The PV system sizing is based on a few ground facts, as we used to call them.  First, the current usage of the structure.  In my case, about 8,000 kWh per year.  Second, the average annual sun hours per day for the location (about 5), and third, the shading, if any and system losses.  Therefore, to calculate the system size, the following method is used:

1. Annual kWh ÷ 365 days = kWh per day
2. Percentage of electricity to offset (decimal)
3. kWh per day ÷ sun hours (about 5 hours in the Hudson Valley)
4. Figure in losses (temperature loss 88%, system derate 95%, inverter 95.5%)

Therefore, my system looks like this:

1. 8000 kWh ÷ 365 days = 21.9 kWh per day.
2. I want to offset 100 percent, so 21.9 kWh × 1.0 = 21.9 kWh
3. I have an average of 5 sun hours per day, so 21.9 kWh ÷ 5 hours = 4.38 kW
4. Calculate system temperature loss, 4.38 kW ÷ 0.88 = 4.98 kW
5. Calculate system derate, 4.98 kW ÷ 0.95 = 5.24 kW
6. Calculate inverter loss, 5.93 kW ÷ 0.955 = 5.26 kW

Therefore, according to this, I would need a 5.26 KW DC rated PV system.  Our system is 4.1 KW DC, which is a little bit lower than required.  I am waiting to see how the micro inverters do with the solar panels.  I will bet they are more efficient than large string inverters and thus, we will get close to the desired number.

Next, things like breaker sizes, wire sizes, voltage drop, temperature de-rate, conduit fill and grounding need to be addressed.  First, there is a three line diagram that shows how the array is wired:

3 line diagram, 4,100 watt PV system using Enphase M-210 inverters

There are two 240 volt 15 amp branch circuits, each one is connected to 10 Enphase microverters.  The inverters are connected in parallel on these circuits.  Each inverter is in turn connected to a single 205 watt Sanyo Hip-205N PV panel.  Therefore, each inverter is capable of 205watts / 240 volts = 0.85 amps.  Maximum branch circuit current is then 10 inverters x 0.85 amps or 8.5 amps.  The NEC states that breakers should normally run at 75% of there rated value, so 8.5 amps x 1.25 = 10.63 amps.  Therefore a 15 amp circuit breaker is satisfactory.

Next, wiring sizing.  A fifteen amp breaker calls for #14 AWG wire.  This will not be satisfactory, however, to deal with the voltage drop between the solar panel array and the service entrance panel.  The distance between them is 124 feet.  Since we paid so much money for the solar panels, I want to keep the voltage drop to 1% or so.  This will ensure that all of the power we generated at the solar array gets into our electrical system and will not be dissipated as heat.  Here is the calculations for voltage drop:

Vdrop = (I x 2 x d) / (1000Ft/Kft) x r

(It is a little hard to write this formula out on one line)

Vdrop – volts lost
I – current
d – distance
r – resistance of wire per 1000 ft (from NEC 2008, table 8, conductor properties)

Therefore, using 14 gauge (stranded) wire:

Vdrop = (8.5 amps x 2 x 124 ft) /1000 x 3.14Ω = 6.62 volts.

In a 240 volt circuit, each leg is 120 volts, therefore 6.62 volts / 120 volts = 0.0551 or 5.51% voltage drop.  Too high for our purposes.

Using 8 gauge (stranded) wire:

Vdrop =(8.5 amps x 2 x 124ft) / 1000 x 0.778Ω = 1.64 volts.

1.64 volts / 120 volts = 1.3%

Therefore, #8 AWG wire is appropriate for this application.

Next, temperature derate.  The wire itself is #8 THHN which is rated for 90°C.  This will be well within our specs, especially since we already accounted for voltage drop, above.  The wire will be in conduit.  In this case, we are derating the conductor for the maximum temperature that conductor is expected to experience.  Since our maximum temperature around here is about 100°F, according to NEC table 310.16, #8 AWG copper wire has a current caring capacity of 55 amps x 0.91 or 50 amps.  This is well above our maximum current of 10.63 amps so that is good.  This step is more critical on roof top installations where ambient temperatures can be very high.

Next is conduit fill.  One can’t just stuff as many wires as one can fit into a conduit.  Generally speaking, the NEC seems to expect about 50% conductor fill in any given conduit.  Thankfully, there are tables that give out this information.  For my purposes, I used 1 1/4 inch schedule 40 PVC conduit.  According to Table C.10 up to seven #8 AWG conductors can be placed in that conduit.  I have two 240 volt branch circuits, including neutrals and a ground wire.  That totals seven conductors.

Finally, grounding.  All non-current carrying metal parts, frames, etc. must be grounded. Fortunately, the 2008 NEC allows us to size our grounding conductor to the size of the over current device (circuit breaker).  In this cast that would be 15 amps, therefore 14 gauge.  Unfortunately, the Jurisdiction Having Authority (JHA) has not adopted the 2008 NEC yet, they are on 2002 which requires the grounding wire to be the same size as the current carrying wire, regardless of the sizing of the wire for voltage drop.  So, #8 AWG ground wire between the array and the service entrance panel is required.  All metal frames and mounting rails of the PV array need to be connected to this ground.  Any ground wire that is not protected e.g. run in conduit needs to be a #6 or bigger conductor.  Also, a separate grounding electrode needs to be installed at the PV array since it is not a part of the existing structure.

Instead of running a continuous ground wire to each module, inverter and mounting frame member, I was allowed to use a WEEB (Washer, Electrical Equipment Bond) type ground.  These little clips go between the modules and the frame, making the mounting frames the grounding conductor.  It saves time.

Front view of 4.1 KW PV system

Ground mounted 4.1 KW PV system

This system is 4.1 KW and should provide almost all of our electric needs once we replace the old refrigerator with an energy star unit.

Enphase M-210 inverter under Sanyo HIP205NHKA5

I used Unirack Sunframe rails to mount the PV modules. The modules are  are Sanyo HIP 205NHK5 Modules and Enphase M-210 microverters.  I like the concept of the Microverter, e.g. each panel has it’s own small inverter.  This allows from some shade tolerance for the lower modules without loosing the entire array.  Also, each panel is matched to it’s inverter at the best efficiency, increasing the overall array output.  Seldom do you get to see the underside of a PV array as they are most often mounted on a roof.

Equally unfortunate is the fact that I jumped the gun on the construction and poured the footings before I had the building permit.  So, once again I rented the Kabota backhoe from the Taylor rental place down the road.  I am on a first name basis with the owner, which is nice, sort of.  Anyway, quick work with a chain and I pulled all six of the eight inch footings out of the ground, dug new holes and place the pre-poured footing in a new whole.  I dumped about 6-8 inches of crushed stone in each hole an compacted it.  All in all, I am only out the one day’s rental on the back hoe, which was not too bad.

Timber Frame for 4.1 KW Photovoltaic system

On to the construction of the frame.  I decided to use 4 x 4 posts and beams, except for the main support beam, which is 4 x 6 inch.  The entire structure is braced with 4 x 4s at all ninety degree meetings.

Corner bracing

Of course, the weather has closed in and I am working outside in the snow and wind.  On Saturday, it was 15 degrees out with a 20 MPH wind.  I don’t know what the wind chill was, I can however verify, it was unpleasant working outside.  That being said, progress has been made.

The frame is mostly up, I need to put the final support beam across the top.  Then I need to put in the “rafters” which will be 2 x 8 x 12 treated lumber.  The rafter spacing will be a little odd, since they are space to support the solar panels according to the panel manufactures specifications.

Hand dug conduit trench

Also completed (before the ground froze solid) is the trench between the house and the support frame.  We dug this by hand, 42 feet long by 18 inches deep, as the current NEC specifies for PVC conduit.

Everything is frozen solid right now, which actually has it’s advantages.  Come springtime, this will be a soupy mud mess.  Once the ground drys out a little bit, I’ll rake it out and plant some grass seed.

The first thing that was needed was the size of the array.  For this system, we will be installing 20 Sanyo HIP-205N modules.  These measure 62.2″ x 31.4″.  I would like these to be installed landscape style, four deep by 5 wide.  The total array size is 311″ or 25.9′ X 125.6″ or 10.4′  I am leaving a little room around the edges for a safety factor, so my support frame will be 27 x 12 feet.

I also want  to tilt the array to latitude, which around here is 42 degrees.  There have been studies that show that the tilt angle is not a critical as once thought, however, since I can do it, I might as well.  Therefore, I will install a total of six support posts, making the structure 26 feet x 7 feet.  The front of the structure will be about 6 feet above ground level, the back will be about 12 feet above ground level.

PV system location marked with stakes

I staked out the frame and aligned it to true south.  It is only a few degrees off from the property line, so it works out well.  Since we have had a lot of rain this year, I decided to dig a test pit to see where the water table is in relation to the bottom of the footings.  Local code requires 48 inch deep footings, my test pit reached 46 inches deep before I saw some seepage.  I left it over night and the next morning there was about 2 to 3 inches of water in the bottom.  Over all, not too bad, I put some crushed stone in the bottom of each footing before I put the form in.

Test pit to see where the ground water table is

It rained most of the day on Saturday, however, I still managed to dig four of the six holes.  On Sunday, I dug the last two.  Then, by this post, I knew that it takes about 2 2/3 80 pound bags of ready mix to fill an 8 by 48 inch sonotube.  I picked up 16 bags of 4000 PSI ready mix.  This time, I borrowed a cement mixer, which made things much easier.  I also used one #4 (1/2 inch) rebar down the middle of each footing, tied to the J bolt on top.  I used 1 gallon of water per 80 pound bag, as the directions on the bag stated.  This made a good stiff mix.

Holes completed, string crosses mark footing locations

To make all of the forms the same level, I used a 14′ 2 x 8 and a level.  Going from hole to hole, slowly putting more packed crushed stone in each hole, I think I got pretty close.  Also, the crushed stone will aid with drainage around the bottom of the footing.  Any differences in level can be made up by trimming the posts.

Footing hole, somewhat deeper than 48 inches

This was a miserable job.  It was wet and muddy all day long.  One of the hole had a lot of water in it, which needed to be pumped out before I could put the form in.  Our soil is thick clay, which caked on everything, shovels, boots, rocks, etc.  The weather forecast was for sun on Sunday, which turned out to be false.  Still, it is done.

Footings completed and backfilled

I was going to use the excavator to dig the trench for the conduit, however, I decided that a ditch witch would be a better idea, less back fill, less mess, etc.  For conduit, I think I will go with two inch.  This system has microverters, which means the feed from the solar array will be 240 VAC.  I could use #12AWG with this and come in at just under 2% voltage drop.  Since I have spools of #8 AWG already on the truck, I will used that cable instead.  That makes the voltage drop 0.6%.  Since there are two 240 VAC branch circuits, plus two neutrals and one ground wire, that makes the total number of conductors 7.  According to the latest version of the NEC (2008), table C.10, 1 1/4 inch schedule 40 PVC conduit is acceptable for this installation.

Once the concrete hardens for a couple of days, we’ll put up the frame.

First of all, as technology often does, newer things are available these days that make a solar system in the North East a better proposition.  Secondly, the solar business has done better than I expected.  As a result, I don’t often have much time to work on household projects.  That means that this years “capital improvement” budget has gone unspent for the most part.  Finally, I would like to offset some of the extra income tax from the profits.  What better way than to invest in the technology myself.  The Federal Government offers a 30% income tax incentive and the NY State government offers a 25% tax income incentive up to $5,000.00.  This will cut the overall cost of the installed system by almost 50%.

There are a number of considerations:

1. How large of a system should be installed.  I decided that I wanted to offset 70-80% of my annual electrical use.  In this climate and environment, that equates to about 4.1 KW DC PV system.  This leaves a little downward room in case I decide to replace the electric stove with a gas unit.
2. Where can it be installed.  Since the south facing roof has the solar hot water system, the PV system needs to be mounted on a sun shade type structure in the yard.
3. What type of technology.  I was initially looking at a grid tied with battery back up, however, after I looked into the newest type of inverter, the Enphase microverter, I decided that this was the way to go.  A battery backup can be added at a later date.

The Enphase microverters are really cool.  The way this system works is every solar panel has its own small inverter instead of one large inverter for many panels.   The advantages of this type of system are thus:  In conventional system, shading of one panel can cause the entire solar array to turn off, making it ineffective.  With the microverters, the shaded panel may turn off, but the rest of the unshaded panels still put out full power.  In the Northeast, trees grow everywhere, it is nearly impossible to have a completely shade free site, nor should home owner’s be expected to clear cut their lots to accommodate a PV system.  The Enphase microverters mitigate some of those concerns.

Also, multiple inverters create redundancy.  Any one inverter can fail, leaving the other nineteen still operational.  There is automatic web monitoring for a small annual fee, or the modules can be monitored manually.  I may write a small web based program to monitor and post my energy output here.  The inverters themselves carry a 15 year warranty, whereas most other inverters carry a 5 year warranty.

Finally, there are no DC voltage losses to account for, making the entire system operate much more efficiently.

In anycase, the order has been signed, checks have been written and the excavator has been reserved for this weekend.  The first step is to dig and poor the footings for the sun shade.

More to follow.