Hacker News Reader

Comment by jackweirdy | Hacker News Reader

Retrofitting new wires to every light fixture and switch in a house is unenviable. In some cases, impossible. Laminate floors can deprive you of subfloor access of a whole floor, running cables through existing stud walls is difficult at the best of times...

Wirelessly linked smoke alarms are becoming extremely common now in the UK because outside a narrow envelope building regulations want linked alarms (linked either wired or wirelessly) on every floor. Imagine converting a basement and attic of a 1800s victorian house, and having to find a way to run a new cable from the roof of the attic to the ceiling of the basement. So instead it's a relatively dumb chip which shouts 'fire' on an unlicensed frequency block and a great deal of complexity is saved.

There is a chance you could do here what powerline adapters do, and use the ground wire to communicate. But they are notoriously inconsistent and there's any number of reasons it wouldn't work

View on HN

Those are all reasonable comments, I have spent the last two years doing up my own house, but that is NOT what robdobson.com did.

Note: "Since Lorenzo (the sparky) was replacing most of the wiring in the house I managed to get him to bring a lot of the control cables together in around five separate locations."

Granted, not ALL the wiring, but enough that there are "at least three (control) units lost in the walls" that don't control anything.

The lack of access in a house to critical infrastructure is why that critical infrastructure has simplicity and reliability as its primary characteristic, and if you are going to experiment, maybe consider redundancy? The writer appears to have a lot of hardware expertise, and almost none about systems. Single point of failure? Check. Closed hardware systems that might be unsupported sometime soon? Check. No diagnostics? Check.

This was a fascinating project and I commend the writer for posting all the hairy details, but I am slightly appalled at the wastage.

> There is a chance you could do here what powerline adapters do, and use the ground wire to communicate. But they are notoriously inconsistent and there's any number of reasons it wouldn't work

This is exactly how Insteon works in conjunction with a wireless mesh. I'm extremely happy with my Insteon setup -- it's just a matter of replacing light switches and outlets with a bit of re-wiring on 3-wire switches (no new wires to run, just bypassing one of the switches so only one controls the light). The Insteon-branded hub is fine for a few light switches, but if you're doing a whole house and need something that can handle more complex configuration I would go for an ISY994 controller (don't know if anything newer exists but it's been rock solid for the last 5 years).

All that said, it was expensive and probably not worth it. I spent $4k on a somewhat large house and I was able to justify it as a learning experience for in-home IoT device use cases -- but I'm not sure there's much benefit to doing a whole house with it over just a couple rooms.

> Wirelessly linked smoke alarms are becoming extremely common now in the UK

Also becoming a baseline feature in detectors in the US, in my experience. New houses have been using wired interconnect for a long time now, but in the last couple years I've helped several older relatives upgrade their smoke detectors and even the cheap ones have wireless interconnect now.

And as I mentioned in another comment, many EU houses have concrete (internal) walls. Tubes for electricity wires come pre-installed in them so together with the other things mentioned, just running some wires can be hard to say the least.

That’s why ZigBee in mesh/star topology exists. No wires except for the hub.

Reply on news.ycombinator.com