Hi Robert, Depends what you mean by real-time, but I would say yes, I am. Here's some verbose debug output from the program running on the raspberry pi base station. This program listens to the remote arduino pro mini (via the RF24 mesh) and then publishes on my LAN to the MQTT broker.
Node 5 sent: temperature : 4.96 C pressure : 1009.25 mbar humidity : 79.05 % temperature at humidity sensor: 0.8 C irradiance : 13 battery voltage : 933 solar voltage : 874 mcu temperature : 32.8 At 2017-02-02 11:19:50 waiting for up to 10 seconds for publication of TIME:0,AMBT: 4.96,BARP:1009.25,RHUM:79.05,HUMT: 0.76,IRRA:13,BATV:933,PHOV:874,SYST:32.80,WIND: 0.0,WDIR: 0.0 on topic weather for client with ClientID: survey Message with delivery token 3 delivered return code 0 rcv Node 5 sent: temperature : 5.21 C pressure : 1009.27 mbar humidity : 77.71 % temperature at humidity sensor: 1.6 C irradiance : 10 battery voltage : 933 solar voltage : 870 mcu temperature : 32.9 At 2017-02-02 11:20:17 waiting for up to 10 seconds for publication of TIME:0,AMBT: 5.21,BARP:1009.27,RHUM:77.71,HUMT: 1.61,IRRA:10,BATV:933,PHOV:870,SYST:32.90,WIND: 0.0,WDIR: 0.0 on topic weather for client with ClientID: survey Message with delivery token 4 delivered return code 0 rcv Those packets are 27 seconds apart. I have the Arduino go to sleep for 8 seconds using a watchdog interrupt timer to wake it up, and hit the snooze button twice, so the total sleep is about 24 seconds. I consider that to be real-time enough for ambient temperature, barometric pressure and humidity. Running the sketch that does this, the remote station (Arduino pro mini, two sensor boards and nrf24l01 radio) consumes about 25 mAh per day, as I describe below. Eventually, I want to be collecting wind and rain data. No sleeping allowed for that data, so I'm thinking of a daughter station using just an ATtiny which collects data continually, and then gets interrogated by the pro mini when it wakes up. The daughter board would use a wired connection, either simple serial or I2C. On Thursday, 2 February 2017 11:05:16 UTC-4, Robert Mantel wrote: > > Hey Bill, quick question for you, are you doing realtime updates with your > system? Also can you tell me how much power draw of the arduino pro mini > and the nrf24 radio? I'm intrigued by the RF24 mesh networking > capabilities and would like to make this my next project, but I want to get > an idea of the power management required for these two boards combined to > make them autonomously solar powered. > > On Monday, 30 January 2017 12:52:06 UTC-5, Bill Morrow wrote: >> >> I have a BMP180 for barometric pressure and a si7021 for temperature and >> humidity connected to I2C on the Arduino Pro mini. >> >> The Arduino uses a nrf24L01 radio >> <https://www.amazon.ca/Kuman-nRF24L01-Transceiver-Raspberry-Compatible/dp/B01C3YNGI8/ref=sr_1_1?ie=UTF8&qid=1485796501&sr=8-1&keywords=Kuman+10pcs+nRF24L01> >> >> and the RF24 Mesh <https://github.com/TMRh20/RF24Mesh> library to squawk >> at the base station. I scraped off the onboard antenna and attached my >> homemade 5 element Yagi. I think this has improved range quite a bit, but >> gigahertz RF is voodoo, so who knows? I want to work on this. >> >> If I was doing it again, I might drop down to the sub-GHz range. For >> example 434 MHz <https://github.com/LowPowerLab/RFM12B> would have much >> better range. Those transceivers draw maybe 30 mA vs. 15 mA for nrf24l01. >> But the radio is rarely active, so should not be an issue. >> >> I send a simple packet of data over RF24Mesh of 32 bytes. The nrf24l01 >> radios are supposed to handle error correction and replies, so I don't do >> anything for that sort of thing. That's the big advantage of something >> sophisticated like XBee or nrf24, you don't have to do the low level error >> handling yourself. >> >> On the base station side, I created a program which is the master on the >> RF24Mesh, and publishes to the Mosquitto MQTT broker. It receives the 32 >> byte packet and converts it into a single MQTT publication that looks like >> this: >> >> TIME:0,AMBT: 7.23,BARP:1001.97,RHUM:70.92,HUMT: >> 4.44,IRRA:12,BATV:974,PHOV:673,SYST:33.20,WIND: 0.0,WDIR: 0.0 >> >> My weewx wxMesh <https://github.com/morrowwm/weewxMQTT> driver >> subscribes to the MQTT weather topic, reads the key:value pairs and . The >> nifty thing about this approach is I can publish on the "weather" topic >> from multiple stations, and wxMesh will merge them all together into loop >> packets. The indoor station sends its data synchronously. >> >> On Monday, 30 January 2017 12:52:41 UTC-4, Robert Mantel wrote: >>> >>> I'm interested in the MQTT standard, didn't really know anything about >>> it, but may migrate to that idea and skip the phant server step altogether, >>> but using the fileparse extention with the jsonp query of the phant server >>> just made things simple for me. >>> >>> As for the lower power radios I did consider using xbees, which I may do >>> at some point, but I'm not sure if they will save me enough power to >>> warrant the extra cost, we'll see how my experiment plays out. I guess I >>> would have to migrate from my BME280 because it's either SPI or I2C and I >>> haven't read anything about how to deliver that over an xbee radio >>> connection. What radio's are you using? >>> >>> On Monday, 30 January 2017 10:50:28 UTC-5, Bill Morrow wrote: >>>> >>>> Very nice. I considered flower pots for the radiation shield also, but >>>> then that wouldn't align with your snowflake theory, would it? >>>> >>>> I like both Sparkfun and Adafruit's ecosystem. Phant looks interesting. >>>> I'm reading temperature and humidity in the house using an Adafruit >>>> Feather >>>> Huzzah publishing over wifi to the MQTT broker. weewx subscribes to the >>>> MQTT broker to get the data. Powered by a wall wart, but the Feather has >>>> the capability of charging a battery. So in theory I could cut the cord >>>> there too. >>>> >>>> My main goal is to harvest power from the PV cell on the top of the >>>> lantern. It maxes out at about 25 mW at 3.2 volts. I should be able to >>>> easily get 30 mAh every day, and a lot more than that on a sunny day. My >>>> station averages about 25 mAh consumption per day. I've made no attempt to >>>> cut power consumption on board, e.g there is a small power LED that is >>>> always on. >>>> >>>> I'm working on a circuit provided by LInear Technologies: >>>> http://cds.linear.com/docs/en/design-note/dn491.pdf, which boosts >>>> almost any voltage to Li-ion charging levels. No success yet, mainly >>>> because of the difficulty of working with parts with millimeter dimensions. >>>> >>>> You're dealing with a much higher power regime. But sounds like you >>>> have some good power management strategies going. You might be able to >>>> reduce your power needs by going to low power radios instead of wifi? >>>> >>>> On Monday, 30 January 2017 11:20:17 UTC-4, Robert Mantel wrote: >>>>> >>>>> Hey Bill, no problem. I can't provide pictures right now but will >>>>> when I get home. I'm using an 8000 mAh battery I got from ebay (I may >>>>> upgrade to a 12000 mAh), that is connected to an Adafruit Solar Lipo >>>>> charger. I also bought a 6 volt solar panel from ebay or amazon, can't >>>>> remember right now, but the adafruit charger works best with a 6v panel. >>>>> The charger board has a connector for the battery, the solar panel and >>>>> the >>>>> load as well as an older mini-usb connector to charge the battery from an >>>>> ac usb adapter if you just want to top up the battery when it's too low >>>>> and >>>>> it's really cloudy. You can also configure the rate of charge with the >>>>> addition of a resistor to boost the charge rate to 1000 mA which is what >>>>> I >>>>> did. Most lipo's can only handle a charge rate of 1C (C being the >>>>> capacity >>>>> of the battery) so I'm well within the 8000 mAh that the battery could >>>>> take >>>>> charge wise. I'm currently experimenting with how many days I can go >>>>> without intervention (hence the reason I may go with a larger 12000 mAh >>>>> battery to get more days of autonomy out of the weather station). >>>>> Ideally >>>>> I would like it to run 24x7 without intervention. >>>>> >>>>> To save power I do with the Particle P1 redboard is that I have >>>>> configured it to run in manual mode, meaning I control the wifi and cloud >>>>> connections. Because I have a local phant server collecting the raw >>>>> sensor >>>>> data from the redboard I have no need to connect to the cloud and because >>>>> I >>>>> have it in manual mode I briefly turn on the wifi every 2 minutes to dump >>>>> the sensor data to my phant server in the house. Wifi is by far the >>>>> biggest power consumer so the more you can control that the better. I'm >>>>> waiting for a 6 dBi wifi antenna to put on the particle P1 because I've >>>>> noticed that sometimes the wifi doesn't connect where I have it placed >>>>> inside my shed. All the sensors are outside the shed of course, but I >>>>> wanted as much of the electronics sheltered from the elements for >>>>> longevity >>>>> sake. The only active electronics I have outside is the BME280 sensor >>>>> inside my own home brew radiation shield I made out of flower pot trays. >>>>> >>>>> While I only post data to weewx every 2 minutes, the station is >>>>> processing data continuously, so rather than feed for example the current >>>>> wind direction and speed, I dump the average wind direction and speed >>>>> over >>>>> 2 minutes to weewx. In a previous iteration of my weather station I had >>>>> it >>>>> sleeping during each cycle but that proved to be a problem for getting >>>>> accurate rain readings because if the station was asleep it would miss a >>>>> bucket tip. So I went with a larger battery and then controlled the >>>>> wifi, >>>>> which saves a ton of power. The system has a rescue mode so that if the >>>>> battery falls below 20% soc it will begin sleep cycles of 5 minutes so >>>>> that >>>>> the solar panel can pump all the power into the battery until it rises >>>>> above that threshold and comes back online. So far so good, hasn't had >>>>> to >>>>> do that yet, but it may tonight because we've had quite a few overcast >>>>> days >>>>> lately. >>>>> Let me know if you want more specifics and I'll help fill in the >>>>> blanks. >>>>> >>>>> On Monday, 30 January 2017 09:59:03 UTC-5, Bill Morrow wrote: >>>>>> >>>>>> Robert, can you share any details on the solar powered Redboard? My >>>>>> weather station is based on an Arduino Pro Mini, powered by an old >>>>>> Li-ion >>>>>> camera battery. The weather station sleeps most of the time, waking up >>>>>> briefly every 10 seconds or so to sample weather. I get about 2 weeks >>>>>> out >>>>>> of the battery, then swap it. >>>>>> >>>>>> It's all housed inside an old solar garden lantern and a container >>>>>> which originally held some delicious carrot ginger soup. It communicates >>>>>> with a Raspberry Pi 1 in the house via nrf24L01 radios, using the >>>>>> RF24Mesh >>>>>> libraries. The radio is pretty power hungry, and the main reason for the >>>>>> large sleep duty cycle. On the left of the case, you can see the antenna >>>>>> pointing at the base station in our house. >>>>>> >>>>>> >>>>>> <https://lh3.googleusercontent.com/-cf4BRy65O2w/WI9S_5UBXLI/AAAAAAAAEfY/YPV_pxfK36cppzaMq9XYPsbg0Lr7F0gKgCLcB/s1600/weather_lantern.jpg> >>>>>> >>>>>> The temperature sensor shield on the bottom is made from some sports >>>>>> drink bottles. >>>>>> >>>>>> My main project these days is developing a circuit which will use the >>>>>> small amount of power generated by the cell to keep the battery charged. >>>>>> I've done some measuring, and it should be enough. I've gone down the >>>>>> rabbit hole of tiny surface mount components, so making slow progress. >>>>>> >>>>>> Oh, and this is all supposed to be done under a zero cost budget! >>>>>> >>>>>> On Monday, 30 January 2017 10:33:58 UTC-4, Robert Mantel wrote: >>>>>>> >>>>>>> Seems like raspberry pi systems are like snowflakes...lol. I use a >>>>>>> solar panel charged battery powered Particle P1 Redboard from Sparkfun >>>>>>> that >>>>>>> I have an I2C BME280 (temp/pressure/humidity/altitude) sensor, a >>>>>>> rain/wind >>>>>>> set from sparkfun as well. The redboard sends weather data over WiFi >>>>>>> to my >>>>>>> Phant server every two minutes, I have a cron job on my Pi that does a >>>>>>> jsonp query to the phant server to retrieve the latest record and parse >>>>>>> it >>>>>>> out into a text file with the "* = *" pairs that fileparse wants. Then >>>>>>> my >>>>>>> weewx system running on my pi archives every 2 minutes. Not realtime, >>>>>>> but >>>>>>> good enough for my purposes. I also have steelseries gauges running as >>>>>>> well. Been a great learning experience and everyone here has been very >>>>>>> helpful. >>>>>>> >>>>>>> Many ways to skin the cat here. >>>>>>> >>>>>>> On Monday, 5 December 2016 04:32:46 UTC-5, Joaquin Lopez wrote: >>>>>>>> >>>>>>>> Hello, at the School we have this weather station >>>>>>>> https://www.raspberrypi.org/education/weather-station/ >>>>>>>> >>>>>>>> It would be possible to configure the weewx system to be able to >>>>>>>> receive the data from the different sensosrs of our weather estation? >>>>>>>> >>>>>>>> Thanks in davance! >>>>>>>> >>>>>>> -- You received this message because you are subscribed to the Google Groups "weewx-user" group. To unsubscribe from this group and stop receiving emails from it, send an email to weewx-user+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
