Main menu:

Position and Monitor solar panels

What IS the right angle??

Warwick Rowell

Several years of discussions with people with many years’ experience in solar work, have confused me.  So I thought I would invite readers to point out any errors in my thinking.   It is a serious invitation…  [Any readers are invited to give feedback to]

My puzzlement is about determining the right angle for solar panels, and it has a number of components.  One is to do with trackers.  Another is about how to maximise efficiency by the range over which you alter the angle of elevation of your panels.  Another is to do with the best angle for totally fixed panels.  The fourth is to do with the efficiency of the panels themselves.

We have been relying on solar power now for four years, and have gone through a number of design upgrades over that time.  And one or two changes that have been at best marginal.  Our observations of the system, and the 128 pages of data we have collected have led us to a number of tentative conclusions.

Gillian and I actually tested the efficiency of the Canon Unisolar 64W panels some years ago.  We measured the loss of power as the panels were tilted from perpendicular to the sun, through until they were at right angles to the sun.  I was tilting the panels by activating the actuator arm of the tracker, measuring the angle of the sun with a large Douglas protractor left over from yacht navigation days, and writing down the time and the angle of tilt. Gill was inside writing down the time and the ammeter readings coming off the panels. I held the panels at a fixed angle for 30 seconds, to make sure we didn’t “desynchronise”.  90 degrees, 10 amps, 0 degrees 1 amp – the data was unequivocal; as soon as you start to tilt, you lose power, and it is a straight line relationship. .. Has anyone done similar figures for other brands of panels?

So we concluded that the more frequently you can tilt your panels to keep them perpendicular to the sun, the higher output you will achieve.  We have a regime of a tilt every five weeks, with six weeks between tilts at the Soltices; June 1, to July 15, July 15 to August 22 etc. Ten tilting sessions a year take about 10 minutes each.  It has become one of the routines; checking the batteries’ water levels, drain the dry composting toilet, check the gutters in winter, wash the panels and dip the water tank in summer, white ant visual inspection.

One of our most important discoveries from all of the data recording off our Selectronic Remote Controller is that the amount of power you get when the sun is below about 16 degrees above the horizon is not worth bothering about. Our theory is that there are just not enough photons penetrating the thicker transect of the atmosphere.

This observation leads us to another conclusion which causes a lot of shaking of heads.  If in mid winter you want to optimise your panels, so they are at their most efficient angle for the longest time, you will have them at an angle of 66 degrees to horizontal (at our latitude), and this requires tilting mechanisms with considerable travel in them.

Equally, we have heads shaking in disbelief at the other end as well.  For optimal solar performance in mid summer, we have our panels as close to dead flat as we can possibly get them.  (Please excuse the geocentric references below; they are not accurate, but it makes for easier description).  This is because here (3339S  11503E) in midsummer, the sun rises from south of east, and keeps heading north, through due east at 0900, when it is 43 degrees high in the sky. At noon it is 82 degrees above horizontal, and then it starts moving south again, through due west and 43 degrees again at 1500, and then keeps going a long way south before setting.  You can work out these angles for your latitude from the chart below, which first appeared years ago as part of a REAC facts sheet.  I would have distributed about 400 copies to Permaculture students over that time.

If you put these two ideas together, you need a tilting mechanism that will cover about 66 degrees.  No, the ubiquitous “they” say; the sun only moves through 23.5 x 2 degrees!  Yes, we say, they are partially right, but that describes the path of the midday sun on the globe as the seasons change; we need to generate power for the whole day, and the sun doesn’t behave the way we were taught at school, and have taken for granted ever since. ….Insert diagram 2 here….  We all “know” ( after someone points it out to us) that south walls that aren’t shaded by too many trees can get lots of hours of sunlight in summer.   We can get nearly eight hours of sun on the south walls of this house, as we have very clear distant horizons.

Parenthetically, what we often don’t realise is that this is why passive solar design suggests we build houses long east to west and shallow north to south.  This design strategy aims to minimise the surface area and so the heat load on the east and west walls in summer, when, depending on the eave depth, the sun could be on the east wall from 0500 until 1130 and on the west wall from 1230 until 1900.  This need, more that the need for windbreaks, led the original settlers to plant trees to the E, SE, SW, and W of their houses.

When I put my idea about this wide angle of tilting to one tracker manufacturer, he was quite derogatory.  I told him I used all five changes possible from the six holes provided, for tilting his tracker between the solstices.  No, he said, you’re wrong; they’re drilled for latitudes from 3-4 degrees in the north of Australia, to 50 degrees, for New Zealand.  Sure, but you’ll optimise your performance when you use them all, says I.   No, No, you only need two at most three changes a year.   Hm, says I.

Another argument the manufacturers of trackers won’t like is that their pretty graphs showing 30-45% average more power are very misleading, because they are talking about the average increase of performance over a whole year.  But again, if you look at the sun angles, and recall my comments about how there is power loss as soon as you move from perpendicular to the sun in the first paragraph, you will quickly work out that panels on a tracker are about 80% more efficient in summer, and only about 10% more efficient in winter.  Go back to my earlier comment about how angles below 16 degrees generate almost no power.  Go to the midwinter line on the solar angles chart, and work out the angles either side of north where you have effective solar generation in midwinter.  And you will see that tracking only gives you a little extra efficiency in midwinter, and maybe none if your panels have some of what I have come to call “polar flexibility” – I am sure there is a more technical term for it, but I use it to mean only marginal drops of power in solar panels as the sun moves the first ten or fifteen degrees from perpendicular to the panels, not the linear drop we measured on the Canon Unisolar panels.

So unless your demand is much much higher in summer than winter, tracking may not be as much of an advantage as it is made out to be.  Sure, for summer solar water pumping, a tracker would be great.  But for domestic use, where the extra lighting load in winter probably just about balances the extra refrigerator load in summer, the tracker might not be worth much.  We dump power from September to April, with an extra fridge, breadmaker, sprinklers off the pump, paper shredder.. We need an occasional boost from April through to September.  Each winter so far we have used about 100 litres of unleaded fuel to run a 24 amp truck alternator connected to a 5HP Honda motor.  With a tracker.  This winter we try more panels but no tracker for the first time. We shall see what happens with a system which now has 1.25 KW of panel power.

If I haven’t been able to talk you into having a tilting device, and you are going to have a fixed array, then I think it is your monthly demand for power over the year that should be the dominant factor in determining its angle.  You need to do some local weather analysis, because the next most important factor is the amount of cloud cover and how it varies over a month, and perhaps even by each week through the year.  If we were really sophisticated, we would keep track of the differing cloud cover over the eight to ten hours of daylight of (here) the winter months, and get our own “average solar hours” chart for each month.   For instance, from the last four years of observations, we know that we will have one or two spells of a week or just over of continuous cloud cover sometime each winter, but most of the time we will have two days of 5-6/8th cloud as the fronts go through and then five days mainly clear as the high pressure systems dominate.   So we can use a generator for back up.  If we had continuous cloud cover for a month or more, as can occur 60kms south of us, then we might need to think about even more panels or wind, or something, as the cost of continuous motor generation is not worth it.

By the way, has anyone noticed how when you have skies washed by rain, with minimal dust, and lots of fleecy white cumulus, you sometimes get higher power peaks than out of a clear sky?   It seems the extra light reflected off the edge of the clouds onto the panels augments that streaming down onto the panels directly from the sun.  This surge has caught several of our people out; they needed a higher Amp contact breaker between the panels and the inverter and batteries. And they have made sure that it is in a frequently visible position, so they can see if it has tripped.

The next step is to map demand variations against supply variations.  The back of an envelope will give you enough data.  The major outcome will be how much bias you might give to winter efficiency.  Remember the caution about how little solar power is generated until the sun is reasonably high in the sky.   In the cloudy winter seasons here the sun will peak at between 32 degrees at June 21, and it will be 45 degrees at the start and end of winter; say April 21, and August 21; two months either side of the solstice.  Some simple maths shows that the average maximum will be about 38 degrees.      So maximising efficiency by setting the panels between this and 16 degrees gives an optimal setting of 27 degrees.  And this is NOT the latitude above horizontal (32), nor even the latitude plus ten (42), which are the most recommended settings.

Now lets look at what happens in summer, if we set our panels for a winter optimum; compare our setting with both the recommended ones.   The difference in angles are marginal in terms of the extra power loss from optimising your winter performance, in my view.  In summer, we will have generally clear skies, and by my rough calculations, at least three times the solar hours that you will have in winter.  If the optimum angle for summer is flat and your panels are “linear”, with no polar flexibility, then your panels fine tuned for winter would generate about 40% of their maximum power in mid summer.  This is more than enough premium for the extra fridge load, particularly if you do something sensible like making sure the fridge is clean, and there is a lot of air flow around it, and plenty of space above it for the heat to disperse.

So there you have it: My opinion is that the optimal angle for fixed panels is 27 degrees (for Lat 32S or N).  If you tilt them, you will get best performance by tilting them over a range of 60 degrees or more about ten times a year.  Use trackers for water pumping in summer only, after doing your sums on panel cost v tracker cost and thinking about vulnerability to wind and rock throwing children.   Your comments would be most welcome….

Warwick Rowell is a management consultant and a Permaculture Designer. He lives near Dunsborough in WA.  For more details, ring 0447994885 or email

Write a comment

You need to login to post comments!