Solar · Mounting and tracking
Mounting decides four things at once: which direction the array faces, what angle it sits at, how hot the cells run, and whether anyone can reach them in February. On a property-scale system, all four are worth deciding on purpose.
What is at stake
The array on the sizing page assumed a panel pointed usefully at the sun. Mounting is where that assumption either holds or quietly stops holding.
Two of the four consequences are obvious. Direction and angle set how much energy the array intercepts, and the sizing arithmetic falls apart if they are wrong.
The other two get overlooked. Panels lose output as they heat, so airflow behind them is worth real watts on a summer afternoon. And an array nobody can safely reach is an array that does not get cleaned, cleared, or repaired.
Photovoltaic modules produce less as cell temperature rises, at a rate given by the temperature coefficient of power on the datasheet. It is a modest effect per degree and a real one across a hot roof in July, which is why gaps and airflow behind panels are part of the design rather than an afterthought.
Three configurations
Uses space that already exists, needs no foundation, and keeps the array out of the way of everything else the property does.
Inherits the roof's pitch and direction, runs hotter for lack of airflow, and cannot be cleared of snow safely. Also ties the array's life to the roof's, which matters if the roof is older than the panels' warranty.
Free choice of direction and angle, full airflow behind the modules, and everything reachable from standing height. Seasonal tilt adjustment becomes practical.
Costs foundations and racking, occupies ground the property may want for something else, and needs enough row spacing that panels do not shade each other in winter.
Follows the sun through the day, and on dual-axis units through the seasons, raising daily harvest substantially over a fixed array of the same size.
Introduces motors, drives, sensors and controls, all of which can fail and all of which must survive wind loading. The next section works through whether the gain is worth it.
For a property-scale off-grid array, ground mount is the usual answer. Not because roofs are bad, but because the two things ground mounting buys, free orientation and safe winter access, are exactly the two things a system sized against December depends on.
Tilt and direction
A grid-connected system optimises for total annual production, because surplus in June is sold back. An off-grid system optimises for the worst month, because surplus in June is wasted and a shortfall in December is a problem. Those two goals point at different angles.
The balanced year-round setting, and what PVWatts assumes unless told otherwise. At 43 degrees latitude that is a 43 degree tilt.
Steeper, facing a lower winter sun more directly. At 43 degrees that is about 58 degrees. Gives up some summer production, which a worst-month system has in surplus anyway, and sheds snow far better.
180 degrees in the northern hemisphere. East or west facing costs roughly 15 to 20 percent against south facing. Note that true south is not magnetic south, and in some regions the difference between them is wide enough to matter.
PVWatts accepts tilt and azimuth as inputs, so the honest way to settle this is to run your location twice, once at latitude and once steeper, and compare the December rows rather than the annual totals. That comparison takes two minutes and is more useful than any rule of thumb, including the ones above.
Tracking
Tracking works. That is not in dispute. The question is whether it is the cheapest way to buy the extra energy it delivers, and for most properties the answer is no.
Gains are commonly quoted between 25 and 45 percent, varying with latitude, climate, and whether the tracker moves on one axis or two. Take the middle of that range and apply it to the fifteen panel array from earlier in this section.
Fifteen panels, 6,000 watts, with a 30 percent tracking gain produces what a fixed array of about 7,800 watts would.
Twenty fixed panels, 8,000 watts, gets there and slightly past it. So the real choice is a tracking mechanism against five more panels.
Five panels have no motors, no controls, no bearings, no wind-load failure mode, and nothing to service. They also keep producing when one of them fails, which a single-tracker array does not. Panel prices have fallen far enough that this comparison, which once favoured tracking clearly, now usually does not.
When usable ground is genuinely limited, when the load peaks early and late rather than midday, or when someone on the property enjoys maintaining mechanical things and will actually do it.
A tracker cannot manufacture sun that is not there. In the overcast December week that sets the array size, tracking gains shrink because most of the light arriving is diffuse.
Snow and access
Not less. Nothing. In snow country this turns the worst month from a dim month into one that can include days at zero output, which is a storage problem as much as a panel problem.
A steep winter tilt does much of the work by itself, letting accumulation slide before it sets. That is the second reason to bias the angle steeper on an off-grid system.
The rest is access. Panels reachable from the ground with a soft brush get cleared the morning after a storm. Panels on a second-storey roof in January do not, and nobody should be up there trying.
A ground mount set too low buries its own bottom edge in the snow it sheds, and in drifted snow besides. Mounting height is a design input in snow country, decided from local snowfall records rather than from the racking catalogue.
Limits
Wind uplift and snow load on a roof array are engineering questions with local code answers. An older roof in particular needs an assessment before anything is fastened to it.
Ground mount footings have to reach below the local frost line and resist uplift. Soil type and frost depth are site facts, not catalogue numbers.
Ground arrays are structures. Many jurisdictions treat them accordingly, with setbacks, height limits, and in some cases review before installation.
Multiple rows shade each other, worst at the low winter sun angle that matters most. Spacing is calculated from the sun's altitude on the shortest day, not from convenience.
Common mistakes
Sensible for a grid-tied system, wrong for an off-grid one. Annual optimisation quietly trades December output for June output that has nowhere to go.
A compass does not point at true south, and in parts of the country the gap is wide enough to cost real production across a whole array.
Rows that clear each other in September shade each other at noon in December, in the exact month the array was sized for.
The gain is real, but it is usually available more cheaply and more reliably as additional fixed modules. Run that comparison before committing to moving parts.
Next
Distance from array to controller is set by where the mount ends up, and that distance drives conductor sizing. These guides are in progress and will link here as they publish.
Conductor sizing over the run the mount location creates, plus overcurrent protection and the permit path.
Short days, snow cover, cold voltage, and the tilt decision made on this page.
What to do when roof space, shade or ground runs out before the array does.
Back to Solar at Property Scale, or to controllers and inverters for the string voltage that shapes array layout.
Sources
Tracking gain percentages describe a range reported across latitudes, climates and tracker types rather than a figure for any particular site. Model your own location before deciding.