Friends, back at the start of September, I noted that it had been a sunny summer and I resolved to add more solar panels to the house in order to increase the solar harvest next year.
I ordered the system just a few days after writing that article and it is now being installed.
In this article I thought I would describe the new installation and how it will (hopefully) integrate with the existing installation.
Click on image for a larger version. The arrangement of the solar cells on the roof of Podesta Towers. The grey panels have been installed for two years, and the red panels were installed this week. I had hoped to fit four panels on the flat roof, but in fact I can only fit three.
The Existing Installation
The existing system was installed back in November 2020 and consists of:
- 12 Q-CELLS 340W DUO G8 panels
- A SOLIS 3.6KW 4G dual MPPT inverter
Why did I select these items? The installer recommended them and they seemed to have adequate performance. And happily, they do seem to have worked OK.
Some key features of these items are:
- 340 watts is the nominal output of a panel illuminated perfectly by sunlight with an intensity 1000 W/m^2 – this is roughly full sunlight on a UK summer day.
- Since the panels are 1.7 m x 1.03 m one can work out that around 20% of the solar energy is converted to electrical power.
- The panel is constructed as two half-panels wired in parallel, each with 60 individual solar cells.
- A silicon solar cell generates around 0.6 V so the 60 cells on a half-panel together generate around 36 V.
- Splitting the panel like this improves the panel performance when one half of the panel is shaded.
- The MPPT acronym stands for Maximum Power Point Transfer and is system for extracting maximum power from solar panels as the intensity of illumination changes.
The quotation suggested that I might reasonably expect 3,780 kWh of generation each year and this year we look on track to exceed that. Last year we generated only 3,517 kWh.
Click on image for a larger version. Cumulative generation from the existing solar panels in 2021 and 2022. The dotted blue lines are based on the expected output according the installer’s initial calculation.
This first installation was done quickly to take advantage of the fact we had scaffolding around the house for the external wall insulation. Because of this, we couldn’t wait six weeks for permission for a larger installation from the local Distribution Network Operator (DNO): these are the people who manage the local electricity networks.
So I opted for a standard installation (for which no permission is required) with a maximum output of 3.6 kW peak and we used the best sites available. I resolved to learn what I could about solar power, and after two years, I feel I served my apprenticeship.
The New Installation
To move beyond the standard system one needs to apply to the DNO, a process that takes about 6 weeks and which was thankfully handled on my behalf by the installer.
My aim was to get as much solar PV on the roof as I could – while not making the house look horrible! For that reason, we avoided using a patchwork of panels across the roof – sacrificing some performance for aesthetics.
Since the best sites had been taken by the first installation – I simply went with what was available.
I had noticed during the summer that in the mornings the Sun rises well north of east, and the east-facing roof of Podesta Towers was in full sun up until solar midday. Similarly, the flat roof was more or less un-shadowed over the same period.
My performance calculations using the excellent Easy PV site were very similar to the suggested performance from the installer.
- The 5 panels on the east-facing roof will hopefully generate ~1,300 kWh/year
- The panels are tilted at ~ 40° and face roughly ~ 20° north of east.
- The 3 panels on the flat roof – might generate ~900 kWh/year
- The panels are tilted at ~ 12° but face roughly 20° east of south –
This would correspond to 2,200 kWh/year, an additional 60% of generation bringing the total close to 6,000 kWh/year. If actual performance gets anywhere close to this I would be delighted.
To put these figures in perspective, we can compare them with household consumption.
- Last year the house used ~5,400 kWh –
- Roughly 3500 kWh of that (~65%) was for day-to-day household ‘stuff’
- Roughly 1,900 kWh of that (~35%) was used for the heat pump.
- The heat pump operated with an average COP of 3.6 to deliver 6,800 kWh of heat.
So the enlarged system will hopefully generate more electricity than we use in a year. Sadly the peak of generation (in May or June) is quite out of phase with the peak of demand (in January or February). But nonetheless, it’s a milestone of sorts.
The new system consists of:
- 8 JA SOLAR 390W MONO solar panels.
- A SOLIS SOL-3.6-S6-DT-DC inverter.
Again, I just accepted the installer’s recommended suggestions.
The Panels.
The new panels are similar to the old ones: the 390 W nominal peak output of the new panels is larger than the 340 W peak of the previous panels simply because the new panels are larger. The efficiency remains around 20%.
Each panel consists of two half-panels, each with 9 rows of 6 rectangular half-cells.
Click on image for a larger version.
When illuminated, each individual cell generates a voltage between 0.5 V and 0.7 V between the top of the cell (the part you can see) and the bottom of the cell (that is at the back of the panel).
Fine aluminium wires cover the top of the cell to collect the generated electrons, and the wires then connect the top of one cell to the underside of the neighbouring cell so that their generated voltages add together. In each half-panel, 54 cells in series generate a voltage ~ 36 V at a current of roughly 5 amps.
Click on image for a larger version. Top: Illustration of the way in which sunlight generates a voltage between the bottom and the top (illuminated) surface of the cell. Right: The fine wires collect electrons generated from within the silicon. The filigree wiring pattern is optimised to collect as many photo-electrons as possible, while not blocking the sunlight. Left: Details of the wiring showing the top surface of the lower scale is connected to the underside of the neighbouring cell.
The two half panels are wired together in parallel so that the peak output of the whole panel is ~ 36 V at a current of roughly 10 amps.
Panels which are similarly illuminated are wired in series in a so-called ‘string’. In this installation, the 5 panels on the east-facing roof are wired in one string and the 3 panels on the flat roof are wired in another.
The inverter design has two independent inputs and the DC currents from the two ‘strings’ are combined to create an AC current at 220 V.
This arrangement works excellently when all cells in a panel and all panels in a string are illuminated similarly. But if one cell in a panel is shaded, then not only does that cell not generate a current, its electrical resistance increases dramatically, and this can restrict the current which is able to flow through the whole string of which it is a part. Fortunately, clever electrical tricks can minimise the shading problem as explained in this excellent video.
Peak Power.
One aspect of the installation which concerns me is whether all the electrical circuits can cope with the sheer amount of power this system might generate.
To estimate this, I downloaded generation data from 22 June this year, a day which was nearly perfect for solar generation: close to the solstice and almost completely cloudless. This data is shown in red on the graph below.
I then made guesstimates of the generation from the two new strings:
- I guessed the 5-panels on the east-facing roof would begin generating earlier in the day and reach maximum power (5 x 390 W = 1,950 W) just before solar noon (1:00 p.m. BST). This is shown as a green dotted line.
- I guessed the 3-panels on the flat roof would generate roughly symmetrically around solar noon (1:00 p.m. BST) with a maximum power of 3 x 390 W = 1,170 W. This is shown as a blue dotted line.
Click on image for a larger version. Graph comparing a perfect midsummer generating day with the existing system (red curve) with the likely generation from the expanded system (purple curve). See text for details.
Altogether (dotted purple line) the total power could potentially exceed 5 kW – a worryingly high power level.
Summary.
Friends, as usual, I have gone on for too long. But this is a significant – and possibly final – step in the house refurbishment.
It offers the possibility of being off-grid for 6 months a year and of generating more electricity than the household consumes (averaged over a year). I think these are significant upgrades.
The cost is not completely clear yet, but looks like it will be just under £5,000. This is more than the initial system (£4,200 in November 2020) but this seems reasonable given the extra scaffolding required.
As I write, the panels are installed but the internal electrical wiring is not complete – but hopefully that will be done soon!
And if you have read this far, thank you! Please allow me to reward you with a video of the installation.
Tags: Solar PV
Protons for Breakfast 
October 28, 2022 at 7:39 am |
Amazing blog. Thanks so much.
There’s one element missing. Have you investigated having the panels tracking the sun?
I have a flat roof and 8 panels but I know I’m missing a lot of power just with a fixed angle. I’ve thought of having two settings. A support strut for summer and one for winter.
Really I’d like a tracking mount.
October 28, 2022 at 5:59 pm |
Nigel,
Thank you for your kind words.
Regarding tracking, it’s a really interesting optimisation problem, but I think it is a very difficult problem. I categorise it along with the idea of non-focussing optics – such as the use of a mirror at the back of a panel to reflect light back.
The fact that there are almost no tracking solar systems around tells me that they are unlikely to make financial sense – you and I can’t be the first people to have thought of the idea! But if you have an easily accessible flat roof then your suggestion of changing the orientation or tilt of the panel once or twice a year might help. I guess you are thinking of having panels at a high angle in winter and low angle in summer. My guess is that – since winter generation is poor in any case – that it won’t make much difference – maybe 10%?
The problem with tracking mounts is that they cast a shadow and so arrays of trackers cannot be very close together. And similarly with non-focussing mirrors. But I think that a single mirror at the back of the panel furthest from the Sun must boost production somewhat.
In any case: good luck with your endeavours!
M
November 6, 2022 at 1:09 pm |
Hi Michael I subscribe to your blog. Ihave no knowledge about solar power and do not really understand the detail you go into on your blog but I consider it must be educating me at some level as I have taken the leap to have solar panels fitted. I am seeking your advice if you can give me any on how to deal with someting very cosmetic, the cabling for the battery.
The other day the installation was due to happen but I had a problem with how the electricians, working for the installation company, proposed to run the cable from the consumer unit to the loft. This was to trunk the cable across my living room wall to the outside of the building. Oh no! I do not want a piece of plastic crossing my living room wall to come into vision every time I use the room!
The electricians said it was unclear if taking up the landing floor and stair/s would give an alternative route for running the cable from/to the loft and I needed to discuss with the company as chiselling out the wall to hide the cable had not been priced in and it would be major upheaval for me – which it would (and probably take too long in relation to the schedule they seemed to be on.)
I apologise if you might have covered this issue in your posts, but I would be most grateful to know how you managed this bit of installation in your house and for any advice you may be able to offer. I am in a terraced house unfortunately!
Kind regards Diane
>
November 6, 2022 at 2:21 pm |
Diane
Hi. This is a real problem which I am facing myself at the moment.
I have just had extra sola panels installed and the DC cable route from the panels to the inverter (in the loft) is pretty straightforward. But the (mains AC) cable run from the loft down through the house is problematic.
Our first set of panels were fitted while we were having external wall insulation applied and so it was easy to put the cables down the outside of the house and then cover them with insulation. They then came into the house in the porch and under the flooring to the consumer unit.
We could do that again with the new wiring but that would then require repair of the external wall insulation which I think would look bad. So we are having to find a new route through the house.
What I did was to take photographs of the route from the inverter in the loft to the porch and draw on where I wanted teh wires to run. It’s not the easiest route, but it’s the best I can think of. I then sent this to the electricians in advance of their visit (in 10 days time) hopefully that will allow me to get what I want – unless it’s impossible for some reason.
I’ve deleted your e-mail from your post but will contact you privately to send the document.
It is very stressful and I sympathise – but I hope you will feel better about it when the sun starts shining again!
Best wishes
Michael
December 13, 2022 at 6:02 pm |
Very interesting post. I have a 3.5KWp system on a south facing roof and have just collected a complete set of daily data for consumption and generation over 2021/22. I generated enough electricity for the house demand on 89 days of the year. My lack of battery meant I still had to buy electricity on those days, a 30kwh LFP battery is going in next year together with an extra 13KWp of ground mounted panels in a field at the rear of my garden. By scaling the new plant from last years daily performance I think I should get to only needing input of electricity for 100-120 days a year.
As an ex refrigeration engineer I would like to go down the heat-pump route but all the time COP’s are 3 or below during the cold months and electricity is over 3 times the price of gas per KWh it makes more economic sense to burn gas.
December 13, 2022 at 7:54 pm |
Robert, Wow! A 30 kWh battery is quite a monster. If those extra panels have good exposure that I think your estimate looks about right.
Two comments.
1. Regarding the battery, is it AC-coupled or DC-coupled? I don’t know if this generally true but following Nichols Howell on Youtube (https://www.youtube.com/@nicholashowell ) he can’t quite completely ‘off-grid’ because the battery management system he has is slow to kick in when household loads switch on and off. This means he consumes perhaps 5% of electricity from the grid even on days when he is nominally “off grid”.
As long as your system is grid-tied you can never get to zero consumption but with the best systems it is possible to get below 1%.
2. Regarding Heat pumps, in the recent cold weather COP has dropped below 3
But these cold days are relatively rare. Most of the winter I am buying cheap rate electricity and running the heat pump from it. It is *dramatically* cheaper than gas. On my most recent bill, average cost of a unit of electricity was 12.4p/kWh of electricity. So the cost of a kWh of heat is less than 12.4/3 ~ 4p/kWh of heating. Dramatically cheaper than gas.
All the best
Michael
December 14, 2022 at 4:16 pm |
Hi Michael,
The batteries are going to be Seplos Mason enclosures with 16x280AH Eve LiFePo4 cells which Costs are looking at under £400 per KWh useable including the Inverters.
Looking at Nicholas Howell’s setup he has left his house mainly connected to the grid and only made some circuits off-grid presumably because he is a bit down on battery power to serve the whole house and his inverter can’t cope with start-up loads without using the grid. I looked at this approach but thought it was cleaner to go completely “off grid” for the house and to have the added benefit of a whole house UPS capable of 2 to 3 days running if rationed in the event of supply outages. I intend to be able to switch the house consumer unit between grid and off grid for maintenance and resilience.
The existing Sunnyboy grid connected inverter will be used to supply the immersion heater in the hws tank with surplus power being diverted to battery charging the new system. This should not cause any changeover issues as the main circuits are completely off grid. The Victron Inverter systems have good management and control and the inverters can supply 2xrated output for motor starting and inrush.
My aim with the oversized array is to pad out the spring and autumn performance and to negate shading problems with my old array at low sun angles.
Apart from Bitcoin mining I can’t think of a sensible use for the excess power during the summer peak so the old system will export to the grid and the new system will just dump.
I have also got a used 2KWp wind turbine to try out which may help a bit in the winter, but on-shore small wind is very intermittent so I’m not expecting great things but every little bit helps.
To cope with the winter shortfall I will use a variable rate tariff and buy electricity to charge the battery as you do.
Could you point me at any of your blog posts that describe your insulation and heat-pump projects.
December 14, 2022 at 11:48 pm |
Robert,
Wow! Your project sounds awesome.
Two points.
Regarding blog articles. This one is a talk I gave to Friends of the Earth in February 2022 about the whole project.
This one describes how the EWI works.
For heat pumps, try this recent article.
In general just use the ‘Search Box’ it’s pretty good. It’s how I find stuff I wrote ages ago!
Regarding your concept of ‘over’ generation, I think it is called ‘superpower’. Build 2 to 3 times the required renewable generation and add storage in order to have a renewable resource with high availability and low down time. And then figure our what to do with the excess energy – there could be many MWh – I am sure you will think of something!
All the best
Michael