Main Site Documentation

My New Project Idea: Solar powered temperature/environment monitoring in greenhouse


#1

I’m going to build a temperature logging system to monitor temperatures over the day in the little seedling raising greenhouse I build from a recycled window frame. I want to make sure I’m not over-cooking the seeds (which I think batch 1 is already crispy) so I want to measure ambient inside the box, outside the box, and in the soil in one of the seedlings; log this over time and check it regularly. Eventually this device might get promoted to

For temperature I’m going to use the MAX6675 thermocouple sensor and K-type thermocouples that I have for a coffee roasting project that is on hold (due to needing to build a much larger shed to store the roaster :slight_smile: )

For power I hope to use Li-Ion 18650 batteries, charged from a solar panel. I might go to larger batteries (say a 6v gel cell) if I find I need to run a fan or a motor of some form to open the lid of the box and ventilate things.

I have a “12v battery charging solar panel” that I have no real information about - all I know is that it has some internal circuitry, it’s not a raw cell, as it has at least an LED that flashes when there’s sun out even if there’s no load on the panel itself. I measured it’s open circuit voltage in sun and it reported over 20v.

This will all be running a Panda with SD card.

This is just a thread to get input on the power aspects of the project. I have little/no knowledge of these aspects so I’m looking for ideas - thanks in advance for your comments!


#2

Have you measured the panel current? If you can find out the wattage rating of the panel, you’ll then have a starting point for figuring power needs. For instance, I have a 6V 2W panel that will give me a theoretical max of 333mA. In reality, the panel has never gone over 300mA. Most ballpark estimates take 80% of the max as a realistic number. I’ve found this to be ‘close enough’ when figuring how long it will take a battery to charge.

[edit] Panda I does not have RTC… I would suggest using a board that does to take advantage of hibernation to save on power. If you solar panel and battery are big enough, this may not be an issue.

I have a cold frame currently setup with greens in it (mesclun, spinach, lettuce). Soon, temps here in VT will be getting below freezing regularly. I’ve got a small ceramic heater on a relay hooked to a Panda ready to keep my salad from freezing. It will be interesting to see how long I can keep them growing before the salad gets too expensive (to heat) or the cold frame gets covered with snow!


#3

These measurements are recorded based on connection to a standard (cheap-ish) multimeter.

it’s short-circuit current is showing .08A in FULL sunlight, at optimal direction to the sun (as determined by me, with no accuracy at all :)).

it’s open-circuit voltage is showing 22v in same conditions.

so is that 1.76W? (probably less under operating conditions, so probably call it 1W).

Opened the back of it to see what it’s construction is like.

It looks like there’s one long series of moncrystalline cells with a +/- at opposite ends of the panel (they look like they are just soldered onto the back of the conductive material on the panel, that is covered with an insulating plastic film). The positive side has a diode protecting the panel from reverse polarity connection to a battery or any power source.

There is also a LED that shows when the panel has power on tap and no load attached - so in the open-circuit test shows flashing, short-circuit test it’s off, panel in the dark it’s off. When I connected it to a 12v windscreen wiper motor (which I knew it would not have the power to drive) LED goes out. The LED connections are specific, they run from the - wire on the panel to a third pad on the far end of the cell chain, which looks to be on the first cell.

So my first thought was can I divide the cell up to increase the current but that doesn’t look possible/easy as I can’t see any way to “tap into” the panel; there’s only one “wired” circuit. Perhaps soldering additional pads onto the back along the length? Then again, perhaps I should just lash out and buy a panel I know more about :slight_smile:


#4

80mA at 20v isn’t a lot at all. How big is this panel physically? Are you sure is is not 0.8A? Even the smallish 12V panels are good for 5W.

There will be two ways to get the 20V down to the 3.3v the Panda actually needs. One is to waste the power with a linear regulator and the other is to use a switch mode regulator which will be much more efficient.

A pretty efficient way would be to use a lithium charge chip that can handle high input voltage.

To be honest, the easiest thing to do is to get one of those solar chargers for cell phones with USB 5v output. They have a built in battery and you just plug it into the usb port or 5v rail on the Panda.


#5

Yeh I am sure it’s .08A, that’s 80mA, at the BEST. Like you said, it’s not much :frowning:

Physically its dimensions are about 200mmx100mm (maybe slightly less) so I’m not that surprised. I think I got it as a freebie with something else I bought on ebay years ago - they called it a car battery top-up charger, but some simple maths says this wouldn’t recharge a battery that had been used for a few minutes with the interior light on the car left on.

Heading back to eBay now :slight_smile:


#6

With an 80% efficient switching regulator you can get 350mA out to charge a battery. Even without hibernating you might be able to stretch that to do your thing, if everything is optimal.

What you could also do is to hook a I2C time keeper chip to the panda and use the time keeper’s alarm output to wake the panda up to do measurements.

Linear Tech has some nice solar switch more down convertors, with Maximum Peak Power Tracking to get the most out of the solar cells. You can maybe look at http://www.linear.com/product/LT3652. I have never used it, and LT stopped shipping samples to my country, so I never will try it.


#7

My first attempt at doing FEZ solar was similarly frustrating. I scavenged some small panels that really sucked. Be wary of eBay. Because solar is “in vogue”, there is a lot of junk out there, even on reputable sites. Adafruit has some nice smallish solar panels,only in 6V though. They take the time to make sure what they sell is the best value for the buck. They also have a lipo charger specifically designed for solar. They are having trouble keeping it in stock as they make it in house, but it rocks. Lots of solar info on the site, too.

Ditto what Errol said on adding a timer chip. Hibernating the board is the best way to be frugal with power.