In your first post you did not mention anything about cost restrictions. If you now apply such a filter on top, obviously the possible range of products will be limited accordingly. But you should not let it go that far so you would jeopardize safety. Ready made products for European market must be CE certified and also fulfill other regulations. When you do DIY involving 240V AC , well....I would not
EDIT: I do have 2 pcs of Shelly H&T and I do not think they are reporting H correctly so I only use the T part in one them to measure non-critical room temperature in one room (battery life time is not so good either and those batteries are pretty expensive 3V lithium)
DIY is a great option for sensors. The D1mini is a fine choice and cheap. But don't go for the DHT sensors. Here are some alternatives.
BMP280 Temperature/Pressure
GY-BME280 Temperature/Pressure/Humidity
SI7021 Temperature/Humidity Sensor
DS18B20 Temperature Sensor
1 Like
Would be interesting to know what sensor chip they use.
Humidity measurements are VERY difficult and should be used merely as a guide. They are highly dependent on temperature and airflow and at least some of the DHT sensors badly drift over time. Certified sensors should have been calibrated but I have my doubts and are expensive.
You are lucky to get within 10%RH of any kind of true value. Though really, you only need to know (for domestic use anyway) whether RH is >60 or <40 for extended periods. Anything over 60% for extended periods in living areas is considered unhealthy.
So is too low humidity, if it is too dry inside it is not healthy either.
When you have some kind of measurment ongoing, for instance to check if the climate is "healthy for your house", like in the ground and non-isolated (cold) attic, it is not enough to just measure the H level. You have to calculate how far you are from the actual dewpoint since wet environment is what you should fear. So it can happen you have a very high H level but still safe from the actual dewpoint. I have this kind of measurement since years. It was when I needed to replace my sensors I found out that the H&T was too far off when I compared with other brands I have had for long time (Oregon)
Sorry, yes I did reference that but didn't mention it since, for most of us, that will never be an issue. Especially here in the North of England. 
<40% long term is also considered unhealthy for living areas.
Yes, indeed, though this is somewhat a parallel issue. High humidity can lead to mold growth but condensation can and will as well.
Again, a major issue in the North of England (and the UK in general actually). Especially in older housing stock like our >100yo Victorian house. We had major recurring mold problems in our living room because of condensation on outer walls. Eventually we sacrificed some space for internal wall insulation to prevent this.
And unfortunately, there is no easy way to get an accurate reading to compare against unless you can lay your hands on a certified measurement device and know how to use it (because the local environment massively impacts readings). Best not to over-think it in my view, get a couple of reliable sensors such as the Bosch ones and compare the readings from 1 location, take an average and apply some correction factors to get them reasonably close. Then avoid the temptation to believe the sensor specs ("accurate to 0.2%RH") and assume you will never get more accurate than +/- 5-10%RH. This is still enough for useful measurements, just output a traffic-light on your dashboard though, not the exact figure.
1 Like
Shelly devices do have MQTT but I found they ones I tried insufficiently documented, buggy, and not supported. On the other had you can OTA flash them to Tasmota.
I started before the big splash about Matter. I expect Matter will end up like the internet browsers which were supposed to follow standards, but can't agree on which standards to follow or what to do with edge cases. It will work if you keep to one eco-system then they will screw with you. Look at the Ring door bell as an example.
I went with MQTT to communicate between devices:
- I believe with a properly configured home network it will be more secure
- Tasmota does a wonderful job supporting it (including security)
- Shelly does the bare minimum, but it does work out of the box.
- You can purchase pre-flashed devices on Amazon, or direct from boutique shops
- It is agnostic to controllers. I use node-red, but others work with it just fine
- It works very well for me with nothing but a Pi3b+
I have avoided zigbee / zwave because of expense, lack of devices, and belief it will fade away before the onslaught of Apple and Microsoft. The perfect example is the X10 ecosystem. Google may even jump in, muddy the waters with one good product, then abandon it like they usually do.
I've never had a problem with any of mine. Quite the opposite, I've only ever used the native firmware MQTT. There were some limitations in the original v1 devices but long since resolved I believe. I use MQTT control every day and have no issues.
I installed one of these Opentherm Gateways last year and they can be Network, WiFi or serial connected. Using an ESP8266 they are very solid.
HTH
My 2cents
I (as in me myself) would love to have all my devices controllable locally with out need for T'internet and someone else's cloud server BUT.....
in the real world, these cloud connected devices work quite reliably and can be easily controlled by my muggle residents using voice to Alexa, or the Alexa app or the device apps.
Whereas my own NR stuff fails daily for many, many reasons - not least of which is that I keep messing with it
3 Likes
Thanks. Unfortunately my boiler doesnāt have OpenTherm anyway. Iām on the waiting list for an EBUSD adapter, although it looks a bit more complicated than that. I think I have to install a daemon that connects to the adapter and it turns the values into MQTT values for which there is an integration in Home Assistant and then youād need to convert the difference between target temperature and current temperature into a power state.
PITA to be honest.
1 Like
MQTT worked for the dual bulb, but only to turn it on and off.
I tried to adjust the brightness and found the documentation unclear. It mentions no security is available for MQTT.
I created what I thought was the correct payload and sent it. Crashed the bulb so hard I had to factory reset.
I reported the bug but they were not interested, and the latest firmware hadn't been updated in more than two years. MQTT was obviously not their priority.
Why would you need additional security for MQTT when everything anyway is in your, as I assume, secure home network?? Also assuming you have a local MQTT broker.
1 Like
"Defense in depth."
No home network is 100% secure.
Wifi security has been broken many times.
TLS has been broken a couple of times.
1 Like
Well, the nature of my job makes me pretty paranoid. But unless you live in a high-risk area or are involved in high-risk work, taking sensible precautions will cover most things.
If you do want higher security for home automation, then you will want to segment your network for starters. Create a new Wi-Fi network for IoT devices with no DHCP and a limited IP address range.
Treat your internal network as unsafe and turn on TLS for MQTT and enforce authenticated clients for everything that supports them. Limit other devices to specific topics.
Turn on TLS encryption for all other inter-LAN comms as far as possible and keep anything that cannot do that on separate network segments. Use an internal router (separate to your Internet router) or firewall to enforce segmentation.
Lots of things you can do to make your home network more secure if needed but overkill for most people in most circumstances - as long as you are protecting the edges of your network. If you really live in a highly dangerous area, you probably shouldn't be using IoT at all to be honest.
Sure, but sensible choices of Wi-Fi AP are pretty good. Good enough for nearly every purpose outside of high-sensitivity areas.
That is mostly a case of keeping your config current. Restricting to TLS 1.2+ and turning off known weak encryption. All easily looked up.
As I say, if you can't trust your home network to a reasonable degree, you can segment and secure it and individual IoT devices with less security can be kept away from other segments of the network. Internal routing/firewall can enforce inter-segment traffic rules. But really, if you are that concerned, perhaps IoT and home automation is the wrong thing to be playing with.
Oh, and I forgot to mention that the Shelly's also have a REST interface which does, I believe have some additional security available?
1 Like
Lots of things you can do to make your home network more secure if needed but overkill for most people in most circumstances - as long as you are protecting the edges of your network. If you really live in a highly dangerous area, you probably shouldn't be using IoT at all to be honest.
I thought the same for a while. I figured why would anyone want to hack my web server, when everything on it was published freely?
I got hacked. It turned out my machine was being used to get into the ISP cisco router it was connected to.
I watched my publicly exposed systems logs and saw intrusion attempts every few seconds. Most for the SQL worm that came out decades before. Nobody told the infected systems it didn't work any longer.
So, as far as I am concerned, everywhere is a dangerous area.
Shelly's also have a REST interface which does, I believe have some additional security available?
If I remember correctly It goes through their cloud service. Which doesn't keep my home operating even if the internet connection is down.
1 Like
Sounds like a typical bot attack of a known vulnerability. These are not personal typically, they are looking for resources to deliver further attacks.
Of course, there are many modes of attack and I wasn't saying you shouldn't ever take precautions, as we've seen with the recent hacks mentioned in the forum, exposure is a risk - always.
What I was talking about is what causes exposure and for most people Wi-Fi isn't a big risk as long as you are following sensible precautions - mostly using latest Wi-Fi security and a complex Wi-Fi passcode.
I don't think so, it is direct to the device. The cloud service is separate I believe.
This NEO Coolcam Temperature and Humidity 3in1 Alarm Siren (NAS-AB02W) is a Wi-Fi USB powered sensor. Iāve got a few of them. The first two, I was able to get Tasmota on but the last two wouldnāt flash for me. So, Iām using the Tuya Web services API for those two. Iām not that happy with the performance but it is the only pre-packaged plugin Wi-Fi temperature sensor that I could find.
I use an rpi3 with mosquitto and node red which gets data from / sends data to several Shelly, Sonoff and home made interfaces based on esp01 and esp12.
Enough to build, program and test:
Shelly and sonoff for most of the lighting.
Outside light measuring, battery powered, esp01, measures daylight every 15 minutes to switch lights on when getting dark.
Three esp01 measure temperature using 18B20 and switch heating on&off for three small greenhouses.
One esp01 grabs data from a p1 port of a smart powermeter, decodes it and sends it to the rpi, showing the used gaz and electricity in graphs.
An esp12 in an old fridge which has also a small heater inside to guard fermenting beer sends data to rpi4 to show graphs.
An ispindel to measure the fermentation, sending data to rpi4 to show graphs.
