Thoughts on Solar

Our first three installments saw us secure our rebate reservation, successfully pull our permits, and deal with a host of challenges on the ground.  But now the real fun starts - in this episode we will document the heart and soul of this project, “Up on the Roof!”

Projection

Solar projects actually begin on a computer screen as the designer tries to map what is known about the roof, the utility service, and the client’s needs into a coherent proposal.  As the project progresses through the rebate and permitting processes, that design is refined - and as we have seen, sometimes altered.  But the trick of any implementation is to go from the designer’s plan to an actual system on the roof - starting with getting the attachments in place.

Nearly ancient methods, tape measures and chalk lines, are the essential tools in this process.  Since a solar array is essentially a grid, the trick is to project what is on the plans into a corresponding grid on the roof.  Precision and accuracy are the key to making this work, but roofs are notoriously inconsistent places!  What seemed to be square, isn’t always.  What appeared to be flat, actually has its own peaks and valleys.  While our projection onto the roof proved easy enough, we were about to discover that what you see - or were told - isn’t always what you get!

There’s What You Plan and What You Get

Our biggest design challenge had been the need to account for the somewhat unusual roof construction that we had to accommodate.  In particular, our underlying roof structure was a  20 gauge, type B steel deck, overlayed with multiple layers of plywood, foam insulation and roofing materials.  Given the thickness of those layers we had determined that we would need to use four,  8-inch-long, self-tapping screws to secure our “FastFoot” anchors to the roof.  We had purchased hundreds of those screws - along with a top-of-the-line Hilti cordless driver - to do the job.  But something wasn’t right.

As we started making our first few attachments it was clear that not all of them were reaching the steel deck!  Apparently in some places the actual thickness from the roof to the deck exceeded the 8″ reach of the screws.  Visual inspection from a scissors lift inside the building confirmed what we suspected - clearly not all of the screws were penetrating the deck, yet in other places, all four screws penetrated without difficulty.  There was only one solution to the problem of our inconsistent roof - longer screws!

Fortunately, we were able to order some 9″ screws from the manufacturer - the longest that they made.  They did the trick - now we could be certain that every FastFoot plate was properly secured.

Lean on Me

The changes to our plans imposed during the permitting process meant that we were very tight on space.  At the top of our array we had to install 3 sub-panels, each of which had to handle three branch circuits that made up that sub-array.  Our original plan was to build a triangular cross brace out of unistrut to support the sub-panels. Unfortunately, given our close quarters, the solar panels needed to come right up to the supports for the sub-panels - a cross-brace system would take up too much space.

Instead, we designed a set of steel braces that were bent at precisely the angle that we needed - 103° - to allow our sub-panels to be perfectly vertical on our 13° sloped roof.  The design was easy, but could we get them fabricated fast enough to keep the project on track?  We knew of a small metal shop near our offices and we took the design to them - yes, they said, they could produce the six parts that we needed for $100 and they would have them in the morning - would that be soon enough?!!!

This turned out to be a very elegant solution to our problem.  Using two FastFoot anchors, we attached unistrut to them and then bolted our braces to that.  When combined with the rigid conduit feeding the sub-panel, we ended up with a very solid solution. 
Next problem!

Need a Lift?

The roof of our building was reachable by a series of three ladders traversing three different roof levels.   While this was acceptable for getting personnel to and from the roof, it would never work for transporting hundreds of feet of rails, to say nothing of 209 solar panels!

Enter the boom lift - the same one that was unceremoniously dropped off for us by parking it under a No Parking sign!

Whether transporting rails, solar panels, or the Enphase micro-inverters as you see in this picture, the boom lift provided us with an efficient means of moving large amounts of gear up to our work site on the roof.  Operating a device that articulates in multiple dimensions in relatively tight quarters takes skill and great attention to detail.  (It also makes for some pretty cool looking photos!)

Connect the Dots

Once the rails were installed, the Enphase micro-inverters could be mounted and the process of running a continuous ground wire and the creation of the Enphase map could begin.

Since this was an Enphase system, we would be able to monitor the performance of the array down to the individual solar panel/micro-inverter pair.  (Indeed, this monitoring ability was a key selling point for the system to the school as it nicely meshed with the school’s educational mission - more on that in our upcoming video!)  Each micro-inverter has a serial number that was carefully peeled off and affixed to a “map” that showed where each inverter was located on the roof.  As part of the commissioning process, we transferred the map data onto the Enphase website and built a true representation of how the system was laid out on the roof.

Now all we needed was to install the solar panels themselves!

Careful attention to detail during this last step is rewarded with an array that aligns precisely and fits as planned.  Using the Enphase Envoy and a laptop computer, we could verify that each and every panel was properly connected and functioning as expected.  We could be confident that there would be no surprises that would need to be resolved later!

That gave us one last task for the boom lift - finished photographs.  Here’s our favorite:

Three inspections later - fire, building & safety, and PWP - and we were ready to officially go live.  Here’s how the system appeared on the Enphase Enlighten website one recent sunny day:

The students at Westridge will be able to analyze the performance of this system for years to come, providing a first hand experience of how renewable energy works and can make a difference in our lives - what a great lesson to learn!

Chapter the Last

Which brings us to the end of this series - almost.  In our final installment you will see the video that we have produced for this project and you will hear from the Head of Westridge School, Elizabeth McGregor, Facilities Director Brian Williams, and three wonderful students talk about how this project plays into the larger mission of the school.  You won’t want to miss it!

In parts 1 and 2 of this series, we documented what went into securing the rebate reservation and the process by which we secured our permits.  Now the actual work could begin - and that work starts on the ground.  So in this Part 3, we will look at the staging that was required for this production and take a close-up look at some of the heavy lifting that was needed far below the array itself.

A project the size of what we were going to install at Westridge - 52.3kW - involves thousands of parts, all of which not only must go together properly for the system to work and be safe, but they must arrive in a timely fashion!  For example, here’s just a sample of the parts that were needed for this job: hundreds of FastFoot plates, thousands of screws, hundreds of flashings, standoffs, and flange connectors, dozens of rail sections, mid-clamps, end-clamps, ground lugs and splices, to say nothing of 209 micro-inverters and solar panels!  Collectively these products came from five different distributors in four different states.

Needless to say, not everything goes as smoothly as you might like when you are pulling together all of these pieces.  UPS likes to brag about Logistics, but we found some of their logistics to be highly illogical.  Such as their sending two shipments that were sitting in an LA warehouse on a frolic and detour down to San Diego for the weekend, instead of driving them the twelve miles up the road to our job site!

Equally baffling were the folks who delivered our boom lift to the job site late on a Friday evening without even a phone call and just parked it out on the street - in front of a No Parking sign!

Seriously…

Would you leave this…

Here?  Right - neither would we!  (And yes, the keys were in it!) The unscripted appearance of our boom lift prompted a puzzled call from the facilies director at Westridge:

Brian: “Were you folks expecting a boom lift to be delivered?”

RoS: “Yes, they are delivering it tomorrow morning.”

Brian: “Well, it’s here - and they left it on the street next to a No Parking sign.”

RoS: (Eek!) - “Really?  We’ll be right there!”

Like I said, not everything can go exactly as planned, but soon enough, everything arrived and in good condition.

Our staging area was set with:

LG Solar Panels

Enphase Micro-inverters

Unirac Solarmount (Evolution) racking parts

And lots of wire!

Transformation

Our first main task on the ground, now that everything was at hand, was to install our transformer.  This project required a transformer to step-down the voltage from the utility service (480 volts, three-phase) to the voltage that would be used by our micro-inverters (208 volts, three-phase).

Setting the Stage

Our transformer was a 700 lb beast that had to be installed on a concrete pad (that we had to pour) in the equipment storage area on the East side of the building.  To secure the transformer to the pad, we would imbed bolts into the pad and then maneurver the transformer on top of the bolts and anchor it with washers and nuts.  Two key challenges there - first was to guarantee that our bolts were precisely positioned in the pad since the transformer gave us very little margin for error.  Second was to get the transformer in place on top of the bolts without damaging them.

We solved the first problem by drilling into the existing concrete and securing our bolts into the ground with heavy duty anchors - as you see here with the framework for the pad surrounding them.

Then, when we were ready to fill in the form with concrete, we added some framing at the top to try and keep the bolts as plumb as possible, as you see here:

The Big (Not So) Easy

That took care of problem number one, but what to do about problem number two? Now that the pad was dry, the challenge became getting our transformer into place.

Our solution would make any student of ancient cultures proud - we crafted a wooden platform over the pad and slid the transformer from its pallet onto the platform.  Then we lifted each edge, one at a time, and placed blocks of wood under each corner.  That allowed us to remove the platform and then begin lowering the transformer over the bolts by carefully removing a block at each corner.

As we removed each block, the transformer came closer to the bolts protruding from the pad.  We could push the transformer - gently - so that it aligned with the bolts.  Ultimately, the last block was removed and the result was a complete success with just the right amount of angst along the way!

But as you can see, we really didn’t have much margin for error!

The Art of Conduit

In addition to our transformer, there were several other pieces of gear that had to be mounted on the ground including a 200 Amp sub-panel, two disconnect switches and a performance meter.  Linking them all together is the conduit through which our conductors would be pulled.

Pasadena requires rigid metal conduit (instead of EMT) to be used for solar power systems wherever it is accessible on the outside of a building.  That offers some additional safety, but it comes at a cost - especially given that we were using 1.25″ conduit for most of our runs.  Rigid conduit of that dimension is heavy and cannot be bent by hand.  Instead, a motorized pipe bender was the order of the day - and it took some really skilled craftsman named Don and Josh to get our conduit in place and looking good.

It really is an art, as much as a science, and when done with care and precision, the result is quite appealing!

Pulling it Together

Our final ground-based task was to pull the conductors through the conduits.  Our longest pull was 245′ - not quite a football field, but close!  Moreover, that longest pull had multiple bends as we routed the conduit to make it as invisible from the ground as possible.  (To complete the task of making the conduit “disappear” to the greatest extent possible, the client painted the conduit to match the walls and the trim!)

At the end of a very long, drizzly Saturday, we were rewarded with having our conductors fully in place from the utility disconnect switch and performance meter socket:

… to our disconnect switch adjacent to the transformer:

Our penultimate installment will take you to the roof where the real action takes place.  So buckle in, the next chapter isn’t for the faint of heart!

In Part 1 of this series about Installing Solar at Westridge School, we looked at the process of putting our materials together for the rebate application.  With the rebate safely reserved, it was time to turn to pulling the permits for the job.  A solar project of this size involves two separate permits - building and electrical - but four points of inspection - fire, electrical, building, and utility.  We had already provided the utility, PWP, with the materials they needed but now we needed to load up for the permit center.

Assembling the Necessary Materials

The permit process addresses an entirely different need than does the rebate application.  The permit process is intended to guarantee that the proposed system, as designed, satifsfies all applicable codes and standards.  In theory, once you have successfully pulled the permit, the inspection process should simply be a matter of showing the inspector that you built the system as it was approved when you pulled the permit.

It looks conventional enough!

This project presented one signficant challenge - the actual attachment of the system supports to the roof.  While the roof looked conventional enough, that was not a wooden truss underneath those shingles.  To the contrary, our roof was built from a 20 gauge “Type B” steel deck with two layers of 5/8″ plywood, followed by 3″ of solid foam insulation, followed by 3/4″ of plywood to which the roofing materials themselves - membrane, felt and shingles - were attached.  So the question arose: what would be a sufficient way to attach our standoffs to this roof to provide the requisite resistance to wind loads - the effect of which had recently been demonstrated in Pasadena in such a disastrous fashion?

Unirac Fastfoot Attachment

To help answer that question we turned to the structural engineer (SE) who had originally done the load calcuations for our building.  Could we use a “FastFoot” and simply put multiple screws into the wooden decking materials?  Surely with enough screws - the FastFoot will allow for up to eight - we could reach the required pull-out resistance.  Unfortunately, that wouldn’t work since the engineer could not guarantee the manner  by which the plywood materials were secured to the underlying steel deck.  In other words, while we could be sure that our array would remain attached to the plywood, we couldn’t be sure that the plywood would remain attached to the building!  Images of Wizard of Oz roofs flying through the air filled my mind - clearly we would need another way!

The engineer suggested that we could use carriage bolts that ran all the way through the steel roof and were bolted together on the back side.  Certainly such an approach would guarantee that our array and the roofing materials stayed connected, and indeed, you would have to separate the steel deck from the steel framework of the building for that method to fail.  Unfortunately, that wouldn’t work either since there was no way to access the back side of the roof in order to complete the connection.

“Nine-Inch Nails” Meet 8-Inch Screws!

There was one other approach - a company by the name of Triangle Fasteners sells some very strong, very long, self-tapping screws (called “Concealor screws“) that could drill their way into the steel deck and provide us with the required pull-out resistance. The bad news - our distributors only sold screws up to 7″ long - and that would not be long enough to guarantee that our screws made it through the decking. A call to the manufacturer revealed that in fact, they did make 8″ screws, they even made 9″ screws!  Excellent!  We now had a solution that our SE could bless.  It was time to go pull our permits.

Fear and Loathing at the Permit Center

Anyone who has ever pulled a permit knows the combination of emotions that you encounter upon entering the building: fear that something you haven’t considered will suddenly become A Really Big Deal, loathing for the interminable waiting, and of course, the pain of paying for it all.  Dentists’ waiting rooms tend to be cheerier places.

Pasadena’s permit center is certainly better than most: it is a comfortable old building across the street from the beautiful City Hall.  They have a clever scheduling system that routes you among the different windows: Building and Safety, Zoning, Historical Preservation (very big in Pasadena but not a factor for solar projects), Fire, Permit Processing and, last but certainly not least, the Cashier.  A solar project applicant must navigate their paperwork through every one of those windows before exiting with your Grail - a stamped set of plans and a bright Yellow permit folder where inspection sign-offs will be recorded.

First stop - Building and Safety.

Building and Safety

The building and safety folks are responsible for reviewing your plans for conformity with state and local codes and standards - a really important task.  First, however, you have to speak with someone who knows what you are showing them and on our first trip to the permit center, no such person could be found!  The gentleman behind the B&S desk was very polite, and you could tell that it pained him to inform us that after our thirty minute wait, he couldn’t help us.  Moreover, none of the people who “understood solar” were available - we would have to come back tomorrow.

Tomorrow dawned cloudy but we were determined to press forward.  This time our 35 minute wait was rewarded with an appearance before someone who was prepared to pass judgment on our plans!  We walked him through each of our sixteen 24″ x 36″ pages, explaining as we went exactly what we were doing and where the answers to his questions could be found.

All seemed fine, but then he started throwing us some curves.

Our SE had done his calculations for a basic wind speed of 85 mph - the same wind speed we had always used for load calculations in Pasadena.

“No,” said the man behind the desk, “You have to use 100 mph.”

“Really?  Since when?”

“Since the windstorm in Pasadena at the end of November,” we were told. (Never mind that the wind speed never reached 85 mph in Pasadena, let along 100 mph, during that terrible event.)

“Really?  Where was that published?”

“It wasn’t,” he conceded, but simply told us that we needed to revise our calculations for 100 mph or he wouldn’t approve them.  That meant another iteration with our SE and another trip back to the permit center.

Now the good news here is that we were certain that our system would easily handle 100 mph winds (or 120 mph, for that matter) so this change in policy did not pose a danger to the project going forward.  But changing the basic wind speed for an area from 85 to 100 mph is something of a big deal and will add to the expense of many projects that need permitting.  Shouldn’t there be a more public process before such a change is implemented?

The other curve sent our way was really just odd.

We did a detailed drawing showing our attachment method as it penetrated the various layers of roofing materials and made contact with the steel deck beneath.  We drew that straight up on the page and included multiple elevations  in our sixteen pages that showed the pitch of the roof and indicated that the array was installed on top of our attachment method, parallel to the roof.

“Not good enough,” we were told.

“Why?  What’s missing?”

“You need to show the attachment at the slope of the roof.”

“Really?  We show you the slope of the roof, we gave you the detail of how the attachment connects to the roof and we told you that the array is parallel to the roof.  How is that not sufficient?”

“You need to add a drawing that shows the array attachment and which reflects the slope of the roof.”

“Really?  So what you want is for me to rotate the image of our attachment 13° to reflect how it will be pitched on the roof?”

“Yes.”

Sigh.  Ok, back to the drawing board (or more accurately, the computer screen).

Fortunately, our SE was able to redo his calculations in short order.  And not surprisingly, it was also pretty easy to take our attachment image and rotate it.  We printed up the revised plans and headed back to the permit center.

Surprise - there was yet another person behind the counter this time.  Whereas his predecessor seemed to be actively looking for little things to complain about, this fellow could not have been more helpful. He looked at our revised load calculations - veryifying that they had been done for 100 mph and that the SE had concluded that all was well - and then proceeded to stamp our plans.  (I had pointed out our added, rotated drawing, but it was clear that he wasn’t interested in that at all.)  After he stamped our plans, he then took them himself to the zoning and historical preservation desks and secured those sign-offs as well! Wow!  He saved us an hour of waiting in those queues and he seemed genuinely helpful and concerned.  What a pleasant contrast!  We were well on our way with just one real substantive hurdle remaining - the Fire department.

Fire

The California State Fire Marshall developed a set of guidelines that provide guidance as to how fire departments should permit and inspect solar installations.  The guidelines call for space to be set aside for pathways around the array and for venting of smoke in case of a fire.  The guidelines call for different restrictions based on the size and shape of the roof and whether it is a residential or commercial building.

(While the document from the Fire Marshall is labeled “guidelines", most localities seem to treat it as gospel.  Even more curious, the guidelines clearly say that they are just that, guidelines that do not have the force of law until a local jurisdiction passes an ordinance adopting the guidelines as regulations.  We have yet to see such an ordinance.)

Our building plan included a three-foot set aside around both sides of the array and from the ridge, and was augmented by automatic smoke ventillation devices already built into the roof.  But that was not sufficient - the fire official wanted us to provide a four-foot clearance on all three sides.  Yet another trip to the computer.

We returned with our revised drawing, showing four feet of clearance as requested.  But now there was another concern - the same fire official now wanted us to open a walkway in the middle of the array.  (We already had access paths for potential maintenance, but they were not wide enough to be considered a walkway.)  No matter that our roof was not at all like the flat roof with parapet shown in the guidelines, we still needed to provide a walkway.  There was only one way to do that - take out a column of panels.  Together we X-ed out seven panels and thereby created a walkway.  The fire official was now satisfied - she signed off on our plans.

Done

And just like that, we were done.  Well, not quite - there was still the little matter of paying for all this.  Here we made out surprisingly well.  Unlike some cities that gouge solar applicants (and you know who you are!), Pasadena’s fees were quite reasonable.  Total cost for our now 52.25kW solar project?  $732.  Sadly, we know of residential projects one tenth that size in other cities where the permit fees have exceeded $1,000!  (But that’s a story for another day.)

Altogether, it took us four separate trips to the permit center, three plan revisions, and a little over $900 in expenses to secure our permit.

Now all we needed to do was get the materials to the job site on time, and complete the installation in the two week window that we had to mesh with the School’s schedule.  The real work was about to begin…

In November of 2011, Run on Sun was hired by Westridge School for Girls to install a 54 kW solar system on the roof of the school’s Fran Norris Scoble Performing Arts Center (the “PAC” as it is known on campus), and that project was just recently completed.  This multi-part series will document the process by which we went from a signed contract to a signed-off solar power system.  Not surprisingly, there were a few twists and turns along the way that had to be resolved before we could deliver a successful project, and this series will showcase those developments in the following five parts:

The Rebate Application

The rebates being offered from Pasadena Water & Power (PWP) for this non-profit project were scheduled to step-down on December 1, 2011.  Indeed, this was a substantial rebate reduction - 26% - such that failure to secure the existing rebate rates would have amounted to a hit of tens of thousands of dollars for our client.  And PWP had made it very clear - unless applications were 100% complete and correct, they would be rejected and when resubmitted would be subject to the reduced rebate rates.  Clearly the pressure was on to get this right the first time!

The application package consisted of eight parts - most of which were straight-forward, but a couple required substantial work to guarantee that the application as submitted would be acceptable the first time.  Here are the parts that went into the rebate application: 1) Signed Rebate Application (PWP’s form, signed by client and Run on Sun under penalty of perjury!); 2) Single Line Diagram for the electrical components of the system (more on this below); 3) Site Plan; 4) CSI Report (as produced by the California Solar Initiative’s rebate calculator); 5) Shading Analysis (i.e., a Solar Pathfinder report to support the shading values used to create the CSI Report); 6) PWP’s Net Metering Agreement (executed by the client); PWP’s Net Metering Surplus Compensation form (for AB 920 compliance); and 8) Installation Contract between the client and Run on Sun.  Also, since this was a non-profit client, proof of non-profit status was also required.

Shading Analysis & CSI Report

PWP wisely requires the submission of a shading analysis in addition to the output from the CSI rebate calculator.  Since the amount of shading at the site directly impacts the performance of the system - and hence the CSI AC Watts of the system (or the predicted annual energy output in the case of a PBI rebate) - it really doesn’t make sense for a utility to simply trust that the installer is telling the truth about shading.

The output from the Solar Pathfinder proves that the shading numbers claimed are the shading values present at the site.

Site Plan

The site plan needed for the rebate application is a much simpler plan than what will ultimately be required for the permit, really only requiring an indication of where the various components of the system will be relative to the overall site.   However, our system occupies three different areas of the PAC: the roof where the array itself is located, a ground-level storage area where our step-up transformer will be, and the utility switchgear, located on the far north end of the building.  Thus our site plan included drawings for each location.

The array drawing showed the three sub-arrays and the clear space allocated for fire department access.  Each sub-array consisted of three branch circuits, each of which was “center-tapped” to reduce the voltage drop in the associated branch circuits.  Each branch circuit landed at a sub-array service panel which then fed a master “solar-only” sub-panel in the transformer area.

The transformer area drawing detailed the conduits coming down off the roof (one each from each sub-array sub-panel), the master sub-panel which feeds our step-up transformer (to change the 208 VAC three-phase power coming from the roof to 480 VAC three-phase supplied by the utility service) and then a safety disconnect switch located adjacent to the transformer.  From the safety switch a fourth conduit carries the required conductors back across the roof to our service switchgear area.

The service panel area drawing showed the placement of our lockable PV AC Disconnect, the associated performance meter, and our circuit breaker for the system located in the existing service switchgear.

Single Line Diagram

Our most significant deliverable in the rebate application packet was the single line diagram (SLD) for the electrical circuits.   Since this diagram shows how all of the electrical components of the power generating system interconnect - including the tie into the utility’s grid - we knew that this would be the most closely scrutinized piece of the submission.  To be sure, PWP has a generic SLD that installers can use (in fact, we helped develop it!) but that drawing does not cover the use of Enphase Micro-inverters which we were featuring on this job, nor does it allow for a step-up transformer.

Fortunately, we had developed a very flexible SLD format from prior jobs that we could readily adapt for this project.  However, before we submitted it to PWP, we forwarded it to the application engineers at Enphase Energy to make sure that they were comfortable with what we had designed.  Enphase was more than accomodating - given our tight time frame they bumped us to the front of their engineering review queue and came back promplty with the good news - the design was good as we had drawn it and no revisions were needed.  Of course, that was no guarantee that the utility would agree, but it is always nice to have a P.E. on your side!

Included in the SLD preparation was a complete set of voltage drop calculations.  This was complicated by the fact that we had 9 different branch circuits, three different sub-panels and two different operating voltages!  Good design calls for limiting total voltage drop to less than 3%.  To keep our worst case scenario within that limitation (covering the branch circuit farthest from the main “solar-only” sub-panel) we ended up with 4 different gauge sizes of conductors at different legs of the run: #12 in the branch circuit cables (supplied by Enphase), #8 from branch circuit jbox to sub-array sub-panel, #2 from sub-panel to main “solar-only” sub-panel, #3/0 from that sub-panel to the transformer and then #2 from the transformer back to the service equipment area.  (One change that occurred during the install process increased the length of some of these runs - and that necessitated some wire size changes to insure that we stayed comfortably below our 3% limit.  Those will be discussed in future episodes.)

One Big Present

All of those documents, plus pages and pages of cut sheets describing all of the key products being used, were then submitted to PWP - one day before the deadline!  With no margin for error, our submission had to be perfect.  Thankfully, it was - PWP gave us their official blessing to proceed three weeks later, just three days before Christmas.  One big present, indeed.

Our first hurdle successfully surmounted, it was time to prepare for the most nerve wracking part of the process - pulling the permits!  That’s the subject of our next installment - stay tuned!

Yesterday we wrote about our most recent foray into the California Solar Initiative (CSI) data and how that data revealed trends regarding the costs of solar in SCE’s service area during the first half of 2011. We continue today with a look into the equipment that was specified for these projects and explore who’s hot and who’s not.

CSI Data Generally and Equipment Specifically

As a reminder, our data set for this analysis consists of an extract from the CSI Working Data from 8-24-2011 that includes data for SCE installs where major status activity took place during the first half of 2011.  That data set consists of a total of 6,306 projects of which 698 are “delisted” (meaning the project’s rebate reservation has been cancelled for some reason), 3,131 are “installed” (completed or in some stage of rebate payment) and 2,477 are “pending” (in some stage of the process from initial rebate application filed but no rebate claim yet filed).  For today’s analysis, we will exclude the “delisted” projects from our data, leaving a total of 5,608 projects to analyze.

CSI tracks data about equipment used on projects in great detail.  In particular, for every project, CSI allows for up to seven different panel manufacturers and ten different inverter manufacturers!  So how many of our projects use multiple panels or inverters by different manufacturers?  We would expect not many, and the data supports that surmise.  Only 11 projects reflect two different solar panel manufacturers on the same project and in most of those cases the installer has substituted one solar panel for the one originally designated.  (Indeed, one project reflects four different panels being identified to CSI for the same project, finally settling on what appears to be two 180 Watt panels plus six 175 Watt panels feeding a single string inverter - curious design, that!)  Similarly for inverters, only 24 projects have two different inverter manufacturers specified and no project reflects more than two.  Given that, our analysis will only look at the first specified panel and inverter manufacturer.

Solar Panel Trends

So what is happening with solar panels?

Overall, there are 85 different panel manufacturers included in the data; however, most of them account for very few projects.  If we apply a reasonable filter to this data and only look at solar panels that appear in 50 or more projects, the number of represented manufacturers drops from 85 to 14, and the total number of represented projects falls from 5,608 to 5,079.  In other words, those fourteen manufacturers account for 90.6% of all of our projects, as demonstrated in this first graph. In fact, the distribution is even tighter with only five manufacturers exceeding 10% of the total: Suntech Power (18.7%), Sharp (14.8%), SunPower (14.3%), Kyocera (12.8%) and Yingli Green Energy (10.8%).

Solar Inverter Trends

There are twenty-three different inverter manufacturers represented in the CSI data, reflecting the greater complexity of inverters and the more rigorous path required to bring an inverter to market in the U.S.  Filtering for manufacturers represented by ten or more systems cuts the list from 23 down to just 13.

SMA America is the runaway winner in this competition.  Under their own label, they account for 35.4% of all of these projects.  Moreover, the majority, if not all, of the “SunPower” inverters are actually re-branded SMA inverters.  When the SunPower inverters are added in, SMA accounts for a whopping 48.2%.  That leaves only six other manufacturers to exceed even 1% of the total: Fronius USA (19.8%), Enphase Energy (15.2%), PV Powered (5.5%), Kaco New Energy (3.1%), Power-One (3%) and SatCon Technology (3%).

Two things of interest in those last numbers - the inroads of relative newcomer Enphase Energy (which was only founded in 2006), and the inclusion of SatCon, since alone among that list, it only sells central inverters for the commercial market (where it is dominant).

Popular Pairings

That is how the different manufacturers stack up head-to-head, but what about combinations?  Are there pairings of  panels and inverters that are most commonly preferred?  The data reveals five combinations that account for more than 5% of the total: Suntech panels with SMA inverters (13.5%); SunPower panels with “SunPower” (i.e., SMA) inverters (12.4%); Sharp panels with Enphase micro-inverters (10.3%); Yingli panels with Fronius inverters (8.2%); and Kyocera panels with Fronius inverters (8.2%).

Winners and Losers

While certain pairings are popular - are they cost-effective and how well do they perform together?  We decided to look at system combinations from an average $/Watt perspective and from an average CEC efficiency perspective to see what jumps out of the data.

Here’s the first thing that struck us - picking a system with lower-tier panels does not guarantee a lower installation cost.  In fact, many of the bottom-tier panels (none of which made the cut in our discussion of panel manufacturers above) had install costs well above our overall average for the data set ($6.37/Watt).  For example, we found a handful of systems using solar panels from such luminaries as Apollo Solar Energy, SET-Solar, and REC ScanModule where the average installation cost was more than $10/Watt!

Who was on the very low end of the install cost curve?  Gloria Solar, Suniva, Kaneka, Silray and Solaria each had a handful of installations that were below $4.50/Watt.

More significantly, how did our most popular pairings perform?  Here’s the data:

What is up with the Sharp & Enphase combination?  While Enphase installations are known to cost a bit more than a comparable string inverter installation (confirmed by our own experience), they certainly don’t cost $5/Watt more!  Rather, it turns out that the overall average for all Sharp-based systems is $8.53/Watt (nearly $2.00/Watt above the average) with prices ranging from a low of $6.17/Watt (when paired with a Solectria inverter) to a breath-taking high of - are you sitting down? - $19.30/Watt when paired with a Sharp inverter.  So who installed that system, you ask?  You’ll learn all about it (or at least all that we can tease out of the data) later in this series.

Shifting our attention to efficiency, thin-film module maker First Solar gets the highest overall ranking, 91.7%, thanks to its extremely high STC to PTC ratio.  On the more embarrassing end of the scale, Sunlan solar brings up the rear, averaging only 80.7%.  From our list of the most popular solar panels, Sanyo (long a Run on Sun favorite) does the best, averaging 89.3% across a variety of inverter combinations.  The rest of the top five are: Canadian Solar (87.6%), SunPower (87.5%), Suntech (87.2%), and Schuco (87.0%).  The bottom-five of our best selling panels?  That dubious honor belongs to: Sharp (86.0%), BP Solar (85.8%), ET Solar Industry (85.6%), Trina Solar (85.5%), and REC Solar (85.1%).

As for our five most popular pairings, here is the data:

That is a pretty tight grouping, with a total range of just 2%.  To break out of that mold with a conventional panel/inverter pairing, the Sanyo & SMA combination is your best bet, weighing in at 89.5%.

Who Uses What?

Finally, we decided to see what equipment combinations are preferred by the biggest installers in the market.  The following table lists the top-five installers and reports the number of projects in the data, their most frequently chosen solar panel (and % of times used) and their most frequently chosen inverter (and % of times used).

Collectively, these 5 installation companies accounted for 42.8% of the projects in the CSI data.  Certainly companies this large must have some real clout when it comes to negotiating prices, thereby allowing them to pass along those savings to their many customers. 
Or do they?

Find out in our next installment!