SONOFF ZBMINIL2 as no-neutral “push button” for a 230 V staircase timer trigger

Hardware
, ,
1

I am trying to add Home Assistant control to an existing staircase lighting timer without modifying the existing push buttons.

The existing installation is a conventional 230 V staircase timer setup. The timer is a CRELI 01.17 06 DIN-rail staircase timer (0.5–10 min).

Each wall push button has only two wires :

  • permanent Live
  • trigger wire going back to the staircase timer

Pressing the button simply connects Live to the trigger input momentarily.

The push buttons are also illuminated when not pressed. The small indicator lamp is effectively connected across the push-button contacts, so when the button is open a small current flows:

Live → indicator lamp → timer trigger

When the button is pressed, the lamp is bypassed, the trigger receives Live directly, and the indicator goes out.

I would like to install a SONOFF ZBMINIL2 behind one of these push buttons, in parallel with the existing mechanical button, approximately like this:

Live → ZBMINIL2 L-In
ZBMINIL2 L-Out → staircase timer trigger

The existing physical push button would remain connected directly between Live and the trigger as it is now.

In Home Assistant I would only turn the ZBMINIL2 ON momentarily, e.g. for 0.5–1 second, and then turn it OFF again. In other words, I am not trying to control the staircase lights as a load; I only want the ZBMINIL2 to simulate pressing the existing button.

Why I think this might work

The ZBMINIL2 is a no-neutral device, so while OFF it powers itself by allowing a very small current to pass through the load. I have seen its OFF-state current quoted at approximately 1 mA .

That seems interesting in this application because the staircase timer is already designed to tolerate leakage current through illuminated push buttons .

For comparison, common staircase timers explicitly support this. For example:

  • Finder 14.01 allows up to 30 illuminated push buttons at approximately ≤1 mA each.
  • Other staircase timers specify glow-lamp loads in the tens of milliamps.

So, from the staircase timer’s point of view, an additional ~1 mA flowing into the trigger while the ZBMINIL2 is OFF seems potentially similar to adding one more illuminated push button.

Also, my existing CRELI timer already works correctly with illuminated push buttons, so its trigger input clearly tolerates some continuous current without interpreting it as a button press.

What makes me unsure

The biggest concern is that this is not the application the ZBMINIL2 is specified for .

SONOFF specifies the ZBMINIL2 for normal AC loads and gives a minimum load of around 3 W . The specifications/examples refer mainly to lighting/resistive loads.

A staircase timer trigger input is obviously not a normal resistive load.

So although the ~1 mA OFF leakage may be perfectly acceptable to the staircase timer, I don’t know whether the timer’s trigger input provides a suitable current path for the ZBMINIL2 itself to remain powered and operate reliably .

Possible issues I can imagine are:

  • ZBMINIL2 not receiving enough current to stay powered while OFF
  • resets or Zigbee disconnects
  • failure to switch properly when ON
  • unexpected/false staircase timer triggering
  • the timer input having an impedance/electronic circuit that does not behave sufficiently like the kind of load the ZBMINIL2 expects

There is also the more general issue that the ZBMINIL2’s published ratings are for controlling AC loads rather than using its output as a control signal for another electronic device.

What I would like to know

Has anyone used a ZBMINIL2, or another no-neutral smart relay using the same leakage-current principle, across a 230 V staircase timer / impulse relay / control input in this way?

More specifically:

  1. Is my reasoning about the OFF-state leakage current sound?
  2. Is the timer trigger likely to provide enough of a current path for the ZBMINIL2 to remain powered reliably?
  3. Is there anything about the ZBMINIL2’s output stage that makes switching a 230 V electronic trigger input fundamentally different from switching a small resistive load?
  4. Has anyone measured the actual current drawn by a typical 230 V staircase timer trigger input?
  5. Even if it works experimentally, is there a technical reason why this should not be considered a reliable permanent installation?

I know that the conventional solution would be a powered relay with a dry contact , e.g. at the electrical panel where neutral is available. I am specifically interested in whether the ZBMINIL2 approach is technically viable because there is no neutral behind the wall push button , and the existing two-wire illuminated-button circuit already seems conceptually similar to the leakage-current mechanism used by a no-neutral smart switch.

2

Might work, might not, depends on the relay circuit. ZBMINI itself doesn’t care from where the supply current comes from.

But why wouldn’t you place normal smart relay next to the staircase relay where you do have neutral? Or replace the original with smart one if preferred.

3

Hi :waving_hand:

To better be able to help you, I think it may be worth stepping back from the proposed solution for a moment, because the best way of doing this probably depends on the actual objective.

What is the end result you are trying to achieve by adding Home Assistant control?

I don’t mean “make the ZBMINIL2 simulate another push button”. You’ve already explained that part clearly. I mean: why do you want the equivalent of another push button in the first place?

For example, is the objective to:

  • operate the staircase lights manually from an HA dashboard or phone;
  • add another wireless physical button somewhere;
  • trigger the lights from a motion sensor, door sensor or another HA automation;
  • have HA take over or extend the timing/control logic;
  • or something else?

The best solution could be quite different depending on that answer.

Karosm’s suggestion may actually be the simplest answer if the requirement is simply for HA to initiate the existing staircase timer. At the timer you should already have Live, Neutral and the trigger wiring available, so a suitable smart relay there could provide the trigger without needing to make a no-neutral device power itself through an electronic trigger input.

On the other hand, if there is a particular reason the smart device needs to be behind this specific wall push button, then the ZBMINIL2/no-neutral question becomes much more relevant.

There is also one potential weakness in the proposed ZBMINIL2 approach which may be worth considering separately.

You say you would turn the ZBMINIL2 ON in Home Assistant, wait 0.5–1 second and then turn it OFF again to simulate a momentary button press.

As far as I can establish, the ZBMINIL2 doesn’t have a local inching/auto-off function. So the “release” of the virtual push button depends on a separate OFF command actually being issued and received. What happens if the ON succeeds but the OFF doesn’t?

For example, HA could restart between the two commands, the OFF Zigbee command could be lost, or the ZBMINIL2 could be controlled by a directly bound Zigbee device while HA isn’t involved at all. In those circumstances the ZBMINIL2 could simply remain ON.
I don’t know how the CRELI behaves if its trigger input is held continuously at Live rather than pulsed, but it might be worth checking that before relying on HA to generate the momentary pulse.

If the aim really is to emulate a push button, another option might be to use a device that can enforce the pulse locally, rather than relying on separate ON and OFF commands from HA. Something with a proper inching or auto-off function could do that. The obvious catch is that devices such as the Sonoff ZBMINIR2 need a neutral, which may bring you back to Karosm’s suggestion of putting the smart relay at the staircase timer instead.

So before getting too deep into the ~1 mA leakage-current question, a bit more background on what you’re ultimately trying to achieve would probably help point towards the most suitable solution.

4

If so, I wouldn’t even consider to use it here or in any application where “momentary” output is required.