Smart curtain switches: no manual control / troubleshooting odd behavior

[hheimbuerger (Henrik Heimbuerger)]

I have some apartment window shutter blinds, that I would like to make smart (eventually get into HA, via cloud would be fine for me for now). Instead of installing something (like a Shelly) behind the existing mechanical wall switches, my idea was to just replace them, as they anyway aren’t great and I wouldn’t mind replacing them with something slightly more modern looking.

The shutters originally came with very simple mechanical switches (by Gira), which simply passed current from L on to L1 or L2, depending on which of the old switches (UP/DOWN) was active. I don’t know anything about the motor that is installed.

I found a cheap “Smart Curtain Switch with percentage display”, it looks like this:

So I installed the device, and the manual shutter control immediately does not work.
The device powers up, and I can connect to my WiFi and the Smart Life cloud/app/ecosystem without issues in a few minutes.
But pressing UP/DOWN, either on the touch glass, or in the app, gives two audible relay switches (<50ms apart) and then it shows the roller as fully open or closed.

It sounds like it immediately determines that the shutter has reached its final position.
That seems like it would just not be compatible with the shutters I have installed here. If it wasn’t for… the calibration mode!

Because when I open the calibration mode in Smart Life, it makes the curtain move up and down for a couple of seconds: in some kind of odd sequence, maybe: “30cm down, 20cm up, 30cm down, 20cm up, 30cm down, stop” (I did not do a precise write-up, and had to uninstall the device for now)
During this sequence, I believe the relay can only be heard switching once at every step, which would make sense.

So it is able to affect the shutter in theory! Just at no point did I manage to make the manual UP/DOWN buttons have any effect on the shutters.

Any ideas? Are there different shutter technologies, different ‘protocols’? (I don’t see how, honestly, given how simple the existing switches are.) What’s different between the manual mode and the calibration mode, that could make the former fail and the latter work?

Has anybody experienced something similar? I mean clearly the device is just cheap and crappy, but it also just needs to switch two different relays, so the operation is also quite simple. If I order another similar device, should I expect different results?

The (mis)behavior is so oddly specific, that I feel like the root cause should be discernable from the observations.

[Karosm]

@hheimbuerger

which simply passed current from L on to L1 or L2, depending on which of the old switches (UP/DOWN) was active.

Did you measure that? Maybe it was just a trigger, maybe it was switching neutral
Also, you forgot to post the wiring scheme

[hheimbuerger (Henrik Heimbuerger)]

No, I indeed haven’t measured it. The wiring seemed so simple and straightforward (circuit closed = shutter moving, circuit not closed = shutter stops) that I figured I’ll just give it a try.
But of course, the devil might be in the details…

The wiring scheme of the original non-smart switches (working right now) looks like this:

(“AUF” and “AB” = “UP” and “DOWN”)

And here’s the wiring diagram of the new controller (the one that only affects the shutter during the “calibration” run, but the manual buttons seem to immediately get the feedback that the final position is reached):

[Karosm]

I agree, wiring looks simple and correct.
Try just with one Load output connected. Maybe it’s not interlocking correctly.

[hheimbuerger (Henrik Heimbuerger)]

That’s a smart idea, thanks!

[hheimbuerger (Henrik Heimbuerger)]

Note: I’ll leave this here, in case someone else stumbles upon it and may find it useful (although I cannot imagine how…). I’ve pretty much given up on this device. I might try another, or give up automating this hardware altogether.

---

I’ve run the test Karosm suggested, and it behaves basically just like before:

Let’s say I connect the “up” wire, but not the “down” wire.

Now pressing the up or down buttons on the device itself (or via the app, just the same), I hear the relay clicking twice. No difference whether I press the connected or the disconnected function. The curtain is not moving in manual control under any circumstances.

Running the “calibration” function in the Smart Life app, it once again begins calibrating when I’m still on the “Prepare calibration” page of the wizard, disregarding the fact that the instructions on that page claim that I am supposed to “click Next to start calibration”.
During calibration, the motor does move! Just like before. It is doing the same inexplicable two-steps-forward-one-step-back dance as described earlier. Just that now, it skips the direction of the wire I disconnected.

For example, having connected only the “down” wire, I get the following sequence: relay click, curtain goes down for 1s, relay click, curtain stops, relay click, 1s pause (here it would move up if the wire was connected), relay click, 1s pause, relay click, curtain goes down for 1s, relay click, curtain stops, relay click, 1s pause (here it would move up if the wire was connected), relay click, 1s pause, relay click, curtain goes down for 1s, relay click.

That’s all on the “Prepare calibration” page of the assistant. When I click on “Next Step”, I get to the “In the calibration…” page and nothing happens. Then when I click “Next Step” again, I usually get “Calibration succeeded”. (Once I get “Calibration failed”, but I think that was because I kept it open forever on the first page while writing this post. Probably just an internal timeout.)

Sending commands via HA or via the “Device Debugging” page of the Tuya IoT Platform has the same effect: relay clicks twice, the curtain doesn’t ever move.

Lastly, I’ve tried setting up LocalTuya, but this one device, I cannot manage to get autodiscovered by LocalTuya and I didn’t manage to retrieve the localKey for manual entry either.