Willis heater UFH. Heaters in parallel or series?

[Russdl]

We have two Willis heaters plumbed parallel for our UFH. Generally only 1 is required to maintain the desired temperature in the 'off peak' window when the UFH is on. The idea was that if one failed we could turn the other one on for a seamless transition and keep the heat flowing.

 

There was no way to isolate the Willis heater that was 'off' so I essentially had a mixer where half of the UFH return water would go through the 'on' Willis heater and be heated, the other half would go through the 'off' Willis heater and not be heated.  When the two flows rejoined for another trip around the UFH coils the overall temperature would be reduced after this mixing.

 

To counter this I fitted a couple of isolation valves so I could isolate the 'off' Willis heater and send all the return water through the 'on' Willis heater. The obvious problem with that solution is the risk of turning 'on' an isolated Willis heater (hopefully the built in over temperature protection would save the day!)

 

My plan now is to plumb the Willis heaters in series. This would:

 

Remove the mixing effect.

Remove the need to isolate the 'off' Willis heater.

Still allow system redundancy.

Still allow 2 Willis heaters to be run at once if it got particularly cold and the sun buggered off for a protracted period.

 

What are the flaws in that plan? @Onoff could you elaborate?

 

 

[Onoff]

12 minutes ago, Russdl said:

What are the flaws in that plan? @Onoff could you elaborate?

 

Only what I've read elsewhere, maybe I've got it wrong. Seems sensible just to isolate each with valves (like on a CH pump) should you need to replace one then you can work on that without losing the good one for a period. 

 

[JohnMo]

2 hours ago, Russdl said:

What are the flaws

Not an easy way to take the failed unit out of the system to replace when in series.

 

Are you over thinking it? Why not just simply one Willis on line the other offline, out of the flow stream (valve isolated). The one on line just set the temperature and timer/house thermostat and let it go. Dump all the mixing stuff. 

 

Do you install two boilers, which are complex and way more likely to fail?

 

Have you done a heat loss calc and do you need more than 3 kW?

Edited Sunday at 12:02 by Nickfromwales
Typo

[Mattg4321]

Changeover switch would solve the problem of the isolated one being turned on

[Russdl]

1 hour ago, Onoff said:

Only what I've read elsewhere, maybe I've got it wrong

 

Thanks for that, I don't think that document is relevant for the Willis heater arrangement but I appreciate the 'heads up'.

[dpmiller]

just set the stats differently so one will switch off when the flow temp starts to come up?

[John Carroll]

3 hours ago, Russdl said:

We have two Willis heaters plumbed parallel for our UFH. Generally only 1 is required to maintain the desired temperature in the 'off peak' window when the UFH is on. The idea was that if one failed we could turn the other one on for a seamless transition and keep the heat flowing.

 

There was no way to isolate the Willis heater that was 'off' so I essentially had a mixer where half of the UFH return water would go through the 'on' Willis heater and be heated, the other half would go through the 'off' Willis heater and not be heated.  When the two flows rejoined for another trip around the UFH coils the overall temperature would be reduced after this mixing.

 

To counter this I fitted a couple of isolation valves so I could isolate the 'off' Willis heater and send all the return water through the 'on' Willis heater. The obvious problem with that solution is the risk of turning 'on' an isolated Willis heater (hopefully the built in over temperature protection would save the day!)

 

My plan now is to plumb the Willis heaters in series. This would:

 

Remove the mixing effect.

Remove the need to isolate the 'off' Willis heater.

Still allow system redundancy.

Still allow 2 Willis heaters to be run at once if it got particularly cold and the sun buggered off for a protracted period.

 

What are the flaws in that plan? @Onoff could you elaborate?

 

 

With the 2 in parallel, say 3kw each, assuming a total flowrate of 12.5LPM, 6.25LPM through each, gives a dT of 6.88C, a return temp of 42C gives a flow temp of 48.88C. 2 in series, flow through both is 12.5LPM, dT through each is 3.44C, it enters first heater at 42C, leaves at 45.44C, enters the second at 45.44C, leaves at 48.88C. Switch one off (still in series) dT through first (12.5LPM) is 3.44C and 0C through the second, flow temperature 45.44C.

 

2 in parallel, one switched off, dT through switched on one, (at 6.25LPM), 6.88C, flow temp 48.88C, dT through second (switched OFF)  0C, flow temp 42C, 6.25LPM at 48.8C mixed with 6.25LPM at 42C gives 12.5LPM at 45.44C.

Edited Sunday at 13:02 by John Carroll

[Nickfromwales]

@Onoff the reason for fitting these in parallel is explained to the T in your link, thanks.

 

For the benefit of other forum members:

 

When I fit Willis heaters in parallel I make sure the pipework to each unit, in and out, is hydraulically equal (same pipework distance and size). This does indeed maintain equal flow through both units, and the notion of any problem created by “mixing” within ‘the heater’ is a dreamt up issue at best. 

 

You don’t see domestic clients pull apart a new W-Bosch combi and start adding valves to their plate heat exchanger internals, saying “some pathways are slightly hydraulically different”, do you? Fact is, cold water goes in, and hot water (the product of the various mixing characteristics of the PHE internals) comes out the other end……🤔 Literally couldn’t be easier…..

 

The input temp vs output temp is what you reference; simply getting the sum in / out values and indeed then allowing the heaters to do the work as they were designed to do so. If however the design is hydraulically imbalanced, then flow would go through one heater more than the other, which is why I don’t do that.

 

A lot of micro-managing type folk have discussed this with me, in other instances, unfortunately LONG AFTER THE DESIGN AND INSTALLATION HAS TAKEN PLACE AND THEY’VE HAD EVERY OPPORTUNITY TO ASK QUESTIONS UP TO THE POINT OF COMMISSIONING, often to criticise or question (more so when another plumber has been approached in the interim for a random uneducated opinion on this AND HAS GOT IT WRONG).

 

More annoyingly, in another instance, a complaining client I installed such a system for had posted on a public forum to say the system that I installed was working “too well” if anything, but then went on (when bored and over-thinking everything as usual) start attempting to pick faults that didn't actually exist, complain about hypothetical problems, ahead of them ever occurring / or having had actually occurred, but this is almost always simply the unexplainable and irrational nature of some people.

 

Hydraulically, a pumped system, such as the OP’s (may possibly(?) be) will be primarily piped in 22mm copper, and the Willis heaters are 15mm in/out, therefore these would then be a choke point; if, by design, the system installer had to account for the system to be able to cope with any adversities (and be reliable) then adding 2x in parallel would give 30mm of hydraulic pathway visible to the 22mm pump body therefore allowing it to enjoy minimal resistance and live a long happy life.

 

I assume any “pro” plumber, perhaps not the original installer, if the aforementioned alterations were not bodged by the homeowner, would have made this observation before hacking into a perfectly working already installed system, and there would now be a PBV teed into the flow out of the pump and back to return to cope with the change in hydraulic difference, so as to not labour the pump. Not heard any mention of that, just isolating valves, so this is probably now to the detriment of the originally design-in longevity and reliability. Cool. Only joking, completely uncool. Irony is, that would then constitute a mixing effect, but if that design came from a “pro” plumber….then that would be ‘all fine and dandy, I expect.

 

Just grabbing at straws here btw, so please don’t take any of this as factual, and please don’t quote (re-quote) me here as I have no clue as to what I’m doing, (via my inconsequential 30+ years experience of plumbing / heating et-al), just pure guesswork which you should all ignore  

 

I once was asked, in another instance, about doing this, and I went on to explain in detail why putting 2 in series was (IMO) a terrible idea, and that person agreed and I carried on about my business. System still works perfectly well to this day without complain or defect (Passivhaus type dwelling with an insulated raft / airtight / MVHR etc etc).

 

As per the internal PHE of a brand new combi, in another similar instance, I suggested that the client leave it the fcuk alone and enjoy the heated water that comes out of it reliably, and as per its intentional design, as it’s worked for everyone else that has one like that…..

 

“If it ain’t broke, then don’t fix it”.

 

3 hours ago, JohnMo said:

Not an easy way to take the failed unit out of the system to replace when in series.

I disagree.

 

When I plumb in Willis heaters like this, I have 22x22x15mm compression Tees where the 15mm spurs face forward. The Willis heaters compression fittings push back onto these spurs and job done. In the event of failure the 2x 15mm compression nuts get loosened off the T, the new Willis adopts the 2x reusable spurs and the lot then gets remade back as original (no cutting of pipe etc, just 2 or 4 new 15mm olives needed at most). 30 mins drain down / swap / back to work is my design methodology. 

 

When I design whole of house M&E systems for self build clientele I always factor in this discipline, eg I plumb it to make downstream service / inspection / replacement of components that we all know will eventually fail, easy to do and 99 times out of 100 without anything other than a towel and a spanner, just as any “pro” installer should.

[Nickfromwales]

5 minutes ago, John Carroll said:

With the 2 in parallel, say 3kw each, assuming a total flowrate of 12.5LPM, 6.25LPM through each, gives a dT of 6.88C, a return temp of 42C gives a flow temp of 48.88C. 2 in series, flow through both is 12.5LPM, dT through each is 3.44C, it enters first heater at 42C, leaves at 45.44C, enters the second at 45.44C, leaves at 48.88C. Switch one off (still in series) dT through first (12.5LPM) is 3.44C and 0C through the second, flow temperature 45.44C.

 

2 in parallel, one switched off, dT through switched on one, (at 6.25LPM), 6.88C, flow temp 48.88C, dT through second (switched OFF)  0C, flow temp 42C, 6.25LPM at 48.8C mixed with 6.25LPM at 42C gives 12.5LPM at 45.44C.

Yup. We’re just wasting time discussing a problem that doesn’t exist, unfortunately.

[Nickfromwales]

1 hour ago, dpmiller said:

just set the stats differently so one will switch off when the flow temp starts to come up?

Then the fail rates change and it becomes a “lead & lag” arrangement, meaning one heater is doing the brunt of the work always. Not how I design these things.

 

I design this so that the 2x Willis are “the heater”, with a 22mm flow in, and a 22mm flow out, and end of.

 

I’d have to ask why anyone would want to complicate such a solid, simple, reliable setup which already promotes huge longevity? Especially when it’s in, fitted, commissioned, working, and not defective at all?

[Russdl]

Wow, there's a lot of words up there.  ⬆️ And some shouting. Just for the record I have no criticism of the physical installation in any way, it's a work of art.

 

~~~

 

I'm not sure how many people on here use Willis heaters for their UFH? I don't think there are that many of us? We only need one 3kW Willis heater to heat the house. We have two 3kW heaters plumbed in parallel.

 

As someone who has lived with Willis heater UFH for a few years our experience is that having the water flow through 2 Willis heaters (one 'on' one 'off') mixes down the temperature of the water exiting the 'on' Willis heater. It would, wouldn't it.

 

The water flows through the Willis heater at a certain rate (not exactly sure how many litres/min) but definitely not slow enough to get to any significant temperature before it's on its way around the system again. The highest exit temperature I've seen with a thermal camera is 34.2°C and that's after the UFH has been running for the best part of 6 hours (I set the thermal cut out to 40°C). At the beginning of those 6 hours it would have been around 22-23°C. 

 

The house copes very well with the one Willis heater providing the heat and, despite this January being unusually cold for us soft southern bastards, we've had a useful amount of sun to offset that and just running the one Willis heater has been fine. 

 

Anyway, as the prophet said "He's grown man so let him just carry on about his business please" and, as a grown man - I think I will 😃

[ProDave]

9 minutes ago, Russdl said:

The water flows through the Willis heater at a certain rate (not exactly sure how many litres/min) but definitely not slow enough to get to any significant temperature before it's on its way around the system again. The highest exit temperature I've seen with a thermal camera is 34.2°C and that's after the UFH has been running for the best part of 6 hours (I set the thermal cut out to 40°C). At the beginning of those 6 hours it would have been around 22-23°C. 

The overall flow temperature you are seeing is the result of the return temperature, the flow rate, and the power you are putting into it with the heater.

 

I would wager if you valved off the "off" heater, you would see exactly the same figures as you are still putting in the same amount of heat.

[SteamyTea]

16 minutes ago, ProDave said:

would wager if you valved off the "off" heater, you would see exactly the same figures as you are still putting in the same amount of heat

You should get a slightly higher temperature because the pipework surface area is decreased.

Probably too small to measure though.

[Mike]

4 hours ago, Russdl said:

As someone who has lived with Willis heater UFH for a few years our experience is that having the water flow through 2 Willis heaters (one 'on' one 'off') mixes down the temperature of the water exiting the 'on' Willis heater. It would, wouldn't it.

 

The water flows through the Willis heater at a certain rate (not exactly sure how many litres/min) but definitely not slow enough to get to any significant temperature before it's on its way around the system again

I'll qualify this by saying that mine is still running via a (suitably rated) extension lead, so I only have the temperature gauge on the manifold & finger on pipe to work with.

 

So, based on my limited experience (the end of last winter and this), I suspect that you're just running the pump too fast. After experimenting manually with different settings, the pump tells currently me that it's running at 2W, pushing between 100 & 200 litres / hour, which seems to work well; It has no trouble reaching it's maximum of 38°C within a short space of time and without the heater cutting out.

 

FWIW I corresponded with Elemex (one of the manufacturers of Willis heaters) on the topic when planning my system, who advised that "circulating of the pump /flow rate and heat transfer would ideally want to be smartly controlled to find the most efficient flow-rate : heat transfer ratio". Which I hope to do next winter.

 

I do have twin heaters (pic), but only one active. That's mainly so that I can switch power from one to the other if/when a thermostat packs up; they're both permanently in the flow. And maybe so that I can get to temperature more quickly from cold / to take advantage of cheap-rate electricity.

[John Carroll]

7 hours ago, Russdl said:

 

 

As someone who has lived with Willis heater UFH for a few years our experience is that having the water flow through 2 Willis heaters (one 'on' one 'off') mixes down the temperature of the water exiting the 'on' Willis heater. It would, wouldn't it.

 

Yes, it would and will, but the outlet flow and temperature will be the same, whatever configuration you use or way you install the heaters.

 

[Nickfromwales]

6 hours ago, SteamyTea said:

You should get a slightly higher temperature because the pipework surface area is decreased.

Probably too small to measure though.

Only if the defunct heater is physically disconnected, or else it’ll inevitably heat up from the physical connection of the heated pipework; it’ll just turn into a radiator at that point, no?

[Nickfromwales]

9 minutes ago, John Carroll said:

Yes, it would and will, but the outlet flow and temperature will be the same, whatever configuration you use or way you install the heaters.

Yup, exactamundo.

 

You can lead a horse to water, but some are just destined to become glue.

[Nickfromwales]

As a potential “prophet”, I would attempt to reach out to people and tell them to get off their uneducated horse and rely on good peoples work ethics, their industry experience of 30+ years, and their zero fail rate. 

 

When someone has taken the time and effort to design a system for a client, in good faith, they would hope that they would be rewarded for their efforts….not second-guessed incorrectly and their efforts kicked to the gutter, which is something that happened to me, in another similar instance, where the client decide to retrospectively scrutinise works undertaken (after stating it was working perfectly well) and most annoyingly when they know the square root of fcuk all about plumbing; the reason they sought the services of one in the first place I suspect.

 

IF someone was suitably educated, they would surmise that if 2 Willis heater were installed hydraulically equally, then NO FECKING mixing could occur through the other unit IRRESPECTIVE IF EITHER WERE ON OR OFF. Flow through each unit would occur, as designed, eg equally.

 

If one heater is off, in this scenario, and the other on, then the sum out of the “heater” would be 3kW of heat output.

 

I guess we’re off to the glue factory….

 

[Russdl]

16 hours ago, Mike said:

It has no trouble reaching it's maximum of 38°C within a short space of time

 

That's interesting, I've never seen temperatures even close to that after 6 hours of running. I've turned the pumps down as well to slow down the rate of flow, perhaps I need to turn them down further? I do have a 7°C temperature drop between flow and return which I believe is optimum.

 

Out of interest, what size is your UFH system? Mine is 824 meters of pipe (no idea how many litres that equates too).

 

@John Carroll Thanks for those figures. What is the maths, I can't work it out?

 

That and @ProDaves comment 

 

21 hours ago, ProDave said:

I would wager if you valved off the "off" heater, you would see exactly the same figures as you are still putting in the same amount of heat.

 

certainly indicate that there is no benefit from isolating one of the Willis heaters, but the lived experience is different for whatever reason. 

 

 

[SteamyTea]

1 hour ago, Russdl said:

What is the maths, I can't work it out

Specific Heat Capacity times Mass Flow Rate time Temperature Difference.

That gives you the power required.

For just the energy you use Mass, rather than Mass Flow Rate.

It is always best to use SI units (J, kg, K, s) rather than imperial or derived units. Convert all temperatures to kelvin, K, it can get funny when passing through zero sometimes, especially when dividing (after rearranging formula for instance).

[John Carroll]

3 hours ago, Russdl said:

 

That's interesting, I've never seen temperatures even close to that after 6 hours of running. I've turned the pumps down as well to slow down the rate of flow, perhaps I need to turn them down further? I do have a 7°C temperature drop between flow and return which I believe is optimum.

 

Out of interest, what size is your UFH system? Mine is 824 meters of pipe (no idea how many litres that equates too).

 

@John Carroll Thanks for those figures. What is the maths, I can't work it out?

 

That and @ProDaves comment 

 

 

certainly indicate that there is no benefit from isolating one of the Willis heaters, but the lived experience is different for whatever reason. 

 

 

 

My maths?.

 

I've allways used kw = (flowrate (LPM) x 60 x temperature rise, dT (deg.C)/860

so, kw = LPM*60*dT/860 or any variation of this once you know two of the values

I used a total flowrate of 12.5LPM and 2 X 3kw heaters.

From the above, kw = LPM*60*dT/860

dT = kw*860/60/LPM

One 3kw heater, with a flowrate of 6.25LPM (12.5/2), will give a temperature rise, dT, of, 3.0*860/60/6.25, 6.88C, if the heater inlet temp is 25C, then the outlet temp is, 25+6.88, 31.88C, two heaters (on) in parallel will give a flowrate of, 2*6.25, 12.5LPM at the same dT of 6.88C with the same outlet temp of 31.88C.

Two heaters in parallel with one switched OFF will give a 0.0C, dT , the other (ON) will give a dT of the above 6.88C. You now have 6.25LPM at 25C mixing with 6.25LPM at 31.88C.

Let Tm = mixed temperature

So, 12.5*Tm = (6.25*25)+(6.25*31.88)

So 12.5*Tm = 156.25+199.25, 355.5

Tm = 355.5/12.5, 28.44C

If one heater is OFF and the water flow through it shut off then 12.5LPM is flowing through the switched ON heater with a dT (kw*860/60/12.5) of 3*860/60/12.5, 3.44C, to give a outlet temp of, 25C+3.44C, 28.44C, the same as the above with the flow split between the two heaters.

 

Two heaters in series, both ON, flowrate 12.5LPM.

dT in both/either  heaters will be, (kw*860/60/LPM), 3*860/60/12.5, 3.44C, outlet temp from bottom heater, 25C+3/44C, 28.44C, outlet temp from top heater, 28.44C+3.44C, 31.88C (same as with both parallel heaters ON, above)

Two in series, one off, one ON, dT through ON heater, (kw*860/60/LPM), 3*860/60/12.5, 3.44C, outlet temp, 25C+3.44C, 28.44C, dT through top (OFF) heater 0C, outlet temp, 28.44C. (same as with one heater only on in parallel).

 

You have ~ 800M of UFH piping, a loop might consist of up to 100M of piping with a flowrate of 2.0LPM giving say 6C to 8C dT at a flow temp of ~ 40c/45C, so each loop (assuming a dT of 7C) will emit, (LPM*60*dT/860), 2.0*60*7/860), 0.977kw, say 1.0kw, you should have ~ 8 loops so, at least 8kw if running at the above conditions, 45C/38C, so obviously 3kw will give a very much reduced output (it will give exactly 3kw if the heater stat never cuts out) because of the very low loop flow/return temps, each loop might/should have a flow meter tube showing the flowrate in LPM, if so, just add them to get the total flowrate.

Edited Monday at 17:54 by John Carroll

[SteamyTea]

3 minutes ago, John Carroll said:

kw

What is this?

[John Carroll]

kw is a unit of power, a gas boiler might have a output of say 25 kw (my upper/lower case sense of "kw" might be incorrect but doesn't matter from a thermodynamic point of view) so you have a 25 kw Boiler, etc.

[JohnMo]

1 minute ago, John Carroll said:

kw

May as well write it correctly, your posts look much more informed, and generally could be taken as way more factual. At least you show you know what you are talking about.

 

kW. (k represents 1000, W represents Watts (as in a surname)

[SteamyTea]

3 minutes ago, John Carroll said:

my upper/lower case sense of "kw" might be incorrect but doesn't matter

It should, units are important for clarity.

kW

kWh.