Jonathan >

A solar-pv system was installed on the croft (connected to the electrics at Carrick – The Blue House, and thereby the national grid) in the spring of 2015.  There are 16 panels in all, totalling 4kW maximum output. Half of these are arranged along the roof of the croft store – which faces almost due south : this is referred to as the ‘roof-string’. The other eight panels are arranged in two rows, fixed to frames of galvanized steel, bolted to the bedrock : together, these two rows are referred to as the ‘ground string’. So far, so unremarkable.

It was a couple of years ago that I first noticed how the ground string consistently under-performed the roof string – despite having exactly the same orientation and tilt, and particularly so in winter.

Solar PV ground string – back row removed

Eventually I got myself organized and undertook experiments by rearranging the connecting cables so as to try and isolate where the problem lay. What I discovered was that the fault was not with one or more specific panels, but with the back row of the ground string : if I swapped panels between the front and back rows, the front row still performed well, the back row poorly.

That wasn’t entirely a surprise to me.

Due to limited space available, the two rows had been erected less far apart than is recommended for our latitude, and it was understood that the front row would, in the low winter sun, cast a shadow across the lower regions of the back panels, the effects of which varied from nothing at the spring and autumn equinoxes, to all of the lower half of the back row at mid-day, mid-winter.  I theorized that this arrangement might, at worst – at mid-winter, reduce the total output by up to a maximum of one half of four of the sixteen total panels, ie by one eighth.  The reality, however, was a reduction of about one third of the total output.

The shortfall was, logically, worst at mid-winter, though as output then is very small (lowest sun, weakest sun, most often cloudy), but unfortuantely it’s also when any electricity we produce is most needed – heating and hot water being by electricity.  We needed this fixed, but getting a specialist to come to the island would woud take a huge chunk out of future profits, and then there would be the cost of any remedial work.  I needed to understand what the problem was in order to evaluate the best way of getting the problem fixed.

For a long while, the technicalities were swirling around in my head, round and round, never getting forward, and eventually I realized I needed to take a more practical approach – to experiment further. I had something specific in mind.

Rearranging cables so as to bypass the back row of panels entirely , the result was immediate, unexpected, and absolutely decisive.  With the four partly-shaded panels completely disconnected from the installation, one would of course expect that output would decrease.  In fact, it increased : the ground string (reduced to only the front four panels – unshaded) was now underperforming the roof string (and its full complement of eight panels – all unshaded) by exactly 50% : which is, of course exactly what logic would expect.  With the back row connected again, the output of the ground string was about one sixth that of the roof string.

Wow!  That could only mean one thing : that the back row, when receiving less sunshine than the front row (due to the front row shading the back row) was absorbing energy (generated by other panels) instead of generating it (and presumably disspating that energy as low-level heat waste). Why it would do that I don’t know for sure, but my guess is that a component of the panels called a blocking diode does not do the job it is supposed to do.

As a temporary solution, I removed the four solar panels of the back row and stacked them in the croft store. The remaining 12 panels (four on the ground, eight on the roof) would go on to consistently generate more electricity than the full complement of 16!!  That being the case, finding a permanent solution was not a high priority – I had more urgent matters to deal with.

Another 18 months or so later – early this spring, and having at last ticked off many of the bigger never-ending jobs about the house, garden, holiday homes, croft, websites – etc etc, I turned my attention back to the problem of getting full output from all 16 solar panels – and the original galvanized steel support frames, which by now were corroding alarmingly in the salty sea air.

This slideshow requires JavaScript.

My mind was going round and round in circles (again!), when it struck me how I could use the existing front row frame as a temporary support whilst I built a new timber structure around it. This would allow me to get the panels in place and producing electricity in just one morning : and so it proved to be. Completing the structure and cladding it to give a reasonably pleasing appearance was completed over 14 weeks, working three or four mornings a week.

I’ve yet to apply wood stain/preservative, and admittedly that will have to be repeated every few years, and even so the wood (particularly that closest to the ground) will deteriorate with time ; but it’s been assembled with stainless steel bolts – not nails or screws – such that the entire structure can be dismantled component by component, and repairs or replacements made at moderate cost and by my own efforts, without the need to pay for any tradesperson.

This slideshow requires JavaScript.

It might have been a while a-waiting, but worth it for a solution that pleases me so much – and cost so little.