Thanks again @Transparent; that is very clear and I see the advantage of the String Combiner box for allocating where best to direct the PV input. You have an impressive spread of panels!

Did you design and assemble the String Combiner box yourself, or is this something that can be obtained ready made? There are so many variables that I suspect each one would need to be tailored to the application.

I have a number of other questions but will suppress those for now while you develop the diagram.

@dnshorto 

Yes, that’s my own design/assembly of a String Combiner.

You can buy them off-the-shelf, but even when sourced straight from a Chinese supplier they are not well thought out, and still pretty expensive for what’s included.

The one I’ve picked to show here has a pair of fuses for each string array, but only one overall mcb/isolator.

Despite the label, the fuses are not ‘Circuit breakers’. You can’t disconnect a ‘live array’ by opening the fuse-drawer because you’ll get a huge spark and risk welding the fuse to its contacts!

And if you look carefully, the Supplier has incorrectly labelled the 2-pole isolator as ‘Lightning protection device’, which doesn’t inspire confidence!

I bought the bits I wanted from (Chinese and German) component suppliers and assembled them into a DIN-rail enclosure myself.

We can return to this topic later. At the moment we haven’t even concluded that you’re going to need the ability to switch where PV arrays are sending their energy!

You are quite correct @Transparent, this thread opens up so many interesting peripheral topics, but I must stick to the main subject and explore those later. 

The present set-up works so well, say April to September that I would be reluctant to change it during those months; the meter hardly moves!

In these 6 winter months, it’s obviously useful to collect what solar energy there is available but the reduction in grid input is much less valuable and provision for potential power cuts becomes more of a priority. The battery rarely gets a full charge and is quickly depleted in the evening so is far less useful in the winter months.

Setting up a separate Off-grid inverter to run a few essential circuits (CH boiler, fridge/freezer etc) sound a perfectly workable idea but is there a way to use the same battery for both purposes, or maybe alternate the battery’s function every 6 months?

@dnshorto 

And I’ll now be able to pick up that last point @dnshorto by presenting the 3rd iteration of the diagram, which has now considerably grown in complexity:

The appliances and devices to the right are illustrative of the technologies which can be operated off-grid.

Some off-grid equipment must use the Inverter to generate 240v with a pure 50Hz sine-wave. These predominately have pumps/motors which rely on the mains frequency for their (smooth) operation. Under no circumstances should you attempt to run these using the much cheaper pseudo-sinewave inverters.

The rest of the off-grid devices can be run directly from the battery. After all, they would normally have a cheap-and-nasty transformer plug to power them using DC derived from the mains!

Many of these devices are likely to require a small DC-DC converter to create the required voltage. We’ll return to that later in the discussion.

Here’s a table showing appliances which are most likely to be operated using the three categories of power now available in the home:

Once again, these are illustrative. You may have your own ideas either because your priorities differ or possibly because you can’t face the technological barriers to achieve that mode of operation.

I’m well aware that we will need to discuss further how and why I’m able to put LED-lighting in the off-grid category. It’s a comprehensive subject in its own right and I’m tempted to start a completely separate topic to address this.

What do you think @Tim_OVO and @Jess_OVO ?

All perfectly clear, thank you @Transparent.

I am very familiar with low-voltage DC lighting systems. More than 40 years ago, long before LED became so cheap and efficient, I installed 12v back-up lighting throughout my whole house during the 1979 “Winter of Discontent” when power cuts were frequent, and retained the system for decades afterwards as “night-lights”. 

Your diagram and tables are fully understandable and as drawn would work on a year-round basis, so that the off-grid, pure sine wave-dependant appliances would always use that inverter. 

Is my idea for a 6 monthly changeover workable, and how could it be accomplished? 

In the 6 sunny months my system would remain as installed; the battery supplying power when the sun wasn’t and excess solar energy passed on to the grid.

In the 6 darker months, when there was little spare battery power available for evening domestic loads, the battery would be isolated from the on-grid inverter and would be kept topped up by occasional sunshine, with the option of manually charging from the grid when needed. It would then only power the CH boiler and the fridge via the off-grid inverter, plus optional DC circuits as outlined by you. 

Is this achievable?

@dnshorto 

So great to see such valuable knowledge sharing going on here - @Transparent and @dnshorto.

Can’t say I’m 100% following, but I’m sure those researching their solar storage options will find this thread a great help.

We’re always keen for a separate more detailed guide on any energy subject (but off-grid lighting does sound like a winner!), @Transparent - let us know if you’d like us to move any existing comments over or you’d rather start the topic afresh...

Just slightly amended this topic title and added some tags in the hope this will help others find it.

What do you think of making this more of a discussion topic than a question with one single best answer? (Feel like  all of your comments deserve equal consideration here, @Transparent and @dnshorto)

OK with me @Jess_OVO . Whatever is most useful to most people. @dnshorto 

So this brings us to the next step:

How to select the required inverter.

Each solar-inverter specification shows the voltage and current range of the PV panels which can be connected to it. Here’s the relevant portion of the spec for the Growatt SPFx000TL range which featured in the earlier YouTube video

Looking at the right side of the table for an inverter connected to a 48v Li-ion battery, there are three parameters which we need to note:

1. The inverter can only use its Max Power-Point Tracking (MPPT) electronics whilst the operational voltage range is between 60-115vDC.
2. The maximum permitted (open circuit) voltage must not exceed 145v when I first connect those panels, otherwise the inverter may be damaged.
3. The maximum current which the panels can deliver in sunlight is 30A

Those three parameters dictate which solar panels can be used and in what arrangement of series and parallel.

For example, here’s the specification for the high-end Panasonic HIT panels I have on my main (upper) roof area, compared with a more common Longi panel of similar overall power:

So if I were to put two of my Panasonic panels in series (71.3 + 71.3v) they are only just within the maximum 145v permitted for connection to the Growatt SPF-series inverters. But their peak current of 5.7A falls a long way short of the 30A allowed.

If I’m going to install solar panels which can be switched between a grid-connected inverter in summer and an off-grid one in winter, then I need to start by choosing an arrangement which gives me the required flexibility.

The above arrangement shows three solar arrays. A & B will be permanently linked to the two MPPT inputs of a grid-connected inverter. Array-C can be switched between an off-grid inverter or added in series to Array-B.

The alternative is obviously to keep two completely different sets of solar panels for grid-connected and off-grid use. That depends on available roof-space and how much you’re prepared to spend.

Thanks @Transparent. I have to confess that I struggled at first to comprehend your latest post, but repeated readings helped and I now see generally what you mean.

For these chats to be relevant to other enquirers in I understand the need to keep the information general, but of course I also value help with my specific installation. For comparison with the table of specs for the Growatt inverter you mentioned, here is the plate on the side of mine:

How does this compare?

You also mention the possibility of dividing up the output of the solar panels, or adding more. That isn’t an option for me as all my south-facing roof area is taken up with the 28 existing panels, but they are apparently in 4 strings of seven, even though there are 6 cables coming down from the roof, which is a puzzle.

Any personalised advice you are able to offer specific to my particular installation would be greatly appreciated, including a suitable inverter and how I might redirect the available solar input to achieve the off-grid power I’m seeking.

I hope that this too might be useful to others with similar problems.

@dnshorto 

Yes I think this topic has now reached the point where we need to consider details of your own installation @dnshorto 

I’ll assume at this stage that you have no approved Load Diverter or export limiter at your house. That may not be the case because your initial post for this topic said:

On any reasonably bright day d….] the HW cistern is heated

Leaving that point aside for the moment, let me just dissect your diagram (above) and show the current/power of these different elements, based on a nominal mains voltage of 230v AC.

To have both inverters grid-connected requires the installer to have applied to your Distribution Network Operator (DNO) for a G99 certificate (see explanation here).

The default situation is that a household may not export to the Distribution Grid more than 16A per phase, which equates to 3.68kW.

The G99 export assessment process adds together your export/generation devices on the assumption that solar-output and battery export is simultaneous. So you should have a G99 certificate which grants permission for 8kWh + 3kWh. That is considerably in excess of the ‘standard’.

You put your finger on a long-standing problem there @Transparent . Last year you asked if I had a G99 certificate. I asked the Solar Panel company, who assured me that they had the certificate and would forward it to me. I naively believed them.

Many months later, after repeatedly but unsuccessfully requesting the G99 I learn that they have closed up shop. Thankfully a company called Flexi-Orb have picked up the cudgels for me as I have been in and out of hospital rather a lot recently and they seem to be making progress with the DNO, so I hope to have it soon, albeit a year late. The surveyor has apparently already checked the situation and sees no problem in granting the certificate.

There is no “load diverter” or export limiter, but a Solic 200 Solar Immersion Heater Controller diverts power to the water heater when there is enough PV energy to spare (a solution which I remember you described as slightly “selfish” in a community setting)!

@dnshorto 

I think that Solic 200 Solar Immersion Heater Controller falls within the group of devices which we’re calling Load Diverters here on the Forum.

At the moment the DNOs only allow Load Diverters which send energy to immersion heaters. We need to make representations to the Electricity Networks Association (ENA) requesting that Storage Batteries can similarly be tested and approved to act as diversionary loads.

Tell us the size of your immersion heater, which I guess is going to be the usual 3kW.

The Load Diverter should be referred to in the G99 documentation. Even so, you’re still above the normally permitted export capacity.

8Kw (solar) + 3kW (Pylontech batteries) - immersion-heater

Quite correct @Transparent, 3kW immersion heater, but of course only drawing power while the water is cooler than the thermostat setting, so on a sunny day the load diversion is intermittent.

On the vexed topic of my G99 I finally received the relevant paperwork this morning, so my installation is officially recognised and approved, even though it has been sending energy into the grid for over a year.

However, the immersion heater diverter was fitted only a few months ago, so is not referred to on the G99, an omission which I will now take steps to correct, thank you.

@dnshorto 

Let’s take this analysis of @dnshorto ‘s installation a bit further.

It is a reasonable proposal to have one or more of the PV strings switched from the existing Solis inverter to another off-grid one, especially during winter when you’re less likely to want grid-export.

I don’t know the exact specification of the 340w panels in the diagram, so let’s make a sensible guess that their peak output is  34v @ 10A.

A series string of seven can therefore supply 238v @ 10A max.

It shouldn’t be too difficult to find an off-grid hybrid inverter with an MPPT input which could take Solar Panel export at that level. It would probably need to be rated at 350v minimum in order to allow for the higher ‘open-circuit’ voltage when the panels are first connected.

This would be easier to achieve if the four rooftop arrays arrived in a String Combiner box. But the original photos at the top of this thread indicate that they enter a relatively small metal box, just 4-modules wide.

That tiny enclosure doesn’t have room for a pair of fuses and a double-pole trip for each of the four PV strings. Whatever is in that box won’t allow a string to be isolated from the others in the event of a single panel being faulty.

I’m sure that can be changed to a more flexible arrangement. But Dave mentions there being only six wires (not eight), so I’m guessing that 2 strings share a -ve cable between them.

If so, then it would be necessary to switch a pair of PV strings from the Solis to an off-grid inverter rather than just one at a time.

The video clip shows a double-pole change-over trip in my own String Combiner box during construction. The mechanism permits only one pair of contacts to be ‘on’ at a time.

It’s important to note that this is a DC-trip. Each of the four sections has a spark-quenching ‘ladder’ to prevent the switching-arc from welding the contacts together! For that reason the terminals are marked + and - so that the arc gets drawn into the quencher by a magnet.

Polarity is important when switching high-voltage DC.

That is very much appreciated @Transparent , thank you. The video clip makes that switchover sequence very clear.

Referring back to my own installation, the situation is actually worse than is at first evident. What looks like a tiny junction box is in fact a bundle of cables.

The interior of the box they are routed into (P V System. Main AC Isolator) 

is as below:

I would go along with your recommendation to fit instead a String Combiner box to allow flexibility and to intelligently divide up the incoming PV energy. I can make room for it and the off-grid inverter to the right of the present clutter of items.

Presumably, if this String Combiner box were to be designed for the purpose, a greater or lesser number of strings could be diverted into the off-grid inverter depending upon seasonal PV generation, domestic load and the likelihood of imminent power cuts! The inverter would be chosen to deal with a maximum input from say 3 of the 4 strings. Is this feasible?

Would the battery also be independently divisible between the existing 4 units?

Once again, many thanks for a very helpful discussion.

Dave

@dnshorto 

Thanks for those close-up photos @dnshorto 

I no longer believe that the small metal 4-module enclosure is the one where the panel cables enter. I can see a ‘mains’ earth wire, and the trips are clearly AC rated.

I think the red 100A ‘Switch Disconnector’ isolates all of the new installation for both inverters. The two blue MCBs are most likely one for each inverter.

The mandatory safety label is therefore correct.

On closer examination of your circuit diagram, you actually have three AC breakers, which I’m now labeling AC1, AC2 & AC3

If I’m right, then AC1 is the rotary one beneath the Solis. This style is specified within the MCS accreditation for approved solar panel installations.

As I can’t see another rotary switch, I’m inclined to believe that the 4-module enclosure is what the diagram refers to as AC2 and AC3.

The cables entering the Solis inverter at bottom left have MC4 connectors, which you’d only find on the single-conductor cables used for solar panels.

That means we still haven’t located the DC isolator referred to on the diagram. Is it possibly in the attic rather than the garage?

I rather think @Transparent that some of the breakers and isolators shown on the schematic diagram may be mythical. The drawing was done 12 months after the installation by a director of the sub-contractors who hadn’t visited site, and probably represents an ideal layout rather than one based on reality. Certainly, there is no wiring in the loft, everything is as photographed in the garage.

The installation was done in two phases. On 26 Oct 2020 26 JA Solar 345 panels were installed, and 3 Pylontech 3.5kWh batteries. Then three weeks later a further 2 identical panels were installed (space for them having been achieved by the removal of an unused cement flue pipe), and a fourth battery once I had observed how the panels easily charged the first 3 batterie. The enthusiastic but technically challenged salesman had warned that 2 batteries were probably all that the panels would keep topped up, whereas I wanted maximum storage capacity for future power cuts! :-)

So the inaccuracies of the schematic caused by lack of direct knowledge makes me question the legend: “4 strings of 7 panels”. There are 6 identical cables coming down from the roof and they all feed directly into the bottom of the Solis inverter as you have arrowed in blue. I don’t think the DC isolator exists. Is this a serious omission? 

I guess if a String Combiner box is to be installed the 6 incoming MC4 connected DC cables will have to be rerouted through it so siting of that box will need some thought. 

 Your wisdom in all this is very useful, thanks.

 Dave 

 @dnshorto 

OK, so I think we can assume that the circuit schematic you have isn’t entirely correct.

To better understand what the real situation is, I’ve started by downloading the Installation Manual for your Solis 5G-8k PV inverter from the Chinese manufacturer here. Here’s part of the specification table towards the end:

This tells us the maximum parameters for the panels which can be attached. It’s slightly complicated by the fact that there are three pairs of input sockets, but only two MPPT (max power-point tracking) systems.

As I currently understand this, it means that MPPT system-1 can take 520v at 12.5A in peak sunlight.

MPPT system-2 can take two strings of panels (in parallel) totalling 520v at 25A max.

You have 28 panels, which we believe to be rated 340w each (total 9.5kWh). We can download the specification sheet from the Indian manufacturer, JA Solar. These are the maximum parameters for a new unit:

This shows us that the maximum number of panels in series would be 14 if you are to remain within the input voltage range of the Solis inverter (14 x 34.63 = 484.82v).

So my first guess is that you have a series-string of 14 panels connected to MPPT-1, and two series-strings of 7 each connected to the pair of inputs to MPPT-2.

How does this sound to you?

It sounds spot on @Transparent. I should have looked under the Solis inverter before, as we may have found the missing DC Isolator too. 

There are 4 DC cables from the panels connected to the 4 terminals marked DC2, and 2 cables connected to the 2 terminals of DC1.

I think you are quite correct. The system has been working well for over a year, so if the installers had overloaded DC1 by connecting more than 14 panels we would have had trouble before now.

@dnshorto 

Well I don’t want to jump to too many conclusions about the conguration of your Solar Panels yet. The three pairs of downleads could equally well be connected to arrays of 10, 9 & 9 and still be within the specifications.

The roof layout does indeed suggest 14, 7 & 7. But it also depends on how those last two Panels were connected in when the flue was removed.

Nevertheless, let me present the following diagram as a ‘working example’:

The left-side of the diagram shows the three strings in a 14, 7 & 7 arrangement.

I like using unique colours for each separate array. I mark the downleads and trips to match. Purple, orange, pink and light-blue are the least common colours in PVC electrical tape, but available separately on ebay. So I won’t get these DC leads confused with anything else in the house installed by a qualified electrician!

To right of the panels is the basic layout of fuses and trips which I would expect to have installed. This allows strings with faulty panels to be isolated and some measure of protection against surges.

Yes, I’m suggesting both fuses and MCB trips. You shouldn’t be opening the fuse-carrier on a ‘live’ circuit. That’s what the manual lever on a trip is for!

We here have basis for a design of a String Combiner box, albeit without any two-way switching or lightning protection at the moment.

To the far right is the DC isolator. Yours is black and not easily accessible beneath the Solis inverter. It’s a safety device to switch off all electrical input. As such it should be readily identifiable… by a fireman for example.

I’ve drawn a 6-pole rotary switch which needs contacts rated 20A min.

The one on the inverter may be a more common 4-pole version if Solis have decided that this is the point where the two strings connected to MPPT2 get placed into parallel.

(As you know, firemen are more familiar with a single pole)

I’m hoping that this gradual series of steps in the analysis will allow others to follow the discussion in future.

The litmus test for this is whether @Tim_OVO and @Jess_OVO are still keeping up!

Your step-by-step approach @Transparent is clearly the most likely to be useful to most people and I am finding it very easy to follow, thank you. I’m sure the admin team will too if they have time to read it. Other users will have allied or similar questions in the future and will find this discussion helpful, even though it is specific to a particular installation.

To establish whether the 3 strings are actually arranged as 14, 7, 7, would it be feasible to measure the incoming voltage on each pair? I was thinking that the MC4s could be disconnected after nightfall and then measured with a multimeter early the next day.

I like the idea of clearly differentiating the separate circuits with unique colours and wiring, and marking the switchgear etc with matching tape. I also favour fitting a 20A 6 pole rotary switch directly on the incoming DC leads.           In practical terms this could be mounted next to the Solis inverter without the need to extend the 6 leads.

From this new switch the appropriately identified leads can pass into the string combiner box, though the lightning protection etc and on to the two-way switching elements of the box. I revisited the earlier page where you included a photo of the box you designed and assembled and am of the opinion that I could probably tackle the sourcing of bits and putting them together but would need assistance with getting the design and components correct. I have used RS at odd times in the past for not dissimilar projects and you may well have found other suppliers to be useful.

We haven’t yet discussed the possibility of also dividing the battery modules but am confident that the step-by-step approach will get there eventually!

Thanks again

@dnshorto 

Yes, I would certainly start by disconnecting the MC4s and measuring the voltages. Just remember to mark the cables unambiguously before you start disconnecting.

MC4 connectors are polarised so there is no danger of reconnecting a pair the wrong way around. But you could inadvertently get two negative leads the wrong way around, for example. Each positive and negative must remain as a pair.

I don’t think you need to wait overnight to do this test. Just choose a time of day that’s not full bright sunshine.

If you turn the DC isolator to ‘off’ then no current is flowing on any of the six cables, so you could disconnect and test a pair at a time for its voltage.

Remember that you are measuring this ‘open circuit’, without any load being imposed by the MPPT electronics in the inverter. Open-circuit voltage is specified at 41.36v per panel, so even a 7-panel string would really hurt!

DC voltages are actually more dangerous to us than AC ones because the contact ‘sticks’ to the skin when it zapps you.

I will, of course identify where you can buy components to put together a String Combiner. I buy most of these on AliExpress direct from China. They simply clip onto a standard 35mm DIN rail, which most often comes fitted within a box, but can be bought by the metre.

The enclosure itself needs to be as large a DIN-rail box as you can fit somewhere on the wall! Due to its size you want to source this closer to the UK.

It doesn’t need to be made of steel, like the mains consumer units must be. That means there’s still a possibility of picking up a plastic one which a distributor has got left on the shelf.

Start looking in the Clearance Bargains of suppliers such as CPC, Rapid Online, and even ebay. (I’ll need to change that last link to something more generic later).

@Jess_OVO - I’m aware that there is a lot of practical content emerging here which others might want to find if they’re reading the general topic on PV Panel Installation.

Let us know if you want things moved, or whether we periodically insert links to direct people here.

The largest enclosure I can site @Transparent  in a reasonably accessible position would be 600 high, 400 wide X 200 deep, with a hinged door. There are a few plastic examples available in the UK but generally just with a steel backing plate for attaching the rails, rather that with DIN rails incorporated. More flexibility of layout that way and easier to fit non-compatible items. I am assuming that most or all switchable items would only be operated inside the box, rather than by external switchgear.

Looking again at the rotary breaker you have labelled AC1;   by pulling aside the 6 DC cables laid across its top I can see the AC output from the Solis inverter just above passing into AC1 on the left, but cannot see where the R/H cables go

Would they be routed up to the Growatt inverter and thence to the white “Main AC isolator” box above? I don’t yet fully understand the interaction or relationship between the 2 inverters. I know my system was too much for a single hybrid, and believe that the Growatt is to send and receive 48(ish) volt DC power to and from the battery and pass 230 (ish) volt AC on to the domestic consumer units and then the grid, while the Solis receives the DC from the panels and converts it to 230v AC.  But is all the Solis output channelled through the Growatt? On a sunny day the Solis becomes almost uncomfortably warm to the touch and on really bright days it hums. The Growatt seems to stay cooler.

@dnshorto 

I know the sort of enclosure you’re thinking of @dnshorto - I use them for 24v distribution boards in my house.

For your purposes you really need the trips and the lightning suppressors to be visible. If there’s a fault you should be able to see it without opening a door and then having direct access to all the high-voltage wiring!

I did a Google Image search for “DIN rail Enclosure” and found these two:

Such a format would enable you to dedicate a row to each of the three arrays.

The CEF one is 500 x 355 x 140mm  (H x W x D) and can accommodate 16 modules per row. CEF is a nationwide electrical supplier.

The Thompson/ebay one is 470 x 307 x 100 and holds 12 modules/row.

A DIN-rail module is 18mm wide.

You need four modules-width to accept each panel-array incomer (2 fuses and a 2-pole MCB trip), plus lightning suppression. A double-pole changeover switch occupies 4 modules. So the space soon goes!