I’m already doing a similar route to what you imagined, so this is super helpful to run across!
The problem I run into is that the depth of the box my deadbolt strikeplate is attached to is exactly the right depth for the bolt itself. I may be able to glue a magnet to the end, but no means of putting the reed switch at the end of it, nor spring contacts given that the box itself is metal.
If you can tinker/solder, adding a remote wired Reed to any device is usually very easy. Just solder wires parallel to the stock Reed, and you are good to go. Ecolink has wire screws to make it even easier, but by no means is this trick limited to ecolink devices. A reed switch is a reed switch, and you dont even have to desolder the original as long as you keep it away from magnets.
On a side note… if you glue a magnet to the tip of the deadbolt, are you also drilling it a little to give clearance? I built my doors with pretty tight gap, and wouldn’t be able to do that without putting a blind hole in the bolt (otherwise the strike plate would knock the magnet off). I’m curious if putting it in a hole would reduce the magnet’s effective strength. I had similar issues when adding a reed to my metal breaker panel… the metal case made it where the reed had to be much closer to the magnet than normal for it to work (paramagnetic effects).
Interesting topic. I did the lock detection a little differently. I used two spring contacts screwed to the back of the cavity the dead bolt goes into. When locked, the deadbolt will connect them, closing the circuit. This is used like a normal binary switch. It’s a EU type dead bolt, which has different dimensions from the typical US ones, but the same principle applies. I’ve used this setup for a little over a year now and it works perfectly.
For the actual door open / close sensor I used the Aeotec Door / Window sensor 7. It’s very small and has been working flawlessly so far. I use the door and lock states extensively in my automations and also in my security system. I use exclusively ZWave devices for anything remotely security related. I don’t want to put the safety of my house (that’s the whole point behind a security system after all) in the hands of some cheapo $3 Chinese sensor that might or might not be reliable. Cheap 433 MHz stuff is fine for non-critical systems like lighting. But when it comes to things like doors or locks (or even climate control), it’s ZWave / Zigbee with certified and CE/FCC compliant devices all the way.
Nice post, I’m sure it will encourage others with mortice type locksets to give it a go.
Regarding security of wireless devices, you may want to take zigbee off your list. Zigbee was shown to be vulnerable to remote exploits not long ago (~6mo or so IIRC, when the first article was published). The exploit can be done “down the street”, and can give the hacker full access to the rest of your HA device (by extension, possibly your network as well). Not sure if other hubs have patched this vulnerability. I don’t think HA has patched it yet; in fact I bet HA will be one of the last ones to get patched for that. Regardless, this is a big one for those using zigbee with alarm systems… sadly it was a deal breaker that forced me to replace 3 zigbee door/window sensors. I still use those on ‘non-critical’ stuff, but yeah never in places where security is a thing.
So really, that leaves us with just zwave or wifi for security type stuff (IMHO). Wifi of course won’t be useful for high traffic battery powered sensors for the most part. However if it plugs in or only sees occasional traffic, I usually prefer wifi over zwave (mainly for the price, an esphome device and do tons for much cheaper than zwave).
Wasn’t aware of a Zigbee exploit, but I’m not too surprised. Just shows how IoT security needs to be taken seriously by manufacturers as people will be actively looking for exploits. Just as for any other IT system.
Now about home security. IMO, if someone uses these kind of exploits on your IoT infrastructure, which not only implies that they have the technical know how to pull this off, but also that they know a lot more about your IoT systems than they should, then this is not a typical crime of opportunity. It’s a well planned and targeted attack specifically against you. At that stage we’re waaay past the DIY home security area and step right into professional perimeter defense and patrolling armed guards type of territory 
I would still not trust those cheap 433MHz things for security related devices, also because they’re often really unreliable. I’m also a bit weary of both Zigbee and Wifi. Wireless security cameras have become really common lately, so it wouldn’t be too far fetched for even opportunistic burglars to carry a 2.4GHz jammer. ZWave is much more obscure. The low battery usage and the better signal penetration in the lower frequency band are both a big plus, I’m pretty happy with them so far. High price is an issue though, I agree.
As always, if you can, keep everything wired.
You can find all kind of reliability inside the 433mhz devices range as for Zigbee or Zwave.
You can have very robust and well made sensors and other weaks and bad power designed. But the main difference is that you will find most of them for less than 10 euros.
I figure wifi is secure enough with wpa2 and a strong password. Lots of folks do secured work on wifi… definitely includes folks that have the armed patrols etc. In fact I would tend to trust wifi more than zwave kist on the fact that many more people rely on it being secured, and there is a lot of interest in keeping it that way (more so than zwave anyways).
I don’t think I need any more security than the average middle class suburban home in my area. Problem is our homes are the ones the pros like to target… the kind that typically clean out the house without incident, appliances, jewelry, and all… well organized and trained criminals. I think any diligent pro robber has zigbee ‘wardriving’ down to a system, or are learning how to do it.
Oh yeah, it is. But this is not about the inherent data security, but about integrity of the physical link. If your battery powered Wifi sensors are in deep sleep (off the network) until they get triggered, then someone just has to spam the Wifi channel they use (or all three center frequency channels) and all your sensors will be rendered blind, as they cannot connect to the network anymore when they wake up. And worst of all, you cannot even detect that something is wrong, as this would require a persistent connection of the sensor to the network. If a burglar knows where a door sensor or PIR is located, then that jamming can be highly localized around the sensor. Even a phone wifi is strong enough to do this if relatively close to the sensor. The only way to mitigate these kind of attacks is to use wired everywhere.
Of course all this is true for all wireless devices, including ZWave. It’s just that the 2.4GHz band is so much more common.
In the end I think that anyone who thinks about building a DIY security system needs to assess his own personal threat situation and plan accordingly (wired vs wireless, cost-driven DIY vs professional, and so on).
The main difference is that Zigbee / ZWave / Wifi reports back their state and will retry until the message is delivered. 433Mhz devices will not, there is no way to know if the signal made it through or not. They’re unreliable by design.
So your point here is more the reliability of the communication and the fact that 433mhz door sensors are most of the time mono directional if I get it, and not the sensor reliability itself?
Both, actually. The unidirectional protocol used by most of these devices is the main culprit, but some of the really cheap 433MHz stuff you can buy on Aliexpress or similar is just incredibly bad quality. And if you combine a low quality transmitter (out of tune center frequency due to cheap components / carelessness, bad antenna, …) with a ‘fire and forget’ protocol without error correction, then things can become messy. They’re also highly prone to interference. If two modules send data at the same time, both will be garbled. That doesn’t happen with better bidirectional protocols.
Now don’t get me wrong, I use a lot of 433MHz modules (mostly ones from Chacon / DIO) for non-critical functions like lighting and I like them a lot for that usage. They’re cost effective and have great range. But you can never really trust them.
I have to agree with the idea that wireless at the end of the day can be fairly easily ‘swamped’, which could be an effective way of muting a wireless sensor. I mean, I’m sure we could design a directional antennae setup and some shielding to reduce that attack surface, but won’t be as good as wires ever. That said, when we continue this discussion we will always arrive at the same universal truth… if it was made by man, it can be broken by man… even wired alarm systems, bullet proof windows, uncuttable locks, unbreakable encryption… nothing is 100% guaranteed… except death and taxes.
Yes, external reviews need to be consulted before buying cheap stuff, especialy when buying tens of them. But you can find also cheap ones with quality design and build.
Well, that’s true, 433mhz sensors try to compensate that by sending a burst of messages.
For my personal knowledge do you have 100% of the events caught with your Zwave devices?
But at final it is also a question of value added versus money spend, and the owner requirement level in terms of sensor availability and communication reliability.
As far as I know, yes, as long as they have adequate signal strength in the mesh network. ZWave has a transport control layer, so the transmitter will know whether or not his message was correctly delivered, and may resend as needed. The packets are also encrypted and so there’s implicit crc checks, the receiver knows if the received data is okay (and he can ask the transmitter to resend if it’s not). Of course all that comes with added complexity (all nodes need bidirectional transceivers) and higher prices.
Absolutely.
For quality and usability, go for Aqara sensor. Has been using Aqara sensors for more than 2 years now. Recently I skipped the Xiaomi Hub/Gateway setup and use zigbee2mqtt. The result is amazing!
@seikan I’ve got a bunch of Aqara door / window sensors and they have been great. Since we switch off our wifi during the night, I ditched the Aqara Gateway device, and replaced it with a Conbee 2 (via Hass.io addon). It’s been working really well and since the Conbee2 USB stick is plugged into my HA machine (Dell laptop with Ubuntu), all my sensors are available regardless of not having wifi during the night (thanks to Zigbee).
I know this post is almost 1 year old but I hope the author can tell me how (s)he get the Ecolink Door Sensors to work with HA.
I have no problem including the Ecolink sensors (to Aeotec Zwave Stick 5) but that is as far as I can get. None of the entities (contact, tamper, burglar) report anything with the exception of the battery. The initial state of the contact entity is “Off”. If I remove/replace the battery this may switch to “On”. I have then taken this same device and added it to my Smartthing hub and there it works perfectly. This is fine for now but my goal is to eliminate ST and use only HA.
Any help with this would be greatly appreciated.
Currently running Home Assistant 0.116.2 but this problem started way before this.
SOLVED: Dont add this device via the Z-wave stick include option. Instead add it via the HA interface Menu option Configuration/Integration/Zwave/Configure Put the Ecolink device into include mode and select Add Node Secure Once included this device works properly.
Maybe this topic is too old, i I bought myself some DIGOO window sensors and I don’t get how to distinguish between the tamper button and the reed contact:
Magnet ON:
{“Time”:“2021-02-10T16:40:15”,“RfReceived”:{“Sync”:12930,“Low”:440,“High”:1260,“Data”:“1B7849”,“RfKey”:“None”}}
Magnet OFF:
{“Time”:“2021-02-10T16:44:00”,“RfReceived”:{“Sync”:12940,“Low”:440,“High”:1260,“Data”:“1B7843”,“RfKey”:“None”}}
Tamper OPEN:
{“Time”:“2021-02-10T16:44:00”,“RfReceived”:{“Sync”:12950,“Low”:440,“High”:1260,“Data”:“1B7843”,“RfKey”:“None”}}
Low battery: (powered with 2.2 V instead of battery)
{“Time”:“2021-02-10T16:51:12”,“RfReceived”:{“Sync”:12930,“Low”:440,“High”:1270,“Data”:“1B784D”,“RfKey”:“None”}}
The 'data ’ for Magnet OFF (aka door open) is the same as the data for Tamper Open.
Is that ‘sync’ field useful for something? Can I use it to distinguish between the 2 states ? i powered the sensor a few times and the sync number changed every time, It seemed to stay consistent within one power cycle. I don’t really know anything about the protocol that’s why I hope to find someone to answer my question
I think the so-called tamper function is not really useful as it means the same as 'the door has opened ’
I found the low battery warning useful though.
Has anyone had this issue?
i thought to share also the current when powered at 3V:
Idle : 0.27 uA
Transmitting: 200 mA
forum won’t let me post more than 1 image
I confirm this, see this topic :
You should not, it can vary from one signal to another. Just use data field.
