My solar power experiences

I found that designing a 10 kW solar power system that was much larger than would fit on my available roof space at Clare in Mid North South Australia was not a simple matter. My aim was to put some money into a good financial investment and, at the same time, do a little bit to combat climate change.

This page started with that plan, then went on to become a record of the various installations that I've been involved in.

The rules for getting a workable price for the electricity that one feeds into the grid vary greatly from place-to-place and time-to-time, this page deals with South Australia and Western Australia.

In the end we had to scale down from 10 kW to 1.5 kW because the South Australian Government changed the rules on its incentive payment for solar electricity.

Over the next few years we added another 14 kW of panels in three locations.

This page created 2010/07/18, last edited 2023/03/31
Contact: David K. Clarke – ©

Much has changed since 2006/2010

I write this note in March 2021. Large scale solar (and wind power) are now the cheapest forms of electrical energy. Solar panels are cheap enough so that one doesn't need to worry about placing them on the best possible angle, if it is more convenient to lay them on a gently sloping roof little is lost by doing so.

As I write several workmen are placing 6kW of solar panels on my near-flat carport roof.

Our other solar power system
Solar power and heat
A part of the solar setup on our home.
The PV is only a 1 kW system, not the main subject of this page.
On February 8th 2006 my wife and I were among the first in Crystal Brook to put a solar photovoltaic system on our home (photo at right; we put the solar water heater up thirty years earlier). In late June 2010, because of changed solar incentive offers from the government, reduced prices of solar power systems, and good electricity feed-in tariffs, we decided to set up a much larger 10 kW system on a hobby farm at Clare. The system was to be connected to the power grid (the design considerations of a stand-alone system are quite different).


Our Site, and supporting structures

The site available to us was fairly open with few tall trees (a dry climate has some advantages) and sloped toward the north at a useful 5°.

The only suitably located building available to us had a roof much too small for the required 54 or so panels (each around 0.8m × 1.6m). We thought at first to place the panels on frames on the ground, but were told that the frames would cost something like $10k and that we might be better off to consider building a simple, large shed so that the panels could go on the roof and the space inside would be useful for storage.

We had been thinking of building a pergola for summer shade near the barbeque, so I started looking into designing a suitable pergola that would also serve as a frame for the photovoltaic panels. The problem here was that such a pergola would be very big and high as well; it was difficult to imagine something that could be built for a reasonable price and look acceptable.


Factors to consider in siting photovoltaic panels

  • The site must be open to the sunshine for the maximum possible period of each day and throughout the year;
  • With a large system, the front rows must shade the back rows as little as possible, at all times of the day and year;
  • The panels should be aligned as close as possible toward the north and at the optimal tilt angle for the latitude (the optimal tilt for us was about 30° at our 35° S latitude). Panels that are fixed at low angles are likely to collect dirt and dust and require cleaning, those at steeper angles are largely self-cleaning;
  • The optimal angle for a stand-alone power system, in which one needs the maximum amount of power in winter, when the sun is low in the sky, is steeper, perhaps 50° for our latitude. A lesser tilt is better for a grid-connected system because it increases the amount of power being injected into the grid in summer when days are long and the sun high overhead;
  • They must be close to the point where the power can be fed into the power service;
  • It is important to keep the panels as cool as possible; in practice this means maximising natural ventilation; see Temperature of panels.
Placing panels on a flat, or nearly flat, pergola roof at our latitude (35° south) would require that they be tilted up on edge and arranged in a number of rows (see box on the right). If the rear rows were not to be significantly shaded by the front rows early in the day, then there would have to be a large gap between the rows; and providing such a gap would make the pergola unacceptably long from north to south. The other alternative was to tilt the pergola roof at an angle steep enough to efficiently collect solar energy. The best compromise angle seemed to be 20°, but even this lower than optimal angle led to a roof that sloped up to an unacceptable height (if the low end was 1.7m the high end would be about 4m).

What to do? The best answer seemed to be to place half the panels on a smaller pergola and the other half on and near the shack. There was room for 21 on the roof and I could build a veranda for the other six. But politics were to intervene.


The power from solar panels is low voltage direct current (DC); it must be converted to higher voltage alternating current (AC) before it can be fed into a 'normal', 240V AC, power supply. This is handled by a black box called an inverter. Surprisingly, two five-kilowatt inverters (or three three-kilowatt inverters) are cheaper than one ten kilowatt inverter. This meant that it is quite economically viable to have two (or three) more-or-less independent smaller solar power systems, rather than one large one.

Requirements for Solar Credits

In order to receive the full Federal Government incentive in Australia, what they call Solar Credits, the solar power system must be on premises which can be a house, church, school, or are associated with your business activity (eg. a dairy in the case of a dairy farm, a shearing shed in the case of a sheep grazing property). I had to use the last alternative, my premises are used for the storage, racking, bottling, labelling, and capping of olive oil from my 600-tree olive orchard; the processes of bottling, labelling and capping were accepted as processes that occurred on my premises involved in running the enterprise.

Permission to feed into grid

Depending on where your place is, the local power grid may not be capable of handling the output from a 10 kW 'small embedded generating unit'. I had to contact ETSA (the organisation in SA that runs the poles and wires) and apply to connect. Some rural power supplies are not capable of handling as much as 10 kW of power, especially SWER (single wire earth return) supplies that are a long way from the main power lines. I was fortunate in having installed a two-phase connection and being only 100m from a high-voltage three-phase power line, which itself was within 100m of two branch lines.

Minimising the operating temperature of the panels

Solar photo voltaic panels produce less electricity as they get hotter; the difference is very significant. (I believe that thin film amorphous silicon panels suffer less from this effect than do crystalline silicon panels.) Therefore it is very important that panels attached to a roof should have a goodly space under the panels and above the roof to allow the free flow of cooling air between roof and panels. My information suggests that at least 80mm should be provided, 150mm would be even better.

My panels were installed on a free-standing frame, giving more ventilation than would be achieved on a roof (see photo).

Solar installers

I chose to use one of two accredited local solar power installers, either Clewers (of Clare) or Luke O'Dea (with connections to Clare, Crystal Brook, and Adelaide).

Clewers would use two CMS 5000 (5000W) inverters and 54 × 185W Enertech monocrystalline panels having a total nominal capacity of 9990W. O'Dea would use two Enerdrive GT 5.0 AU (5000W) inverters and 51 × 190W Phono panels having a total nominal capacity of 9690W.

The major hiccough, referred to elsewhere, changed things, but eventually I went with Luke (although I'm sure that Clewers would have done the job very well too).

Quotes, things to think about

One of the initial quotes included trenching, the other did not; but as there was only one trench about 25m long involved I felt I could dig it myself without much difficulty.

It was suggested by one of the installers that the brand of panel used by the other was of inferior quality. I could see no way of checking the truth of this, other than asking that installer, who (as you would expect) held that the panel was of high quality.



It seems that everyone who is making any forecasts in Australia is predicting steeply rising electricity prices. This must happen for several reasons:
  • The Australian power system is in dire need of upgrading, and the upgrading will be very expensive;
  • We have to change from fossil-fuelled to renewable power;
  • We have to change the flat fee for power to a price that varies with supply and demand;
All of these changes will have to be paid for, and much of the cost will finish up in the price we pay per kilowatt-hour for our electricity.

The result of this should be to increase the value of existing solar power systems in the future.


I expected the whole thing, before the major hiccough, including building the pergola, to cost about $50k for 10 kW. I use only about 2.5 kWh of electricity per day (about 1.3 kWh during daylight hours). The 10 kW solar power system should generate about 42 kWh per day, of which about 41 kWh will go into the grid each day (about 15 000 kWh per year). The government payment in SA was, at the time I was planning this, 44¢/kWh, electricity retailers pay an about (depending on retailer) an additional 8¢/kWh. Income should be near $8000/yr, paying back the cost of the system in around 7 years and paying about 16% on the money invested. ("The best laid plans of mice and men aft gang aglee.")

When figuring how much it is all going to cost you, you should not forget that you will probably need a new electricity meter; ours cost us $652 (our daughter in WA only had to pay $250 for hers).

WA instead

In the end my wife and I were not able to install our 10kW solar power system, due to the major hiccough discussed elsewhere.

Instead we installed a small (1.5 kW, about $3k after Fed. handout) solar power system on our own place on 2011/01/18 and paid for a larger (5.3kW, about $23k) system to be installed on our daughter and son-in-law's house in Western Australia on 2011/01/25. In its first full day of operation, the 26th, this installation generated a promissing 40 kWh.

In return for paying for the WA system we will get half of the money that our kids save on their power bills. (I've estimated that from the present roughly $2100 annual total payment for power they should go to a refund of something like $1500.) The calculated annual return for us is around $1800, or 8% on our investment, the kids should save another $1800 in power bills.

And Adelaide, South Australia

Later we paid for a solar installation, I think it was about 3.4kW, on our son and daughter-in-law's flat in Adelaide.


Before late August 2010 the South Australian Government payed 44¢/kWh for solar power fed into the grid (a 'net payment' system). Following that date they made themselves look good (superficially) by increasing the payment to 54¢/kWh, but at the same time greatly reduced the number of people who could qualify for this, and greatly reduced the maximum size of the solar power system that would qualify.


A carbon tax

What our governments should be doing rather than fiddling with rebates and payments to owners of solar power systems, is making the greenhouse polluting fossil-fuel users pay for the damage they are doing to the world, so that non-polluting sustainable generating systems become economically viable in their own right. As well as greatly reducing greenhouse gas production it would be much simpler and lead to fewer distortions of the economic system.

Major hiccough in our plans!

When the government changed the rules for people who want to receive the electricity feed-in tariff to specifically exclude anyone hoping to make a profit and anyone who already has a solar power system, no matter how small, they made my proposed 10 kW system unviable.

Inquiries with the Dept. of The Premier confirmed that our system will not be eligible. So, instead of installing a 10 kW solar power system on my own place I'm arranging to pay for a 5 kW solar power system on my daughter's place in WA and have a 1.5 kW system put on my own place.

We will not get the 54¢/kWh feed-in tariff for our little 1.5 kW system any more than we would for our originally proposed 10 kW system, but we should still get about 8¢/kWh from the electricity retailer, and feel that we are doing a bit for the planet. We were advised not to put anything as large as a 10 kW system up in WA because of the perceived probability that the WA Government were intending to change the rules similarly to SA.

Finally! Up and running

Our solar power installation
Our first solar panel installation at Elysium
They provide shade to our north-facing wall and have excellent ventelation underneath.
Photo 2011/01/21
In the end, even though I had no expectation of getting more than 8¢ per kilowatt-hour for the power I would be feeding into the electricity system (because of the Government's changing of the rules) I ordered a 1.5 kW system (8 × 190 Watt panels) from Luke O'Dea at a cost, to us, of a bit less than $3000 (the Federal Government's shonky solar credits scheme paid $6200).

We also had to pay $652 for the new (3 phase) electricity meter (a single phase meter is cheaper, about $400). Our daughter in WA had to pay only $250 for her new 3 phase solar meter. I have asked ETSA why the difference, and written a letter to several newspapers posing the same question.

The 1.5 kW system should generate about 2300 kWh per year. This was a much smaller system than the 10 kW that my wife and I were originally planning but at least it was an investment toward reducing our greenhouse footprint and it was not a huge financial burden.

There were several reasons for placing the panels on the frame on the northern side of our shack rather than on the roof.

  1. The roof had a slope of only 11°, not sufficient for optimal solar power generation at our latitude; the frame gave a slope of 26° to the panels;
  2. The shade from the panels would keep the hot summer sun off our northern wall and passively keep our temperatures down;
  3. The panels would be able to get the maximum cooling ventilation, minimising the operating temperature and maximising their efficiency.
The cost of the steel for the frame was about $500, and $80 for the (dry-mix) concrete in the post holes.

When the photo was taken (2011/01/21) everything but the inverter was in place, the panels were installed on the 18th. I had to wait until 2011/02/02 for the inverter to arrive and be installed; the new meter (capable of measuring feed-in as well as consumption) was installed the next day.



Installer's forecast

The installer's figures indicated that I should expect an average energy generation of about 8.7kWh per day for February. This was calculated for a tilt angle of 30° and at Adelaide's latitude – 34.9°S. The tilt angle of my panels is 25°, which should very slightly increase the generation in February, and my latitude is 33.8°S, which also gives a very slight advantage compared to Adelaide.

Measured characteristics

Based on data recorded on 4th and 7th Feb. 2011, the power output of our system fits the multiple linear equation P = 0.0107 L - 4.71 t; where P is power in Watts, L is light intensity in Lux, and t is panel temperature measured by my infra-red thermometer in degrees above 23°C. This roughly means that it generates 11 Watts for every thousand Lux, but that this is reduced by 4.7 Watts for every degree C that the panels are above 23°C.

Seven years of operation

Output in the first seven years:

The capacity factor is the power actually generated as a percentage of what would be generated if the system operated at full label capacity 24 hours a day and 365 days a year.

The first full day of operation, 2011/02/03, the maximum power achieved was about 1100W (Nominal power of our system was 1520W); and 5.3kWh was generated. The maximum ambient temperature was around 38°C and the panels got up to around 65°C (measured from the back with an infra-red thermometer and demonstrating the need for maximum ventilation). There was cloud cover for more than half of the day.

As one would expect, there is a linear relationship between light intensity and power output; with an increase of about 11W for each thousand Lux.

The power output of silicon solar panels, especially crystalline panels like ours, decreases with increasing temperature, and an analysis that I did on 2011/02/15 indicates that the output of our system decreases by 4.7 Watts for each one degree increase in temperature.

By 2011/02/22 we were averaging 6.8kWh/day at the inverter, significantly less than the installer's forecast (see box on the right). The short-fall would probably be due to there being a lot of cloud cover for the time of year. The maximum power that I had recorded to then was 1326 Watts.

Lux against Watts
The graph on the right has the power (in Watts) generated by the panels plotted agains the light intensity (in Lux). As you would expect, in general, the brighter the light, the more power. But the data do not fall exactly on the trend line. The scattering is due to the variable temperature of the panels during data recording.

Measuring Lux
Measuring Lux. The light sensor is facing the same direction as the panels

Lux against Watts
This graph shows the same data as in the above graph, but with the temperature factor mathematically removed; the much closer match between data and trend line is conspicuous. This strongly suggests that the predominant factors affecting the power production from our solar panels is light intensitity followed by temperature.

This is roughly what the power output of our system would look like if the panel temperature could be held to 23°C.

Generation from our second solar system
Early data from our solar power system at Clare are shown in the graph on the right. The blue points represent the power generation as recorded by the inverter, the pink is the power that we have imported from the grid, and the yellow is power that we have exported to the grid. (Import is shown on our ETSA meter as item 03, export item 09.)

It is pleasing that we are exporting three or four times as much power as we are importing; a good result considering we only have a 1.52kW system. It was less pleasing that our system generated an average of 7.04kWh/day in February; the installer's figures indicated 8.7 as the expected average for February. (It generated 5.99kWh/day in March, against the stated 7.7.) But there was probably more cloud in this February and March than average, time will tell if it is as good as it should be.

Our average electricity consumption before installation of the solar system was 2.24kWh/day.

The WA installation compared to Clare

As mentioned elsewhere, the solar power system that we had installed at Clare was rated at 1520W and the one installed on our daughter's house at Mandurah in WA was rated at 5280W. Our unit has eight 190W mono-crystaline Phono panels, a 1.5kW SMA Sony Boy inverter and was installed by O'Dea Electrical on 2011/02/02. Our daughter's unit has 48 Kaneka 110W hybrid panels, a Xantrex GT5.0-AU inverter and was installed by The Solar Shop, Mandurah on 2011/01/25.

To 2011/02/17 the greatest output I have recorded from our unit is 1326W, that is, 87% of its rated capacity. My son-in-law recorded an output of 5079W, 96% of the rated capacity of the Mandurah system. You'd have to say that, on this measure, the Solar Shop installation is at present looking better than the O'Dea installation.

Our solar power at Crystal Brook

Solar power at Crystal Brook
Solar power and heat
A part of two of the six panels on our place at Crystal Brook; solar water heater in the background
Electricity record at our home
Electricity graph
On 2006/02/08 my wife and I had six solar panels rated at a total of about 1kW installed on a shed at the back of our house. At the time a rebate of $3600 was available from the Federal Government; we had to pay about $12 000 ourselves. (Compare this with the new 1.5 kW system – the main subject of this page – that cost us about $3000.)

The graph at the right shows a record of data from our electricity meter. The blue line records that amount of electricty that we have taken from the grid on our main circuit. Our power consumption is greatest in summer due to two things: the refrigerator and freezer have to work harder to maintain their set temperatures, and we use a ducted evaporative air conditioner for cooling the house. The pink line shows the amount that has come from the grid at off-peak times for the electric booster of our solar water heater. The yellow line shows that amount that has been fed into the grid from our solar panels.

Power from the solar panels is fed into the grid only when its amount is greater than the power that we are consuming in the house.

Out with the old, in with the new at Crystal Brook

On 2021/03/01 we had our old 1kW solar system replaced with a new much bigger system.

At 4pm (central daylight savings time) the day before I measured the output of the old solar system as 1.9A at 245V, so 466W. The sun was on an angle of perhaps 45 degrees to the panels at the time.

The old system had worked well. There was a failure at one time, but it was fixed at little expense. Considering that the old system was over 15 years old and still operating when it was superseded this seems quite good.


Summary of meter reading from the old system

The figures below are averages in kilowatt-hours per day

Main circuit
'J' circuit
2006-August 2017
August 2017-April 2021

The 'main circuit' is power imported from the grid
The 'J' circuit is power imported for water heating overnight (we have a solar water heater so we only need electricity to heat the water in overcast wintery weather).
The 'Feed-in' circuit is power exported to the grid.
The power meter was changed in August 2017 and again in April 2021.

I've written a section about the new installation below.

Effect of replacing the old freezer

We bought a new, more efficient, freezer on 2010/06/30; previous to this we had a freezer that was more than 30 years old. This significantly reduced our electricity consumption and increased the amount our solar system was able to feed into the grid.

Our average power consumption on the main circuit for the two years before this was 7.02 kWh/day and dropped to 4.38 kWh/day for the period 2010/06/30 to 2011/01/16 (a decrease of 38%). The average amount of power that we fed back into the grid in the two years before the new freezer was 1.39 kWh/day and this increased to 1.96 kWh/day from 2010/06/30 to 2011/01/16 (an increase of 41%).

I have calculated that the $1100 new freezer will pay for itself in four years due to the lower power consumption and the greater payments for the power that we put into the grid.

This section added

More solar at Crystal Brook

Solar at Crystal Brook
The old solar PV panels, installed in 2006, are in the foreground, the new panels, installed in March 2021 are on the left, the solar water heater is on the right. (This solar water heater was installed about 2010, replacing one installed about 1980.)

Evolution of solar panels

Our first solar system (installed in 2006) had six panels each of 166W - a total of 996W. The new panels each are capable of generating twice as much power as the old ones. They are about 25% bigger in area.
As I write this four men are covering my carport roof with solar panels. They are installing 18 panels each of 330W, a total of 5.94kW. The inverter is 5kW, so the maximum power that can be generated by the system is 5kW. As the panels are lying on a nearly flat roof (very slight slope to the SE) they would not generate their rated capacity in any case, especially at our latitude of 33.4°.

It will be interesting to see how it goes.

Our first 1kW system at Crystal Brook had no information available from the inverter at all. I could measure the current output and I could read how much power from the system was being fed into the power grid on our electricity metre. I could not measure how much power from the solar system we were consuming at home.

On 2021/04/06 a man came to install a new meter so that our new panels could export power to the NEM (national electricity market). The SMA inverter apparently was transmitting data (the meter installing electrician said so), but I was unable to get any sign of operation from the SMA app that I installed on my iPhone at the time the panels were installed. The power meter indicated that power was being exported.

The SMA inverter in the new system is supposed to connect to my iPhone through either an app or via a QR code and provide current output as well as a record of solar generation. Sometimes the QR code system works, most times it does not. The app system doesn't work at all. I still have to rely on the meter in my home meter box to check that the new solar system is working.

I called the installer for help and they dropped in twice to try to sort out the problem but as of late May 2021 I was still rarely able to communicate with the SMA inverter. SMA's system is very poorly arranged.

Another 5.2kW

New solar installation on our shack
New solar
The top of the chimney needs some work!
Photo 2014/02/18
On 2014/02/18 we had another 5.2kW (20 × 260W panels) installed on the roof of the shack.

The angle is not optimal, being only about 14°, although it does face almost directly north.

This job was done by O'Dea again.

So as of 2014/02/18, my wife and I have paid for the installation of:

  1. 2006/02/08; 1kW of panels on our house at Crystal Brook
  2. 2011/01/24; About 5.3kW on our daughter's house in Mandurah
  3. 2011/02/02; 1.52kW on our shack at Clare
  4. 2011/06/29; About 3.4kW on our son's house in Adelaide
  5. 2014/02/18; Another 5.2kW on our shack at Clare
That's a total of about 16kW.

After the 5.2kW system had been in place for a full year (2014/02/18 to 2015/02/18) our power meter showed that we had exported 19 times as much power from our shack as we had imported. Over the year we had exported 10.2MWh. That would have displaced fossil fuel electricity on the national ower grid and might have reduced greenhouse gas emissions by anything up to about ten tonnes.

This section added

Total solar installations

With the second installation at Crystal Brook the list of all the installations my wife, Denece, and I have paid for as of 2014/02/18 is:
  1. 2006/02/08; 1kW of panels on our house at Crystal Brook
  2. 2011/01/24; About 5.3kW on our daughter's house in Mandurah
  3. 2011/02/02; 1.52kW on our shack at Clare
  4. 2011/06/29; About 3.4kW on our son's house in Adelaide
  5. 2014/02/18; Another 5.2kW on our shack at Clare
  6. 2021/03/01; Another 5.94kW on our carport at Crystal Brook
That's a total of about 22kW. Figuring on a capacity factor of 18% and our solar power displacing gas-fired generation at 0.5 tonnes of carbon dioxide per MWh the solar that we have installed would reduce emissions by about 35 tonnes per year.

This section added

A move to Western Australia

Due to declining health and feeling that we wanted to minimise driving long distances my wife and I moved to Western Australia in 2022. Public transport in both Clare and Crystal Brook was minimal; there were good public transport options in Mandurah, WA. Together with our daughter and her family we bought a house big enough for the six of us.

The house had what must be about 3kW of solar power. As we had ample roof space I considered adding another 5kW or so. On inquiries I found that if I wanted to increase the size of our solar power system I would have to either remove the existing system or forego a $3000 government subsidy on a new solar power system.

To the present I have not made any changes to the solar power on our shared house. (We did replace a gas water heater with a heat pump water heater.)

Related pages

General solar power links

The Rainbow Power Company has a page on Solar Panel FAQs.

Wikipedia, Solar Power in Australia

Solar panel comparison links


The rating we need

Buyers of solar power systems need to know how many Watt-hours they are going to get for each dollar that they spend. Obviously this will depend partly on their site. Unfortunately rating systems such as the PTC one do not take price into account.
Panels seem to be evaluated by giving them a PTC rating. This is a contraction of "PVUSA TC", and appears to me to be quite technical.

GoSolarCalifornia has a large page in which PTC ratings are given to many makes and sizes of solar panels.

WholeSolarPower, a comparison between some 25 panel makes.