Here is my new solar airplane (the previous version is here). 

As before, it uses the Parkzone Vapor's radio and motor, plus five Powerfilm solar cells.  In parallel with the solar cells, I include a tiny 20mAh lipo battery to keep the radio happy.  However before I hook up the solar cells, I make sure that there is not enough juice in the battery to run the motor (no cheating).

The wing is a vast improvement over the previous version.  It is stiffer torsionally, has more span, more area, and a better elliptical planform.  The wing and tail are covered with mylar from one of those emergency survival blankets.

Here are some specifications:

Wing span: 35in (89cm)

Wing area: 165 sq in

Total Weight: 42g

Last Saturday was the first calm and sunny day all year (Seattle).  Although conditions were not perfect (low sun angle, a bit too windy), I still managed to sneak in a few flights (see the video below).  I think with the right conditions, this airplane could fly for hours.  Finding a day with the right conditions might be challenging, though.  It is often very calm early in the morning, but the sun is too low.  By mid-day, when the sun is high, thermal activity tends to kick up random gusts.  I'll keep waiting for that perfect day... 

I have been tinkering with solar powered RC planes this summer. The photos above shows the latest incarnation.

This model uses the Parkzone Vapor's radio gear and motor ( http://www.parkzone.com/Products/Default.aspx?ProdID=PKZ3380 ), plus five Powerfilm solar cells ( http://www.jameco.com/webapp/wcs/stores/servlet/Product_10001_10001_228056_-1 ). The motor/prop are quite well matched to the cells, delivering about 2.5 watts in direct sunlight. Here are some test data:

All up weight is 38g, 18g in solar cells. It has a tiny 20mAh battery (1g) in parallel with the solar cells that acts to smooth out the voltage. Before I flew, I ran the motor using battery power until the voltage dropped too low to run the motor, then I plugged in the solar cells. That way, when I flew, I knew I was really running on solar power alone. I flew it for about 5 minutes or so, but it was a pain to fly because the wings were too torsionally floppy. I am planning the next revision that should address some of this design's shortcomings.

Related Work

There have been many successful solar airplanes. I found the following efforts particularly interesting because they are not military or NASA scale projects, and they are well documented.

Andre Noth's (ETH Zurich) 3.2m span Sky-Sailor, which flew 27 hrs in 2008

www.asl.ethz.ch/research/asl/skysailor

Carl Engel and Adam Woodworth's (MIT) 3.1m Aphelion, which has flown over 7 hrs

http://www.rcgroups.com/forums/showthread.php?t=572000&pp=15

Solar Cell Selection

In many applications, solar cell performance is measured by efficiency. In my experiments with the Powerfilm cells, I got about 70 watts per square meter. Using a fairly standard irradiance value of 1000 watts per square meter puts the cell efficiency at something like 7%. This is not particularly good compared to the state of the art (see e.g. http://en.wikipedia.org/wiki/File:PVeff%28rev100414%29.png).

The Aphelion uses an array of 18 A-300 Sunpower Solar Cells. The total encapsulated array weighs about 350 grams and can deliver 55 watts of power. These cells are reported to be upwards of 20% efficient. BTW, if you dig deep enough into the Aphelion thread referenced above, Carl reveals his secret supplier of these cells:

<quote>

Also, here is the contact info for our solar array supplier, now that he has given permission for it to be posted. One note that he wants to be made clear: he is very busy, and it's a small operation, so there can be a serveral month lead-time on ordering an array. With that, here is his info:

SunCat Solar is the name of the comany, and our contact in Alain Chuzel(pronounced "Allen"). His email address is
ahmchuzel<at>aol<dot>com. (replace the <at> with @and <dot> with . )

He can fabricate pretty much any size or confguration of array, given a good drawing (and enough money )

</quote>

The weight of solar cells accounted for 47% of the total weight of my airplane. Therefore, I would argue that for a solar powered airplanes, an important performance measure of performance is watts per gram. I measured about 0.27 watts per gram for the Powerfilm cells. This compares favorably to the Sunpower cells at 0.16 watts per gram.

Some other attractive things about the Powerfilm cells is that they are reasonably priced, and readily available in small quantities. I have not contacted Suncat Solar to determine the cost/availability of the Sunpower cells.

Scaling Arguments

I think it would be very cool to build a solar powered drone that could stay up all day, and possibly all night. My little airplane has essentially no payload capacity. Presumably, a larger airplane would be required to carry up a GPS, IMU, two way radio, etc.

As the airplane gets bigger, we have more area for solar cells, but the weight gets larger too. So how does the power budget work out with a bigger airplane? Here are some scaling arguments to help answer that question objectively.

Definitions:

Preq: required power

Pavail: available power

Rp: the ratio of available power (Pavail) to required power (Preq)

D: drag

V: flight speed

W: weight

g: acceleration due to gravity

rho: air density

S: wing area

CL: lift coefficient

Kcell: power output of solar cell per area

LOD: lift/drag (or glide ratio)

WOS: wing loading, weight per area

Relationships:

(1) Pavail = K * S (assuming the entire wing is covered with cells)

(2) Preq = D * V

(3) V = sqrt( 2 * W * g/ (rho*S*CL) )

(4) D = W * g / LOD

Combining (2), (3), and (4) gives

(4) Preq = W * g/ LOD * sqrt( 2*W *g/ (rho*S*CL) )

Combining (1) and (4) gives

(5) Rp = K * S ^ (3/2) * LOD * sqrt(rho * CL) / ( (W*g) ^ 3/2 * sqrt(2*g) )

Finally, simplifying (5) gives

(6) Rp = K * (1/WOS) ^ (3/2) * LOD * ( sqrt(rho*CL/ 2) * g^(-3/2) )

There are four terms in (6). The first term depends only on the solar cell. The second term depends only on the wing loading. The third term depends on the aerodynamic efficiency of the airplane, and the fourth term we can treat as a constant. Thus, starting from my little airplane, as we increase the size of the airplane, if we use the same solar cells, maintain the same wing loading, and maintain the same LOD, then we should still have enough power to fly.

The bad news is that as airplanes are scaled up, their weight usually goes up faster than their area, thus their wing loading goes up. Stated another way, bigger airplanes usually have higher wing loading than smaller airplanes given similar structural technologies.

The good news is that as we increase the size of the airplane, we can expect the LOD to improve. The LOD of my airplane is around 7. Doubling or even tripling this value should be possible ( see e.g. http://www.rc-soar.com/tech/perfanal.htm ). The wing loading of my airplane is 540g/m^2. If we double the LOD, we can increase the wing loading (WOS) to 850g/m^2 (i.e a factor of 1.58) and maintain the same power ratio.

Also, as we make the motor and propeller larger, we might expect an improvement in propulsive efficiency, meaning that we fly with a reduced power ratio.

Considering these factors, building a bigger solar airplane is probably harder than building a small solar airplane. Based on the successes referenced above, I must conclude that it is possible, but not easy (or cheap).