Home automation is a popular project to undertake but its complexity can quickly become daunting, especially if you go further than controlling a few lights (or if you’re a renter). To test the waters you may want to start with something like this home safety monitor, which is an IoT device based on an Arduino. It allows remote monitoring of a home for things such as temperature, toxic gasses, light, and other variables, which is valuable even if you don’t need or want to control anything.
The device is built around an Arduino Nano 33 IOT which has WiFi and Bluetooth capabilities as well as some integrated security features. This build features a number of sensors including pressure/humidity, a gas/smoke detector, and a light sensor. To report all of the information it gathers around the home, an interface with Ubidots is configured to allow easy (and secure) access to the data gathered by the device.
The PCB and code for the project are all provided on the project page, and there are a number of other options available if Ubidots isn’t your preferred method of interfacing with the Internet of Things. You might even give Mozilla’s WebThings a shot if you’re so inclined.
You’re going to be lucky to get a couple of hours out of this thing before the battery is exhausted. I would look at using a hard-wired power source or radically redesigning your power delivery.
A 6LR61 9 volt battery only has a capacity of about 500 mAh (depending on manufacturer, chemistry and discharge rate) but an AA battery is about 2000 mAh.
9V batteries are not great – then you’re throwing away half that power with a 7805.
A buck-boost regulator would allow you to get 5V from 4x alkaline AA batteries, or a 3.6V Li-ion, or 4x NiMH cells, giving you more efficiency, more battery capacity and rechargeable chemistry options.
Then there’s that gas sensor heater – basically an always-on 28 ohm resistor drawing 180 mA constantly on your 5V rail.
Use a FET to turn it off, only turn it on when you want it on.
Keep the whole system in a low power sleep state, and use a low power timer to wake up every 5 minutes to transmit a sensor reading – Texas Instruments has some good options like the TPL5110 for this purpose.
And the 9V battery in the picture is particularly bad. I once bought a pack of them from eBay because they were cheap. Half were already empty, half had about 100mAh of capacity.
That circuit can be made to sleep at around 120 microAmps. A 7805 has an idle current draw in the milliamps – you wouldn’t use one for an IoT sensor anyhow – better linear regulators come with 10 microamp draw.
If you only need to measure stuff and connect to the wifi network for 10 seconds every hour, and the circuit draws 100 milliamps, then you need about 9.6 mAh per day and a single 9 Volt battery will run you for 52 days. 80% of that power draw comes from the need to connect to wifi – otherwise the circuit can turn on for a second and just shut back down.
With some further optimisations, this thing can easily be made to run for a year on a battery like that.
>A buck-boost regulator would…
…have a much higher constant current draw when it comes to running electronics down to very low idle power requirements. At the very lowest current drain, a linear regulator wastes less.
You are not wrong, but you do get pretty great buck converters. I am using the TPS62740 buck in a sensor module, it’s Iq is 360nA.
There’s a trick to that. The Iq is reported for the off-state, when the converter is not switching. In reality, it switches all the time at some low duty cycle to keep the output up.
For 10 uA output current, you’re still wasting at least 1 uA to run the converter, and at 100 uA you’re wasting 5 uA to the converter, so it’s not any better than linear regulators of similar specs – except that you get the voltage ripple that interferes with stuff.
And don’t forget the leakage current of the input and output capacitors, which is greater when there’s a ripple going through them.
I agree with you. The Gas sensor, furthermore, has a coil heater that draws a lot of current and need to be properly heaten (=some minutes) in order to work properly. For me is not a good idea to use a 9V battery with a gas sensor on board. Maybe some LiPos are a better choice, but is good to evaluate also usage of a solar cell for the charge.
9v * 0.5a = 4.5Wh
1.5v * 2a = 3Wh
So, 9volt battery has 50% more energy.
I read “HiWaste” on that battery…
You are not the only one!
I initially read “Ubidot” as “U b idiot”
I have to say that I really don’t want to disrespect that build! I live the good part of the year in my truck (half workplace and half RV of sorts) and a good monitor/warner for stuff like carbon monoxide and gas leaks is quite important. Right now I go for the standard stuff, tho. A combination CO/propane/anaesthetic gas warner which is certified by my insurance company costs about $50. It is sealed, disposable and works about eight years with the built in lithium batteries. That’s a deal good enough for me.
Great project, but in general the progression of IoT/networked widgets (that rely on the assumption that the US’ marginal consumer-level data infrastructure will be utterly reliable) often neglects the usefulness of an attached bright light and loud siren to be useful when there’s no Internet for the Thing, which around here is not uncommon.
After he have experience using this device he will learn more about building a better power system.
Ah, more “Internet Of Shitty Things” frameworks. Install dibble, to interface with the dabble account, which needs puffle to compile druffle, and a babble account to communicate with a bibble server, passing data to the shibble processor, which parses the data to the tibble format, which then gets formatted into the shittle format, which needs a dapple account, to communicate with fabble, to enable puffle to fluffle the socksifier.
Something that can be done with a raspberry pi zero , some other components [ I’m thinking a nice spi adc ]. and done.
for that, I get a nice system I can ssh into, with ONE and only ONE open port on the firewall, on NO THIRD PARTY SHIT/SERVICES.
Another AD, for the another SHIT IoS [ Internet of Shit ] framework.
Okay, this design isn’t going to sip power in the manner it needs to for a 9V battery… but a bit of work on the PSU end of things and it could be a _lot_ more efficient than a Raspberry Pi could ever be, with faster start-up times and a more robust architecture to boot.
Nothing stopping the replacement of the IoT service with something that works on-premises either. It has Bluetooth, so maybe do something over BLE to a server running on a small computer (maybe a Pi).
Then the sensor placement is only restricted by radio range and the locations where it will best detect the gasses of interest. The computer collecting the data can be placed anywhere where networking and power are convenient.
How is relying on someones s*tty framework, on their s**ty servers “more robust”.
For the ultimate in simplicity, an ESP8266, or an ESP32, or
Arduino + HC05 module. Done. No IosT [Internet of Shitty Things] required.
Logging to SD cards can be done reliably too, without giving access to all your data to some IosT “broker”.
Use a length of wire instead of a fragile chain of arcane, buggy, insecure, and unreliable online services?
How…inelegant! Hardly worth mentioning…
If you have any real concern over toxic gases an iot Arduino is the last thing you should turn to. A toxic gas hazard is sil3/PLe territory.
It might be the same logic as putting can openers on pocket knives.