not a heat pump.

3 minutes ago, sharpener said:

but the UFH thermostats click on and off every 30 mins

That would indicate your flow temps are way to high.

[JohnMo]

9 minutes ago, Beelbeebub said:

dT is 7c (or maybe 5c, I forget

Both are correct, over 30 Deg flow it's 7dT, while under that is 5dT.

[Dillsue]

4 hours ago, sharpener said:

 

 

 

The table has fixed rad sizes(existing??) showing the output at various flow temps to guide on choosing the minimum flow temp. I'm working on fixed flow temp/heat loss with a calculation to show dT50 equivalent rad output for running at 35c. I'm wanting to keep flow temps low to get a good COP so it's the rad size I'm interested in rather than the flow temp needed for a preset rad size.

 

For your bedroom 2 the rad output shown falls way below the heat loss. Are you running like that??

[sharpener]

40 minutes ago, Dillsue said:

The table has fixed rad sizes(existing??) showing the output at various flow temps to guide on choosing the minimum flow temp. I'm working on fixed flow temp/heat loss with a calculation to show dT50 equivalent rad output for running at 35c. I'm wanting to keep flow temps low to get a good COP so it's the rad size I'm interested in rather than the flow temp needed for a preset rad size.

 

For your bedroom 2 the rad output shown falls way below the heat loss. Are you running like that??

 

That table was purely to show you how Heat Engineer tabulates the output at various MWTs, to help you scale your rads for a 35C flow temp if you can't get that info from the rad mfr's data sheets. The relative data are what is important, the absolute values less so.

 

No, those were the original rad sizes. B2 is the one they wanted to upgrade to a 600 x 1200 K33 but the compromise was a K22 with TBOE and upping the MWT to 47.5C. The original B2 rad went into the bathroom where there was none at all. Previous owners of the house had some funny ideas, e.g. the rads are mostly on the far side of the rooms from the windows but taking up all the floors to re-site them was too much. Many were too small even with the oil boiler, I put 3 additional ones in before the HP was thought of.

 

 

Edited January 29 by sharpener