Remote Weather Station

[parkview]

What does the words: Remote Weather Station conjure up in your mind? In mine, it's a weather station way out off the main road on the Nullarbor desert.

This will be a many phased project - each phase building on the previous one as I learn and experiment along the way. Something like this:

Beta: just build a quick PCB to test out various sensor. Incorporate LOTs of test-points, and places/jumpers to measure voltage/current. Based on a STM32L151RE due to max programming space. Test this at home in the back yard. Allow for external sensors, but provide footprints for all on-board sensors too.

V1: The idea is that the first version would be setup in a remote farmers field. Will use LoRa tech to transmit the collected data back to me.

V2: The Second version would incorporate the Satellite data feed and be tested in a more remote place and be optimised for low power usage

I am initially imagining this to be the size of a 2L milk carton, and will be held up off the ground by maybe 10cm to 20cm on 4 legs. These legs might contain sensors to monitor soil temperature and moisture. I would like these to be as cheap and robust as possible, so I can place quite a few of them around WA.

All of the above will be powered from a LiFePO4 rechargeable battery. I haven't used these yet, but there are now four of these in transit for me to experiment with. I also have two different kinds of small 6V solar PV cells on their way too. There will be one mounted on each of the sun facing surfaces East-North-West. At the top there will be the a LUX and UV sensor, along with a GPS antenna and the LoRa/Satellite antenna.

Some of the data I would like to collect:

• Air Temperature
• Air Pressure
• Air Humidity
• sunlight LUX and UV index
• Soil Moisture
• Rain/Dew indicator
• Lightning count
• GPS (location/Datetime)
• Accelerometer (is it still standing upright?)

A stretch goal later on might be to include rainfall, wind speed and direction.

When I start a project, I create a Project folder and a README.txt file where I write down my ideas, things to explore, tests to be run, things to purchase and a project journal. I might spend a day or two researching my ideas and making lists and notes etc. Next up is usually followed by designing up some test PCBs, as thats my thing, and I enjoy the process. This also might include some bread board tests and testing some software etc.

[parkview]

While I started designing up a 10cm x 10cm cheap Beta version with some weather station sensors and a STM32L151RE MCU, I quickly hit the silicon shortage snag and found that if you can find MCUs, prices have gone through the roof. A MCU I purchased back in 2019, is now five times the price. I now have to settle on a 64pin or 100 pin MCU for around $11 to $18. Luckily I have a few 48 pin MCU's in stock, but they don't have much Flash/SRAM and I would like a lot to start with, as I haven't done much programming on a STM32 platform. Etra Flash/SRAM allows me to write bloated code, that once I get working well enough, I can then explore the process of cleaning it up and shoehorning it into a smaller and cheaper MCU.

I did purchase some STM32L151 and STM32L433 development boards to experiment and test code out with my sensor boards. Generally the STM32L433 is way over blown, however it seems to use 1/2 the power of a STM32L151. Maybe my solar power budget might be enough to run the STM32L151 for a day of five on battery power - time will tell all when I run some battery and solar power tests.

In the mean time, I did research some solar PV to LiFePO4 batter charger IC's. Initially I was going to use the LTC4162 IC, which has built in Maximum Power Point Tracker (MPPT), but I could only find one project based on it, and it does cost around $13.26ea and in short supply! Stephen mentioned the open source https://hackaday.io/project/20909-lifepo4weredpi. A very nice project. I see that it (and some others), are using the CN3801 IC and also includes a built in MPPT. It didn't take long to whip up a small PCB for it:

solar-pcb.jpg (44.57 KiB) Viewed 5356 times

Note: the two solar/battery connectors and Inductor are not shown.

I placed an order for the IC's today - before they sell out. Now I can get started on some software testing - or rather learning how to write some STM32 code, or maybe STM32duino might be a quicker way to go

[parkview]

The PCB arrived the other day and the charger IC a week later from China. It didn't take long to make up the PCB, but I did need to print out a new battery holder, as my 18650 LiFePO4 (LFP), batteries don't have protection circuit built in, so they are 2mm shorter that my Li-ion batteries.

Charger.1_sml.jpg (65.34 KiB) Viewed 5187 times

Today was the big test, I connected the battery up and all ok. Then went outside into the mid-afternoon sun and angled the 6V PV panel a bit and connected it up. I measured 4.2V coming in from the panel, which from memory, is about the optimal power point for the small PV panel. The battery was maybe around 14% charged, so getting down there in capacity. It started charging the battery. Looking at the photo below, I now notice that the protective cover is still on the PV panel

solar-LFP-Battery-Charger.1_sml.jpg (101.83 KiB) Viewed 5187 times

It's hard to see out in the sun, but the D4 'charging' LED is lit.

[parkview]

Today was a nice warm day, topping out at 34 deg C, so a great day to test out multiple solar PV panels. I used my Ideal Diode test board: viewtopic.php?f=12&t=2473https://forum. ... =12&t=2473 and configured for dual input, with one output. I made up a number of adaptor and test cables, including these PH female to Pin-Header to PH male connector:

plugs_sml.jpg (24.6 KiB) Viewed 5168 times

These cables allow me to quickly measure the current travelling through the cable. Left cable is in measurement mode, right cable is pass through mode with a yellow jumper.

I spent the morning running various tests of the setup and at around 2pm, I plugged in the solar LFP battery charger board and around 240mA at 4.26V was going into the charger board. One issue I noticed was that sometimes the Ideal Diode circuit didn't kick in properly when a panel went into shade. Mostly it did, but sometimes it didn't. This needs further investigation.

In the photo below by 3:30pm, I had one PV panel on concrete/wood chips to kind of test out how PV temperatures vary. With a background temperature of 58 deg. C for concrete and 70 deg. C for the wood chips. The PV panel on the concrete/wood was running at 72 deg. C and on the white board at 69 deg. C. The white board was a relatively cool 47 deg. C:

solar_LFP_Charger_sml.jpg (153.89 KiB) Viewed 5168 times

By 4:30pm, both PV panels where in shade, but still charging the battery! Next up tomorrow morning I will finish the battery charge and make sure it cuts off at around 3.6 volts. I will also design up a special PCB with four inputs through ideal diodes and one output. I will also add on some INA219A voltage current monitoring of the input/output voltages and currents. The sensor PCB design still needs to be finished off.

Note: the charger and battery where only placed in direct sunlight for the photo. They normally sit in the shade.