Weather Station and Solar tracker

Hi everyone... I've been snooping around for a few weeks and casually reading and building some pretty simple circuits based on an intro book. I've come to the conclusion that it's time to build something for real!

I'm interested in having a weather station, and could use some advice on feasibility and scope.

I saw a youtube video (Sun Tracking Solar Panel w/ Arduino - Powers ITSELF!!! - YouTube) and it made me think that is a good idea. The sun changes position throughout the day, and also throughout the year, so I figured a double axis solar tracker might be cool to try. The guy in the video used 3 photo resistors and calculated difference between the center and each one on the side. Perhaps it would be possible to place the 3 in triangle and calculate a difference not only in the x-axis, but also the y-axis. Two servo motors could be assembled in a similar fashion to what was shown to rotate it literally in any direction. If 3 resistors wouldn't work, maybe 5 assembled in a + sign could be used to calculate differences. Thoughts?

The first issue I run into is how I am going to power it (the tracker and the weather station). Not being familiar with weather station sensors and servo motors, will I need two separate power supply sources (e.g. 2 separate solar panels/batteries)? I don't want to have it plugged in, so that way I can place it wherever I want. I was thinking a solar panel might be good - I live in Houston, TX and it's pretty sunny here. That coupled with a rechargeable LiPo battery might do the trick? I have a nice 9V@1W solar panel from Parallax, and about 5 or 6 of those little garden light solar pannels that, if wired in series, might be able to get me up to 9V in peak day light (maybe 4 or 5V indoors) total.

The second issue is weather I should buy individual sensors and solder them to a breadboard on my own (which has the sense of ownership factor) or should I buy sensors that are already assembled and working with proper circuits? Examples might be the weather station by sparkfun Weather Meter Kit - SEN-15901 - SparkFun Electronics? and https://www.sparkfun.com/products/10586, or perhaps a complete weather station by Oregon scientific which transmits data via RF 433MHz (http://us.oregonscientific.com/shop/browse.asp?cid=2&scid=48&page=2). I like the idea of purchasing an anemometer because, to be honest, I'm not sure how I would do that myself. But the temperature, humidity, UV, etc. would be easy enough and considerably cheaper to simply buy the sensors and assemble it myself. I also don't necessarily like the idea of all the data transmitting via RF because then I need to have a RF receiver set up somewhere to receive the data, and then send it to the internet. I Think it would be better in the end to invest in a WiFi transmitter and send it directly through my wireless network to the final data repository.

Thanks in advance for reading this long post. Any advice or suggestions on things to consider would be greatly appreciated. I hope to document everything from start to finish so others can easily figure this out. If it's been done before and I missed the post, please point me to the place!

Robert

Given the date and time and your location it is possible to calculate where the sun is.
If you use photoresistors you may find you are tracking the brightest part of the sky which, when there are clouds, may not be where the sun is.

I think I read somewhere that Norsemen used polarising crystals that let them navigate by the sun even on very cloudy days - not sure about that though.

Unless this is a remote unmanned station in general I don't think it's worth automating the second axis. You can adjust it manually every 1-2 months.

Using any type of light-based detection has issues which need to be handled, like what if the sky is a constant brightness due to cloud? You need to apply sanity tests to make sure the thing doesn't hunt all over the place.

Personally I would use an RTC and a small lookup table of sunrise/set times for the year. Probably just one entry per week or even less.


Rob

geofzx:
The first issue I run into is how I am going to power it (the tracker and the weather station). Not being familiar with weather station sensors and servo motors, will I need two separate power supply sources (e.g. 2 separate solar panels/batteries)? I don't want to have it plugged in, so that way I can place it wherever I want. I was thinking a solar panel might be good - I live in Houston, TX and it's pretty sunny here. That coupled with a rechargeable LiPo battery might do the trick? I have a nice 9V@1W solar panel from Parallax, and about 5 or 6 of those little garden light solar pannels that, if wired in series, might be able to get me up to 9V in peak day light (maybe 4 or 5V indoors) total.

LiPo batteries are a possibility, but they aren't as temperature tolerant (scroll down to "Temperature Effects" section), as some other battery chemistries. If the battery is going to be outdoors it should be an at least somewhat sheltered place, but still with at least some potential air flow (e.g. under a building's eave or deck). So it isn't exposed to the worst extremes of heat and cold. Of course if your system has the batteries inside a temperature controlled room, that would be even better.

On the other hand, while there are still limits on operational temperatures a sealed "gel" lead-acid battery would better tolerate exposure to both higher and lower temperatures, and would have significantly less risk of fire, etc... if it severely overheats. An appropriate SLA would be both heavier and bulkier than a LiPo battery pack, but reasonably sized (1 to 2 lbs and ~96 in3) 6 V and 12 V batteries are available in most hardware stores and some department or hobby stores. I'd recommend something like that for a "remote" weather station, even if "remote" really means a couple of yards from the nearest heated building in an unshaded case.

In any event, you'll probably need to regulate the power coming from the battery / solar panel so to a predetermined and known level before it gets to your electronics. The linear regulator most Arduinos would work with voltages that are ~7 VDC to 12 VDC, but it wouldn't work at all with a nominally 3.7 VDC LiPo pack and it wouldn't work well with a nominally 6 VDC SLA. Depending on the exact levels involved you might need to either step-up or step-down the voltage. To go from a range of lower voltages to a higher one (step-up) you'd need what's called a boost converter. To go from a range of higher voltages to a lower one (step-down) you'd need a buck converter, buck converters are also more efficient and generate less heat than linear regulators. There are boost-buck converters (a boost and a buck converter, and possibly some control circuitry, in the same module) that can do both, but they are generally more expensive and usually avoidable through proper planning and design.

Thanks for the replies everyone! The discussion has given me new things to think about.

@Radman: Good idea, I'll look into the polarized lenses. Could easily get a cheap pair of sunglasses and fit the lens over the resistors and compare observations with / without. I'll look into that more - thanks for the idea.

@Graynomad: It will be "remote" in the sense that is will be in my backyard, probably under a roof overhang inside an enclosure, but located not very close to any AC outlets. My options then would be to drill a hole into the garage and run an extension cord in (I don't want to do that) or to use strictly DC power through batteries or solar or some combination. Because the arduino and other sensors will be under a roof overhang, I was thinking a U shaped PVC arm could be built to take the solar panel past the overhang and up higher than the roof.

I agree about the sanity test. 1 or 2 days of constant rotation would probably wear a servo motor out. Perhaps some code logic could be used to say something like, if I start a timer when it starts moving, if it has 2 minutes of continuous rotation, then go to a set position and wait a couple hours? I'll look into the tables you speak of and see how I might be able to apply that to a servo to match sun position.

@Far-seeker: That's a good point about the temperature of environment affecting batteries. Houston is usually hotter than it is cold, but there are a couple weeks of the year where we might see freezing temperatures and that could be bad for the batteries. I absolutely don't want to have a heated enclosure - don't want the fire risk of a heating element. I'll start reading about SLA batteries - thanks for the tip.

On the suggestion of a voltage regulator: I agree that's a good idea and I'm not familiar with how they work. If I have a solar panel that outputs 9V in daylight, that voltage level will obviously fluctuate depending on time of day and clouds, etc. The 9V should be good for an arduino to run on, but do arduinos handle the voltage fluctuations? Would it be better to regulate it back to a constant 6V? Or perhaps run everything off of the battery and only use the solar panel to recharge batteries during the day?

Thanks everyone, this has been a huge help!
Robert

will be in my backyard,

So it's no problem to adjust the angle every month or so, then you don't need to automate the second axis, fun to do but no necessary.


Rob

I'm making a similar project, but without the wireless part. I really wanted wireless but the climate here just won't allow it. It can stay below freezing for months in the winter and I know even with a SLA battery I'd have power issues in that kind of prolonged cold. Plus, the sun can disappear for weeks here. Therefore I ended up running 100ft of cat5 to my outdoor sensor station, and dropping the clock rate for the sensor communication, given capacitance issues and such caused by the long cable.

I'd also be careful about program memory. Mine disappeared fast when adding an ethernet shield and code to communicate with the multiple sensors, and I had to upgrade to a mega, which I am waiting to arrive before continuing my project. You might want to just get one of those instead of the uno and save the trouble of running out of space.

I really wanted wireless but the climate here just won't allow it. It can stay below freezing for months in the winter and I know even with a SLA battery I'd have power issues in that kind of prolonged cold. Plus, the sun can disappear for weeks here.

I'm sorry, but I had to laugh at this. What useful data are you going to get that visual observations can't provide? It's damned dark out there, is the same when you look outside or you look at your weather station.

It's too damned cold out there is the same whether confirmed by the weather station or not.

Not that I'm making fun of your efforts, or plights, but collecting data about a situation that you can control seems reasonable. Collecting temperature data when it has been too damned cold to go outside for the last 6 weeks doesn't strike me as useful.

Umm, some people do like actual data. Otherwise, there wouldn't be a market for weather stations at all, nor would there be ASOS stations all over the country.

The guy in the video used 3 photo resistors and calculated difference between the center and each one on the side.

I tried this method but found that the all photocells are not created equal in that their resistances are not uniform under the same light levels thus if one photocell had more light hitting it, it could still have a higher resistance than one with less light hitting it. I thought of mapping the ranges of each photocell to values of say 1 to 10 but then came across a method which makes it 'obvious' which direction is lighter or darker. It's done by placing a tall light barrier between two photocells (e.g. vertical barrier between the two photocells used for east-west). If the two photocells are not directly facing the light source, one of the two photocells will be in the shade of the barrier, making the difference in resistance between the two 'obvious'. The inherent differences in resistance between the two photocells under the same light source can be ignored by throwing away any values that are relatively close and only paying attention to values that greatly differ.

  • Scotty

@geofzx: I mentioned Norse navigation just because it was a fun idea, I did not think you might take me up on it!

The idea that the Norse used 'sunstones' to find the location of the sun on overcast days came from Norse sagas.

I thought this was an untested hypothesis. Fortunately if this link is accurate Viking Sunstone Navigation | The Mary Sue it sounds as if the idea has been tested. You may need to rotate the polariser and take readings at different points in the sky.

I would locate and read the paper before spending time on it though.
If you do find it works please post your solution, it would be interesting.

geofzx:
I'll start reading about SLA batteries - thanks for the tip.

They are also referred to as VRLAs, the most common type you'd find in hardware/department stores would be as replacement batteries for emergency or outdoor security lights.

geofzx:
On the suggestion of a voltage regulator: I agree that's a good idea and I'm not familiar with how they work. If I have a solar panel that outputs 9V in daylight, that voltage level will obviously fluctuate depending on time of day and clouds, etc. The 9V should be good for an arduino to run on, but do arduinos handle the voltage fluctuations? Would it be better to regulate it back to a constant 6V? Or perhaps run everything off of the battery and only use the solar panel to recharge batteries during the day?

There are two main types of voltage regulator you'd use on small to mid-sized electronics projects, linear regulators and switching regulators. Both can usually handle a varying input voltage, provided it stays within the operational input voltage range for the specific regulator.

The official Arduinos, and most of the derivatives, use linear voltage regulators to provide the proper voltage for the micro-controller and pins to power off-board devices (i.e. the ones marked 3V3 and 5V). These are effective, but inefficient because the decrease in voltage from the input to the output + the voltage dropout (basically the voltage drop necessary for the regulator to function) has a proportional relationship to the heat generated by the linear regulator. The amount of waste heat generated, therefore how inefficient the regulator's performance is, depends on the excess power it has to dissipate and that's determined by the decrease in voltage at a given current level. So for your application this heat is going to be wasted energy, which will decrease the effective battery life for your project, and will increase the temperature within whatever encloses the electronics.

For example, if you had 9 VDC coming into an Arduino's barrel jack while the board is drawing 250 mAs, the power needed to be dissipated would be ~3 VDC (for the sake of argument I'm just using a voltage dropout of ~1 VDC the exact value will vary with the current) * .250 A = 0.75 Watts. If you had 12 VDC as the input the power to be dissipated would be about twice as much.

In contrast, switching regulators use components like capacitors and conductors to store energy and release it at specific intervals to either raise or lower the effective output voltage. They are very efficient (over 90% efficiency is very achievable) and generate relatively little heat, but can cause some ripple on the output voltage and other electromagnetic interference issues. Many pre-built modules have capacitors or filters to reduce or eliminate these unwanted by-products. Although if you get one without these, or they aren't sufficient, you should add them to your design. Here's an example of a step-down regulator module on a small PCB, and here's an even more compact one in a potted package.

Finally, one slightly different alternative I didn't mention is to just have a step-up or step-down (whichever is appropriate) transformer. A transformer can convert between two voltage levels, but it doesn't have a fixed output level. Instead, the shift is proportional to ratios between the primary and secondary windings. So you would still need to use a voltage regulator for the Arduino (but the built-in one would be fine). If the battery would need one to charge would depend on the type, for a LiPo I would want some voltage regulation but a SLA would work provided the voltage highest voltage is a volt or two above the battery's nominal voltage. Also, a the power on each side of the transformer will be equal (minus some internal losses because we are dealing with real world devices) so for a given voltage increase or decrease on the output side, the current will change inversely as well.

Can someone suggest me how to calculate sun elevation angle.

I wanted to know basic steps using latitude and longitude how they calculate the zenith and azimuth angle and there after they calculate the elevation angle. i searched it google but i could not able to understand.

Please let me know simple way of learing it

AMPS-N:
I searched it google but I could not able to understand.

It took me about ten seconds to locate the formula in Wikipedia and the formula seems simple enough. If you don't understand it then I suggest you look for forums specialising in solar power for a simpler explanation; it's nothing to do with Arduino, anyway.