Solar Tracking system

[timex83]

Hi all. I intend to build a Solar tracking device. For this project i intend to use light dependent resistor and ADC. The LDR will be connected to d ADC so that I can compare the intensity of sunlight. I will positionally place 4 LDRs on the sides of the platform that will track light position for highest intensity. I hav knowlege of 8051 micrcontroller.

I need advice on what to do next to implement this project. Thanks.

 

[dknguyen]

You are probably going to be using RC servo motors to move the platform which will require your 8051 to be able to produce timed pulses or PWM. Maybe a make each LDR be part of it's own resistive divider and use an op-amp buffer to read the voltage into an ADC to figure out the resistance of the LDR.

 

[crutschow]

How were you intending to determine the direction to move the tracker?

One way is to mount four vertical shades at right angles (a small square box) and have the sensors at the outside bottom of the box at the center of the sides. That way when the box is pointed at the sun all sensors will be equally illuminated. If pointed away from the sun one or more of the sensors will start going into shade.

Of course in operation you would have the microcontroller adjust the tracker until the output from all the sensors is equal.

The height of the box will be a factor in determining the pointing accuracy of your tracker.

For increased pointing resolution you may also want to design it such that when pointed directly at the sun all 4 sensors are partially in shade.

 

[Boncuk]

Hi timex82,

you don't need to mount the sensors far apart from each other. The only source of absolutely parallel light is the sun, just because it is far bigger than our globe.

Sticking the four sensors into PVC tubes of variable length (telescope) you have the option of super accurate tracking and fine adjustments.

The drive motor power doesn't have to be much. The sun apparently moves at a speed of 0.25degrees per minute, which allows to use a high ratio reduction gear.

Depending on your location don't forget to consider the maximum elevation travel angle of the sun. (At my location the sun travels beyond the 12 o'clock position and seems to shine from the north during the hot season at high noon. That means tracking over the top.)

It is also a good idea to include a homing circuit for the tracker, to have the solar panels return (including the sensors) back to the sunrise position after dark. If left in the sunset position the sensors "don't know" where to look at the next day.

Boncuk

 

[Pommie]

Have you seen the RedRock Site. It's a slow loading page but has a lot of useful information about solar trackers.

Mike.

 

[dknguyen]

What's more accurate? Dividing shades or tubes of variable length? Is there supposed to be an arrangement for the tubes (like tallest one in the middle surrounded by shorter ones, or what? I can imagine halfway how it would work, but I can't seem to make it to the other half)

EDIT:
Do the tubes of variable length actually move to face the sun (as I think most of us are imaginging?) Or do they just sit there completely stationary facing up? If that is the case, then I might see how they would work.

 

[nickelflippr]

Or how about a shaded pole effect, with an adjustable piece of pipe, all-thread, or long screw in the middle of your closely spaced ldr's.

Better yet, a Taos light to voltage sensor instead (i.e. op-amp inside sensor).

 

[crutschow]

What's more accurate? Dividing shades or tubes of variable length?

I would think a square shade might be more easily aligned but that's just my opinion.

Is there supposed to be an arrangement for the tubes (like tallest one in the middle surrounded by shorter ones, or what? I can imagine halfway how it would work, but I can't seem to make it to the other half)

EDIT:
Do the tubes of variable length actually move to face the sun (as I think most of us are imaginging?) Or do they just sit there completely stationary facing up? If that is the case, then I might see how they would work.

I would think you would use four equal length tubes, arranged in a square, pointing at the sun, but they would have to be slightly tilted out, so that as the array became off center one tube would see more light and the other directly across would see less. If they all point in the same direction there would be little difference in signal between sensors as it moved.

 

[timex83]

thanks for all your submission to the topic. The motor aspect wouldn't be a problem. The problem is how to effectively track light source, not necessarily sun light. If LDRs are mounted on the middle of the lengths and breadths of a rectangle, how can one effectively track where light's intensity is highest using analogue- digital converter (ADC) and a micrcontroller?

 

[gramo]

digital converter (ADC) and a micrcontroller?

Almost all PIC's have a bunch of ADC channels, and controlling a stepper motor would be easy if coupled with a ULN2003 or a SN754410 for a standard DC motor

 

[Boncuk]

What's more accurate? Dividing shades or tubes of variable length? Is there supposed to be an arrangement for the tubes (like tallest one in the middle surrounded by shorter ones, or what? I can imagine halfway how it would work, but I can't seem to make it to the other half)

EDIT:
Do the tubes of variable length actually move to face the sun (as I think most of us are imaginging?) Or do they just sit there completely stationary facing up? If that is the case, then I might see how they would work.

Hi dknguyen,

the tubes have to be fixed to the solar panel carrier. The longer the tubes are, the more accurate will be the tracking. I use four 10mm clear red LEDs in pairs of two connected antiparallel. In bright sunlight the voltage output of a single LED is approximately 1.3V. Antiparallel at the same view angle the total voltage output then must be zero V. One pair is mounted side by side for the azimuth and the other pair vertically for the elevation.

Imagine the tubes in a fixed position. There will be only one point when both LEDs receive the same amount of light - exactly pointing towards the sun. As the sun moves (which is not true, but the earth rotates underneath the sun) the view angle of the LEDs changes and will cause either a positive or negative voltage output which can be used to drive the tracking motor and gear (not directly of course).

Have a look at the drawing. It's self explanatory.

Kind regards

Hans

P.S. The sunrays are painted blue. Eagle doesn't have yellow.

 

[mneary]

If you want to track the sun you don't need sensors. The movement of the sun is predictable, and the formulas have been known for centuries.

 

[dknguyen]

If you want to track the sun you don't need sensors. The movement of the sun is predictable, and the formulas have been known for centuries.

You have to know exactly where you are don't you?

 

[crutschow]

Originally Posted by mneary
If you want to track the sun you don't need sensors. The movement of the sun is predictable, and the formulas have been known for centuries.

You have to know exactly where you are don't you?

Yes. And you have to adjust for the day of the year.

It would also require accurate position sensors on the tracker to know where it was pointed. All in all much more complex than simply tracking the light source with sensors.

 

[crutschow]

Have a look at the drawing. It's self explanatory.

Now that I see your drawing I understand how it works. The cut angle of the tube top is critical to getting a proper differential signal.

Also like your idea of two back-to-back diodes. Gives an automatic null signal when pointed at the source.

 

[Boncuk]

Now that I see your drawing I understand how it works. The cut angle of the tube top is critical to getting a proper differential signal.

Hi crutschow,

the cut angle isn't as critical as you imagine. It just takes care of the tilt which would have to applied in a more elegant way: The LEDs can be mounted on a straight surface. The steeper the "slope" the more inaccurate will be the angular measurement. A five degree slope gives very accurate results since the output voltage starts to move in either direction at half of that angle.

Also like your idea of two back-to-back diodes. Gives an automatic null signal when pointed at the source.

This applies only if the LEDs have the same properties. Small differences can be compensated easily by nulling the output voltage using a 4.7KOhm trim potentiometer the same way as used for offset trim.

Regards

Hans

 

[Boncuk]

If you want to track the sun you don't need sensors. The movement of the sun is predictable, and the formulas have been known for centuries.

Do you assume this as a better solution?

Even living on the equator the elevation angle is variable +/-23.5 degrees during a year.

Even feeding half of the Air Almanach into the controller won't do any good.
Not knowing where the panel is directed will certainly cause another problem.

Maybe there are formulas being known for centuries too how to determine position and movement of the tracking platform.

Boncuk

 

[mneary]

Each solution has its advantages. Using the sun itself to track is good because it doesn't need to be calibrated, and it's portable. It accounts for atmospheric refraction, and earth's orbit eccentricity. I would worry about the challenges presented by rain, clouds, birds, and dirt. These are all probably manageable since the solar cells themselves need to be kept clean anyway.

I only suggested the calculation method because if you actually do know your location (most of us do) and orientation (compass), it doesn't require optical sensors. You don't need to feed huge tables into the controller; it's trigonometry.

 

[dknguyen]

Each solution has its advantages. Using the sun itself to track is good because it doesn't need to be calibrated, and it's portable. It accounts for atmospheric refraction, and earth's orbit eccentricity. I would worry about the challenges presented by rain, clouds, birds, and dirt. These are all probably manageable since the solar cells themselves need to be kept clean anyway.

I only suggested the calculation method because if you actually do know your location (most of us do) and orientation (compass), it doesn't require optical sensors. You don't need to feed huge tables into the controller; it's trigonometry.

I thought it was quite a bit more than just trigonometry (and why do you need tables if using optical sensors? THat's just trig too isn't it?)

 

[mneary]

Sorry I wasn't clear. An earlier comment suggested that a huge table would be needed for the calculation method. I looked it up on Wikipedia, and although it's complicated it appears to be trig. There are some javascript calculators that just need lat, long, date, and time of day.