Solar Off Grid Garage
Stephen Bosbach, 2005
Dscfsolarmodules1.jpg Design Sequence:
Balancing loads and input / Balancing storage and recharge
Garage loads- door openers, lights (interior florescent, attic DC, closet DC) intermittent power tools, vacuum, etc.
due to this being a rental unit - functions can not be compromised due to weather except in extreme circumstance - garage loads must function for 7 days without sun and draw down battery less than 60%
Sun Insolation - Austin receives approx. 4.8 KWH/ sq. meter/ day, or 4.4 hours/ day average generating capacity with photovoltaics (much of this in the afternoon)
Garage orientation - standard gable roof slopes SW and NE and is free of shadowing except for a power pole shadow in late afternoon - typical weather pattern in Austin is cloudy morning, burning off to partly cloudy by 10 AM and remaining so until midnight when coastal fog and clouds move in again. Major fronts and highs will upset this pattern, but it is typical. Given this pattern, the SW orientation works.
Money - the biggest variable. I spent more than anticipated on the system, but wanted more than a quick system that worked. It needed to be salable as part of an investment project and reliable with very little compromise. It needed to be safe enough for renters to interact with without concern and require very little maintenance. A 10% tax rebate was available as this is business property. (this is now 30% for 2007 / 2008)

Project power options:
-Bring power from duplex - trench through tree roots, around septic, cross water line est. cost = $950.00
-Power drop from bare pole with PEC - New service fee, two drops, two meters, stand pipes, etc. est. cost = $1600.00
-Solar off grid - 2 panels, 2 batteries, charge controller, battery monitor, inverter
est. cost est. = $2000.00
-Actual solar installation - 3 panels, 3 batteries, charge controller, monitor, inverter, module rack, closet construction, conduit, wire, boxes, and misc. = $3083.00

I had just finished the Solar Energy International PV design course and played around with various configurations in the system sizing worksheet to get a feel for what would be required. I totaled all the garage loads, estimating amperage and time on per day. This gave me an average daily load in watts split out in DC and AC.
Batteries were sized based on how long I wanted to maintain the load without sun.
Array sizing was estimated based on the average load per day, sun insolation, battery and inverter efficiency
Basically I wanted to run the garage year around without pulling down the batteries too drastically (no more than 60%) and be able to do this for 7 days of little sun.

Planning started around the available space for the garage and a quick, unsophisticated method to determine if the site was appropriate for solar. Knowing the height of the garage, the height of the trees, and the position of the sun at winter solstice, I could predict how long the sun would impact the panels.
Before the garage was built, the area was cleared of trees and I did a quick visual estimate of where the sun would be on winter solstice. To do this I found a table that showed both the summer and winter sun altitude in degrees across the sky. I then checked the degree altitude of the tree line from the height of the solar array using an inclinometer on a high step ladder to simulate the position of the array. By including the tree horizon on the sun path chart, it was easy to see how many hours of sun would fall on the panels (theoretically). The chart showed a maximum of 5 hours of winter solstice sun and 9 hours of summer solstice sun. All I needed was about three to four hours of sun to keep up with maximum expected loads. This was just enough data to show that the site would indeed work as a solar site. Panel inclination was set for one month each side of winter solstice ( about 48 degrees ) to maximize recharge when the loads were anticipated to be greatest, in winter. What I didn't anticipate was how much this would limit morning solar gain at other times of the year with a southwest orientation. I don't get sun on the panels in mid summer until nearly noon! This isn't really an issue however, as loads that time of year are minimal and days are long.
Once I knew solar could work on this site, I set out to see how many (or how few) solar modules and batteries I would need for the job. Given the intermittent nature of the loads in a garage, this is one of the easiest buildings to power. Even power tool use is infrequent and often limited to short bursts of activity. Lights are usually on for short periods when entering and exiting. Garage door openers were the big load (5 amps cont.) , but operate for only 12 seconds each time the door is raised and lowered. By adding up the loads and multiplying times the number of minutes used each day, the total power used each day could be estimated. To keep up with this demand, I would need an after loss solar production equal to this load. This said, I could do the project with one large panel of 125 watts (barely) or easily with two panels of 75 - 80 watts. I chose the latter, then when some extra cash came unexpectedly, I purchased a third panel just in case! This turned out to be a good decision, as winter sun is often minimal and when off grid you need all the winter charge you can get.
The garage needs to function when there is no sun however, and the battery bank determines for how long. A large battery bank will have a lower percent of discharge for the load design and therefor a longer life cycle, but it would also require more input to get the occasional deep discharge back up to 100%. I still chose to error on the conservative side and designed for three 150 amp hour batteries. This gave me 7 "days of autonomy", or days without solar input. Looking at meteorological records for Austin shows very few instances of seven days or more without some solar input. I know from experience, however, there can be weeks of cloudy weather in Austin where there is some sun, but very little useful solar charging possible. I wanted enough capacity to handle most of these instances. After a winter of very rainy weather I added another battery to the system. Since this scenario would probably happen in winter when insolation was minimal, I realized that the batteries may need extra help recovering after a hit like 10 days of no sun. Rather than build in mega capacity to the solar charging, I use a portable 12v. charger in those rare instances to get the batteries back up. This same procedure is often done in other off grid applications where a generator does double duty and provides the charge needed for equalizing the cells of the battery.
After checking several web sites for lightning protection, my analysis of this site indicated the risk of lightning was very low and I decided not to include an arrestor in the system. Omitting the lightning arrestor is not an option if your system is grid intertied or on a occupied dwelling, but in this case it did not have to meet codes and it was just not worthwhile.

Construction Sequence:

1. Confirm suitability of site
2. Determine interface of PV and structure - roof rack, attachment points, reinforcing necessary, wire entry, ground rod position
3. Choose components to meet the demands of the system loads
4. Design physical layout of the components - secure access? ventilation? insulation?
5. Install components - loads, breakers and fuses, modules, charge controller, metering, batteries, inverter, final high amperage connections
6. Test system and measure performance over time

My contractor built the power closet with a separate, lockable entry door from the outside. This is finished off on the inside with sheet rock and painted. I installed a 3' × 5' panel of Hardi-Backer mounting board to the wall as additional fire protection and to make placing boxes more flexible. Plywood could have been used here as well.
The first stage was to construct the module rack. This was fairly straight forward, but did require some forethought about spacing to accommodate both the modules and the standing seam spacing on the roof. Once completed and pre-fabed on the ground, the entire rack was lifted into place on the roof. Even on this 22 degree slope roof, the metal roofing was slippery and I needed a safety rope and belayer (thanks Crystal!).
The panels were hoisted up one at a time and attached to the rack. Connections were made between panels and everything was protected in flex conduit. Using conduit might be a little overboard, but the cost was actually cheaper than using UV resistant PV cable of a smaller gauge. Watch your wiring here, as the current is flowing unless you have the panels covered or there is no sun. My plan was to work my way from the roof down in electrical connections, completing the module to charge controller path, then make the connection to the battery via the meter. I had previously wired the AC loads and installed the breaker box, with one 15 amp breaker for each side of the garage.
SolarGaragePowerCloset.jpg The batteries were placed on cinder blocks and wired in parallel with 4/0 fine strand welding cable. This is very flexible and nice to work with, but I understand is not to code for residential installations. Again, with the short runs I had this size is a little overkill, but it was called for in the inverter installation instructions and I felt it couldn't hurt to use the largest size I could to keep resistive losses to a minimum when the inverter had to start both garage doors at once. Now the batteries were charging from the panels.
Once the feed from the panels was complete, the inverter rack was installed and the inverter set up. I played with the layout for several days and pre set components temporarily on the mounting board to see just how much 4/0 cable would be required for each high amperage connection. Make these measurements twice and check again! Once the cable is cut and connectors are crimped on the end, that's it, you can't stretch this stuff, and if you leave it too long it becomes awkward to connect. I left the smaller components like the fuse and battery disconnect switch loose until the cables were connected, then anchored them down to the wall.
Since this is in a non-conditioned space and will have significant humidity, I coated all the large low voltage connections with NoAlox (an anti corrosion conductive grease) to prevent corrosion. Yes, even copper to copper develops a patina that increases resistance in the contact. Small low voltage connections were treated with DeoxiT for the same reason. Ground connections also should be treated and preferably painted or coated with Plasti-Coat. All my DC wire (except for the 4/0) to spade and ring connectors were soldered. The final connections were the big 4/0 cables to the inverter. It's a big moment when you finally throw the disconnects and connect the loads.

-3 solar modules BP 80 from Alternative Energy Store $853

-Module array rack (owner built) from Lowes $75

-roof mounting clamps from Meridian Solar $60

-Used Inverter -Xantrex UX1400 and remote - 1400 watt nominal
constant load with up to 4200 watts capacity for up to 30
seconds - very low stand by draw of less than one watt, from
Meridian Solar $603

-Surplus Batteries - 3 (eventually 4) Exide Marathon M150 12v., 150 Ahr,
AGM (absorbed glass mat) batteries from Meridian Solar $375

-DC fuse 300A $60

-Cables and connectors #4/0 from Az. Wind and Sun $70

-Overstock E meter, shunt and box (now Xantrex Link 10) from
Meridian Solar $200

-DC fuse block (low amp.) from Meridian Solar $30

-BZ Prod. charge controller MPPT200, 16 amp, temp. compensated,
pulse width modulated and maximum power point tracking charger
from Alternative Energy Store closeout $70

-DC disconnect (boat battery switch) from Az. Wind and Sun $60

-AC fuse panel and breakers from Home Depot $80

Misc. ground rod, wire, conn., conduit, boxes, Home Depot $200
-Closet construction $400
total $3083

Cost was more than originally estimated, but will be amortized with the construction and maintenance cost as a tax deductible item on investment property. It will not add to property value as per Texas law. Even without the tax advantage, the system will pay for itself in less than 20 years with predicted rises in electric rates. There is presently no state of Texas rebate or PEC rebate for solar installation.

Testing / Operation:
My first connections were to charge the batteries from the modules through the charge controller. This worked fine, with the E meter keeping track of the charging progress. When the Trace UX1400 was connected, it failed to go into search mode and would only put out 108 volts AC. Garage doors failed to operate, even when the search mode was disabled. A replacement UX1400 from Meridian Solar in Austin worked as specified and everything operated normally. The acid test was a full on scenario where all lights and both garage doors were operated simultaneously. I almost expected a hiccup from the inverter, but it managed the loads without a complaint. I've also plugged in power tools such as circular saws and weed eaters to the outlets and it managed the loads fine. There were a couple of unexpected problems, however. The flood light outside would not function correctly, and then it dawned on me that although the transmitter in the sensor head had it's own battery, the receiver in the light did not. The receiver in the light needed current all the time just to receive the signal from the sensor head! This low draw is not enough to take the inverter out of search mode, and if I manually defeat search mode the inverter uses more current. I looked at a lot of alternatives to remotely trigger a light using the system in low draw search mode and each had it's draw backs. A completely separate light powered by it's own small panel is the answer, and what I eventually installed.
Another issue was the garage door openers when the inverter is in low current drain search mode. Just as with the motion sensor light, the receiver for the garage door opener must have constant current (or nearly so) for the pulse from the hand held opener to be sensed. This means the inverter needs to be left in the high current drain setting (specified at 5 watts) for the doors to operate from the remotes. After of winter of very rainy weather, I installed another battery to increase the number of days I could go without a recharge. After almost a year of operation, I purchased a good clamp meter that measured DC amps as well as AC. With the modules disconnected and all other loads off, I measured a draw from the inverter of 1.1 amps in active mode, with search mode only drawing .3 amps! This is better than specs., and yet I still had significant battery depletion in the winter!
I'm very happy so far with the performance of the system. I check the system about once per week in the summer (more to check for loads left on) and more often in the winter. Solar gain is regulated to 2-6 amps during the day, with battery draw-down nonexistent during the summer. During the winter, I had only two incidents where battery draw down would have exceeded 50% if I had not recharged from the grid. If I had been totally off grid, a DC generator would have been used for recharging, as in most residential off grid systems. The system has been up and running now for about a year, and so far there haven't been any problems with the operation that required modification of the system. I would be happy to talk with anyone planning a similar project.

Stephen Bosbach
September 2004
Revised September 2005

Sources and contacts:

Meridian Solar, Austin solar and wind system contractor

Home Power Magazine renewable energy magazine

Arizona Wind and Sun RE retailer

Solar Energy International renewable energy workshops

Texas Solar Energy Society educational organization

Database for State Incentives
for renewable energy

Austin Sustainable Building

Alternative Energy Store RE retailer

Solar Knowledge. Com educational web site

Add Energy Consulting cost / benefit analysis of solar install

Austin Energy rebate programs and education

Delta Lightning Arrestors lightning arrestor manufacturer

Harger Lightning Protection manufacturer and dealer - risk assessment analysis

CAIG Laboratories corrosion breaker and contact enhancer for low voltage connections - DeoxiT

Welding Depot source for 4/0 battery cable

Electrical and Construction
Bookstore source for electrical code manual

*Astronomy: SBD Observing Bench*

Materials list:
½ sheet of hardwood faced 1/2” plywood (4 ft. x 4 ft.)
Two 1” hardwood dowels 36” long
One 1.25” hardwood dowel 36” long
One small bottle of yellow wood glue
Brass flat Phillips wood screws #6×1”
One 9.5” folding brass leg brace, with two #8× 1/2” pan head wood screws and one #8×1”brass round head bolt with lock nut for pivot
One 5” gate hinge and three #10 × 1/2” pan head screws
plus three #10×24 oval Phillips head brass bolts and nuts (I cut down longer bolts)
One closed cell foam stadium cushion

Tools needed:
1” , 1.25”, and 1.5” hole saws
jig saw
If you have a table saw, it improves the straight cuts and square corners, but isn't absolutely necessary. Access to a drill press will make the lightening holes easer to bore as well.

Cut Pieces:
Main section:
Main seat back
2 Main seat standards
2 lower rear braces
2 rear leg pieces
2 upper leg spacers
2 middle leg spacers
3 lower leg spacers
2 1” dia. oak half rounds for feet
1 1.25” dia. oak half round for rear foot

Main seat piece
2 lower seat supports
lower seat cross brace
Two 1” hardwood dowels
Adhesive back Velcro for pad attachment and closed cell foam stadium seat cushion pad

Foot rest:
Two side supports
Two 1” hardwood dowels
Horizontal cross brace
vertical cross brace
One 1.25” hardwood dowel

Cutting and Assembly Instructions

The layout of the various pieces is done to minimize cuts and make it as easy as possible to cut long straight lines. Straight lines are righteous, but curves are sublime. Lots of curves were used in this design, and not simply for aesthetics. The curved corners and shaped braces distribute load and save weight. The total weight of the stool with padded seat and footrest is 14 lbs.

When cutting curved sections of twin pieces, it's best to cut the straight parts of the piece first, then stack the pieces and cut both curved sections together. They may not be perfect, but they'll be better than if cut separately. This is especially so for the scalloped edge on the back of the rear supports. Just take your time, as cutting two thicknesses of material will tend to skew the blade and create some inaccuracy in the cut, especially if you cut too fast. This also applies to drilling out holes for the dowels. Drill alignment holes with twin pieces stacked so bores made with hole saws line up precisely. This can be done with a 1/8” pilot hole, then follow with a 1/4” alignment hole. The drill bit for the hole saw will fit this alignment hole. Be sure to clamp your work or temporarily tape it together with masking tape to keep the work aligned. After the alignment holes are drilled, bore the lightening holes with a hole saw half way through from each side of the piece, ending the bore in the middle of the material to minimize splintering of the oak veneer.

It's also prudent to pre-fab the assembly with screws and no glue, just to be sure something isn't drastically out of alignment or needs further cutting. Once the unit is assembled and checks out, unscrew pieces and re-assemble with glue. I know, you want to skip this step. Measure twice, cut once. Enough said.

Hardwood faced ply is nice stuff, but it isn't perfect, and has a tendency to splinter easily. Also, slight bows and warps in the wood are OK, as the box construction of the chair will hold the longer pieces true. There will be voids in the wood in places unless you're buying marine grade lumber, and these should be filled with wood filler after your cuts and prior to sanding. This stuff hardens fast, so work quickly and wear rubber gloves. When cutting across the grain of the surface, covering the cut line with masking tape will prevent the veneer from splintering. It's more important on the blade exit side of the piece, but I did this on both sides of the cut. You'll be able to see your line right through the tape. Peal the tape off carefully, toward your cut, and soon after you make the cut, or the tape will pull off surface material from the veneer, leaving a slightly rougher appearance to the taped area.

The scale drawings are done 1/4” = 2” Lightening holes are your option, and can be left out if you want a simpler project. The larger holes in the chair back were first marked as 3” circles, with pairs of them drawn tangent to each other down the center line, then joined with straight lines from the perimeters. This makes an oval that's easy to cut out. These ovals were then spaced 3 inches apart down the front.

Once all the pieces are cut out, rough shape with 80 grit, smooth sand with 120 and final sand with 220 grit until smooth. Some folks stop at 120 grit, and your project will be presentable at this level, just not glass smooth. I found the inside of the holes and the scalloped cuts sanded easily with a sanding drum in a drill. I use a little Black and Decker “Mouse” sander and it makes quick work of the flat pieces. I also sand exposed edges and round over the sharp edges and corners. I don't sand edges that will be joined together, as your nice straight cut will be compromised by sanding. If only using a jig saw, purchase or fabricate a guide arm to attach to the base plate of the saw and guide off the outside edge of the piece. I've tried to arrange the pieces in the pattern so you could continue guiding off the last cut. When cutting a curve free hand, go slow and split the line with the blade.

Drill the screw holes as indicated. I used a 3/32 bit to drill the tap hole for the screw through both pieces. I then separated the pieces and drilled out the upper piece to 1/8” to accommodate the screw shank so the screw will seat and pull the pieces together. I then used a countersink bit to flare the top of the drill hole so each screw head would set below the surface. Some of the parts, like the rear leg and long front section, will be difficult to hold in place while holes are drilled through both pieces. In this case, I drilled the long upper sections first, then clamped the braces in place temporarily while drilling through the lower piece.

Part of the prefabrication is assembling to see how things fit together. The seat should sit fairly level. If not you may need to increase or decrease the spacing between the stationary dowel and the removable dowel. Decreasing it is easily done with a piece of 1 inch plastic tubing split lengthwise and slipped over the dowel. The rear leg needs to be checked for proper length after the feet are glued on. The rear leg should not touch the ground (or just barely) when the chair is folded and stood up. Aligning the drill holes for the rear leg was done by first attaching the long side of the T hinge to the leg, then after centering the leg in the main chair section, I taped the hinge down in place and lifted the leg to expose the screw holes for marking and drilling.

After you have prefabricated the chair with screws only, it's time for glue. I find it easier to use standard yellow wood glue, as it has a reasonable working time, and cleans up easily with water on a paper towel. As the work is screwed back together, the excess glue will squeeze out of the joint. Wipe off the excess quickly with a dry paper towel and then go over the area with a wet paper towel to clean off any glue residue that might interfere with the final finish coat.

When gluing the stationary dowels in place, I shifted them off center by 1/2” and applied glue all around the circumference. I then twist the dowel back to center, allowing the glue to work into the joint. Don't be afraid to use plenty of glue here, the excess cleans up easily as long as you work fast. Remember to leave the two rear most dowels free to be removed, so the seat and foot rest can be removed from the main piece of the chair for storage.

The feet at the bottom presented a problem. I would rather use a plastic foot that is water proof and could be easily replaced. I'm sure a local home improvement store would have something that could be adapted, but I opted to use some of the doweling material instead, as an oak foot would hold up much longer than the open grain ends of the plywood legs. I first split a one inch dowel down the middle on a table saw, 3” long cut , and cut those pieces off. On the table saw again, I cut out a groove along the length of each piece on the flat side, 1/2” wide and 1/4” deep, to just fit over the bottom of the feet. A dado blade would make short work of this, but I just made successive passes with a standard table saw blade. I glued these feet in place. The single foot on the rear leg is made from a 1.25” dowel, split in half, 1 and 5/8” long. This is left flat on one side and drilled through the center from the curved side and countersunk. This is then glued and screwed on to the rear leg.

After the glue has set, you may want to go back over some joints that don't quite match up with a sander to even them up. Inspect the entire piece at this point to look for veneer splintering or rough hole edges that can be cleaned up before a finish coat is applied. Once all the sanding is finished, wipe it down with a damp paper towel followed by a tac rag. I'm always surprised what the tac rag picks up, even after I've wiped the work down with a damp rag already!

I apply the finish coat of polyurethane with a brush, the old fashioned way. Finishing after final assembly is a compromise, as ideally one would finish the surfaces separately and then assemble, mainly to keep runs and sags to a minimum by finishing surfaces in the horizontal position. In this case there will always be parts that are vertical when others are horizontal. There are practical issues with glue bonding and time constraints that make prefinishing impractical for this project. After all, it's an observing chair, not heirloom furniture. We want a durable finish that will stand up to bumps, scrapes and the dew of the night. That said, I'm careful to keep the coats light and wipe over the edges of the horizontal surface just done with a paper towel to pick up stray drips and sags. It will pay you to take your time at this stage and just do horizontal surfaces, turning the work as it dries. Keep the light in front of you so you can see the reflection of the light in the fresh coat, and don't rush the brush. Plan on three coats of poly with light sanding of 220 grit between coats. Just be sure to allow adequate hardening before sanding and clean off the grit well with a tac rag. Your last coat will be very smooth indeed.

After the last coat hardens, I apply adhesive backed Velcro to the corners of the seat and attach the stadium seat. Enjoy!

Stephen Bosbach, Austin, TX