giggsie

They explain the cable size thing/voltage drop very clear in the SANS code. There are a lot of factors to consider. Like, is the cable in open air, is the cable in a conduit, is it underground, is it in concrete, length of cable, copper or aluminium cable. Then very important the length, voltage and amps.
In South Africa we work with mm2. Over your distance, out of experiance I will use 16mm2 red and black single core house wire and 10mm2 earth wire in a 25mm conduit. Please stay away from normal Twin and earth unless you want to replace it in 5 years time. If you need to shoose between twin & earth and surflex, surflex is the much better option.
Surflex has a metal sheet around the live, neutral and wire. The purpouse of that sheet is to protect the live and neutral from moisture as well as act as a type of earth in case of cable damage. Picture yourself that if you cut away the insulation on twin & earth and you also cut away the insulation of the red/live wire alone. Now if you touch that bare live copper, what will hapen? The only thing that will trip is the 10 or 20 amp breaker. Then you will have a nasty shock already.
If you damage the surflex you first have to cut through the metal sheeting then you will reach the live wire making a short already. In that case it will trip the earth leacage immediately.

One spike, 6mm insulated (green with yellow tracer) in pvc piping is good. 

Introduction
What is presented here is my experience with solar over the past 2+ years.  I hope this write up will be of value to readers and assist with their journey, its become a hobby.  Feel free to comment.
The section on Payback represents a situation faced by most homeowners and reflects an approach which I thought appropriate.  Our unit cost is relatively low, and the same applies to savings.  Data on the number of hours without power, cost of generating power using a generator is not readily available. There is recognition of increased cost of power during an outage, but the approach is debateable.  I’ve steered away from opportunity costs.
 
Objective
1.       To provide power during load shedding and frequent extended outages experienced due to unreliable and worsening municipal electrical infrastructure (also not keen on using a generator, bit of a closet greenie).
2.       Provide power to keep lights, fridge freezer, router, office, (desk top PC, screen, printers) as well as DSTV decoders and TV’s going.
3.       Eliminate (or reduce) frustration experienced every time power goes off.
4.       Recoup investment in solar in as short a time frame as possible and reduce reliance on ESKOM.
 
The system
·         Goodwe 5048D-ES 4.6 (current DOD while on grid 5%, off grid 90%)
·         17 x 365w Canadian solar panels = 6205w (8 facing North, 9 facing North West) at 11 degrees from horizontal
·         2 x 3.5kw Pylontech batteries current SOH 99% (changed from 100 to 99% shortly after 18Dec 2021 after 23.5 months of use).
·         Installed Jan 2020.
Photos of installation in separate post
General information on electrical set up
·         Household 4 adults, residential and home office.
·          750w heat pump, water is heated to 42 degrees, no cold water used when we shower.  1 degree increase in water temp takes 6 min. Starting water temperature in summer about 26 degrees, winter about 20 degrees.  Heat pump needs to be started a couple of hours before planned shower.
·         750w pool pump used for 2.5 hours per day from Oct to March.
·         Replaced electric hob with gas.
·         Replaced 50-amp municipal CB with 40-amp CB.
·         Due to extent of power failures all plugs, lights pumps are run through the inverter, only oven directly connected to municipal supply. 
·         The inverter capacity is 20 amps + 10 amps for 10 seconds before it trips.
·         Tripping of the inverter can be avoided by using 2 or less appliances at the same time, especially when the dishwasher or washing machine (max power use while heating water of 2400w) is on.  Training ongoing, it is easy to forget…
·         Normal background use of lights, fridge freezer, PC etc about 500 to 600w, this can be reduced to 350w in extreme cases if outside lights are turned off as well as PC and screen.
·         Power is sold and bought at the same rate (municipality under administration).
·         Only 40% of PV is used with current set up (self-use), rest provided to grid.
·         Electricity meter read at month end and provided to municipality.  They use the information!
Cost in December 2019 + fixes to DB later

 
Monthly power produced and consumed based on data from inverter and meter readings

Yellow = 5 months with lowest PV produced and highest meter reading.  Green = 5 months with best PV produced and lowest meter reading.
I suspect actual consumption is higher and saving would be higher than shown.  The Goodwe does not keep track of power supplied from the battery (both on and off grid).
Summary data (gradual improvement)

 
Payback
The variables used to calculate the original payback period changed.  The solar system produced less power than anticipated (635w vs 900w) and cost more.  Both would increase the payback period.
At the moment payback will be about 10 years based on the following assumptions.
·         Continue to sell to back power at same rate as purchased
·         Eskom increases for 2022 = 9.61% and from 2023 will be at least 6% per annum
·         No major breakages
·         At least 10 hours of monthly outages that equals 20 kWh at, at least 4 times the going rate per kW.  For March 2022, 10 hours x 2 kw per hour @ R1.51 x 4 per kW = R121/month.  I believe this assumption is conservative.  My calculations show kW cost from a generator is far greater than R10 per kW
·         No interest added to original investment of R150k.
Should increases be 8% per annum, power outages double and cost per kW increase from 4 to 8 times the going rate (see example above), payback will shorten to 7.5 years (so 5.5 years more to go).
SARS allows a proportional depreciation over 3 years for a home office, effectively reducing the original investment by about R25k.  This will reduce the payback period by at least another year from 7.5 years to 6.5 years.
Perhaps financially not the best investment decision (but better than Steinhoff) and to date happy with the outcome.
 
Lessons learnt, observations
·         Educate yourself about solar, speak to many people that have installed solar, join, read, and ask question on powerforum!
·         Understand what your objective is and what you want to achieve.
·         Get at least 3 quotes, make sure you deal with a reputable company and that the inverter has been “accepted” in the larger metropolitan areas of the country and that the batteries are compatible and supported by the inverter.
·         Give preference to systems that allow incremental additions e.g., adding extra battery capacity or extra inverter. I can’t do the latter.
·         The company should provide information on expected production, make sure they disclose the variables on which it is based and expected production throughout the year. If they don’t, move on.  Only one of the 3 companies that quoted provided this information.  Some also appear to be stuck on a single range of products.
·         Make sure to get a COC for the solar installation and the DB. Useful for insurance in case there is a problem (lightning, wind damage, surges)
·         I obtained theoretical production values from the website below for my location, and while the monthly values varied a lot, I’m at 96% of the predicted production of 16 513w for the 25 month period https://re.jrc.ec.europa.eu/pvg_tools/en/tools.html
·         Somehow everyone, forgot that parts of the country experience cloudy/rainy conditions and that the 30kwH per day or 900 kwh per month is probably only relevant in the desert.  Actual production of the “900 kw system” turned out to be 635kWh per month.  Don’t like me accept information at face value.  The screenprint below is an example of the type of information out there, ready to confuse the novice.…

·         Be aware that system losses can be between 14 to 25%.  Reasons include converting from DC to AC, temperature, inverter efficiency, etc.  For my system and location predicted losses were expected to be 25.6%. 
·         Bear in mind that your panel mounting areas largely dictate production.  In my location optimal angle = 35 degrees, however the roof angle is only 11 degrees.  Production in Gauteng will be better as solar radiation is higher.
·         Shop around, one supplier is seldom cheapest on all the major items (panels, inverter, mounting structures, batteries).
·         Include in the quote to fix, or have your DB checked beforehand to prevent surprises. I did not and probably paid R8+k extra.
·         Understand what items you want on the backup (battery) circuit.  In my case initially lights, fridge, freezer, TV, decoder, router.  Had it changed later to include all plugs, heat pump and pool pump as with frequent outages while the sun was shining, we were unable to consume the power being produced. Obviously cost extra.
·         Ensure there is good airflow between the roof and the panels, heat reduces production of power.
·         Plan to be able to clean the panels by leaving a walkway between sets, dust, grime, birds can reduce PV production by 5%.  At 11 degrees angle my panels don’t self-clean!  I use a soft broom with a plastic extension to reach 2 meters and a hose pipe.  There are small yellowish sticky deposits that even resist a high-pressure cleaner.
·         I can only communicate with the inverter via wifi. I installed an UPS to power the ONT box and router in case there is a need to change settings on the inverter.
·         Understand the limitations of your inverter, I am unable to use a generator with the Goodwe and hence unable to charge batteries in periods of ongoing power failure and consecutive cloudy/rainy days.
·         Specify that a bright LED be added where it is easily visible, (probably in the kitchen to indicate a power outage). This to prevent use/waste of battery power when there is no sun. Backup systems are so efficient you are often not aware that loadshedding or a power outage is occuring.
·         If in future unable to feed back into grid I will have to increase battery capacity and the payback period will increase.
·         Don’t wait, invest now.
22 March 2022

Just to provide some feedback on this topic. I have just returned from the COJ office in Roodepoort after enquiring about cost to convert to prepaid.
Information provided by the agent who assisted me:
Account must be in good standing and up to date There is a form that needs to be completed and submitted to begin the conversion process. City Power will then come to site to conduct and inspection and provide a quote for the conversions. If you have a Smart Meter then the cost to convert is R1 208. (This is a standard rate and the above quote is only applicable if there is no smart meter) The process takes between 14 and 90 days after the quote/fee is accepted and paid. Service fees and network charges are removed from the monthly invoice. The Agent Informed me that the entire section relating to electricity is removed from the invoice completely. I am sure the process of feeding the meter and paying down the post paid costs incurred will still apply as per Post by @Bobster For my situation where we are on single phase supply and a monthly average consumption of 660kwh that cost us +- R2000 this makes sense as out power cost will be reduced to around R1250. This is before any additional price increases.
I hope this helps any one browsing the forum looking for answers. 

Just to provide some feedback on this topic. I have just returned from the COJ office in Roodepoort after enquiring about cost to convert to prepaid.
Information provided by the agent who assisted me:
Account must be in good standing and up to date There is a form that needs to be completed and submitted to begin the conversion process. City Power will then come to site to conduct and inspection and provide a quote for the conversions. If you have a Smart Meter then the cost to convert is R1 208. (This is a standard rate and the above quote is only applicable if there is no smart meter) The process takes between 14 and 90 days after the quote/fee is accepted and paid. Service fees and network charges are removed from the monthly invoice. The Agent Informed me that the entire section relating to electricity is removed from the invoice completely. I am sure the process of feeding the meter and paying down the post paid costs incurred will still apply as per Post by @Bobster For my situation where we are on single phase supply and a monthly average consumption of 660kwh that cost us +- R2000 this makes sense as out power cost will be reduced to around R1250. This is before any additional price increases.
I hope this helps any one browsing the forum looking for answers. 

Hi Giggsie, it is on hold at the moment. Will let you know if it changes.

Technically there is a fuse in the inverter (200A I think) and usually a way to turn off the battery via a button or switch, but these are expensive to replace and clunky to operate in an emergency (trust me I know), so it is best to have an external fused DC breaker. I can't comment on the compliance aspect, but for safety it is definitely better.

Yeah afaik it will work perfectly and also adds all the data, graphs etc. you would ever need. There seems to be some work ongoing in the king firmware for shoto support so it might come in the future. I just run my king on USE mode with voltage settings and it works perfectly. SA will help if you want to control things remotely and run certain things based on SOC.

I think that is a good choice and will be nice and neat as they are all rack mounted. I would personally go for a rack mounted shoto battery because of the C rating (and because it's what I chose), but it is not as 'trusted' as the Pylontech or Dyness and doesn't talk directly to the inverter yet, but that doesn't bother me as much as others. My system is DIY and I don't have a COC, so warranties didn't factor too much into my decision, but it might be more important for your install at a school and as you mentioned you don't want to take a risk with other people's money.
I can't comment on the other solutions as I have no experience with them, but having a chance to test them first is great. If you need something quiet for an office then you need to look for inverters with heat sinks instead of fans, there are a few of them around.

Don’t forget you need to know how much power those PoE switches are using. If you want to save some money, I would recommend buying a kill-a-watt for R400-R500 and conducting an energy audit. Figure out how much power you are actually using. For example, one of my computers has a 750W power supply, but it only uses about 250W most of the time. If I based my spec on 750W I would end up buying a battery 3x larger than I need, and the batteries tend to be the most expensive part of any such build.
 
Once you know how much concurrent power you need in watts, multiply that by the number of hours you need it for. So say it’s 1000W and you need it for 4.5hrs, that’s 4500Wh or watt hours (aka 4.5kWh). I would add a good 20-30% on top for safety and efficiency losses.

Also consider the difference between how much you’re actually using and how much you could theoretically use. Computers use way more power starting up or shutting down for instance. Your 1000W system might pull 2000-3000W at peak, so make sure to size the inverter accordingly.
I recently did the calculations for Stage 6 myself. It gets a little bit complicated if you are trying to work out the minimum requirement. If you are going off worst-case (ie a full 4.5hr outage instead of 4.1hr or whatever) and you need it to always stay on, you need to figure out the weird parts of the schedule. Particularly where one day ends and the next begins. For example, you will usually have at least 4.5hrs between outages, but every few days (where I live) we get a 4.5hr outage either 2.5hrs before or after a 2.5hr outage. So you have to figure that you either need to be able to recover for or from a 4.5hr outage within 2.5hrs.
 
One method is to get batteries that can charge that fast (honestly most rackmount batteries can do 0.5C which means that they can be charged in 2hrs with a sufficiently powerful charger/hybrid inverter) or to get enough headroom that whether you have 2.5hrs, 4.5hrs or whatever, you aren’t going to be so close to the line that you need to worry. The latter is the more expensive option clearly. Note that most rackmount batteries are 48V, I think the Pylontech UP2500 is a notable 24V exception.
 
Also note that some non-rackmount inverters frequently get placed on their back in a rackmount configuration but that isn’t really ideal as they are designed to be vertical for cooling purposes. If you are going to mount them in this configuration, make sure the room has adequate cooling.

The suggestion of 5kW is purely based on two important reasons. The 5kW King rack mountable solution is quality. I don't offer anything less than 5kW because quality of the inverters tend to drop in the sub 5kW range & of course 24V batteries imho also start to get into the lower quality range also. 
As explained above the amps are lower in the 48V range. 

Another thing to consider is the voltage of the system and the load on the battery. A 1kw load on a 24V system will pull 40A from the battery, whereas a 1kw load on a 48V system will pull 20A from the battery.

If your requirement was for a solution that could be rack mounted then the Rack Mount Axpert King would be a good candidate. That King would need to be paired with a Lithium battery. 

thanks. this looks good. its a little pricey though. Will read up on them. Would be good if they have replaceable batteries though.

I came across this site. They have a 48V DC Mini UPS which could help to power your switch and IP Cameras. They also have some case studies which give different scenarios which might also give you an indication of run time dependent on load.
https://tech-buddies.co.za/

I came across this site. They have a 48V DC Mini UPS which could help to power your switch and IP Cameras. They also have some case studies which give different scenarios which might also give you an indication of run time dependent on load.
https://tech-buddies.co.za/

Just an update guys, went through numerous installers and all were pathetic other than Gideon at solarpower.co.za I can recommend him and his company without a shadow of a doubt, amazing prices, quality workmanship and great service and supervision on site!

My 2 cents: Looks like all the other scams

Looks like a scam site. 
 
Also, I see http://solarenergyrsa.co.za is now dead.