I have some very old touch lamps on nightstands that I'm looking to bring into the smart home and have yet to find any solutions that I'm happy with. The hard part is that I'm not willing to sacrifice the touch lamp feature. Was hoping I'm not the only one that's faced this issue and looking for other options.
So far, the best thing I've come up with is a cheap On-Off (non-dimming) touch lamp sensor connected to a Shelly Wave device
It took a bit to get the shelly to work the way I wanted, and it's mostly functional. My main problem is that it's unstable when the touch switch is in the "on" state. The shelly randomly thinks the switch state has changed so the light keeps cycling.
Anyone else come up with something better? I'm not happy with this enough to put it in my own nightstand let alone my Wife's.
What Shelly do you have? Does it have way to flip the relay when it gets 120v from the touch switch? Why didn't you use the lamps existing touch sensor switch?
I have used one of these for a make-it-smart application.
It senses 120v turned on to flip the relay. So you could connect the touch sensor 120v to the S1 terminal to trigger it on/off. The lamp plug wire from the outlet goes to L and N. Then it switches L to the lamp.
See the "High Voltage Lamp Wiring Diagram" that is on the Amazon listing.
The problem is that I'm unwilling to sacrifice the touch-lamp feature. I'm willing to replace it with a different touch mechanism, but I want to be able to roll over in the middle of the night and tap the lamp base to turn it on.
I'm using the Shelly Wave i4 because it's the only Z-Wave device they had in stock at the time. I tried a Shelly 1 but couldn't get reliable feedback from the switch status in Habitat.
As I dig deeper into the issue, I'm starting to believe my problem is in the touch sensor. The triac they use in those become unstable without a minimum load on the output. Since I have the touch switch only connected to the shelly there's no electrical load, causing the unstable state.
Hoping someone has ideas on another way to do the touch sense on the lamp base.
I looked up the Shelly i4. That does not sense 120v applied to it from the touch sensor. It only senses a circuit closing, by a switch or a pushbutton.
My point is the existing touch switch, when you turn it on now, is just supplying 120v to the bulb. Take that 120V that the touch switch supplies when it is on, and connect it to the S1 terminal on the relay I posted. Instead of supplying 120v to the bulb, it supplies 120v to the relay, which senses it, and turns on the lamp.
Unless you have something against Zigbee.... this relay has been working great on my Stove Fan Hood, it senses 120v when I press the touch switch on the fan hood controls to turn the light on and off with the existing touch switch.
No objection with ZigBee at all. I run both Z-Wave and Zigbee devices in my home. It's the WiFi connected devices I try to stay away from where possible.
Just ordered a couple of the Tuya modules you recommended. Fortunately, they are pretty cheap. I'll swap it out with the Shelly device I'm using now and report back if it makes a difference or not.
Ok. Is the touch switch you bought different than the existing switch? I think all you need is the original touch switch, to turn 120v on and off to the relay instead of the bulb.
Might be the Shelly input is too high impedance for the output of the touch switch hence prone to spurious triggering or the touch switch needs a minimum load to be stable.
As an experiment put a low wattage (say 5-10W incandescent) bulb on the Shelly input to neutral. Its only a test to prove the impedance theory. If the circuit becomes stable with this load on the Shelly input you could experiment with fixed resistors to find the largest value that makes the system stable but doesn't draw much power in the resistor. Be sure to compute the V^2/R power dissipation to be sure the resistor isn't overloaded (wattage).
You can also probably play some games in your rule to debounce the Shelly from spurious inputs. Using delays and maybe some sort persistence counter scheme you could create a rule to ignore unwanted triggers or quickly reverse them if detected as rogue triggers.
