temp/humidity monitoring project, to cover 22 rooms. Needs to talk over wired Ethernet (No wifi or other wireless) and have the option for battery power (or PoE) for some rooms. Total cost for the first 22 rooms has to stay under $1000. I was going to use either the sparkdata service or openhab for data collection, so something that works with MQTT.
What board do you suggest I use?
The LONG version
I am tasked with creating an inexpensive (we will get to what that means) environmental (again, definition coming) monitoring system for the entire (wait for it) building at work. If it proves useful, it will be expanded to the second building.
Inexpensive means that covering the whole building should involve an initial outlay of less than $1000.
Environmental is taken to mean temp and humidity from each room, plus the AC vent in each room (two locations in each room)
The entire building means 22 rooms.
And it needs to be wired. No wifi/zigbee/bluetooth/etc… And must be possible to be powered from a battery. And the battery needs to last a month. My first thought was to use an arduino as the boards are cheap, the sensors are cheap and the programming is easy. But the ethernet shield is expensive ($44).
While there are many other boards out there, not all/many of them have ethernet, fewer have ethernet and deep sleep capbility, and I am familiar with almost none of them beyond Pi and Arduino.
So I am turning to the community to see what your collected knowledge can bring out as the best price/feature trade off for gettign this done.
I intend to use the OpenHab or data.sparkfun.com (except run in house) to log the data and make use of it. This means MQTT, so please keep that in mind.
If I search for “ethernet temperature humidity sensor” I get a bunch commercial solutions. Have you requested pricing quotes from those? At least to get a ballpark idea of the unit cost. Battery powered and requiring ethernet connection does not make much sense. PoE is definitely a smarter idea in that case. You don’t want to go out on a chase for units with empty batteries. As undesirable as a wireless unit that can’t connect because of bad signal strength. PoE would require adding the cost of a 32-port or 24 port PoE switch, or a generic one with as many individual PoE power injectors. Old 10Mbit would probably suffice if that makes it cheaper. Either way that will probably take up a considerable part of that budget which you seem to not have included. And don’t forget the effort and added costs of laying out the cables, guides, etc.
I already have 48 ports of empty POE (old equipment, long since replaced, but still serviceable for this) that I can use to create a separate network for these devices.
The commercial stuff usually costs a LOT per unit and then charges more for their central control (~$200 for the POE version of the watchdog unit).
I decided to power most of them off USB chargers, as we have hundred (literally) of them around and they all have 5v 1a output. The network ports and the power plugs are near each other but nowhere near the AC unit, and I need to take temp of the ac air and the room. So… long lead?
These devices will not be secure in any sense of the word, so they should end up on their own VLAN and have no connectivity to the internet except… it would be SO much easier to have them all dump data to a google sheet (found that here https://www.open-electronics.org/how-se … readsheet/ ) and then use something like appsheet.com to change it into a mobile readable “app”. Given my personal drive to reduce attack surface, it will probably end up with a locally hosted version of openhab or sparkfun data and I will simply need to put more time in on it.
Being an occasional repeat beginner (I occasionally return to arduino to start yet another project) I find myself relearning what I learned two years ago. For instance, did you know its possible to burn a thermister by not paying attention to what you hook it up to?
Ill return here occasionally and let you know what else I burn.