This page details a sun simulator that I use for testing solar thermal collector designs. The simulator provides light levels up to one full sun for a collector up to about 2 ft wide by 3 ft high. The simulator allows collector design changes to be evaluated under consistent conditions, and also allows easier instrumentation of the collector to get a better idea what's going on in the collector in terms of temperature distributions and flow patterns -- this can point the way to improvements in the design.
There are lots of sun simulators out there. What I was aiming for in this one was a simulator that was good enough to do good engineering work on collectors with a budget of a few hundred dollars. The final word is not in yet on whether this simulator get there, but things are looking good.
This page describes the "final" design -- there is some material on the earlier trials and tribulations here... I want to thank all the folks who sent in suggestions and comments -- many of those suggestions were helpful in getting to the current design.
The simulator uses six 400 watt metal halide lamps as the light source.
This is the spec for the lamp... From 1000bulbs ... (I don't usually recommend companies, but I have to mention that 1000bulbs.com has been really helpful through two orders of lamps and ballasts and odds and ends -- real humans who know what their products plus good prices and fast shipping). The lamp is 1.8 inches in diameter by 9.75 inches long -- 10 bucks a lamp. The lamps produce light by maintaining an arc through the arc tube that you can see inside the outer glass envelope. The arc length is about 1.5 inches. Each lamp has a ballast that provides the right arc starting voltages and maintains the arc current at the right level once the arc is established. The arc operates at about 135 volts once established.
The magnetic ballasts I used are essentially a transformer and a capacitor. To save money, I bought ballast "kits" which have the ballast transformer and capacitor, but no case. There are more sophisticated digital ballasts, but for this project, the cheap magnetic ballasts appear to work fine. The ballast kits for 400 watt lamps cost about $40.
The six lamps are mounted in an array that is 2 lamps wide and three lamps tall. Each horizontal pair of lamps gets a shallow parabolic reflector behind it. The reflector for each pair is 15 inches tall and about 30 inches wide -- so, the three pairs stacked gives 45 inches high by 30 inches wide.
A "light tunnel" surrounds the array of lamps and reflects light rays that are headed off for the sides and would not hit the collector back toward the collector. The light tunnel is 31 inches wide by 47 inches high. The inside of the tunnel is covered with reflective aluminized Mylar.
The basic idea is that the parabolic reflectors get all the light from about the back 220 degrees of the lamp and reflect it forward, but not necessarily right at the target. The idea of the tunnel is that light that would miss the collector gets reflected by the tunnel back toward the collector. In practice, the light tunnel makes for a significant improvement in light levels on the collector.
The reason for choosing a shallow parabolic reflector is that as you make the reflector deeper, the focal point moves closer to the apex of the parabola, and pretty soon the apex interferes with the lamp envelope. By using the small diameter version of the metal halide lamps and a relatively shallow parabola, I can place the center of the arc tube right at the focal point of the parabola -- this results in the light being reflected straight toward the target. Of course, the arc tube is about an inch in diameter, so not all of it is at the focal point.
This arrangement achieves full sun light levels about 13 inches from the mouth of the tunnel. The light levels drop off as the collector is moved further away.
The table below shows the light reading in foot-candles on the glazing of the collector. This is with the collector 13 inches from the end of the light tunnel.
The center point is the center of the glazing, and the other points are basically 6 inches in from the edges of the collector glazing (sides and top).
Light levels on glazing surface in footcandles.
This table shows the light level at each location as a fraction of the mean
light level of all the cells.
So, the top 6 points look good -- all within about 5% of each other.
The bottom row looks a bit dim -- about 7% down from the mean except for the bottom right corner, which is 15% down from the mean.
I'm going to work a bit on figuring out why the bottom row comes in lower, but I think the distribution is good enough now to do testing.
The reflectors were made from 0.018 thick sheet aluminum to which aluminized Mylar was attached.
Forming the segmented parabolic reflector.
The reflector after rough forming.
The three reflectors are supported by these two template
boards, which get mounted to the light backboard.
Reflector supports mounted to the light backboard.
Reflector mounted to the form, which enforces the
With the aluminize Mylar applied.
This shows the light warming up. You can see the lamp's
reflection in each segment of the parabolic reflector -- I'm
taking this to mean that the shape of the reflector is good.
Testing with two of the reflectors done.
The ballasts mounted on the back of the light board.
The 6 junction boxes on the edges go to the 6 lamps.
The junction boxes in the middle are switches to
turn the lamps and the light tunnel ventilation on and off.
I had to split the 6 lamps onto 2 circuits because all 6 on one
circuit would exceed the breaker rating.
Another view showing the light tunnel.
Shows the perimeter frame around the reflectors that
will support the light tunnel.
The tip of the drill bit is on the perimeter frame.
The light tunnel sides are light weight quarter inch plywood
with 1 mil aluminized Mylar help in place by 3M 77 spray
The light tunnel in place.
The collector is to the left of the light tunnel.
The ventilation plenum and ducts are explained below.
View into the business end.
(the lower right lamp is was on back order at time of picture)
A bit more from the side. Note all the distinct lamp
reflections on the light tunnel side panel.
(again, one lamp is not installed yet in this photo)
Just a quick check on performance shows light levels around 100K lux (full sun) at 13 inches from the end of the light tunnel. This is just OK as it leaves just enough room to play with and take measurements on the collector during testing.
The temperatures around the reflectors are surprisingly low -- no problem to touch anything and leave your hand on it. I left ventilation space behind the collectors, and I have a salvaged bathroom vent blower ventilating the light tunnel. Standing in front of the light tunnel fells a lot like standing in the sun on a sunny day.
The prototype collector is 2 ft wide by 3 ft high and 6 inches deep. It will be used to look at variations in solar air heating collector designs.
The collector is made so that both the front (glazed surface) and the back (insulated surface) can be easily removed. This will allow quick changes to the collector internals, and also allows glazing that is IR transparent to be substituted for the regular glazing so that the IR camera can be used to take pictures through the glazing of the absorber.
The collector is set up such that it can be rotated quickly to point sideways rather than straight into the light tunnel to take the IR pictures.
The collector frame with the insulated back panel in place.
The aluminum angles hold the glazing frame in place.
The supports for the screen type absorber have been added.
The collector with the absorber and the front glazing installed
the glazing comes off with a few wing nuts (or clamps).
The IR transparent glazing will be built on a frame that is
interchangeable with this frame.
The stand is just a sort of sled that supports the collector at the right height and can be easily moved back and forth to get different sun levels.
The collector mounted on the collector stand.
The stand distance from the light tunnel is varied to achieve different solar intensities.
Outside ambient air is brought in through a duct to cool some areas of the collector and sun simulator.
The collector glazing has an air distribution tube along the bottom of it that is set up to provide an about 3 mph "wind" at the outside ambient air temperature. The tube has an about 1/8 slot cut along its length for the full width of the collector glazing -- the ambient air exits this slot and flow along the glazing. I check the "wind" speed next to the glazing with a Kestrel wind meter. Since the bulk of the heat loss from solar thermal collectors is out the front glazing, its important to be have a flow of outside air over the glazing at a realistic wind speed.
At the end of the light tunnel, a clear acrylic panel is used to control the IR emissions from the lamps. Without this clear panel, the collector glazing runs hotter than it would in the real sun. I believe that this is because the lamp envelopes run hot and act a bit like a space heater aimed at the collector glazing. The clear Acrylic panel at the end of the light tunnel absorbs most of this unwanted far IR heat radiation. In addition, if the Acrylic panel can be (hopefully) cooled down to a temperature not to far above ambient. With the heat shield cooled down, the collector glazing will be exchanging radiation with a surface (the heat shield) that is much closer to a real outdoor setting -- this should make the glazing radiation losses more realistic.
The heat shield panel runs quite warm if no cooling air is directed over it, so I've arranged for the same blower that supplies ambient ventilation air on the collector glazing to also cool the heat shield. In the current ventilation arrangement (shown in the picture above), the heat shield runs quite a bit cooler, but still hotter than I would like -- I plan to work out a way to increase the ambient air ventilation on the heat shield -- maybe a separate blower -- any ideas?
At any rate, with the ventilation shown, the glazing temperatures on the collector are in the range one would expect for an outdoor test. But, I still have to do more indoor to outdoor comparison testing.
The air supply for the collector intake is just taken from the room. The collector exit air just blows out into the room.
My hope is that by starting a test early in the morning (when ambient temperature is low), I can get a couple points with different sun intensities at low ambient air temperatures, and then by repeating the test later in the day when ambient temperatures are higher, I can get two more sun intensity points with the higher ambient temperatures. I can also just blow room air over the glazing to get points with less temperature difference from inside to outside the collector. Maybe this set will be enough to establish a first cut efficiency curve? Bozeman has lows of 40F or below 225 days of the year, so I should be able to get descently low ambient temperatures most mornings.
The collector support stand allows the collector to be rotated on a pivot bolt to the side so that once equilibrium conditions have been reached with the collector pointed right at the the sun sim, the collector can be quickly pivoted to the side and an IR picture taken before conditions change significantly.
With the normal glazing, this will give a temperature map of the glazing, and with the IR transparent glazing, it will give a temperature map of the absorber.
Collector pivoted to side for an IR picture.
I've not really kept a careful accounting, but this is pretty close:
|Item||Quantity||Unit Price||Total Cost|
|400 watt metal-halide lamps||6||$10||$60|
|Wire -- scrap box||$0|
|Electrical boxes, wirenuts, wiring supplies...||$30|
|Plywood for lamp board and light tunnel||$55|
|Mylar reflector film||$17|
|Vent fan for light tunnel -- scrap box||$0|
|Metal and other supplies||$62|
|Plywood for collector stand||$15|
The cost table does not include money spent on the trial versions (not a great deal).
The cost of the test collector(s) is also not included.
This is not a token cost project, but for school or club or solar nut (like me) its manageable. There is still a lot of room for design improvement on solar thermal collector to make them more efficient and cheaper to build. The simulator (I think) will make it easier to see what's going on in the collector and to test design changes that might improve the collector in cost and/or efficiency.