I know this is ambitious for my very first project but I have a real world problem to solve and I want to do it elegantly.
I'm going to build a tracking solar collector. I have it built with 2 linear actuators and I have tried 2 commercial products and the both suck big time. So I decided that I'm going to do it myself and do it right.
Here are my goals.
Obviously - to track the sun.
I want to add sensors to detect the angle the panels are tilted at and to detect shaking. The idea being that I want the collector to go horizontal at night and in case of wind. But instead of detecting wind I want to detect shaking and assume it's wind.
I'm using a YUN so I want WIFI access to control and monitor it through web pages.
NTP Time access will allow me know know if it's really sundown and not just dark clouds
All powered by the DC from the panels so at night it goes dead
I need to sense current draw by the actuators so that I can tell when it hits the limit switch
Some additional thoughts - a compass sensor might be nice so when it wakes up it knows which way it's pointing without having to see the sun. A display might not be necessary because I could control it from my cell phone.
So - in the spirit of doing it right and releasing my code to the world - but keeping the price reasonable - I'm looking for suggestions and ideas.
Some questions. The 2 failed commercial systems used many LEDs for light sensors. Do they know something I don't or are they just being cheap? I'd like some good sensors and wondering what people would recommend as something good to track the sun with?
Is there an I2C light module where I can connect 4 of them? Photoresistors?
weedpharma:
You say you want to use the power from the panel to run the system so at night it will be dead.
At night you will be facing the wrong way to the morning sun so how are you going to turn the panel towards the sun?
Weedpharma
In the evening when I park it for the night I plan to turn it east before the light goes dead completely. Or - when it wakes in the morning and I get enough sun I'll turn it around. I prefer to pre-position it at night for the next day. but it needs to be capable of moving itself from where ever it is to where it needs to be.
Qdeathstar:
it would seema to me you could use math to track the sun.... we are talking about predictable orbits....
That is just a "thot" though.
The mechanics are not linear so math tracking would be difficult. Unless I could use a compass and tilt sensor and knowing the time of day and day of year.
hmmmmmmmm
Eliminating the sun would be great. Would need a light level sensor just to make sure I had enough solar power to park it at night.
BTW - my system rotates and tilts. Not one that tilts on 2 axis. There are some differences.
LEDs make fine light detectors, but they are very color-sensitive (red ones detect only red light, etc.).
With a good quality RTC module (e.g. DS3231) you will always know the time to within a few seconds. From that and your position, an Arduino can calculate exactly where the sun is in the sky and position the panel accordingly.
mperkel:
The mechanics are not linear so math tracking would be difficult. Unless I could use a compass and tilt sensor and knowing the time of day and day of year.
hmmmmmmmm
We have known how to calculate the position of celestial bodies for centuries as long as we had a good time base to work from, you mentioned using an NTP so you definitely have a good time base, plus you mentioned a compass. Instead of compass and tilt sensor I would use photoelectric detectors and flags or holes as interrupters to zero the azimuth and tilt every morning and rely on the servo or stepper accuracy through the rest of the day and calculate the position of the sun. Think of this as the old fashioned way to due solar work, I think the ancient greeks did the calculations originally and they have only gotten better since. Don't worry I think the church stop burning people a long time ago who could "predict" such things.
I work with sensors a lot but first I ask do I have to sense that phenomenon or can I calculate it, look it up, or real time google it. When I have to sense something I do and have quite a bit but only as a last resort.
Why are you going to do this?
From an economic perspective Solar Trackers dont work.
The improvement gained by tracking the sun , is always less than the cost of adding extra Solar Panels
facing in a fixed direction which compensate for the power lost from not tracking.
Solar Panels are now so cheap, that they compensate for the mechanical hardware needed to build and control the tracker.
Why are you going to do this?
From an economic perspective Solar Trackers dont work.
The improvement gained by tracking the sun , is always less than the cost of adding extra Solar Panels
facing in a fixed direction which compensate for the power lost from not tracking.
Solar Panels are now so cheap, that they compensate for the mechanical hardware needed to build and control the tracker.
I hope this link works.
Zoom in over the fence and look at the PV array. Tracking PVs.
I'll see if I can get some better pictures, I go to do some work there a couple of times a month.
Thank goodness its only a couple of times, I my waist band is already to big.
mperkel:
The mechanics are not linear so math tracking would be difficult. Unless I could use a compass and tilt sensor and knowing the time of day and day of year.
ummm... the Sun and Earth are in a perpetual 'dance' the exact time is known and you can set your watch by the sunrise. it is one of the most fundamental known and reliable things in nature,
that is why it is so easy to track the Sun.
you can set your axis and 'follow' the Sun 24 hrs a day. exactly.
mauried:
Why are you going to do this?
From an economic perspective Solar Trackers dont work.
The improvement gained by tracking the sun , is always less than the cost of adding extra Solar Panels
facing in a fixed direction which compensate for the power lost from not tracking.
Solar Panels are now so cheap, that they compensate for the mechanical hardware needed to build and control the tracker.
economics are in favor of tracking.
motors and gears/belts/ etc are cheap.
a tiny solar panel is much cheaper, but once you scal up the project, the costs make tracking very favorable.
a 20-30% gain in perforance can be expected.
the balance is that of space. most people do not rotate their house and put panels on the roof. economics of rotating the house would favor the reduction in solar panel performance.
there is always a balance. if you have to buy all new and pay to have the gantry/array made, then costs might favor just bying more panels.
but i think there are lots of reasons why following the sun is desired.
