[gneandr]

For a larger location also with different floors installing a system with the need for around 10 switches to control lightning etc the use of RF-sender connected to a Raspberry seems to be a problem to trigger all those switches (rf power level). A configuration with multiple RPI's (pilight ad-hoc network feature) isn't an alternative here .

It seems the handheld remote controls (like ELRO) are sending with more power so it will be more secure to control the switches. (Note: the Elro remote controls work with a 12V battery!).

With that in mind a connection of the Raspberry GPIOs to the ELRO remote control is planned. It would be a straight forward connection of 4 +2 lines which could simulate to press the 2* 4 buttons. Another 2 (or 3) lines could be used to modify the 'house code' which would increase the 'simulated' key pressing to 16 (or 28). Pretty enough
(See attached pic to show the modification of an ELRO remote control).

Now for simulating a key press always two lines need to be set to GND for a second or so (like sending the rf signal with pulse codes). If the 'house code' is used one or two additional lines need to be switched. So a max of 8 or 9 switchable lines.

To get it working with the pilight concept/software system a different 'protocol' is needed. To operate an ELRO-switch the pilight would need to pull specific GPIO lines and for that a protocol could look like this:

Code:

{
        "name": "Livingroom",
        "lampe1": {
            "name": "Lampka",
            "protocol": [ "elrocontrol" ],
            "id": [{
                "systemcode": 17,
                "unitcode": 2
            }],
            "state": "off"
        }
    }

So, the JSON isn't different. An activation of a switch would map the systemcode to the "DIP" lines/GPIO pins, the unitycode *\ relates to the button row which connects to the "ANx/ANy" lines/GPIO, the 'state' would be mapped to another two lines/GPIO.
*\ valid values would be 1, 2, 4, 8
(See attached pic to show the modification of an ELRO remote control).


Would it be possible to implement such a concept? What support/details are required to get it done?

[curlymo]

1. You can make a protocol for this, however, i'm not going to.
2. Have you considered using the pilight ad-hoc network feature.
3. Have you considered running the pi from a 12v source. So the sender runs on 12v and the RPi on 5v.

[gneandr]

1. You can make a protocol for this, however, i'm not going to.

.. Think it will be difficult to do it for myself; I'm not familiar with C programming. Anyone else interested to join this?

2. Have you considered using the pilight ad-hoc network feature.

.. yes, as said with the op already. Not an alternative right now.

3. Have you considered running the pi from a 12v source. So the sender runs on 12v and the RPi on 5v.

.. waiting for delivery of rf devices, scheduled for only begin of April

[curlymo]

For someone who can write PHP it shouldn't be that difficult to adapt the generic_switch protocol.

[koffie]

I'm getting the feeling this is a theoretical situation?
I mean : have you really tested that a RPi can't cover the complete area?
I think you should look for other (better) options, instead of trying to simulate keypresses on a physcial remote.

Can you explain why it isn't possible to stash a small Raspberry Pi somwehre with a wifi connection?
How about a RF repeater somewhere on the other floor ?

[gneandr]

@koffie
It's not about the WLAN RF of the RPi, it's about the rf power of the 433MHz sender.
And yes, I could install a second/third RPI with more 433MHz senders, but that increases the price and the complexity.
curlymo's suggestion to use a higher voltage for the sender could be good alternative.
Nevertheless I'll try to evaluate the "simulate keypresses on a physcial remote" .. that way also includes the ability to manually switch without the RPi .. a fallback in case.

[curlymo]

But as koffie asked. Have you already tried it?

[gneandr]

if you refer to

I mean : have you really tested that a RPi can't cover the complete area?

Yes.

[curlymo]

I'm wondering what the 12v results are. Have you also tried adding an antenna to the sender?

[gneandr]

The RPi with the 433MHz sender has a normal 17.3 cm copper wire antenna as described in the forum and on other places.
The remote control hasn't (see pic posted before).