SOLAR TRACKING SYSTEMS

On Wed, 21 Oct 2015 23:13:18 -0700, Jeff Liebermann Gave us: snip

snip

For the same $100 one could accurately track the true sun position with this:

to orient the panel, while light is not suffisant, and the morning to orient the panel toward east.

Your simplest solution is an equatorial mounted panel with its axis of rotation pointed at the pole star. This reduces the problem to making a stepper motor drive to rotate the thing at a fixed rate of once per 24 hours between the crosses the horizon points and a slow adjustment to match your panel to the suns actual declination which varies seasonally by +/- 23.5 degrees and slowly enough you could alter it manually. (monthly or bimonthly would be good enough )

Making the panel S facing and adjustable in altitude only will help get you get better annual results with relatively simple mechanics. Assuming here that you actually do get to see the sun shine in winter.

You might find the following website with concrete examples and pages that let you put in your data and see graphs helpful.

Jean Meeus, Astronomical Algorithms is the standard reference for doing this sort of thing from scratch although it is aimed at astronomers and isn't an easy read if you don't have the basic mathematics.

One thing you should be aware of is that you rapidly get into the law of diminishing returns and your system has to be robust enough to survive the worst expected wind loading (which can be non trivial).

Misalignment of the panel to the sun by an angle theta loses power as cos(theta) so for small amounts of error isn't worth worrying about.

Even a 30 degree error still gets you 87% of theoretical maximum output according to geometry - slightly less in practice since shallow angle incidence on glass will reflect a bit more light away.

Depending on where you are a non focussing flux concentrator may be a better investment anyway since it is mechanically much simpler and increases output from diffuse light as well as direct sunlight.

Do the sums very carefully or you will be making yourself a maintained trap that will forever be breaking down.

McVeigh's book sun power (at UK latitudes in the late 70's) is a half decent introduction but dated now.

Not worth $60+ for the kindle edition but hardback for $3 is OK.

That is the problem. It will hunt randomly when that condition is not met and with the cheapness of RTC and CPUs now it is a trivial job to make a precision drive that points a telescope at the sun.

Making the mechanicals able to support the panel under all conditions of wind loading or detect problems as stow it are more challenging.

For an equatorial mount you would be right. Indeed it would be good enough to have it just drive at constant 15 degrees per hour and go home to the sunrise position when it goes dark. A second very much slower adjustment is needed to match the suns changing declination.

But if it is on an altaz mount (mechanically much easier to make robust) then you are in real trouble if at some time of the year the sun is near the zenith. The altitude control is fine but at some point the azimuth has to whip round through 180 degrees in a time that gets shorter the nearer the object is to the zenith.

Major telescopes always simulate the observing track in advance to make sure that the observation will not outstrip the drive capabilities. Almost all the worlds largest telescopes are altaz mounted and I suspect that most of the tracking solar arrays are too.

It doesn't strictly need hardware floating point at all. You can use fixed point scaled integer for the sort of trig accuracy needed here.

Setting 1.0 = 100 would be more than good enough for this application and you could fit everything in 16bit integers that way.

Losses scale with Cos(pointing error) so ~1% error is tolerable.

Historical note - the old half mile telescope tracked objects by using a precomputed paper tape from the mainframe that was loaded on a daily basis by observers in the days before true computer control.

For a given fixed location you could precompute the required data on a PC and use a fixed point lookup table in the device.

If it can't sense any difference in light level from left to right, then it doesn't matter. In fact movement should be disabled when ambient light is below a minimum.

I think the light-sensing method is more sensible, especially for a beginner's project. If you build the mechanism, you can upgrade it later with a timed controller, with either no mechanical change or just the addition of a micro-switch to sense a start position.

It's also immune to the effect of having the whole thing moved. (I assume this is a small thing, not intended to power a home.)

So this is for pointing solar panels? If you want to make it energy efficient, then you might want to think about the waste of running the motors all the time. so maybe... only activate system if there is enough sunlight. and second the loss of area you get by a little misalignment will be small for small angles (going as the cosine^2). So it might make sense to just update the angle every half hour or something.

George H.

On Thu, 22 Oct 2015 10:42:15 -0400, "Tom Del Rosso" Gave us:

In the sixties, one would say "Fan out and solid!"

On Thu, 22 Oct 2015 11:51:35 -0400, DecadentLinuxUserNumeroUno Gave us:

Oops... wrong thread.

I think I see a small problem. Can it move a 500 lb solar panel array?

For single axis tracking, a common 24 hr clock would do as well. Just attach the 500 lb array to the clock hands (somehow), install a frictionless air bearing system, hope the wind isn't blowing, and it should sorta follow the sun.

Personally, I prefer the oversized clock escapement mechanism. It has character.

That's for my home. Make your own at: There are some options that might help.

Ummm... no calculation required. The elevation of the panel should be same as the elevation of the sun at all times.

Fun project: I could print the aforementioned plot and attach it to a plywood board. If I mill slots into the board that follow the various curves, I could built a pantograph-like device that would track the sun by simply following the graph.

Do it with microwaves, and ignore clouds and fog.

It doesn't need to be a large board. You could use a plastic kitchen cutting board trimmed to the size of a sheet of paper.

It doesn't need to. Sunlight is _bright_, so you can just make it ignore signals that aren't within a factor of, say, 10 of zenith sunlight. Clouds, flashlights and headlamps become pretty much a non-problem.

Another approach is to use a coarse-fine system, with a MCU doing most of the work and the analogue thing handling deviations.

You need decent mechanics if you do it that way, though, because you have no easy way of deciding when it's correctly aimed. A bit of self-calibration code, where e.g. you move the panel a few degrees either way in both axes and find the point of maximum power output would do much the same thing as the analogue fine aiming system.

Yup. One good hailstorm or flying tree branch is all it takes. Just stowing the panel when it's cloudy or dark is a reasonable precaution--you don't usually get hail or windstorms in clear weather.

Or just point directly at the zenith during a slower azimuthal motion, and accept losing a bit of power.

Cheers

Phil Hobbs

I think it would make a neat project to put a mirror out there in the sun t o reflect light into a North window etc. (or maybe onto a stationary solar panel) I think the mirror then has to move at 1/2 the rate or 90 deg pe r 12 hours. Or if you made a 2 sided mirror it could just spin at 1/2 rev day. Very simple Mark

On Thu, 22 Oct 2015 08:56:00 -0700, Jeff Liebermann Gave us:

Ever heard of a pantograph?

You would use the telescope pointer to move a small arm that sets the position of a controller arm for a PLC which controls the larger mechanical load.

Call it "Big Grin".

Earth's axial tilt is 23 degreees, so you have a 46 degree span for that variable. If you want 5 degree accuracy, it'll need tilting 18 times, higher frequency in the spring and autumn than in the summer or winter. Maybe every week?

It's not all that much of a gain. The area goes as the cosine (I made a mistake before and said cosine^2 oops.) cosine of 23 is 0.92.. you're losing 8%. Given the rather poor economics of solar panels, solar trackers seem like an even bigger waste of money.

George H.

Solar panels are mostly about emotional rewards.

When I see those tracking mounts, I think about what will happen in a good windstorm.