Rooftop PV Panels Point Where the Roof Points

On several previous occasions Euan Mearns has fulminated about this photo he took of roof-mounted solar PV panels in Aberdeen. If you’re going to do something as dumb as installing solar panels at latitude 57N, he argues, at least point them south. Don’t point them east.

So why are these panels pointing east?

Because that’s the way the roof points.

One would think that in Aberdeen, where solar PV load factors are less than 10% to begin with, PV panels would be pointed in the direction that gives maximum output, which according to NREL’s PVwatts calculator (which allows for local weather conditions as well as sun angles) is ten degrees west of south tilted 35-40 degrees south, an orientation which allows the panels to generate 27% more kWh/year than they will when pointed east. But few rooftop PV panels in Aberdeen or anywhere else for that matter are optimally aligned. Most are bolted flat onto the roof regardless of which way it points.

In this post we will take a photographic tour of some of the outcomes.

This array in the Southwest US is typical of many rooftop installations. The homeowners wanted a larger system than one roof segment would accommodate, so they bolted the panels flat onto two roof segments angled at 90 degrees to each other. Assuming that one of the segments points south then the other will be operating at only about 80% efficiency:

Next is a photo of a subdivision in Japan, taken I would guess around the time Japanese utilities imposed a moratorium on new solar hookups to avoid destabilizing the grid. The roofs are plastered with panels pointing ninety degrees apart and the only house with panels that aren’t bolted flat to the roof is the third from the left in the first row:

In this aerial view of houses in Germany every panel also lays flat on the roof:

As do most PV panels in the UK. (These houses are in Nottingham. Panel tilts don’t change but azimuths vary gradually through about 90 degrees. This will smooth out solar generation during the day but still won’t make the sun shine at night.)

Now for some individual cases. This pyramidal rooftop array in Japan makes use of all the available roof space by pointing panels east, south, west and north. According to PVwatts the north-facing panels will generate only about 60% as much power as the south-facing ones, all other things being equal.

Somewhat more efficient is the Atlantic Health Training Center in New Jersey, which at least omits the north-facing panels:

But note the flat panels on the roofs at left. Owners of buildings with large flat roofs commonly lay their panels flat even though it cuts output. According to PVwatts these panels on the roof of the US Department of Energy are 13% less efficient than correctly-oriented panels. And DOE has a website extolling the need for energy efficiency:

The huge solar array on the roof of the Goodyear-Dunlop plant in Phillipsburg, Germany is similarly afflicted:

And then there are the artistic but equally inefficient PV panel arrays, such as the bulge on top of London City Hall:

The roof of the Vatican, which needed specially-shaped panels to make everything fit:

The National Stadium in Taiwan, which “with the assistance of construction geniuses from Japan, US and R.O.C. … has become a proud milestone in green architecture application.” It has a load factor of 12.5%, about 5% less than optimally-oriented panels at this latitude would have:

And the “Photovoltaic Wall” in Barcelona, Spain. The curve is presumably there to relieve the monotony:

The Barcelona example shows us that people don’t stop at bolting PV panels to the roof. They bolt them to the wall too, usually suffering even larger efficiency losses in the process (the estimates below are again all from PVwatts). I can’t get a number for Barcelona because I don’t know which way the panels are pointing, but the panels on the south wall of the Solar Building in Anchorage, Alaska generate only 78% as much power as they would if they were angled skywards at 45 degrees:

The 15kW of south-southwest-facing thin-film panels implanted in the wall of City Hall in Flagstaff, Arizona, generate only 71% of the power they would generate if they were optimally aligned:

Or at least they would if it weren’t for the tree the city has planted in front of them:

As well as being spectacularly ugly the vertical panels on the wall of Colorado Court apartments in Santa Monica, California are 31% less efficient than properly-aligned panels:

And then there’s the CIS Tower in Manchester, which is festooned with over seven thousand 80-watt PV panels on its south (actually south 25 west), east and west sides. Relative to the optimum panel angle the south-facing panels are 70% efficient, the east-facing panels 63% efficient and the west-facing panels 59% efficient, giving an overall efficiency of 64%. With solar load factors averaging around 10% at this latitude we would therefore expect that this £5.5 million, 575.5kW system to operate at a load factor of only about 6%.

But it doesn’t even achieve that. According to Wikipedia and other sources the load factor is less than 4% (575.5kW capacity, average generation 20-21kW), a number low enough to rank the CIS Tower among the world’s most inefficient electric power generating plants. Yet in 2006 the UK Department of Trade and Industry still chose it as one of the “10 best green energy projects” of 2005.

It would nevertheless be a mistake to think that all PV panels are bolted onto rooftops or walls with a cavalier disregard for orientation. Some people get it right:

And those who are motivated to optimize occupant comfort will usually make sure the panels are angled in the right direction:


For those interested in how much difference PV panel azimuth and tilt makes I put together the following table from the PVWatts calculator. It shows the percentage of generation obtained at different panel azimuths and tilts relative to the maximum of 100% at azimuth = 180 degrees (south) and tilt = 30-40 degrees relative to the horizontal. The numbers are for Gatwick Airport (PVwatts insists on a specific location) but they are broadly applicable at all mid-to-high latitudes.

This entry was posted in Energy and tagged , , , . Bookmark the permalink.

44 Responses to Rooftop PV Panels Point Where the Roof Points

  1. I fail to see the point of this long article. The requirements of any religion are adherence to the faith, not science. The panels are a religion and follow the faith so what is the objection?

  2. Yvan says:

    PV cost is now dominated by the support structure and the installation process. This is why it might make sense of now align them optimally.

  3. It doesn't add up... says:

    Those concave panels in Barcelona presumably fry a few birds, like the Walkie-Talkie building. Do they get extra subsidy for doing meals on wings?

  4. Junior says:

    Could you post a more irrelevant article?

  5. ristvan says:

    You are right about roof orientation.
    But miss the compounding effect of panel type efficiency. Which makes the whole thing worse than you have portrayed here. See a recent Guest post at Climate Etc. on solar grid parity for explanations.

  6. Euan Mearns says:

    Roger to clarify my point about the panels in the picture. The same building has a large S facing roof! Orientation of panels in Aberdeen seems random. I’m told it makes no difference since we receive so little direct sun light and they are working mainly on back scattered photons.

    Another favourite of mine was a new housing development on the edge of town with solar hot water mounted on a NE facing roof. Last time I saw these they were covered in moss.

    • Joe Public says:

      Hi Euan
      See my comment below.
      I strongly suspect the panels on the Aberdeen commercial building were installed not simply to generate electricity, but primarily to generate credits to enable the building to obtain Planning Permission, and so be built.

      I don’t know the time or date that photo was taken, but it shows some of its panels in shade from part of its own roof!

      • Euan Mearns says:

        You are probably correct Joe. This combined with cowboy outfits that get paid regardless of the job they do.

        I once had a British Gas engineer try to sell me a condensing boiler on grounds that it condensed out the CO2.

        We still have an old inefficient boiler. Its on a service contract and gets serviced every year. For about the last 10 years we’ve been advised to get it replaced. Its not the cost of the boiler but everything else that has always put us off.

        Last service visit the engineer actually monitored gas flow at our meter when boiler was functioning and found it was not nearly as inefficient as we had been told. What’s more, he told us the new condensing boilers keep breaking down and don’t last. And he told us our cast iron boiler would last forever and that the energy savings we would make having it replaced would never repay the CAPEX at current gas prices.

        A breath of Fresh Air.

    • I suspect what may have happened here is that the building is divided into offices and the panels were installed by the people in the office directly below. That would explain why they are a) pointing the wrong way and b) in the shade.

  7. Joe Public says:

    In the UK, new or refurbished buildings only get Planning Permission if they incorporate certain energy-efficiency measures. The key word there, is “incorporate”. Those measures don’t have to be efficient, effective or used!

  8. Flocard says:

    I would like to comment onf this article.

    I live in the south of France 43° latitude in a sunny village near Montpellier. My house has a large section of roof (average slope 20° as typical of mediterranean houses) plus a large terrace facing south.

    At one point I considered having solar panels to heat domestic water (no electricity) The companies I contacted all wanted to have the panels flat on the roof.

    I pointed out that my needs for hot water were about the same all year round. Thus what was crucial was to have the panels working optimally during the winter season while not producing too much heat during summer, so much heat that I would have to consume cold water and electricity to protect the system from overheating.

    Thus the right inclination was facing south and perpendicular to sun rays at the winter solstice namely an angle of 43-23=20° to the vertical.

    Finally I could not do it even on my terrace because it would have required to protect the panels so that no person could touch it. The cost plus the overcost did’nt justify the investment.

    To conclude, it seems to me that the vertical wall of solar panels on the building in Barcelona (a very sunny city about 38 ° latitude) does not achieve an optimal production year round. What it achieves is an almost FLAT electricity production year round, which may not be too bad an idea as it minimizes storage requirement which are confined to a daily (not seasonal) production.

    • Yvan says:

      Same issue here in Quebec. Solar thermal is better done on wall because they overhead in summer which cook the glycol. THis is also better for the maintenance.

    • PV panels and solar water heater panels don’t work the same way. Water is heated by long-wave (IR) solar radiation and gets hot only in direct sunlight, but PV panels generate electricity from higher-frequency solar radiation in and around the visible light spectrum, which we get during the day whether the sun is shining or not. (My panels start to twitch when the sky lightens in the morning before the sun is even up and continue to pump out electricity, although not much, under a heavy overcast.) Because of this water heater panels generally have to be angled more steeply than PV panels.

      • Yvan says:

        Thermal collector tend to be more brandband than PV that’s all.

        • It’s eleven in the morning here and overcast. My rooftop solar panels are putting out 500 watts. The black tubes in my rooftop water heating panels are stone cold.

      • Luís says:

        Water is heated by long-wave (IR) solar radiation and gets hot only in direct sunlight, but PV panels generate electricity from higher-frequency solar radiation in and around the visible light spectrum, which we get during the day whether the sun is shining or not.

        I am really baffled by comments like these. I take you as an intelligent person, but you often recur to deceiving elaborations to defend your points of view. In this particular case it is totally unnecessary.

        To a great extent this attitude reflects on the way you opted to bash Tesla. Questioning the media hype is absolutely necessary, but the way you did it in the ends serves no purpose.

        • JerryC says:

          Are you saying he’s wrong? If so, what’s the truth? Solar water heaters work the same way as PV panels, is that it

    • Here’s a graphic comparing output from the Barcelona panels, assuming they are tilted south at an average of 80 degrees, with what they would generate if they were tilted at 30 degrees. System capacity is assumed to be 1 kilowatt and efficiency 100%:

      • Yvan says:

        Yes because if you are more tilted you miss the summer peak. However, depending of the local energy consumption you might want less efficiency in exchange of a more uniform production because if you are producing at the wrong time you production might need to be curtailed.

        • Why would you want less efficiency in exchange for a more uniform production? All you’re doing is eliminating any possibility of selling your summer surplus.

          • Luís says:

            Interesting comment. On your posts bashing Tesla you completely ignored this. But now to bash PV in general it suddenly became important.

          • Flocard says:

            It is obvious that moneywise getting the maximal yearly production is least for the producer if there is a favorable feed-in tariff. A large number of ludicrous anecdotes to the effect that money and not environment is the key driver of PV abounds in France.

            On the other hand, if the common interest is taken into account (of course I am dreaming, but let’s pursue just in the interest of the discussion ) flattening the production curve might be an option to consider. As I said the energy storage problem become a daily problem instead of being both a daily and seasonal problem

            For exemple, I believe I heard (this formulation shows that I have not checked what I am going to say) that in Germany on sunny days in order to avoid the frequency to exceed the limit of 50.2Hz, the grid has to disconnect some solar panels. If so I do not know how the grid decides which panels have to be disconnected. I have heard that in turn this suddendisconnection may generate local stability problems when the PV production falls.

            I know (here I know) that in the French caribeean islands where already too many PV panels are installed the local grid (a subsidiary of EDF) disconnects PV panels on the basis “last connected to the grid, first disconnected from the grid when its safety requires it”.

      • It doesn't add up... says:

        Perhaps this influences the demand side of the equation:

  9. Euan,

    You missed one of the prototypes of solar Pv systems very efficiently used for propaganda and very inefficiently used for electricity generation. They are in the roof of the Reichstag. Google ‘aerial view of German Parliament’ and see what they were capable of doing.

    I am very much with some opinions given here that many buildings are looking for construction permissions, by just putting “something renewable” as a hat and they then get without any problem the stamp on the permits by brainless, if not corrupted, civil servants. I am somehow relieved to learn that it is not only Spain, doing stupid things and making wrong decisions.

  10. disdaniel says:

    PV modules represent under 30% of the cost of getting solar installed most places… so at least 70% of the expense is everything else. That means it could well be cheaper to install solar sub-optimally from an efficiency perspective–if in doing so you knock a big chunk out of the “everything else” cost. (I am not saying that is what was done in these cases…some of the photos do make one cringe…)

    I doubt everyone here drives their car only at 55mph because that is the optimal engine speed for efficiency?

  11. Dave Ward says:

    The “Photovoltaic Wall” in Barcelona, Spain is here: Lat 41.565701 Long 2.007040 I make it as pointing @ 152 degrees – use Google Earth and you can see the panels on Street View.

    But there’s another aspect that many of these installations have in common (particularly the first picture): partial shading of some of the panels. If one (or more) panel(s) in a series string is shaded it will cause a significant reduction in the output of the entire string. I’ve seen installations where they would get better results if some panels were feeding different inverters or separate inputs to a multi inverter.

    • Thanks Dave:

      Turns out that the Barcelona array actually points south 30 east, meaning that it will operate with 68% efficiency relative to an optimally-aligned array (azimuth ~180, tilt ~30 degrees south)

  12. roberto says:


    a modern and scientific way of assessing the potential of rooftop urban PV installations is exemplified in this paper:

    … which uses LiDAR techniques to automatically detect the orientation of all roofs and allows the calculation of mutual shadowing.



  13. Dave Ward says:

    It’s obvious that many installers are frequently described as “cowboys”, judging by the ridiculous placing of panels that most of us see on a daily basis. If you put the following into Google Earth: 52.629848° 1.359007° and zoom in, it takes you to one end of a row of three terrace houses. These are former “council” dwellings, currently run by a housing association, although some are now in private ownership. There are no panels visible because the GE imagery is several years old. However this property has them installed on BOTH sides of the roof. The SW facing ones are no doubt quite effective, but the 6 NE facing panels are in shade for the bulk of each year. In winter, by the time the sun has risen above the horizon it is already far enough South that the panels never see it. In summer the large tree blocks it for about half the time it would otherwise be visible… I doubt that side manages even 10% of the output provided by the SW facing panels, yet the cost of installing them must have been the same.

    Note: I have the greatest respect for “real” cowboys and the work they do, often in harsh conditions. It’s perhaps unfortunate that this name is regularly misused.

    • Dave: The table at the end of the post applies in this case. The SW facing panels will be about 90% efficient and the NE facing panels about 60% efficient assuming a ~50 degree roof angle.

      When I put together this table I was surprised that panel angle doesn’t make a bigger difference. I think it has to do with the fact that a lot of the power generated by PV panels comes from light reflected off clouds and not from direct sunlight.

  14. Eoin Licken says:

    I’m surprised nobody has mentioned aesthetics yet. Here in France the law states that PV installations MUST be integrated into the plane of the roof (the law actually states that the roof must lose its waterproofing if the panels ar removed). The idea was twofold: to limit the amount of subsidised installation, and to encourage innovative installation techniques.

    Am I alone in thinking that panels bolted above roof tiles, or raised at angles to existing rooftops, are ugly? Given that PV costs are falling, but that heir lifetimes are 20+ years, I’m happy to sacrifice some efficiency for improved aesthetics. I think that’s what’s happening in several of the photos.

    Another point: using ugly PV to shade existing ugly buildings (the courthouse in Santa Monica) is a case of why not?

    • Flocard says:

      In this message it is said that in France “the law actually states that the roof must lose its waterproofing if the panels ar removed”.

      This law is actually another proof of the stupidity of French civil servants working in the environment ministry whan they prepared the law (which of course later had to be approved by the law makers who rarely read or understand such details)

      As a matter of fact the only law states what you have to achieve at the start of the installation if you want to get the most favorable feed-in tariff (above 6 times the market price). Whether it is still the case after few months is not important Nobody comes to check.

      Clever peoplle have not missed this point. As a result all over the French countryside you can now see hangars (four poles, a frame slanted in one direction (south with the appropriate angle) and supporting a coverture of solar panels. Generally there is nothing stored under the hangar because after few months rain tends to leak between the panels (there is no real roof and the solar panels being used are not mean to be rainproof) . Thus these hangars have no other use than the collection of money from the bill of customers via a tax called CSPE.

      For instance I have a picture showing such a brand new hangar with a sign “for rent”. A proof that the hangar had no use as a hangar proper to begin with.

      At one point it was shown that a Neuilly (a rich suburb of Paris) based company owned over fifty such “agricultural hangars” all over France.

      Similar stories abound, The best one recorded in my region being a “green house” covered all over (roof and three of its slanted walls,) by solar panels which was announced by the farmer to be designed for the culture of asparagus (a plant that does not need light to grow).

  15. Stuart says:

    I imagine the cost associated with alignment simply is not worth the extra efficiency.

    The extra cost of designing and installing bespoke supporting structure for every job and providing a calibrated installation to achieve the optimum angle of incidence would probably more than double the cost of the system. Especially for retrofitted systems.

    Therefore if bolting the panels straight onto the facade is still better than 50% efficient then that is the way to go.

    The only efficiency that matters is cost efficiency because the world is too complex for command and control solutions.

    • The metal framework that points my panels in the right direction maybe added 5% to the cost of the installation.

      The optimum angle of incidence is the same over large areas and doesn’t require that individual installations be “calibrated”.

      With load factors in the 10-20% range PV panels are inefficient enough to begin with. Cut the efficiency in two and they just aren’t worth installing.

Comments are closed.