Fitting a Willis Immersion heater

tnegun

Sorry sure Red is the temp at the top of the willis, Orange is the temp at about the middle of the hot tank and Blue is the temp a couple of inches from the bottom.

To further complicate matters I just discovered the stat on the cylinder immersion had failed! No matter what temp I set it to I could get it to heat and I've swapped it now with a 7" spare I've had for years but what are the chances.

The Continental Op

It then comes back up to the same temp as the tank without any power applied this has to be reverse circulation e.g. hot water from the tank is flowing to the willis do you think is this expected?

I'd expect it. If the Willis can heat the tank the tank can heat the Willis. Its just the water stratifying and the Willis is part of that equation.

John.G

https://www.boards.ie/discussion/comment/119077301#Comment_119077301

What has happened at the start of the dip down (willis top temp) and what has happened at 1847 when it starts to rise again, the willis bottom temp is gradually reducing so how does this equate to actual reverse circulation as one would expect the bottom to start rising as well?.

tnegun

The dip down coincides with a shower being taken and the rise the finish of the shower, only the red line is the temperature of the willis the other two are temps from the hot tank itself.

The Continental Op

Can you measure the temperature just below the highest point where you have joined the Willis into the system? I think that would give a better idea if cold water is coming in via the Willis.

John.G

Also the return to the willis, if reverse circulation then this will too get hot?.

tnegun

https://www.boards.ie/discussion/comment/119078496#Comment_119078496

I can try just need to figure a way to attach to the pipework.

https://www.boards.ie/discussion/comment/119078548#Comment_119078548

Yes this gets quite warm/hot once the willis shuts off - no solar excess or the stat opens.

John.G

Maybe the 50cm max is vital, as pointed out in post#88, forget what yours is, the heater seems to be acting like a gravity fed radiator?.

Is the vent coming off the very top of the HW cylinder?

I would insulate that 3/4 ins Flow piping as it will act as a nice little heat emitter and promote reverse circulation maybe.

I would install that pipe stat now.

Maybe do the obvious and install that pipe/cylinder stat low down on the cylinder, should be able to get the optimum position to give almost that full tank of hot water as the willis de stratification effect especially if the flow piping is lagged may have little effect, insulating it will also reduce losses.

Looking at the above trend, looks like you may be able to locate it not too far above where the blue trend sensor is located?.

Tnegun HW Cylinder.png

tnegun

The willis joins the vent at 15cm above the top of the tank and the vent is coming off the very top of the cylinder. I've added insulation to the 3/4" coming from the wills and fitted the pipe stat here just above the top of the willis on the return piping set to 75c(I tested manually and it was cutting out too quickly when set to 65c and just when convection started to get going) and reset the willis internal stat to 60c.

I didn't fit the stat to the tank as the willis will create very hot water if not controlled more directly. Todays been a bit rubbish for testing but when low levels of excess <1.5kw are available the convection can keep a nice constant flow through the willis as opposed to when it goes higher and starts tripping the stat.

John.G

If you could limit the max diverting kw you might get a optimum Willis input < 3kw. Don't think you can improve on your pipework, I presume the isolating ball valve on the Willis return is full bore as its the only obvious potential restriction.

tnegun

Yes the isolating valve is fully open and that's my next thought to see if the diverter can limit its output, I don't think it can as there's nothing obvious in the config but there is an option to buy a relay board which will run a destrat pump when the immersion stat trips so that could be interesting! Also an interesting graphic below I've overlayed the export the green line on top of the temps. You can see the export jump when the willis stat opens and then disappear once it closes. Its not perfect but not bad and will hopefully perform well on marginal days when the excess is less.

John.G

SO, the willis stat is doing the switching and pipe stat not operating because its set to 75C?.

What is the 0 to 300 scale on the right?.

tnegun

The pipe stat is doing the switching at about 75c, the wills stat is set to 60c with its previous performance, I'd expect that to actually switch out at 85-90c. The scale on the right is the kw export. I messed it up it should be showing 0-3000 and its relevant only for the green line.

John.G

Very good, and its the pipe stat switching back in when the temperature falls 5/6C (stat switching hysteresis) which means the export is diverted to the willis. The pipe stat seems to be switching off at 65C to 68C when the export/diverted is 1.1kw to 1.2kw or maybe that's just lack of solar?.

Re the ball valve, I should have said full bore trim, most ball valves have reduced trim ie a 3/4 ins ball valve may only have a 1/2" "hole" in the ball unless full size trim is specified, it probably still shouldn't make any difference but one never knows.

I'm beginning to wonder just how successful these heaters are/were in view of the deafening silence from the thousands? of users, maybe the stats are just shorted out and the timer determines the volume of hot water as its hard to see how any of them will maintain say a steady 60/65C at 3kw. but then surely someone would have shouted about the 75/80C hot water.

tnegun

Yes I agree, the export is occurring when the stat opens and is 2.5-2.8kw close to the max input of the willis so it's safe to say the solar was providing this level of power to the willis when the stat opened and it can't dump the heat into the cylinder fast enough. Very good point re the ball valve I have it on the 1/2" supply-side so it could well be restricting flow.

I agree re the lack of other reports on success or otherwise I wonder if it's because other users may have a battery and car soaking up excess too but it is very quiet out there regarding them!

John.G

The Willis pre dated solar PV by decades, it was invented I think as a cheap option to replacing a DHW cylinder that didn't have the facility to install electric immersion elements. So assuming they did/do work with 1/2" flow piping it follows that yours has some restriction to flow, maybe replacing that ball valve with a gate valve will help but I wouldn't hold my breath.

The Continental Op

The Willis company still exists in Belfast https://willis-heating.com/about/

The Willis Immersion Heater .

Third generation, Jack Willis, Fred Willis’ son, invented probably the best known of all the Willis inventions, the Willis Immersion Heater.

At one time there was hardly a household in Northern Ireland that did not have a Willis Immersion Heater, and a high proportion still have. As our history shows, the Willis name has always been associated with innovation, quality, service and performance. This remains true of the Company today.

Our success and reputation has ensured repeat business and negotiated orders since 1887, an unrivalled record which speaks for itself.

John.G

That sounds like a ad for the Wallis empire, I wonder was Barnes (Wallis) any relation as he invented that ingenious bouncing bomb that did so much damage to the Ruhr valley dams in WW2.

tnegun

Not a great day for the Solar but the tank soaked up all of the excess with 5kwh dumped into it. On the point of the isolation valve restricting flow I had a look at one, I found in the shed today and this type reduces the diameter by about 40% so I think I'll replace that over the next couple of weeks too.

I've also added a CT to measure exactly what the diverter is supplying to the immersion and willis. KW supplied by the diverter (Eddi) is the blue line below, Yellow the temp at the pipe stat above the willis and red the temp at the middle of the hot tank.

John.G

Some great info there now, what is the capacity of the HW Cylinder?.

tnegun

I think it's about 150l there are no markings on it to say for sure.

The Continental Op

Height and rough diameter and I'm sure we can work it out.

John.G

So ~ 9.2 kwh to heat a full tank, so the diverted 5kwh (assuming little or no HW draw off) should have been fairly close to rising the middle tank temp steeply, be interesting to see when it does.

tnegun

https://www.boards.ie/discussion/comment/119095187#Comment_119095187

Its approx 900mm x 500mm.

https://www.boards.ie/discussion/comment/119095226#Comment_119095226

That's not far off what I was hoping to get into it with the willis. I did some calculations before I bought it and thought if I got 8kwh into it from cold I'd be doing well! I submitted the gas reading this evening too I've only used 1 unit in a month vs 29 last May. I did the electric too 250kwh this May vs 611kwh last year.

The Continental Op

Its approx 900mm x 500mm.

900mm x 400mm is approx 95l

900mm x 450mm is approx 120l

Assuming an inch of foam on the outside of the tank I'd guess its approx 450 mm actual diameter of the tank so probably around 120 liters.

John.G

22/28X.45^2X0.9X1000 is 143 litres.

John.G

Just looking at pressure drops in orifices, a few interesting numbers showed up. The circulating head at a flowrate of 0.72LPM with flow/return temps of 15C/75C is 0.00758M and if one considers the ball valve bore as a orifice then it would have to be greater than 8MM. The circulating head at a flowrate of 1.23LPM with flow/return temps of 15C/50C is 0.003M the ball valve bore would (theoretically anyway) have to be 12MM or greater.

You have a spare, can you measure the bore with drill or a calipers?.

tnegun

Great idea! A 7mm drill bit fits through with a little to spare the 8mm won't. The CT readings are all over the place and not accurate I guess due to the diverter varying voltage, current etc as it adapts to the available excess but it's done a great job today soaking up excess with 3 showers taken and another 6kwh gone into the tank which is about half full now going by the temp readings.

John.G

It can't do any harm to change that sometime to a full bore ball or a simple gate valve, every bit helps.

tnegun

I'm going to go for it. Up to 2kw diverted there is enough flow through the willis not to open the stat any higher and it opens after 1-2 mins then closes again after about the same not bad but it would be nice to have it run at the full 3kw.

tnegun

Picked up the replacement valve today so will try fit it when I've time in the mean time the impact of the willis can be seen here nicely. The tank immersion has priority so it heats first with the red(willis) temp tracking the yellow tank temp. The tank stat opens and the diverter detects this so starts sending surplus to the willis and the red line now climbs. The stat is opening and closing along with less available excess giving the peaks and troughs but the tank temp continues to climb and most importantly the bottom of the tank - blue line is climbing too so the whole tank is heating.

John.G

The blue (bottom) will never rise with the tank immersion/stat as the immersion is above that (and the reason for installing the willis) so what is the yellow trend? if its the middle of the tank temp then that should rise uniformly on the tank immersion assuming long (bath) element in service and no HW draw off?, and is the tank immersion stat set to 60C?. or is it only set to 38C?.

tnegun

Yeah I understand that I was highlighting that the willis was having an effect as warmer water was making its way down the tank as hot water entered from the top. The yellow trend is the middle of the tank and the blue line the bottom. The in-tank immersion is set to 65c on bath but could never heat more than the top 1/3 of the tank sufficiently the yellow line would rise to about 40c and that was it which is what sent me down this rabbit hole 😀 . The tank was almost completely cold this morning and took 7kwh of excess today before the showers started.

John.G

The dual immersion brings its own challenges/problems, If the cylinder is stone cold then it should all (IMO) heat up uniformly with a fully heated volume from the bottom of the bath element to the cylinder top, if hot water is then drawn off (after tank stat switches off) until the layer of hot water just extends to the bottom of the sink element then the stat will not switch back in because the stat rod only extends the same length as the sink element and will only switch back in when nearly all the HW is drawn off?. Now, the willis was tracking the cylinder top temperature presumably and it coincided with the middle temperature because the whole cylinder was cold as you say and at ~ 25/26C when the cylinder immersion came on so if my theory of the whole tank heating uniformly under these conditions then why did the cylinder stat cut out at only ~ 38C??, if it was reaching 65C or so then shouldn't the willis reflect this or close to it with reverse circulation until the diverter cuts in. It maybe begs the question, perhaps the sink element is being utilized and not the bath element?, or maybe faulty stat, if no faults then it should heat ~ 70% of the cylinder say 100L from 25C to 65C, 4.9kwh, if its only heating 100L to 38C then only 1.5kwh, if heating 1/3rd or 50L from 25C to 65C then 2.3kwh. Your ESB meter should tell a lot re immersion consumption if monitored.

tnegun

Its a top entry 24" Shel immersion with an 11" thermostat if it makes any difference. I tested the resistance and confirmed that the bath element is the one wired. The top of the willis is positioned about the same height as the middle of the tank so it tracking the temp sensor at the middle of the tank makes sense to me as with the reverse circulation the temp at the same height in a closed system should be similar. The 38c at the middle is well below the bottom of the tank thermostat and the top of the tank is almost too hot to touch when the tank stat cuts out. The cylinder immersion takes 3-4kwh from cold before cutting out I can see this from the diverter.

John.G

I would have thought that the willis temp with reverse circulation would be nearer the cylinder top temp than the cylinder middle but you could well be correct.

The 24" (610mm) bath element should/might heat 68% of the cylinder vol to 65C, so 101L heated from 25C to 65C = 4.7kwh. Immersion was on for ~ 2.25 hrs & assuming a 2.5kw element = 5.6kwh but you will know its power by the measured resistance. The bath element bottom is ~ 160mm below the cyl middle (temp) so would have expected the middle to be higher than 38C at the end of the heating immersion period. You say it takes 3 to 4 kwh before immersion cut out so at say 3.5kwh = cyl temp of 55C but again assumes that the whole 101L is being heated to at least this temperature but we it isn't as the center temp is only 38C. The 11" stat definitely has a bearing but is a compromise length to satisfy different length elements, one 24" element only with a 24" stat might achieve a greater volume of hot water but debatable whether this is a advantage or not now that the willis is installed. All very interesting.

tnegun

I think its the water density settling rather than full reverse circulation, also adding the insulation to the feed off the top of the willis seems to have helped here too. I'll recheck the bath element of tank immersion is the one wired as I did have to rewire that to replace the stat and when adding the willis. A neighbour has the PV same system and diverter and has similar performance to mine in that he can only get 3.5kwh into the tank but doesn't have any temperature monitoring.

John.G

Something not adding up there in both you and neighbors, if the cylinder immersion is only inputting 3.5kwh into a cold tank (say 25C) then that will only heat 75L to 65C. and a 24" immersion should heat close to 68% of a 900mm tall cylinder which means the cylinder capacity is only 110L, highly unlikely. Also, it would only leave 35L to be heated by the willis which would only require another 1.62kwh, no way obviously. If a 150L cylinder then 3.5kwh will heat 101L to 55C which looks more likely but then (assuming no HW draw off) would only need a additional ~ 3.45kwh to get the whole cyl to 65C. (101L from 55C to 65C & 49L from 25C to 65C).

What bath element resistance did you get and can you also measure the resistance across the live & neutral upstream of the immersion, where they are connected to the diverter or whatever.

The Continental Op

If you can measure around the cylinder and give an indication of the depth of the foam we can be a bit more certain of the size of the tank.

John.G

Standard insulation is 35mm so vol could be ~ 130L and if 5% is allowed for the domed top, as low as ~ 120L??.

tnegun

I'll measure again when I get home from a pic on my phone the tank immersion is a Shel C24DT2 the first google result returns a 24" Shel immersion but the part no isn't displayed so maybe mine isn't that long?

The Continental Op

Its always possible the the plumber tried a longer immersion and found it fouled on the coil so put in a shorter one?

John.G

For the moment, if a cylinder size of 120L is assumed then a 24" 2.5 kw (21 ohm) element that heats 68% or 82L from 25C to 65C will consume 3.8kwh, not too far away from the stated 3.5kwh.

tnegun

I cleared enough junk from the hot press to measure the circumference and its 1600mm with 30mm of insulation also the resistance for the immersion is about 20 ohms it wanders about a bit.

John.G

The above measurements give a cylinder capacity of 135L allowing 5L for the cylinder dished top, the bath element is 2.65kw based on a 20 ohm resistance, so you have a 24 ins, 2.65kw element heating (135*0.68), 92L of water. It requires 4.28kwh to heat 92L from 25C to 65C and will take 1.62 hrs ( 1hr & 37min) to achieve this. OR, if the measured energy consumption is/was 3.5kwh then the final temperature of the HW will be 58C.

The Continental Op

I'd go far more for it being a fairly standard 900mm x 450mm which is often sold as being 120 liters but is normally a tad shy of that at 117 liters.

Look up that size online and you'll find it fairly universally stated to be 117 liters.

As an aside - My guess is that like a surprising amount of manufactured items the machines its made on were set up for imperial measurements making that a 36 inch by 18 inch cylinder.

John.G

Yes, definitely looks like a 900X450MM, 117L does seem to be the universal capacity but even allowing for a wall thickness of 2MM it still comes to ~ 141L NOT allowing for dished top so the manufacturers seem to calculate a reduction of 24L for this which seems a bit high but they would scarcely undersize their cylinders. SO to edit post # 144........

For the moment, if a cylinder size of 117L is assumed then a 24" 2.65 kw (20 ohm) element that heats 68% or 80L from 25C to 65C will consume 3.72kwh, not too far away from the stated 3.5kwh.

HOWEVER, if we subtract the 24L for the dished end then the volume heated is only ~ 56L and should only require ~ 2.6kwh???, time for the bed I think.

The Continental Op

Would the reverse dome for strength in the base account for the extra in the dome at the top?

John.G

The dome in the top reduces the capacity IMO, but I would have thought by only say5 to 10L, , there's definitely a calculation for it, the dome in the bottom further reduces the vol but I suppose the sum of the two dome losses come to that 24L so maybe allow 12L for top dome capacity loss which means the heated vol by the bath immersion is, 80-12, 68L, requiring 3.2kwh, that definitely sounds the more likely case?.

Milsey

A lads I've read all 5 pages and no conclusion on the willis install. I have to commend your tenacity to make it work optimally. I can't wait to hear more.

I've never seen such a productive string. Fair play all that have contributed.