Introduction: Heliostat for Solar Heating

This enables DIY'ers to build a $111 solar house heater and light source. It pays for its parts in 2 winter months and should last as long as the PVC survives the outdoor winter. It provides 4.5 kW for 6 to 8 hours per sunny day, equal to three 1,500 Watt electrical heaters. If it lasts 10 years, this works out to be $0.005 cost per kWh, 30 times cheaper than electricity. The supports are spring-loaded to let the boards fall to the ground in moderate wind. A galvanized steel design would be better.

Click on the first image to see the ANIMATED GIF.

The part sticking out back with the wires does not really help and turned out to not be needed. If it is close to a window like this, just out of the shadow of the roof, a single motor can be used and no adjustment all winter is needed if the axis and tilt off the axis are just right, as described below. Some of the pictures are just past attempts. The main problem for a DIYer is not being able to find the coupling to the screw for the motor (I had to make it from PVC and epoxy) and getting the bolt for the screw at the other end on a secure swivel.

Keep in mind this might burn your house down.

Another warning: all PVC like this results in a "flimsy" design that has a lot of movement in it. It works fine and is acceptable for me. The heat problem I show in the video has been fixed by screwing small wood pieces behind the middle of the 4x8's into the PCV to make them bow out in the middle just a little. It's working get except the wind release didn't work today and I have a little repairing to do.

The circuit diagram has an error: the right side is correct. The left side needs to be wired to the AC voltage supply the same as the right side.

A few notes: 1) the horizontal arm needs to be about 4" shorter so that there is less bending. 2) The PVC sticking out the back with the wiring is not needed. 3) The central two axii need to be different to reduce the large wobbly nature of this PVC design that most people are not going to think is acceptable. I have a design in mind but it's hard to describe with words and I haven't built or drawn it. It's a small 3D truss coming out the back where the horizontal bar will need two axis "couplings" to make connect points for the truss on each side of the middle. The objective is to get the middle T out of the picture because it gives too much.

There are details about construction that I need to upload videos for, probably during this 1st week of September.

It uses two 4x8 insulation boards covered in aluminized mylar and mounted nearly vertically on a pole and rotating to following the Sun so that sunlight reflects into a large window or sliding glass door on the NORTH side of the house. This "heliostat" is placed exactly north of the window about 15 feet from the target window or solar cells (20 or 25 feet needed for more northern houses especially if the roofs are steep or two stories). The aluminized mylar from Amazon or ebay is attached to both sides of the 4x8's (for strength and appearance) first with spray adhesive and then with clear packing tape around the edges (which has lasted 4 years on mine). If a target mirror is placed beneath the heliostat at an angle equal to the latitude, then the heliostat can turn at 15 degrees per hour on a single axis and the target mirror can reflect the light into East or West windows, but more ideally into a South window that is behind the heliostat. Another single axis

I heated our 4000 sq ft house last winter with two of these devices using $200 of materials. It saved about $400, lowering the natural gas bill from $800 to $400 plus reduced electrical costs for lights and the heater blower. Just the extra intense Sun-light coming horizontally down the hallway is worth the effort (more than wonderful) erasing the feeling of winter.

The first image is an animated GIF that shows it in better detail and ends with a circuit diagram that is WRONG. The voltage supply is supposed to be on a common line in between the two symmetric circuits of paired MOSFETs, not off to the side on one of them. It's a very efficient circuit which means you have to wire it right or none of it works, which means you will not be able to isolate a problem with the wiring if all parts work.

Comparison to South-facing windows and vertically-mounted Solar Hot air boxes
A North wall-mounted solar air box could also be used as the target, getting up to 6 times more heat compared to a south wall-mounted solar air box because the light can be tripled and it has tracking. But the best solar air boxes are not as efficient as simply letting the light entering a house window of the same size because little light energy reflects back outside. But triple light concentration is too much for anything that may be flammable. The heliostat needs to be 50% wider to get the most out of the first and last hours of the day, plus twice this to get a doubling of light. Double light intensity is as safe as two mirrors at 30 degrees reflecting light onto a surface facing the Sun.

September 7, 2013 UPDATE (was originally July 28)
I have three PVC heliostats built now and will continue to improve this instructable to make it easy to implement. There are many "snags" that can occur and I would wait until I finish this instructable before attempting. I have made changes to the animation's design as follows: 1) no wires for the back support. 2) 1 steel or thick-gauge copper wire going over the post that supports the top motor and attached to each end of the main arm to keep it from bending down under the weight 3) absolutely need epoxy under the top 1.25" sleeve (orange in animation) instead of PVC glue. 4) Low pressure 1" instead of high pressure 3/4" PVC arms running up and down (4 of them). They are lighter and bend less. 5) it's important but not necessary to find 3/4" steel, copper, or aluminum for the 2 axis (3" pieces needed) instead of the high pressure 3/4" PVC. 6) need stand-offs in the middle and on each side of the 4x8's to spread out the light, especially if $10 spray adhesive is not used to hold the aluminized mylar tightly to the 4x8's 7) I use a small spring to keep one of the front T's in the middle from coming off, which means wind force on the 4x8's can push it off, allowing that side of the 4x8's to fall down. A strong connects that T to a pin (a nail) that holds the other T in place. So they both come off at the same time in strong winds, allowing both 4x8's to fall to the ground.


$26 PVC parts (see below)
$22 two 4x8 insulation boards
$5 PVC glue
$2 two 1/4" eye bolts, 1.5 to 2.5 inches long, 6 1/4" nuts
$2 three foot 1/4" screw rod from Lowes
$5 clear packing tape, the wide plastic stuff like 3M.
$5 good epoxy like JB-kwik.
electronics option 1:$20 for four 9 V 150 mA solar cells
electronics option 2: $4 12 V AC supply, > 100 mA, $2 four IRF530 mosfets, eight 10 Mohm resistors, $3 twenty-four 50 mA 5 mm LEDs of any color or infrared (6 needed to drive turn on the cheap mosfets)
$15 two 12 VDC motors, about 100 mA, vertical motion 20 RPM, horizontal motion 50 RPM. 1 motor is optional if using single axis method I'll describe in more detail later. (The 1/4" screws are 20 turns per inch, so 1 to 2.5 inches per minute. More than 50 RPM is not powerful enough for 1 W motors)
$20 for 4 mil aluminized mylar. ($4 for 4 emergency blankets is possible but not very durable considering all the moving and bumping of the boards)
Free small gauge electrical wire from somewhere. (depends heavily on your setup and my final design for placing the sensors either close in front of the heliostat or beside the target...could get expensive.) Optional adjustment: Only 1 solar cell is needed if single-axis and manually-resetting in morning.

Total Cost:
$123 for 2 axis method, using solar cells
$111 for 2-axis method using electronics

Tools Required

Drill with good selection of bits from 3/32" to 1/4".
Dremel tool or angle grinder to help special PVC cuts and "rectangularize" the ends of the 1/4" screw and the motor shafts so that they do not slip while turning.
Optional: Soldering iron and solder if you use the electronics option.

PVC list:
$2 one 10' 3/4" pipe high PSI (480) for bottom arms and axis, $2 one 10' 1" PVC pipe low pressure (top arms)
$3 one 10' 1" pipe high PSI (450)
$4 10 pack 3/4 elbows (need 6)
$6 10 pack 1" elbows (need 6)
$2 five 3/4 couple
$6 10 pack 1" T
$1 two 1" couplings
$0.50 1.25" coupling

Building the AXIS: 3/4" couplings rotate perfectly inside 1" T's. Connect two 3/4" couplings with 3" of 3/4" pipe to go inside a 1" T, so the 3/4" stuff rotates inside. Now 1" T's, elbows, and couplings can glue to the outside of the 3/4" couplings, VERY close to each side of the 1" T so that they are NOT glued to it. So 1" stuff is connected through the 3/4" axis to other 1" stuff at 90 degrees via a rotating 1" T. To optionally make the 3/4" axis stronger against bending, coat the middle 2" of the 3/4" pipe with epoxy and embed fiberglass or find a 3/4" metal pipe instead of the PVC.

Connecting Motor Shaft to Long Screw: The only hard thing that always seems to be a problem (with DIY's like me who don't weld...yet) is joining the 1/4" screw shaft onto the end of the motor shaft. So far the best I have is to cut 7 of 1/2"x1/2" pieces of PVC from a coupling that is the harder plastic and bend them more flat with pliers. Metal snippers are easier for this than hacksaw or dremel. Then use epoxy, not PVC glue to make a stack (they do not fit perfectly, so PVC cement is risky). This forms a hard block of plastic that you can drill each end, one for motor shaft and one for the 1/4" screw. Use dremel to square off both motor and screw shafts and make 2 side-by-side drill holes in the plastic on each side, to fit each shaft, leaving some play in the 1/4" shaft hole so the motor does not "bind" to a stop. To keep them together and prevent force on the motor shaft, a T coupling or end cap or elbow is fixed in front of motor (not shown in current animation), 3/8" hole drilled for 1/4"shaft to pass through, and a nut on the shaft is epoxied on each side of PVC hole to allow for 1/4" screw force to be on PVC instead of motor, leaving the motor and its PVC stack coupling to the 1/4" screw inside the PVC. Motor most be kept from turning inside the PVC by epoxy if you're sure of yourself or by a screw through PVC side for your first try if you're unlike me and therefor more wise.

Connecting other end of long screw: If a 1/4" nut is welded into the eye of the eye bolt, carefully select the eye bolt from Lowes so that the 1/4 nut can go inside and liberally apply epoxy everywhere surrounding the end of the eye bolt except on the nut's threads. It should be capable of supporting 10 times more force from wind than the rest of the PVC heliostat can support.
Pivoting Motors: See the animated GIF images. I found a small screw I liked and found a drill bit to match the diameter. Then I made two very close hacksaw cuts into the 1" elbow as shown.

Morning reset:
I plan to use the LED sensor or solar cell that would normally go on the west side of the window to do a morning reset. I imagine placing this in just the right spot in the shadow of the heliostat using an extension of 12 AWG copper wire or clothes hanger from the heliostat support base for easy bending for adjustment of position. So when the Sun comes up, the sensor will cause the heliostat to turn East until the sensor is in shadow again. The size of the heliostat and the closeness to the target should allow this to work whereas it would not work in most heliostat situations. Otherwise, you would need to use my single axis design described below for accuracy (long distance) work. As far as I know, the single axis I've described is original work.

Choice of glass for window
It is difficult to get accurate numbers on how much light makes it through typical single and double pane glass, especially at odd angles. Due to the iron in cheap glass (green on the edges) and "low E" glass blocking a lot of light, plexiglass appears to be better.

The lowest theoretical cost of solar concentration with a heliostat using aluminized mylar is limited by wind, a common regulation being 1 sq ft of surface capable of withstanding 10 pounds of wind pressure. Theoretically, I get that at least 1 kg of steel is needed for each (square meter)^(3/2) at 10 lbs/ft^2 front-back pressure based on a cable design that I think is the most efficient (I used 400 MPa for tensile strength of steel). So 1 m^2 requires 1 kg of steel for a safety factor of 2. 20 hours of 1 kW Sun on 1 m^2 is the energy payback for 1 kg of steel.

I used general purpose 3M spray adhesive from Walmart and clear packing tape (for the edges) to attach aluminized mylar ($20 from Amazon) to the front and back of four 4x8 insulation boards. I did the back because it makes the boards REALLY strong and durable. I've used them for 3 or 4 years now with little sign of "wear and tear" and the packing tape still sticks. PhD Russian physicist Y_Po from theeestory suggested the insulation boards, and Nekote suggested the aluminized mylar.

The Sun light that hits a directly-sun-facing 4x8 is 3,000 Watts (3 meters^2) as long as the Sun is not too low in the sky. It is about 2,500 Watts at 9 am. Double-pane windows and the aluminized mylar block and reflect away about 18% of the energy. Framing of the windows block another 10%. After the light enters the house, only about 5% will leave back out the window even if everything inside is a light color. So efficiency is about (1-0.18)*(1-0.10)*(1-0.05)=70%, 2,100 watts per 4x8. Everything else is direct heating, so this is a very efficient use of the Sun's energy. If the reflector is up to 50% wider than the window so that the full window gets the sun in morning and afternoon, this is many times more heating than a south-facing window that suffers greatly from "cosine angle" losses (see tech note below). With the COP efficiency of a heat pump and the cheaper price of natural gas and oil verses electricity, I would guess heating costs to be about $0.08 per kWh when compared to a direct electrical heat-element heater that costs about $0.13 per kWh in US. With 3 kW of Sun for each insulation board and 6 good hours of sun hitting it almost directly and 80% sunny days this method saves about 2.1W*6*0.80*0.08 $/kWh = $0.81 per day per 4x8 which is $244 for two 4x8's on 1 heliostat for 150 days of winter (5 months). My observed savings agree with this estimate, even with oak tree blocking Sun and inefficient reflector aiming.

2 well-placed 4x8s and a big north window are needed for every 1,500 sq ft to cut heating costs in half, or more. The house will reach over 80 F and with good insulation the normal heater may not be needed until midnight. Ours turned on about 9 pm (southern U.S., 1979 house, 50 F nights, heater set to 70 F).

There needs to be a middle support for the 4x8's to bend them slightly outward.

Single-Axis Heliostat (Only good for winter and above 20 degrees latitude)
It's possible to use only 1 solar cell on the East side of the window and 1 DC motor (one turning axis) but it requires manually adjusting the mirror tilt once a week, and manually resetting each evening or morning. Actually, you want the heliostat to point down to the ground at night because the sky (space) will "absorb" a lot more Stefan–Boltzmann radiation (Twindow^4-Tspace^4) from the windows than trees or even snow on the ground because the sky is about 50 F lower in its black-body radiation temperature than ambient temperature ground radiation. If you have a thermal thermometer you can observe this by pointing it into the sky, especially at night.

For U.S. latitudes and the heliostat center at the same height as the window and directly north of the window, you set the
single axis degrees above north horizon = 0.83*latitude - 6.9
This is for less than 0.5 degrees error from 9 am to 3 pm, winter only. The bottom and top of the reflector are perpendicular to this axis (running east-west) but the reflector has to tilt upward on this single axis towards the target window from 35 to 60 degrees depending on day of the winter (Sun's declination) and your latitude. The equation is
Mirror tilt degrees from axis = latitude/3 + 26.3 +15.3*sin(days)
where "days" is up to 90 before or after December 22, the sign making no difference for sin() function. The zero angle is when the mirror is sitting perpendicular on the axis, reflecting south down into the ground. Add 90 if starting position is on it's face on the axis, or subtract the number from 180 if it starts on it's back. I developed these equations, so you will not find them elsewhere. They have about 0.5 degree error.

If the heliostat is x degrees above (or -x if below) target, add 5x/(latitude-15) to 1st equation above and 5x/(latitude-60) to the second. I checked this adjustment for heliostat up to 30 degrees above target, but to get less than 1 degree error, heliostat can only be up to 10 or 15 degrees below target. You want the target window as high as possible relative to the heliostat without the shadow of the roof blocking the heliostat because you lose "cosine angle" efficiency, but a single axis will not work very well if your target window is much more than 15 degrees higher than the heliostat.
1st Example: 30 degree latitude on December 22 gives 18 for first equation and 36.3 for second equation. At noon at this latitude and date the sun will be 36.6 degrees in the sky. The mirror is 36.3 degrees up from the axis, but the axis is 18 degrees below the SOUTH horizon, so the angle up from horizon is 36.3-18 = 18.3, which is half way up towards the Sun as it should be (36 degrees instead of 36.6, 0.0 degrees error).
2nd Example: same example except 45 latitude. Sun is 21.6 degrees at noon. 1st eq: 30.45 degrees tilt beliow south horizon. 2nd eq: 41.3 degrees up. so 41.3-30.45 = 10.85, which is ideal if the sun was 2*10.85=21.7 degrees high in the sky instead of 21.6, so this example has an error of 0.1 degrees.

A problem is that these large heliostats are getting too away from the ideal center point. Fine-tuning your setup to be more accurate based on observation takes a visit in the morning, mid-day and afternoon, for several days. And moving it it is likely. So all in all, 2 axis with letting sun sensors avoid all calculations as I described above may be best. My wooden-structure pictures show 3 axis because 1 axis was for testing this single axis idea.

The pictures show a past attempt, a "stand" that held three 4x8's for a window, but the heat intensity was too dangerous and adjusting the angle of each board several times a day was a hassle. My design was not able to withstand more than 15 or 20 mph WIND.

I am also looking into a water-absorber heliostat design using a car radiator for the target and another for heat release inside the house that will not need the windows and can store energy at night. It will have less wind problem on the ground rather than my high deck. 15 mph was max wind they could take and the wood screws get loose too much. But notice that I've been doing it for several years now and the benefits are great enough that I can't stop. Winter without the Sun in my house has become unthinkable (I work from home, so it's more important to me than to others).

Comparison to South facing windows: 48 degrees is the average of the 3D angle between the sun and a directly south target window (heliostat due north is the vertex, same height as window) from November 1 to April 1 from 8:30 am to 3 pm and latitude 32 degrees (Atlanta). At latitude 41 degrees (Chicago) the average is 41 degrees. The average "cosine angle" light loss of a South facing window is therefore cos(48)=0.66 => 34%. A heliostat that points halfway between the window and Sun is less than half of this due to the cosine effect: cos(48/2) = 0.91 => only 9% loss. But if the heliostat is large enough, about 50% wider than the window, there is no cosine angle loss. The diagonal of a 6x6 window or sliding glass door is 8.5 feet, so a 12 foot wide heliostat across the diagonal is REASONABLE because you really want morning heat. Getting back to comparison to a South window, there is an additional higher reflection loss from a south window from the steeper sun-incident angles if it is before 9 am or after 3 pm, close to 60 degrees incident angle.

Hot Air Box

I have also built hot air boxes (see last picture for how to double light into a hot air box that has greenhouse plastic as the "window". The reflectors on each side are set at 30 degrees tilt back. 6" ductwork into and out of house is the main problem. 50% efficient verses the 70% efficiency of aluminum can designs (not worth the effort) ... one 4x8 needs at least a 100 CFM 4" fan from MPJA or bgmicro. The backing was 4x8 insulation board painted with black Rustoleum high temp paint. The was no plywood backing, just 1/4" playwood (luan) sides with a $2 shelving bracket at each corner, and two 1x2 at 1/3 and 2/3 length of the insulation board for support...weighed only 15 pounds. The mylar-coated reflectors folded over to close the box for transportation. See my youtube channel if you want to get crazy with hot air box calculations and see the single-axis heliostat rotate.

Weekend Projects Contest

Participated in the
Weekend Projects Contest

Arduino Contest

Participated in the
Arduino Contest