So Is there anyone who know how to work a arduino?
Project
Solar tracking rig for maximum sunlight using a arduino.
1. Fully automated sun tracking.
2. I must have 2 axil control.
3. Abel to hold 10 panels.
4. High winds must be taken into account and turn accordingly.
5. Low budget.
So there is my project (Personal project for own home. No power grid connection. Solar only) that im busy with. Point 2 and 3 I can do. Im studying mechanical engineering part time right now. So the rig and strength is not a problem for me to build.
My big problem is the electronics and the arduino. I tend to do 10plus hour days and 18 hour days is not uncommon for me. I dont have the time to turn the rig by hand. The second motivation to the automation is that I get about 150W maybe 200W if its a good day out of a 900W array. And its not the panels. I have tested them. I still need to add another 600W to the current array. To run everything that I need to run at this point.
Later on, I need to add 3 more arrays to give me a total of 6 kW.
Questions.
1. Im looking at the arduino nano r3. Daz anyone have any experience with the arduino? And will the nano be enough.
2. My limited experience say that LDR sensors is the path to go, Mounted in a star to give direction. Any thoughts?
3. Is the LDR sensors sensitive to the angle the light hits them?
4. Any thoughts, comments, advice or criticism anyone can give.?
Thanks.
Other thing I forgot to add is stepper motor requirement and sufficiently sized drivers.
Listen bro, I can't help you with this but I just want to share something I've learned through the years.
You're going for the holy grail right now. Don't. First build a platform that can be turned and skewed with a simple panel. Fine tune the shit out of that, it needs to be absolutely flawless. You'll need an Arduino, some motors and a little pit of engineering.
Once that is all done and is working like a champ, once that is done.. don't you fucking dare build a sensor that tracks the sun.. the sun follows the exact same path every damn day. So you're going to program a cronjob that turns the panels according to the time of the day.
In case of storms it's probably best to let it go back to it's default flat state.. which could simply be done by measuring the windspeed from a high pole nearby the panels.
Anyway, good luck.
You'd have a much easier life by programming it to move according to a timer.
Sensors introduce complication and a damaged sensor can screw over your entire array.
>bird shits over LDR
Another complication you'd encounter is cloud chasing. Your array would get disturbed by clouds since they'd change the values of your LDRs leading to basically your array moving constantly during the day, burning away the little electricity you have.
Now to solve such an issue you'd have to introduce mechanics for dealing with situations that arn't simply a clear sky and the sun. And a way of dealing with those mechanics I found was to simply limit the sensor array to move during specific times, but simply find an ideal position during those short frames. But hey, if you're already moving according to a timer, why not just ditch the liability of the sensors.
The most energy efficient way to do this is have pre-set points for the array (bonus points for scaling according to season) and have the array move in steps to those points.
No constant slow crawling motor.
Step A =single movement=> Step B => Step C etc.
>>1005066
There's also the issue of actually driving the motors.
Ideally you'd want to have all the logic required for moving the motor in an circuit integrated with the motor.
Driving relatively complicated motors while dealing with sensor readings or ideally a clock would quickly become a pain in the ass to do with an Arduino.
Outputs should be at most switching on a Relay and letting the motor circuit do all the thinking.
>>1005051
Oh shit, do we have a bingo? I think we have a bingo! Where's the sheet?
>OP here
All valeted points guys. And yes I am aiming high with thus project. My problems daz not stop at not enough power from the array. Just to name one, is all the baboons. If thy cant steel the damn panels. Thy will throw it with rocks.
Never the less. Yes programing it to step with a clock will be the most simplified way of doing it. And on the array at the farm house, that would be the fastest and best way to do it. I agree. But not with a mobile rig.
Now sensors is defiantly not the best way to go. But will it not be the eases way to go when you nether have the time nor the equipment to position the rig correctly. so that it will move with the earths axial?
Time is also an issue. I ether will have to build it at night after work, or on the one off day I get ever now and then. So I dont have the time to tweak it. That is the other reason after considering the sensor.
>>1005074
bingo?? where??
>>1005051
The mechanics and structural parts of this are going to be most of the work. Once you get that figured out, the controls will be relatively simple.
Stepper motors might be an overkill. On/off control of the elevation and azimuth axis drive motors could be sufficient to 'bump' the array a few degrees at a time. You won't see much of an improvement in panel output with a more precise control.
Using sun sensors (LDRs) will work, but could be confused by overcast, passing clouds, etc. You might be better off just working out the sun's path and having the Arduino look up the elevation and azimuth of where the sun is supposed to be and just point the array there. Some position encoders on the two axis (potentiometers?) provide feedback to the Arduino pointing routine.
>>1005062
So yeah this guy said not to build solar tracking devices and just run a timer instead. Well I read that after I found this nifty little photo diode solar tracker.
http://www.electroschematics.com/10182/arduino-solar-tracker-servo/
>>1005112
>only one square
>>1005107
yes, with a mobile rig too. it will take you about 4 hours of google time (or geography textbook time) to come up with an algorithm that returns the sun't position on the sky when given coordinates, time, date and orientation
>>1005433
No, four.
>What x do I buy
>Arduino
>No money
And maybe a bit of a stretch
>Free energy!
In the form of solar.
>>1005051
You just need to make a light tube, and you dont need an arduino, just 4 photosensors and 4 relays/scrs/transitors. Light tube is just a 4ish inch secition of pvc pipe, with an opaque surface on one side facing the sun. Poke a small hole in the middle of the opaque side of the tube. Then arrange your photosensors like a +, and attach them to the other side of the light tube. Now, the outputs of the photosensors go through your relay/scr/transitor and the out put from that powers the motor to move it in the appropriate direction. No arduino needed. Its like a simple line following robot but with 4 sensors instead of 2
>>1005074
>>1005112
>>1005433
>>1005440
>>1005508
>>1005509
Butthurt will be butthurt. I feel sorry for you.
To all the rest of you. Thank you for the thoughts. Its some things to consider and advise to take into consideration. But for now its back to the farm with no power connection and no internet. Have fun ever one.
>>1005511
You need to handle passing clouds or overcast days. Plus some logic to reset the array back to the East every evening.
You can do this all with discrete logic but it gets complicated.
>>1005051
Another person's thoughts:
1. any Arduino could do this
2. it would be a lot simpler to just move the panel according to a timer. You aren't aiming a telescope, it doesn't have to be perfect.
3. it would be easier to build this using one automatic axis (that tracks east-to-west) and just have the second axis (the north-south tilt) manually adjustable and adjust it by hand every two or three months. The sun's path varies by about 44° over the course of a year. If you manually adjusted the north-south angle four times a year, it would never be more than about 5° off at the most.
4. To move the east-west axis, you would want a linear actuator with a stepper motor (so that you can move it an exact number of steps every hour or half-hour) but you would want to make an absolute rotary encoder for. The reason is that stepper motors lose their tracking if they aren't powered all the time, and you don't want to have to power that motor all the time. This rotary encoder could be made a couple different ways; the easiest is to just use a potentiometer and move the panels every hour. Potentiometers are cheap and easy to work with, and they provide absolute positional encoding.
5. Solar panels are expensive and easy to physically destroy. If you can't protect this project from vandalism, then maybe solar cells aren't such a good idea.
6. Solar might not be the best method to use anyway; most places on Earth that aren't a desert can produce more power from wind than solar. Where are you located?
>>1005062
>the sun follows the exact same path every damn day
no it doesn't
>>1006014
Very true, maximum height in the sky will vary by time of year. It's not terribly hard to work out the elevation though, and then you're back in business.
If you're concerned with accuracy and being mobile, don't be. They're panels with a cosine response (more or less) not a pyrheliometer. 5 degrees won't matter that much, just use a compass to set up at a location.
>>1005062
What are Seasonal Variations....
>>1005051
Make array follow some simple arc, you will blow a lot of energy just fiddling the array around for max output.This is not a radar, it just has to face roughly the right direction. Photocell sensor/timer combo to ensure array is roughly in correct direction and timer delay to make sure it gets sunset. Then array moves back to pickup sunrise. Minimize movements.
Possibly just program in a few dozen (26) paths and every Sunday lets say change to new path and timing for the following week.
Question: You noted you get 150-200 watt out of ( supposedly 900 watt) array on typical day day. When you have time, aim to best angle and measure it's efficiency. If at best angle it does not appreciably improve performance I would suggest you put your effort into more arrays rather than optimizing what you have. Moving arrays are a point of failure, expense and a power consumer.
>>1005051
Arduino will not be suited. Also if you want advanced logic to prevent cloud tracking you want machine vision, not LDR's. A PLU or maybe a Raspi with opencv could do this, it's all down to the code though
>>1005051
My dad actually did this. Just used a clock and programmed it to calculate the sun based on our location.
It's not worth it.
Just add more solar panels and take the 35% hit by not tracking. You'll end up ahead in the long run.
>>1005659
>le wind meme
Unless your in a very peculiar area wind on the DIY level is total shit.
