Update: 4 October, Planning the panel installation.  I laid 4 panels on the floor, and used the center and end clamps to measure the full width.

And these are the M10 wood substructure mounting bolts from Rernusol, the variosole system.

I drilled 9mm holes into the pine beams which gave these bolts a good firm grip.

 

Update: Saturday 5 October - the panel install day:

First we measured where all the mounting bolts must go - this is very much determined by where the roof beams are!  They look quite unevenly spaced - but they are drilled into solid beams.

 

After fitting all four rails to the bolts, they have to be leveled - horizontally, and adjusted to a straight line parallel with the roof.  This takes quite a while to get right, but makes of a very neat installation.  The Variosole  system is quite adjustable and allows for fixing many offsets in different dimensions.

The first three panels were fitted - where we figured out the system of connecting the panels in series as they are laid down, before clamping them down:

First roof done! (8 of 20 panels - but they will be getting the most sun). You can see the copper strip for the earthing of the rails., and the DC cables for the solar power have not been neatened.

After a long stretch through the afternoon, the second roof was also done except for the last panel, which is on its way from Artsolar:

 

 

Edited October 7, 20196 yr by Ironman

Update: 7 October

The last panel arrived, and was installed.  All solar cables neatly routed through 25mm conduit to the combiner box under the roof. The earth cable is not finished yet, still lying in the gutter.  Waiting for sparky to connect earth to the surge arresters in the combiner box.

The enclosure that arrived for holding the 4 * Pylontech US3000 batteries contained one side for 4 X US3000 and one side for 5 X US2000 - so I could not use it.  Waiting for replacement.  In the mean time, I temporarily connected two US3000 batteries just to be able to use more of the current from the sun.

And this is the result after today:

Notes:

- Still only had 16 of the 20 panels connected

- We had an actual power failure from 1am to 5am.  Were not aware of it - the Victron system carried all the frigdes and important stuff while we were sleeping

- Early power use (5:30 to 7:30) will always be a problem here - the school is starting up - and the batteries will battle.

- We had medium cloud coverage from 13:00 to 16:00, connected with very high consumption (solar geysers did not heat up due to clouds, watering garden with borehole) so the battery did not charge fully

- Also still limiting the max current from MPPT to 70 amp due to just 2 batteries.  When all 4 are connected I can open that up to 100 amp.

 

 

Edited October 7, 20196 yr by Ironman
Spelling

I really want to thank @Jaco de Jongh for coming to help me with this system. He programmed the ESS settings into the Venus GX, and lent me his crimp tools for the MC4, earth lugs and the 35mm cable lugs. He also helped with support and experienced guiding suggestions to help me implement all this. Thanks!

Update: 10 October

The replacement parts for the wrong side-plates in the enclosure arrived yesterday - so I assembled the battery enclosure, installed the batteries and connected all up.  Added some more trunking for the earth and signal cables, and this is what it looked like with the battery door not fitted:

From now on - the system will not be limited by small batteries - I will post an update of a typical day's data soon.

Congratulations, looks very nice/neat.

Now the paying starts, 

G

 

ps... instead of dumb timers for those pool pumps and geysers, look at using sonoff 4c units, for the control, switching contractors that with the circuits on/off.

that allows you to easily change the schedules, graphically and remotely via the web interface.

G

 

My largest consumption at the moment is the borehole - it is very dry and hot in Gauteng and our garden is a crucial part of the school. A spot of rain will help a lot!  We have an automated irrigation system with intelligent timers - but it has stopped working due to our dogs chewing the control cables, and other issues that built up over time 😐.  So the garden is being watered manually by staff - starting early morning and continuing most of the day. Unfortunately they are much less efficient because the watering points run much too long.

We are planning to fix all the issues with the irrigation system early December - then I will also be able to schedule the irrigation to run after the solar geysers' top up timers.

This week's consumption graph: 

Monday - Wednesday had good sunshine, but I was limited by only having two US3000 batteries connected, and only using 16 of the 20 panels.  Solar current was limited to 50%.  Battery consumption was also limited.

On Wednesday evening I connected up all 4 x US3000 batteries, and you can see the massive increase in battery consumption on Thursday.  Grid consumption went down from 10.33 KWH to 1.9 KWH!  I also connected the last 4 panels, so all 20 started working and the peak power generated by the panels increased from 4100W to 5200W, but the clouds came Thursday afternoon. Overall solar energy will be way down Thursday and today due to very cloudy skies.

 

Currently - break in the clouds, getting about 4750W from the sun. Total consumption is about 4540W.  So the inverter can only supply 4000W continiously. Drawing just over 500W from the grid, so the inverter is limiting itself to its 4000W rating:

But I keep getting these overload warnings: 

Is it okay to run like that, or should I be worried?

 

 

Edited October 11, 20196 yr by Ironman
spelling

3 hours ago, Ironman said:

My largest consumption at the moment is the borehole

I use a 300W bolehole pump with MPPT & protection built-in, and 2 panels as a standalone system.

No batteries, just panels onto the pump. It pumps away slowly all day, everyday that the sun shines.

If your head is less than 80m, you get away with a 210W pump and one panel.

Pumps are from China (200-300 USD, inclusive of shipping and customs).

6 minutes ago, phil.g00 said:

Pumps are from China (200-300 USD, inclusive of shipping and customs)

Are these DC pumps?  Where is the MPPT?  Do you have a link?

I currently have a 1500W borehole, with a 100L pressure vessel. Open tap -> pressure drops -> borehole switches on. 

If I want to change to a low power continuous pump system I will need a large tank, either mounted pretty high or using another pressure booster pump. It all adds up and will be quite costly.

I think that for me, the solution lies in reducing the borehole water consumption by improving the efficiency of the irrigation process.  

I use the borehole pump to large tanks at ground level, and then use a transfer pump to smaller tanks at height.

So I have water when ESKOM and the inverter fails me. 

I have recently purchased a 300W  solar surface transfer pump from the same crowd. It comes standard with level switches, but it has an external controller. ( Not installed yet).

Actually I bought a 1.2kW solar pool pump as well, also not yet installed.

I won't decommision any existing set-up, that's for when the sun fails me.

https://kairuijidian.en.alibaba.com/?spm=a2700.8443308.0.0.20083e5fCQCWu4

For the borehole pump specs, see the attachment.

 

 

deep well solar pump 2018.11.pdf

Update: 13 October:

Solar combiner box earth connections done using a neutral bar. One earth coming from one solar panel mounting grid, another from another panel mounting grid, and the third goes to a new earth spike - 1M copper, driven into the ground directly underneath the combiner box.  Not connected to house earth.

An you can see the earth connection to the orange surge arrester.  Thanks @Jaco de Jongh for making up the combiner box with all the fuses, DC contactor and surge arrester.

 

I have seen over 40 Amp being drawn from the four strings, and they are wired up with a 10 Amp fuse in the + and - connection for each string.  That is really running on the limit for these fuses - I hope they are slow-blow.

In the graph below, the current runs at 40 Amp for quite a long time - at one time peaking at 41.7 Amp while the MPPT is searching. Peak power is about  5100 Watt: 100 Amp at 51 V.

 

 

53 minutes ago, Ironman said:

Not connected to house earth.

What's your rational for doing this?

Edited October 14, 20196 yr by phil.g00

58 minutes ago, Ironman said:

I have seen over 40 Amp being drawn from the four strings, and they are wired up with a 10 Amp fuse in the + and - connection for each string.  That is really running on the limit for these fuses - I hope they are slow-blow.

One thing we never thought about was to upgrade the fuses for this site. Remember this was build for the other site with the smaller panels and when you decided to keep it for this site, we never upgraded it. I will send you a set of 12 Amps and have these returned. I can use them on another site. 

1 hour ago, phil.g00 said:

What's your rational for doing this?

As I recall it depends on whether your inverter is isolated or not (sometimes also called transformerless... though that term is bandied about so much that it's almost useless). If there is a path from the grid to your PV modules, then the PV modules needs to be earthed to the same earth as your grid connection. So for high voltage PV-inverters (eg SMA) that switch directly from the high voltage DC onto the AC grid, the earth is required.

If there is isolation between the grid and your DC side, as is the case for most hybrid inverters with attached battery, then you earth for other reasons. Your PV modules might require earthing, or you might want to provide an alternate path for lightening to follow and you might want this path to be somewhere else than your main earth. Of course this doesn't mean you cannot use a single earth for both purposes.

I once downloaded some Dehn documentation on how to use their surge arrestors for lightening protection, and if I recall, that also clearly showed using a separate earth.

Edit: I found the old paper again. I was wrong. They show interconnecting to a single earth. Still. I think the reason why people keep them separate remains. They want lightning to go THAT way...

(Of course you need a very good earth. Below 25 ohms or something like that. Which is why you might want to rather use the one that's there already and done properly).

Edited October 14, 20196 yr by plonkster

1 hour ago, phil.g00 said:

What's your rational for doing this?

...What @plonkster said.  My panel mounting grid is for lightning protection. If the lightning strikes close by, I want to try to keep the surges, eddy currents etc isolated from the house 220V and earth as far as possible.  Nothing will protect against a serious direct hit.

 

...Come to think of it - the PV panels are mounted on the mounting rails.  The rails are earthed by my combiner box as per the photo above - to a new earth spike.

But they are also mounted on the steel roof with metal bolts.  And the steel roof is earthed to the house earth anyway. So the panels are indirectly earthed to the house earth.

 

Mmm, there is a misconception about electricity that it goes into the ground and stops, and it does lead to some strange earthing recommendations.

In any sort of fault, power system or lightening, the current is trying to get to the source to complete the circuit.

I believe Cahorra Basa used the ground as one of the DC conductors to SA during periods of sabotage.

Current flows in the ground just like in wires, and following Ohm's law, that dictates there must be a voltage across that ground. The ground is wide as the globe so it has massive girth as a conductor, but not close to the fault site. There it has minimal conductor cross-section. So the the voltage dissipates outwards from the fault site in a gradient getting less and less as the cross-sectional area of the ground increases.

If I stuck 3 earth spikes, say each a foot apart in a row and put a 220V phase live on the furthest one, I would be able to measure a fair portion of a 220 volt drop between the live rod and the second rod. I'd also measure a lesser volt drop between the 2nd an 3rd rod.

This is the voltage gradient as the "fault" current makes its way back to the power system source.

If I now bonded the second and third rods with a copper wire, then there would be now no voltage drop between the rods. But if I put a clip on ammeter around that wire I'd now measure current.

You can do this experiment yourself.  My point is it clearly demonstrates fault current will flow up an earth spike to get to another point in its journey to the source.

If the 2 earths are bonded, then minimal potential difference can build up between them. In other words, current wont come up your earth spike and also wont pass through your plant to get to your house-earthed panel frames.

You have effectively put your plant between the two earth spikes putting it at risk, if you bond the two earths you will ensure minimal voltage build up, thus protecting your plant.

Theoretically, you could be electrocuted opening a farmers gate between fences in the middle of nowhere, but not if those two fences where electrically connected.

I hope this gets my point across.

 

 

 

Edited October 14, 20196 yr by phil.g00

So would you then rather connect the earth from the combiner box to the house earth only, and not to a spike?

Update - I asked my helper to open up some paving and hit the spike into the ground today...

So it looks like I will not be using the spike at all - the area where I wanted to insert it into the ground (under paving) is solid foundation concrete.  Last option - just take the earth into the roof space and connect it to the house earth.

It just seems to be such a long route to take to actually get into the ground:  the house earth all goes to the into main DB, and from there back out again all the way to the front of the property - and I don't even know if it is connected to an earth spike there.  Single phase 80 amp connection - COJ.

 

So long as all your earths are bonded together it doesn't really matter from an electrical viewpoint. 

There maybe regulations that say you are not allowed earth spikes, I don't know.

The urban house earth, (and I stand under correction), has to be less than "X" amount as part an parcel of your certification. So should be good enough.

(Perhaps, the electricians among us will chip in).

I am in the sticks.

In my own circumstance, where I have PV panels on a 10m steel tower in the Drakensberg ( read plenty lightning), the tower footing is bonded to 2 salted earth spikes, bonded to the fence, bonded to the power system earth, bonded to the earth available at the inverter. A later practice I have adopted is burying a length of fencing wire in any trench I dig either for pipes or cabling.

 

13 minutes ago, phil.g00 said:

A later practice I have adopted is burying a length of fencing wire in any trench I dig either for pipes or cabling.

Another added bonus is that you can find underground plastic water pipes that you've forgotten about with your nephew's metal detector.

Early days, I found that low voltage cabling say from the house to the automatic gate  control wouldn't last a single thunderstorm.

This was because the house wasn't bonded to the fence and the insulation on those skinny control wires didn't stand a chance. It took the gate motor every-time with it as well.

 

 

20 minutes ago, phil.g00 said:

PV panels on a 10m steel tower in the Drakensberg ( read plenty lightning)

 

In that case - you would probably benefit from a tall standalone lightning pole - like the ones used next to thatch-roofed houses.

Interesting note: This image comes from the EE Publishers website, but it is labeled dehn-xxx ..... @plonkster ...

 

28 minutes ago, phil.g00 said:

the tower footing is bonded to 2 salted earth spikes, bonded to the fence, bonded to the power system earth, bonded to the earth available at the inverter. A later practice I have adopted is burying a length of fencing wire in any trench I dig either for pipes or cabling.

You have gone to a lot of trouble.  Have you had any hits?  Did it survive? I suppose if you don't get any damage you would not really know how many times the lightning protection did its job.

 

@Ironman. Yes there is a nice sharpened spike as high point on the south of the tower, so it doesn't cast any shade.

The PV tower is a relatively new feature, and to my knowledge hasn't endured a direct hit yet. - It will soon enough.

Although, I know there has been a very close strike, and by close I mean a tree in the garden, since the tower has been up.

The Drakensberg can be like a light show sometimes, we have two other high water towers as well. I can't think that they haven't been directly hit over the years.

Like I say, numerous issues in the early years, but I don't want to tempt fate as say it'll never happen again, but overall I don't think any electronics have been fried in the past 5 years.

I forgot to mention, I also have some overhead CAT 6 runs between some buildings, I break the circuit with two media converters with fibre-optic between them.

I suppose when you have lightning as bad as I do, you just build in the protection as you go along.

I also use ferrite beads on the around the wiring coming from the PV.   Lightening doesn't like inductors.

For the same reason you should never coil you earth wires.

 

 

 

 

Very interesting.  

I never thought of ferrite beads for lightning protection.  Ferrite beads are supposed to suppress high frequency noise on DC or low frequency lines - they absorb changes in current, similarly to how a capacitor absorbs changes in voltage. I suppose a lightning strike would cause a fast change in the current flowing in the PV lines - but by that time the surge arrester in the PV combiner box should also be working. I suppose they all add up to make an effective suppression system related to the current surge.