Ok and for this novice, how does that work with no tag collector access to a Wifi router nearby, just a cellular mast?
Happy to go to any supplier, how much is the 40c per MB likely to add up to with say 6 tags running 24/7 and at 10min intervals?
As an aside. At my home I cant read all the tags I want because of walls and distances.
What do you know about Bluetooth Mesh, that seems the obvious solution?
Regarding traffic cost: each message is like 100 bytes * 6 tags * 6 mess sets per hour comes to 3600 bytes per hour * 17 hours(23:00-06:00) is like 60KB per day! for 30 days consider 100% overhead or more we’re talking like $1.00 . I hope that’s not a problem 
The Rpi talks to the the SIM7000C (or similar ) board which talks directly to the Cell network and sends the message packets onto the internet host.
Well I am rerired but I reckon my pension could stretch that far.
Ok lets carry on and see where we end up.
We heard about it a while ago at a BBKA seminar but it was pretty short on detail then and now although they use the terms AI and intelligent hives a few times with no details.
Oracle say nothing about the costs involved in using the various services they provide.
There has been a commercial system around for some time with weight sensors as well as temp and humidity but its way too expensive IMHO.
Bear in mind a commercial beekeeper near where I live has 2000 hives on 15 different sites.
He is going to want to see a rise in honey production at his honey sites and an increase in queen numbers at his breeding sites.
The hobby beekeeper is probably more likely to engage but a simple turnkey, self contained system that he can interrogate with his phone or tablet and works on batteries would be his choice.
We are no where near that but Meazurem are close, although if my experience of the raspberry pi is anything to go by, I am not filled with confidence.
Tricky little beast with numerous pitfalls.
I visited a commercial apiary Sunday.
Over 60 hives in woodland clearings spread over an area of about five acres.
The only way sensors would work on that scale is a bluetooth mesh.
My own apiary layout is in a row at a field edge and means the collector/client has to be pretty carefully placed to capture the furthest ruuvitags I have in place today.
How to overcome the low range and line of site drawbacks is an issue. IMHO
60 hives over 5 acres, wow!
Regarding your apiary and range: What is the distance from the center of the row to the farthest hive?
If the Rpi is placed in the center of the row it may reach all hives.
There isn’t really an optical “line of sight” issue, rather a radio interference consideration. I believe the ruuvi antenna is omni directional.
More information at https://github.com/ruuvi/ruuvitag_fw/wiki/FAQ:-range:-Bluetooth
Regarding mesh of ruuvi’s: I would be concerned that the battery load for meshing 60 RuuviTags would be excessive as many of them would need to run the radios too much of the time. The radio being a significant load on the battery. Multiple solar powered Rpis each collecting data from a group tags would be a better solution.
The apiary I am thinking of is about 70 mtrs long and using Core Beacons on an Iphone I can get both end tags provided the hive lids are non+metallic; if I put metal covered lids on I lose one or the other end.
I was intending to use this apiray as a test bed as its nearer and handier to get to.
I spent the weekend wrestling raspberrypis into submission and am not all that impressed but I guess set up at home as a plug and play system it might work, except they seem very touchy about being powered down incorrectly.
I was thinking maybe using a zero w in an elevated position (and I understand one can turn the LED and HDMI off to save energy) leaving the BT and Wifi on (for ssh purposes?) and just run the collector on there and the rest on the client device?
Scanning for tags can only really be done on site so access to the Rpi has to be by ssh I guess but as they only need to be on maybe four times per hour in daylight that could be accomodated.
Agree about using more than one collector Rpi and rsyslog looks pretty good to this nooby.
Sorry you had issues with Rpi. What image did you start with? RpiZeroW?
What kind of issues did you run into?
Please see my notes at MyBeacons.info/pi
Re disable Rpi HDMI: Sounds like a good idea. I am not familiar with this. Can you point out a reference.
re disable Rpi LED: I suggest that this is minimal.
re Rpi : WiFI Once deployed it should not be needed.
Connection via cell network should be configured to be enabled only periodically to transmit status OR while ssh connect is active.
Sufficient solar generation and battery capacity will need to mitigate these issues.
re Rpi powerdown: Graceful shutdown should always precede poweroff. It has NOT been my experience that abrupt power has caused a problem (and I have done it a lot). What issues have you observed?
re configuration i.e. include/exclude bluetooth devices. This should be able to be done pre-deployment. I’m not even sure this is necessary since RuuviCollector logs all visible tags and I expect there won’t be any around you don’t want.
Can I see you hives via google Maps??
Issues? SD Card, taking an age to format and burn, getting the headless working, issues with router configs.
StretchLite, NOOBS, Ruuviberry, you name it, I played with them all.
See this re Rpi power consumption.
https://www.jeffgeerling.com/blogs/jeff-geerling/raspberry-pi-zero-power
Understood re Wifi use and ssh over cell network.
Battery and solar capacity will need to be set for the worst case.
I had several issues with no led coming on after unplugging accidentally before a kill command had completed and a change of sd seemed to be the cure.
If the tag ID can be set up in plain English pre-deployment then fine, if not then we are back in the ‘which tag is which’ scenario.
Ive just moved some hives onto a new site at 51.1001621, -0.2719824
On Google you can see the cellular mast and the first batch of six hives are from there to the east along the field boundary.
Did you use balenaEtcher to write SD?
On MacBook?
Wish you’d have let me know. I could have given you a running image for a piZw with ssh, ruuvicollector, influx and Grafana running.
Use sudo shutdown -h now to stop ANY *nux system. With RpiZW it will reboot after a few seconds so you’ve got to unplug it after a count of 5 (?). (kill command??)
I’m looking at
https://www.google.com/maps/@51.0993834,-0.2698815,492m/data=!3m1!1e3
but don’t know what I’m looking for.
That white dot at the upper left hand corner of the field, at the end of the farm track, is probably the cell tower. Look at the shadow it casts to the north: looks like a lattice mast with antennas at the top.
Jeff Geerling reference is very informative.
My pi zero running ssh, ruuvicollector, influxdb and grafana runs 110-190ma.
20ma less with tvservice --off
bvYes to Balena on mac and with 7zip on PC, both seem to be liable to hang up.
Thanks for the heads up and offer, that would be good so I actually know what it should look like.
NoHalfBits has it, the hive site is from the mast to the East.
I have put stands for 8 and 5 hives in place for starters as it may just be a bit windy but we will see.
Some tolerate it, some dont, same with the mast, some say it buggars them up but no one at my club has suffered the issue and there are masts just about everywhere these days.
So working on 200ma plus the SIM board that I understand can be pretty power hungrry so that is something that will need looking at.
It may be the RF output and other stuff can be cut down resulting in power savings.
I dont think 2 amp at 5v (mentioned in the ad) is workable.
How do I send you an image?
PS
Some good stuff here too
What if we can run just the collecter on the Rpi and Influxdb and grafana as fhe client?
Andrew, Have you changed IDs here?
re building image: It’s not necessary for experience my pain in getting all the pieces together (unless you’re into that)!
Not sure what you mean by “just the collecter on the Rpi and Influxdb and grafana as fhe client?”
I will do some testing re cost of running Grafana when not connected and influxdb when not being feed data from collector.
We could use cron to start collector every hour (+/-) and have it only run for a few minutes. Just before running the checkHive.sh script. Or actually invoke checkHive.sh when collector exits.
I think I will order a hologram.io SIM card and either the nova (https://help.hologram.io/hologram-nova/what-is-the-novas-power-draw) or
… They seem to be really into IOT.
You could create a google map and share it like mine at
https://drive.google.com/open?id=1I-HftKIVhJeSiar77nhrdc4ZrTc&usp=sharing
Sorry, changed IDs?
Its a steep learning curve and I just want to understand why I am doing what I am doing.
Somewhere I read the collector and the influxdb and grafana can be run in different locations so my thought was save power by running just the collecter on the Rpi at the apiary and the rest on tbe Mac or PC or whetever in my office.
I have orderd a hologram card anyway.
Was reading about another project and it mentioned sticking to 2g as the USB dongles are far less power hungry due to the lower functionality and USB specs.
I am re-evaluating this apiary monitoring project, in view of your insightful and to me thankfully understandable (largely) comments and will test my thinking on you later if thats alright?
Yep we got to the same place in the end.
thnx, but no intentions to do that (create a google map and share it)… just can’t resist when somebody posts lat/long coordinates 