Towards a better tinyhouse

Inventing to freedom?

Much cheaper photovoltaics part 2

with one comment

I chose this design because it has a more manageable size, could turn upside down to keep the snow off (though I don’t think you really need to given that the upper reflector sort of shelters it and you could make it a bit closer to the ground if you are willing to dispense with that feature), and should be fairly easy to build, including the curved corner reflectors which are just mylar bent into a cone shape.
Sorry the pictures aren’t very good but I didn’t want to waste time with sketchup:

system has a concentration ratio of 6.65 not including reflectivity, or 6.09 at 0.9 reflectivity (the reflector adds 5.65, so add 1 for the fact that sunlight also hits the panel directly) with the reflectors at an angle of 67.5 degrees( relative to the plane fo the solar panel, which goes in the flat spot in the center) and the reflectors reach 1.68 meters forwards ( to get a spot 1.2 meters wide) for the side reflectors and 2.65 for the top and bottom reflectors. The spot is about 10 cm bigger than the size of the panel, so it is tolerant of some misalignment.

There are 4 flat collectors, and the reflectors at the corners are actually sections of cones. The collection ratio would be reduced when needed with a shadecloth wrapped in a roll above, one above and one below the collector, with numerous holes punched for higher transmission sections, or with 2 layers of shadecloth (the stuff sold for horticlture blocks 60% by default), one above the panel one below, and rolled out when needed, you can have several setting, depending on how much the cloth is covering and it’s transmissibility.

Okay I don’t expect you to grok this right off the bat because it is not a very good explanation, though accurate, and most people can skip it, but for anyone who actually might build this and wants confirmation this will work and wants to see why:
To see why this works, viewing the reflector system from the ground looking upwards, imagine a triangle formed with one angle 112.5 degrees, in the plane which is perpendicular to both the solar panel and the left side reflector (and therefore the right too) with the left side of the triangle along the left reflector, and the 112.5 degree vertex is right at the edge of the reflector where it meets the bright spot area, and one side against the solar panel/the surface of the bright spot area. If a row of sunlight rays hits the left side of the triangle, after reflection they will end up on the right side. So you can see you could rotate the triangle around the rightmost vertex while keeping it in a plane perpendicular to the solar panel, the left side would sweep out a cone, and the light reflected from the cone would hit the area swept out by the side of the triangle which is along the solar panel. You can similarly imagine pivoting the triangle around that vertex with the oblique inner angle, which was against the edge of the reflector, and the reflected light would cover a semicircle shape if the reflector was a suitable cone section. Then remove any unnecessary parts of the cone so the spot of light is the right shape..

The actuator would be a winch, and since you only get 1.5 sq meters or so of panel you would probably need 3 of them to get in the range of 5 kWh per day during the worst month, but they can be controlled with the same winch and control electronics, obviously the winch has plenty of torque. Just wrap a rope around the wind a few times, attach the middle of the rope to the winch drum, and then keep wrapping a few times, and you have both ends of the rope, attach them to the collector for motion, apply tension to prevent motion due to wind. The tension if applied to bothe ends of the rope eaqually balances out and exerts no net torque on the winch drum.

The axis of rotation is sort or around the center of pressure, so wind loading shouldn’t exert much rotational force. To adjust it for the seasons, there is an adjustment that you can see in the video.

If you take a line which is perpendicular to the plane of the solar panel, it draws a cone around the axis of rotation, which when the axis of rotation is parallel to the axis of the earth’s rotation which it is, is perfect 1 axis tracking.

There might be light sensors on the panel so the electronics module knows it has things in focus, with 3 sensors at the edge of the panel, they would all be the same when it was sufficiently in focus. There should also probably be a temp sensor (e.g. thermistor) attached to the back of at least one panel in the array ( for 5 kWh per day minimum there would be 4 collectors like this but they can use the same actuators etc.) (they should all be about the same temp after al on the back of the panel. The electronics module would be an AVR and an h bridge.

Okay, I think it would cost, I can get an 8 foot 2×4 for $2 at home depot I think so I dunno, maybe 8 2x4s $16 and $12 for some door hinges and some screws, then $70 for the atv winch (from google products search “atv winch” list by price) and if you get the electronic module custom made at a sparkfun like place maybe $45 with the light sensors and $3 for a tupperware container to put it all in for water resistance, and $30 for a gearmotor to control the shadecloth with one controlling all the cloths on all collectors (use fishing line?) plus $20 for the mylar.

The performance ratio is 2.43 so with unassisted panels it took 1.73 kWh of solar panel at $1.69 per watt or $2925, so this would be $1203 plus $204+$145 for a total of 4 collectors = 1552 , for a savings of $1373 and a cost per watt of $0.90 per watt. There must be a better reflector system though. Also amorphous panels would help, and I haven’t looked into a cooling fan yet.

Rulgert raises the point about ROI, which I was sort of ignoring, but it looks like, actually, if we reduce the max. operating temperature to 85 degrees, that actually makes almost no difference, giving an improvement ratio of 2.4 instead of 2.43. Reducing it to 65 degrees, which is typical for a rooftop gives 2.24.

Clearly it would be a good idea to at least do the financial calculations, and correctly, which includes the interest on the money you saved, rare but catastrophic possibilities like the possibility that the system will be vandalized, or you will have to sell it at a loss etc due to a change of plans all need to be taken into account too, and these things improve a lot with a lower capital cost.

But there doesn’t seem to be any good information that I can find on how lond panels last at higher temps. So we have a choice: fall for fear, uncertainty and doubt and just do things the easy they are always done (which corporations just love by the way), or we can go for it. Your choice, live in fear or live in hope.

Plus if there are problems with financing and so on, a lower capital cost is more than just a good thing from an ROI standpoint, it could be the difference between able to build it or not.

And finally, even if the panels failed in 7 years instead of 25 or 60 (I’m not saying they would, as I said it is unknown, and they might last for 20 instead), does that even matter so much? In 7 years the price of solar panels and other renewable energy sources is going to be a tiny fraction of what it is now. There are already companies like first solar making panels at 90 cents a watt (they don’t sell them to individuals though darn). People make the decision to use propane instead of solar heat all the time for instance, because what they care most about is getting things up and running and can’t be bothered with longer term results. At least this is a more environmentally friendly way to do so, if it were one.

By the way here are some totally random notes that I don’t want to spend the time to clean up but you migh thave use for to invent your own even better ones. I don’t want to edit themm so they are a total mess but there might be some ideas and info for you to extract:

conc ratio improvement ratio solar panels req for 0.9 kW array total aperture price of panels
3 1.85 4.38 13.14 884.43
4 2.14 3.79 15.14 764.58
6 2.42 3.35 20.08 676.12
8 2.57 3.15 25.21 636.65
12 2.7 3 36 606
25 3.04 2.66 66.61 538.22
50 3.22 2.52 125.78 508.14

winches including atv winches which are available for as low as $70
valve actuator $50 probably no good for main tracking mechanism but maybe good for varying the conc ratio
heavy duty rc servos tend to be expensive though IIRC
drapery actuator No idea what the cost
gearmotor of sufficien size these are a little small but maybe something like them from a similar vendor

maybe even stepper motors

Satellite dish actuators are awfully expensive that I have seen, $150 or something each.

I like the idea of using cables to couple the actuator to the main apparatus. Should be easy to do in a do it yourself scenario, you can vary the leverage etc quite easily and so on too. Plus you can connect multiple apparatuses together so they move in synchrony using only one actuator. Also you can use a tupperware container or bucket to put the actuator in if it is not weatherproof by ittself, then run the cables inside. It is also perfect for the winch one. You can have a “balanced” system too to give resistance to wind – suppose you have a winch and there is the connection point on the winch drum somewhere. You could attach one end of the cable/rope to that point and have a winch that can wind in the cable, but suppose you instead attach the middle of the rope to that point. One half of the rope gets wound up clockwise on the left half of the drum, maybe only a few turns but could be many, and the other half get wound up counterclockwise. So now one end of the rope is unravelling while the other is wrapping up, see. Any static force exerted on the cable does not result in a net torque onthe winch which is good.

interesting site with a bunch of info, good starting point for research

coaxial, keeps the snow off the panel too plus shading with “shade cloth” available for greenhouses to reduce conc ratio when needed (shade cloth could be fine for other colllectors)

large Plate reflector at one end of a large parabolic trough collector that is twice or even longer than the solar panels. The panel is located closer to the trough than the focal point so the light is sufficiently diffuse. Bring it closer to reduce conc. ratio. Could be attached to ropes to do so using the cable actuator? Suppose you are looking through a hole in the plate, whan you woulde see when you looked at the trough at an angle is a reflection of the sun in the plate reflector if the rigamarole is oriented right.. You’d see 2 suns right, would sort of approach 2x the concentration of the trough alone. The trough could be vertical, and the flat plate on the ground, or it could be in anothe orientation. In this orientation it might be better to have 1 axis tracking but maybe you could do without.

The plate one, but the areas in between the panels are used by adding a diffuse reflector e.i. like a sheet of paper, conc. ratio is again controlled with shade cloth or by making some of the reelfector panels move so the reflected light form it no longer hits the panels. single axis tracking

like above but at the outer edges of the panels another, slightly curved reflector could be added for more light, sunlight is directed towards solar panel and spreads out due to curve or reflector.

2 plate reflectors could be added to either side of a panel, an then extended a ways beyond the ends of the panel. Then 2 large plate collectors added some ways beyond the ends to direct sunlight towards them, providing maybe 9x, 3 times from the 2 panels (what hits the panel directly counts as 1, add one late is 2, 2 plates is 3) the 3 more from each end plate, single axis tracking?

move the panel rather than moving the reflector might be easier for very harge collectors, could use curtain tracks and drapery actuators but the tracks might rust outside. With an overlength parabolic trough collector just move the panel closer to the mirror when already past the focus for less light, should be fairly uniform I think. Could be used for sun tracking. Problem is that then you loose a lot of sunlight in the morning and evening potentially.

a vertical parabolic trough reflector with a mirror on the ground, which produces a mirror image so is like 2 suns sometimes, vertical axis tracking could be very large for high conc ratios, move panel towards/away from to vary ratio, also add overhang for roof and a bit more reflector

one issue is the need to have the focal point relatively uniform because of the way solar panels work. They are a large number of cells in series or parallel, and if they are in series if a disproportional amount of light hits one cell the extra power is not used optimally, because the maximum power point of voltage-current is not reached. It can’t be because in series the current through all the cells has to be the same, but the ideal voltage-current point changes when the light level changes.
If they are wired in parallel similar idea; the voltages are all the same.

Also when the panel is operating close to it’s maximum acceptable operating temp then you’d have to be sure that excessive light does not get directed to a particular cell or it could be damaged I guess, although I don’t think that 111 is an actual damage threshold, it might be substantially higher. It might only be for purposes of guaranteeing the specified performance parameters/ lifetime of the panel.

After all they are basically giant silicon diodes, and a typical rectifier diode can handle temperatures higher than that I think.

Ideally you would figure out the ideal curves for the shapes of the reflectors, I guess. A set of flat panel reflectors, though, maybe recycled bathroom mirrors or something, actually might be the way to go, as the reflected spot would be automatically uniform.

Another way could be to use a shade cloth material of some sort, with it’s transparency different in different places and just right in the right places so that it subtracts light from the areas that would otherwise be too bright. Maybe use a low density shade cloth and then put multiple layers to dim or punch small holes to brighten, in the right places.

If you have a collection of flat plate reflectors, each directing sunlight to the panel, you could remove several of the plates to reduce the concentration ratio.

problem with the parabolics is that it seems like the center of the spot of light might be brighter I don’t know. Would it be when the spot was outside of the focal plane?

electronics modules:
the red rock ones

programmable logic controllers

can make your own, also if you don’t have the soldering etc. equipment you can get circuit boards custom made relatively easily I think, one company is “sparkfun” but there are probably better ones. Take an AVR and an H-bridge to control the actuator (servos and the valve controller have their own power electronics though) and a suitable sense resistor and voltage divider to measure voltage coming from the panel, measure the amount of power coming from the panel, the microcontroller knows the time of day approximately, then does fine adjustments based on the amount of power coming from the collector, and maybe a ten cent CdS photosensor at the panel surface just for redundancy to protect against excessive irradiance. Put the whole thing in a tupperware container for water resistance. You might need an encoder of some sort to provide feedback about the tracking system’s position.

Another problem is the wind loading during storms. Obviously some of these are pretty big. Still if you can just put it on the ground it’s not any where near as as on a rooftop, where the roof would not be strong enough to stand storm winds.

Also remember amorphous panels would be better, if windspeed is factored i nthat should improve things. also the possibility of a fan is plenty real, the fact that snow might be kept off of the panel could increase real-world gains with a reflector system tha thad taht feature, you save money on batteries too ( I’m pretty sure, have to check how much)
wind speed only went down ot or below 1 meter per second for 24 hours during the entire year so wind is definitely a source of improvement factors that could reduce performance reduce savings that have not been taken into account are:

if life of panel is reduced
potential for damage durig storms but with suitable design should be fine
if system is capable of highg conc ratios potential for malfunction and overheating damaginng panels

Lastly, there must be a substantial amount of money to be made here. I mean think about it for a second – suppose for that hypothetical 5kWh/day minimum system you take a generous $250 in parts to build the collector. How much money did we save again? $2068? And that’s quite a small system, remember, the average American house uses 32kWh per day.

Figure out a good mechanism and you could sell kits, and charge a pretty decent profit on your time, and still leave the customer with huge savings vs. an unassisted array.
I checked out the amorphous panels, looks like this would give an improvement ratio of 3.03! so pretty good

edit: it occurred to me that if you were concerned about the impact of high temperatures on the panel’s lifetime, you could just turn down the concentration ratio during the summer. That would greatly reduce the time it spends at high temperatue, without sacrificing the net cost of the system, because it still produces just as much during the winter, which is the limiting factor when determining system cost.

edit: I found d a paper about a system that in many regards is similar to this. It has a fixed concentration ration, though of 3. The system has apparently beed running fine for 5 years (notice at the end, he seems to be writing this 5 years after first building the system) :


Written by gregor

January 30, 2011 at 20:02

Posted in Uncategorized

One Response

Subscribe to comments with RSS.

  1. you whip stuff like this out in a mater of days….it cracks me up, your smart enough to design actual products for manufacture if put to the task….but you like 2×4’s, and door hinges…oh and don’t forget the tupperware….

    anyway….like ya said ” there doesn’t seem to be any good information that I can find on how long panels last at higher temps.”

    i couldn’t find these listings either…but with all electronics, the max temp is actually based on a structural strain figure ( as is the min temp ) so i would be inclined to believe your reflectors in conjunction with an automated cooling system would pay off in a cooler climate, being that it would only draw power when necessary.


    January 30, 2011 at 23:30

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: