Fitting a Willis Immersion heater

John.G · 12-05-2022 10:18pm

And the Willis stat is set to ~ 60C?.

Circulation is too low in any event at ~ 0.66LPM to give 88C so changing to 3/4" should help.

tnegun · 12-05-2022 10:36pm

Its set to 65c there was a sticker on it saying set at factory so I didn't change it. From the info sheet included with it the immersion is 10", the thermostat is 7" and the safety cutoff is 85c. I ran it again there and the bottom of the willis is 21c, middle 55c and top 84c. If it helps the exact one I have is an Ideal Emerald Geyser. Bought here https://heatingpartsdirect.ie/products/willis-external-immersion-heater-ideal-710536 edit. I linked the wrong heater fixed now

The Continental Op · 12-05-2022 10:49pm

Can you swap the immersion on that one or is it a sealed unit?

tnegun · 12-05-2022 10:52pm

I didn't inspect it closely enough to be sure before fitting are you thinking a smaller element?

The Continental Op · 12-05-2022 11:02pm

I was asking really because I want one that has a standard 2-1/4" immersion so I can take it out to replace it. I know when I was looking around for one there was one model that I read a review where it was suggested the immersion and the copper "can" were a sealed unit. Looking at pictures of the Emerald Geyser I don't see the hex head I'd expect to see on an immersion to screw it in or out.

This was the one I was hoping to get hold of https://www.stevensonplumbing.co.uk/willis-external-immersion-heater.html but I know the local heatmerchants has a few various ones hidden covered in dust on the top of a storage cabinet so I'll see what those are like first.

John.G · 12-05-2022 11:17pm

https://www.boards.ie/discussion/comment/119054530#Comment_119054530

The average of the 3 temps is 53C so the stat may cut out by reducing the setpoint to say 50C but again this is really not satisfactory as if it does cut out and you don't draw off any water, it might take a long time for it to reset except that you are drawing off some HW to short circuit some cold through the Willis. Ideally, the dT through the Willis should be ~ 35C to 40C to allow continuous operation until the return temp is above say 25C to 30C, increasing to 3/4" will certainly help but the horizontal runs are possibly the biggest problem re circulation so try and get them inclined a bit.

The Continental Op · 13-05-2022 12:14am

Odd question OP do you have a Surrey flange on the cylinder? You'd only really see them installed on a copper tank. A doohickey to prevent air getting into the the hot water coming from the tank, helps improve the life of show pumps afaik?. Its something that might affect the flow of water? Didn't think they were common in Ireland but I see screwfix and heatmerchants have them https://www.screwfix.ie/p/salamander-pumps-s-tank-flange-fitting-for-hot-water-cylinder-22mm/76906 & https://www.heatmerchants.ie/1-warix-flange-u80279. I believe some tanks come with a similar fitting as part of the tank?

tnegun · 13-05-2022 9:10am

No surrey flange from what I can tell it looks like a regular take off at the top of the tank, the shower pump is fed from a second 1/2" take off slightly lower the the main one. It's actually this pipework and the location of the pump that cause me the most trouble as I need to work around them hence the 90 degree to a short horizontal at the top of the willis. I'm planning on doing the following!

Lowering the willis in relation to the tank right now the top of the willils is just slightly lower than the top of the tank.

Replace all the return 1/2" with 3/4"

Rise straight out of the willis for approx 500mm and then have the horizontal run at an angle if possible to the expansion pipe

The Continental Op · 13-05-2022 9:36am

It shouldn't make any difference if you connect a bit higher up on the vent pipe if that helps you make a more direct run. Provided the hot water can rise cold water will come in and replace it below so you don't need to connect as close as possible to the top of the tank.

The Continental Op · 13-05-2022 9:56am

Note: Slightly off topic but interesting.

I went digging around for more Willis info didn't expect much as I've done it so many times before but I did come up with an interesting highly technical report on the Willis "Solarsyphon" (not an immersion its a water to water heat exchange which installs in the same way). For anyone interested here. Its all very technical but if you look at the pictures you can see the the flow pipe is connected quite high up to the vent pipe.

tnegun · 13-05-2022 10:07am

https://www.boards.ie/discussion/comment/119055608#Comment_119055608

I'm limited how high I can go by shelving in the hot press, to give the water more room to rise I'm going to drop the willis lower down and have the turn to the horizontal at the top of that, right now the horizontal is almost immediately on top of the willis so the water is having to negotiate the horizontal first before it can rise which might be why its getting so hot.

John.G · 13-05-2022 11:12am

This is fascinating as I've never actually seen any numbers from anyone who has installed one of these, you can see from the attached the tiny circulating heads available to promote circulation, especially as the heat source is more or less at the same elevation as the HW cylinder, you can change the numbers yourself in yellow if interested.

Edit: Slight correction in Rev2.

Circulating Force in Gravity System Extract Rev2.xlsx

The Continental Op · 13-05-2022 1:27pm

Very interesting. It looks that if you double the height you double the circulatory force but how exactly is the height measured?

If you were make the flow connection from the Willis to high up on the vent pipe would that be an increase in hieght? I assume it would because I've seen some old Range (solid fuel cooker) back boiler direct connections that manage to heat tanks at heights very similar if not below the boiler by taking a high flow pipe throught the roof and return through the floor without using any pump.

John.G · 13-05-2022 2:00pm

The circulating height is easy enough if say the heater/boiler is on one lower level and the cylinder on another level, say boiler on ground floor and cylinder on the landing, the circulatig height is the distance between the centre lines of the flow and return on the boiler and cylinder. I have seen references to extending the flow pipe up then back down but don't really know if this increases the effective circulating height. I think mariposa recommend this to prevent reverse circulation and cylinder cooling. I have seen where some recommend extending the (Willis) flow pipe 0.8M above the cylinder top but this is easier said than done in lots of cases.

John.G · 13-05-2022 2:02pm

Should have said Maxipod above.

The Continental Op · 13-05-2022 2:07pm

One old cottage we lived in had a Waterford Stanley No8 with a boiler in it in the kitchen and a galvanised hwc in an airing cupboard in the bedroom. The top of the tank was about the height of the middle of the Waterford Stanley No8. The tank still managed to get very hot and even boil when the range was on all the time during the winter. Flow pipe went up into the roof space and the return was chased out into the floor.

John.G · 13-05-2022 2:13pm

My parents had a back boiler which heated the cylinder directly, both were on the same level (bases) and when you heard the cylinder giving a grunt or two you just hit it a belt of your fist and away it went.

Going back to the Willis extending 0.8M, I suppose that is up and the back down as it comes back down the 0.8M via the vent?.

tnegun · 13-05-2022 2:26pm

I think I can follow what you guys are describing 😂 So my idea of getting the willis lower and creating more of a head should work in my favour? I picked up the piping and other bits at lunch so might get to go at it this evening.

The Continental Op · 13-05-2022 3:15pm

Lowering the Willis can only improve the flow. The pipe size can do nothing to improve the circulating force - its all on the spreadsheet calculation, however increasing the pipe size should reduce the friction in the pipe allowing the best possible flow. The force moving the water is so low that anything you can do to improve it has to be a good thing

I'm getting my head around the circulating force and I'm just starting to understand John.G's comments (and excellent spreadsheet) on how small that force is. I understood the force was low but it was only when I realised that Circ Force in the spreadsheet calculator is actually the Circulating Head that I could see that 1/4 inch of head is tiny. If I'm understanding this correctly 1/4 inch of Circulating Head would equate to a pressure in PSI of 0.01?

John.G · 13-05-2022 3:29pm

Thats correct, seems quite low.

tnegun · 14-05-2022 12:36am

God I'm a sucker for punishment, got in from training this evening and had to wait for the whole house to shower and be down with the bathrooms so it was after 10 before I started! I lowered the willis in relation to the tank and replaced all of the 1/2" on the return with 3/4" I was able to get about 600mm of a rise from the top of the wills before the horizontal to the expansion and the horizontal has a nice angle too it anyway to get to the point after 20 mins of running the willis the max temp of the return was 66.6c

This is how hot it got earlier today when I let it run on the solar excess

tnegun · 14-05-2022 12:38am

Thanks for all the advice and help you've given me. I'll monitor it for a few days and report back I'd really like to see a full tank of hot (not scalding) water the best I've managed is about 3/4s from what I can tell or 6kwh solar excess absorbed. I'm hoping to get to about 8kwh on a cold tank.

John.G · 14-05-2022 6:32am

Well done, delighted that you persevered with the installation.

It may be difficult to get a completely full cylinder as the bottom of the tank will receive heat by conduction from the hot water above it as going by your trend, the circulation is 1.075LPM so once the bottom of the tank reaches 25/26C then the stat will switch out the heater, you may find ironically that you may get a a fuller cylinder in the winter as the bottom of the tank will start from a lower base temperature but of course not as much diverted energy available then.

You could consider going the final mile and replace the return Willis 1/2" piping with 3/4", even though probably a short length, it may increase circulation and knock another few degrees off the dT to allow the heater to run longer.

It will be very interesting to see what the dT will be at various diverter inputs to the heater.

Where are you monitoring "tanktemp1" ?.

tnegun · 14-05-2022 8:57am

All bar the first 20-30mm of the willis return is now 3/4" the first short section that gives me enough room to fit a 1/2"-3/4" compression coupling. Unfortunately, I don't have the diverter input over time just the sum of the power diverted in kwh. When running the willis in boost/manual it takes 2.8-3kw depending on the grid voltage. I've a 3.6kw solar inverter with a base load of 300-400w so if nothing is on in the house and the sun shines the willis should get the full 3kw. tanktemp1 is usually about halfway down the tank but I've moved it for now to just inside the insulation jacket at the top of the willis.

The Continental Op · 14-05-2022 12:22pm

One simple question on the Willis. Do you support it or just make sure the attached pipework is well supported ? Just wondering what support I need if any?

John.G · 14-05-2022 4:14pm

https://www.boards.ie/discussion/comment/119059380#Comment_119059380

Looking at that top trend (after mods) it seems that the heater was on for 5 minutes (2338 to 2343) before cutting out at 63C, it then stayed off for ~ 16 minutes (2343 to 2358) before cutting back in, so apparently on ~ 24% of the time (5/21). If I am interpreting this correctly then fine if mains boosted or if the whole cylinder is hot but in the day time with no usage and if "whole" cylinder not up to temperature then the diverter will not be utilized to its full potential?. Todays trend will be interesting.

tnegun · 14-05-2022 4:27pm

https://www.boards.ie/discussion/comment/119060994#Comment_119060994

I supported the feed pipework by placing a piece of wood underneath it as I didn't fancy having to remove the tank to fix any damage!

https://www.boards.ie/discussion/comment/119060994#Comment_119060994

Ignore that first spike I switched it off after checking for leaks to put the hotpress back together and then boosted again to test temp. This is what it has done today I was out but the kids had 2 or 3 showers today after games. Its still heating up there and the bottom of the tank is now warm to the touch as can be seen with the return now climbing to 80c

tnegun · 14-05-2022 4:28pm

I should have mentioned that the tank has taken 11kwh of excess today however there were some showers so I guess from cold should take 6-8kwh, up from 3-4kwh on the regular immersion.

John.G · 14-05-2022 4:33pm

Why is it climbing to 80C though?.

tnegun · 14-05-2022 4:44pm

I think as the supply is now warmer? When I had the willis out last night the I checked the thermostat and it only goes 2/3 of the height of the willis so it's not detecting the hotter water at the top?

John.G · 14-05-2022 5:05pm

The thermostat pocket and presumably the thermostat rod itself does not extend the full length of the element, also a electric heating element will emit the same heat per unit length (per inch etc) but all have the end of the element bent back up on itself so in effect is producing twice the energy for this portion of the element so perhaps this explains it? The rod measures the average temperature along its full length so its surprising that it's hitting 80C or more with a setting of 65C. Are you happy to let it rise to this temperature or do you intend to reduce the stat setting to operate at a lower temperature? and is it actually cutting out at 80C at all?.

tnegun · 14-05-2022 5:28pm

I'm going to try to reduce it to 50c and see what impact that has, the stat is definitely working as it was cutting out previously and the diverter shows this. The bottom of the heater is still cool to touch so I guess the flow-through is keeping the stat within its limit.

tnegun · 14-05-2022 7:09pm

The tank took a massive 13kwh today and I've now set the stat to 45c, 20c less than it was so I'm hoping that translates to a max output of 70c. Tomorrow doesn't look like a great day for solar so it might be a while before there's enough excess to heat the whole tank again and test but I'll let you know.

John.G · 14-05-2022 9:13pm

Hopefully you will get consistent switching because the circulating head at say 30C/65C is 81% of the circulating head at 15C/65C, if not, you could consider installing a pipe stat on the flow pipe and use the existing stat(s) as supervisory, I have used 16A gas filled EPH pipe stats on a few relation's solid fuel appliances and they give very consistent control with a switching hysteresis of ~ 5C.

I have a 50 year old (Santon) dual immersion and it doesn't use a rod stat, the stat is surface mounted on the immersion boss, electric showers use the same method to protect against scalding, the TCO (thermal cut out) is mounted on top of the heating can and switches off the power at exactly 48C.

Isn't it strange that this problem has never apparently reared its head previously or if so doesn't seem to be posted anywhere.

tnegun · 15-05-2022 10:02am

The pipe stat is a great idea wired in series with the existing stat right? I think I might have turned the stat too low as despite 2 showers last night some reverse circulation/convection overnight has the supply side of the willis warm and it hasn't switched on today despite there being some excess available. I couldn't find much about this online either beyond people recommending it.

John.G · 15-05-2022 11:49am

Granted that there is a temperature gradient through the heater but that stat might just be faulty, it should certainly have switched in after reducing setpoint to 46C?, easy to test when removed but probably unable to withdraw it downwards with upside down installation, pipe stat certainly worth investing in anyway for €15 or so and yes. just wire it in series, only drawback just maybe a little slow to reset despite its 5C hysteresis due its location vs the rod type which is longer and just may reset sooner but no use if not switching consistently if in fact its not faulty.

The Continental Op · 15-05-2022 12:25pm

This is interesting, sorry about the size of the picture which I pinched from Topline.ie where the script shows it a little to small to read easily.

Just wonder why 50cm max?

The inculsion of a ball valve in the feed is also interesting, would certainly facilitate changing the element if the Willis wasn't too low down, otherwise it would still help removing the Willis.

tnegun · 15-05-2022 5:36pm

https://www.boards.ie/discussion/comment/119065709#Comment_119065709

It started heating as soon as there was a shower this am as the willis short circuits the hot feed when the shower is on cold water from the bottom of the tank was drawn into it and the stat closed, I turned the stat up to 50c this afternoon and I'll leave it there for a while and see how it performs. Not least as to get to it now I have to remove the shower pump and the fact its upside down close to the floor makes adjusting even more difficult! Would there be any benefit to fitting a non-return on the return from the willis or would it just impede circulation in the correct direction? My installation is not too dissimilar to the pic above now I just need to source some insulation for the return pipework.

The Continental Op · 15-05-2022 6:09pm

I wonder if it would stop the short circuit if the return for the Willis from the cold feed was taken down low (assuming there is room for that) and then up into the Willis?

John.G has shown the Circulatory Head is so low that I doubt it would be good enough to open a valve of any form?

John.G · 15-05-2022 6:38pm

Couldn't see a NRV working, even if it did it would only prevent reverse short circuiting if this occurs as well?.

Is there much of a short circuit flow when using HW? ie to what degree does it cool down the cylinder HW outlet,

tnegun · 15-05-2022 8:19pm

With the NRV I was hoping to prevent the feed from becoming warm as this is likely prolonging the stat staying open. Right now we know the willis's stat can't measure temps at the top accurately so I've set it low to account for this but then there is no excess solar being generated the flow through the willis stops and heat moves downwards. Then I have the problem that this water is warm enough to open the stat so when excess solar returns the stat stays open due to the reverse flow of the hot water.

No noticeable change in temperature with the short-circuiting at the top of the tank but I can feel it on the feed to the willis that will be warm to the touch and after a few seconds of running a hot tap it cools right down the feed from the willis its self says hot/warm even when it's not heating but I haven't checked say after a shower when its not heating.

tnegun · 15-05-2022 8:25pm

IMG_20220515_202029.jpg

IMG_20220515_202038.jpg

IMG_20220515_202205.jpg

Just dawned on me to post some pictures of it in place note a lot of room to work with!

John.G · 15-05-2022 10:37pm

https://www.boards.ie/discussion/comment/119067905#Comment_119067905

The water within the Willis is probably mixing due to conduction etc but is hardly moving out of the cylinder top and reverse flowing to the return?, if this is correct (no actual reverse flow) then the return pipe should cool down and allow a pipe stat to reset?.

John.G · 15-05-2022 11:36pm

Should have said allow flow pipe to cool down.

O

tnegun · 15-05-2022 11:43pm

I'm not sure but it's staying warm enough to keep the stat open and the feed to the willis is becoming warm despite the bottom of the tank being cold. The willis now needs to drop to below 40c before it will start heating again and I think I observed the following today. The temp dropped enough in the willis to close the stat and it started heating with the excess solar but as the water in the willis and feed from the cylinder was still warm it quickly got to 40c and the stat opened before circulation could draw enough cold water from the bottom of the cylinder to start the cycle properly. I think I may have to go with a pipe stat rather than relying on the one in the willis. I'll monitor it for a few days though before changing anything.

John.G · 16-05-2022 9:08am

EPH Stat.

EPH Gas filled Pipe Stat.jpg

tnegun · 16-05-2022 3:02pm

Definitely needs tweaking / a pipe stat

I've added two more temperature probes yesterday. The willis cutout at 4.2kwh today into a mostly cold/empty tank. tanktemp1 is at the middle of the tank and tanktemp2 the bottom with the willis temp being taken at the top. The dip at 14:30 was me running off some hot water to attempt to jumpstart circulation it worked briefly but there was too much latent heat and the stat opened again. I double-checked and the stat on the willis is approx 48c its hard to be precise as it's upside down with difficult access! This probably explains climbing above 70c.

tnegun · 16-05-2022 3:03pm

I'm thinking tomorrow I'll let the cylinder immersion go first and then the willis second and see what impact it has on temps.

tnegun · 17-05-2022 7:12pm

It was a crap day for comparisons plus I'd use the gas to heat water this am. I've noticed this a few times now where the willis fills with cold water when hot water is drawn off(shower) its short circuiting but not notably so. 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?

John.G · 17-05-2022 7:50pm

Can you just post what each colour is (again) please.

Fitting a Willis Immersion heater

Comments