Design - Solar Collector - Problem-Solving

This is fifth post on the solar collector but, undoubtedly, not the last since I will reporting on its performance over time. The purpose of this post is to report that it has not lived up to expectations and to suggest changes that will make it work.  

Lack of Performance

The glass for the collector was installed by the middle of July just in time for lots of clear skies and 90+ degree temperatures that should have maximized airflow from the collector into the conduits.  I gave the system a few days to rev up, thinking that it may take a while for the hot air pushing up the conduits to overcome the cold air dropping down.  During the early afternoon on a bright and hot day, I checked the air flow from the north ends of the conduits and found that, if there was any movement at all, it was so minimal that I could not detect it.  I waited for another sunny day but one without any wind in order to dangle thin strips of paper from the conduit outlets that would detect any air movement.  Still no detectable convection.

At our latitude, the temperature of the soil where the house sits would be in the mid-60s by mid-summer at the depth of the conduits.  However, since the house temperatures during the past couple of winters have stayed above 40 degrees (despite no insulation), it is reasonable to think that the temperature of the earth under the house surrounding the conduits is higher than it would be if not shielded by the house.  If so, the temperature of the air falling out of the conduits might be as high as the 70s but apparently still too cold for the heated air from the collector to overcome. 

The performance was disappointing but not totally unexpected given the dearth of practical information available in print and on the web, which means that our design had to be largely original.  I look at the situation as just another problem that needs to be solved and reported on just like many other surprises and challenges that we have encountered with such a unique build.

Are the Conduits the Problem?

In a previous post, I listed some of the unknowns that come into play.  "Assuming the design of the collector is adequate, its function is still at the mercy of many unknowns about passive air flow through the conduits.  Will 4" diameter conduits be the optimal size for sufficient airflow?  Are conduits that are nearly 90' long from collector to daylight behind the house too long to expect passive flow?  Do they angle upward enough from 10' below floor level when they leave the collector to a depth of 3 or 4' below floor level at the back wall of the house and then make 45 degree turns to daylight?  Will using the corrugated (rather than smooth) piping under the house (the intent for which is to cause turbulence in the air flow and thereby improve heat transfer to the soil) slow the flow too much?  Will the cooler soil during the first winter and, to a lessor extent, after each succeeding winter, cause cool air to flow backwards towards the collector to the extent that the warm air from the collector cannot passively reverse the flow?"   Of all of the items on the list, only the one in italics can be addressed at this point -- the rest are what they are.  At this juncture, I would add one other possibility.  The conduits terminate with two 90 degree fittings in order to keep rain out.  Perhaps if the conduits pointed straight up, there would be less resistance to passive flow.

Is the Solar Collector the Problem?

The problem could also be in the design of the collector rather than in the conduits.  Maybe one layer of galvanized roofing is not enough; maybe multi-layers are necessary.  Maybe there is so much space between the metal and the glass that the volume of heated air is insufficient for spontaneous escape up the conduits.  Perhaps the collector is not large enough to supply nine conduits passively.  Maybe it will be necessary to add to the system one or two what might loosely be called "solar chimneys" whereby the conduits would be brought together and exit to daylight through a common chimney, with or without the assistance of a fan.

At first I assumed the problem lies with the

Termini of the nine conduits.  The one in the middle was
modified to accept a vacuum hose for testing.

conduit portion of the system instead of the collector which means there is only one factor I can test -- the one in italics above. I manipulated the airflow in one conduit to see if it could be jump-started to overcome the effect of the cool soil by cutting away the double 90s in order to fit the end of the conduit with an end-cap having a hole the same size as a vacuum hose.  I pulled air through the conduit with a vacuum for a couple hours hoping that, when the cap was removed, I would feel warm air, or any air for that matter, coming out of the conduit.  Such was not the case. 

The next probable cause for under-performance that could easily be investigated was to measure the amount of heat the collector was producing.  Having assumed that the temperatures would be too high for plant growth, 

Thermometer resting on the metal is maxed out.

I began to suspect that heat generation was insufficient when a couple of plants sprouted along one edge of the collector.  I pulled the plants and placed a thermometer of the common type with a scale to 120 degrees inside the collector.  It recorded temperatures approaching 100 in the early morning when only the west half of the collector was sunlit and the thermometer was shaded by the east wall of the collector.  As the sun reached the glass fully, the temperature readings quickly rose and stopped at the maximum capability of the thermometer somewhat above 120.  And further plant growth has been non-existent.  So, pending the purchase of a thermometer with a higher range, the initial readings are encouraging enough to look elsewhere for ways to make hot air flow through the conduits. 

Reconfiguring the Terminal Ends of the Conduits

It is becoming obvious that the conduits will require redesigning at their terminal ends.  Instead of nine conduits exiting to daylight independently, I am now convinced that they need to converge into one or two solar chimneys fitted with a solar fan(s).  Since each conduit is nearly 90 ft long, I think two chimneys with four or five conduits each would be more efficient than one chimney located 40-50 feet from the termini of the most outlying conduits.  At the time of this writing, it is mid-September and completion of exterior trim for the house is the highest priority.  Reconfiguring the conduits will have to wait until spring.

Construction - Solar Collector - Safety Fence

The previous post covered the functioning part of the collector.  This, the fourth post on the collector, describes the safety fence surrounding it, the design for which needed to meet

The framework for the safety fence around the collector shell.

several criteria.  It needed to be tall enough to meet code for handrails, it needed to be as attractive as possible considering its prominence in front of the house and it should not block the view from the house any more than necessary.

(Reminder: click on any picture to enlarge it for better viewing.)

Upgrading the Top of the Wall
When the collector shell was built five years ago, the top of the walls were covered with 2 x 12 pressure treated planks that were anchored with bolts embedded in concrete.  While they protected the top of the walls, they left much to be desired aesthetically and begged to be replaced.  It took only a few minutes at the landscape supply store to decide in favor of concrete pavers instead of pricey capstones.  I dry-fitted them to determine where the posts for the fence should be located in order to minimize laborious notching of the pavers.  I then installed the posts and notched the pavers using a diamond blade in the radial arm saw.  I also custom crosscut a few pavers as necessary to filled the gaps that would not accommodate full sized units. 

Post Placement
Like the solar collector framework, the pressure treated lumber for the fencing was stickered and air-dried for several months so that it remained straight and would accept and hold paint.  Then, before assembly, it was undercoated on all six surfaces and final coated on at least three surfaces.  The 4 x 4 posts extend below the ground to the depth of 12 - 16" so they needed to be ground-contact rated; the pressure treated lumber for the rest of the railing system needed only to be rated for above-ground use.  The corner posts were "V"-shaped at the bottom so as to rest on the top of the wall and straddle both surfaces of the wall below.  Two long and robust Tapcon concrete screws through each surface was more than adequate to anchor them firmly and, with the sometimes help of composite shims, hold them plumb.  Instead of "V-ing" the bottoms of the intervening posts (south and north sides), they were reduced in thickness by a half so as to rest on top of the wall and extend downward where they were anchored by four screws each.  Backfilling and tamping the soil around all of the posts will help to support them as well.

With the posts in place I could lay the pavers in mortar.  The weather was hot and I got in a hurry to get done so I laid them without the benefit of a mason's line.  The result, I am sorry to say, looked pretty amateurish.    

East and West Sides 
The wood framework for the enclosure supports a wire grid fashioned from cattle fencing that I cut from panels that come 4' high and 12' long.

The fencing for the east and west sides of the enclosure presented a challenge in that it had to be stepped to follow the contour of the stepped walls and its top had to slant to follow

Temporary layout board.  A few of the cap stones
 are yet to be mortared in.

the slope of the stepping.  Since the cattle panels were 4' high, I installed a temporary board 4' above the lowest step and leveled it.  Off of that I could measure the amount that the fence panel would have to be reduced over each step uphill from the lowest step. These measurements, along with the length of each step and the slope at the top, were laid out on a cardboard pattern.  After dry-fitting and tweaking the pattern, I laid it on a section of fencing that had been cut to length and marked the wires to follow its edges.  An angle grinder with a metal-cutting diamond blade easily handled the 4 gauge wire.  The cardboard was accurate enough to be used as a guide for cutting and fitting the support board at the top of the fence.  I used 2-by blocks under the wire panel to hold it off the wall slightly as I attached the board at the appropriate height to catch enough of the top of the fence board for secure fastening.

Dry-fitting the cardboard pattern to the east end of the 

The cardboard pattern in place over the cattle fencing
for the east end of the collector.

collector showed that, with one minor adjustment, it could be used for cutting the east fencing panel. I added the top board on the east at a height that matched the board on the west.  Confident that metal panels were installation-ready, I set them aside so that they would not interfere with laying the pavers with mortar.  

North and South Sides
Compared to the east and west sides, the north and south sides were easy.  For the north side, I installed the top boards at a height that matched the northern ends of the side boards and the bottom boards at 1 1/2" above the pavers.  I added 2x2 nailers to the sides of the corner posts to receive the vertical edges of the

Fencing completed except for caprails.  The ladder
used during construction is still in the collector.

fencing panels.  As an cosmetic touch after the panels were fastened to the top and bottom boards and to the posts, I added  1-bys on the collector side of the entire framework.

I decided to delay the fence for the south side until the glass was installed on the collector in order to leave plenty of access with the heavy glass panes.  I entertained the idea of building an access gate into the south fence but decided against it based on looks, which means that we will have to drop an extension ladder into the shell when access is necessary for maintenance.  The design of the south fence was then identical to the north fence.

Post Caps and Cap Rail
Finally, I added store-bought pressure treated post caps to the tops of the posts -- for aesthetics and to protect the end-grain of the 

In lieu of a gate in the fence, a ladder will have to
be used for servicing the collector

posts from deterioration.  The caps for the top rails will be 2 x 6s that I customize with bi-sloped tops to shed moisture eventually.  Their addition is being postponed while more urgent projects on the house are handled.