Double push after "All Off"

[Warichet]

If I get you right, that's what I've done so far.
As mentionned, maybe it doesn't work in my particular case.

Should I activate the write flag (E) on the actuator's status KO as well ?

[Warichet]

Hi,

I've been implenting the trick to put the actuator's status GA behind the actuator's switch object.
All of a sudden, the HomeServer goes berzek. Also, I can't switch off certains lamps anymore. HS sequenzes work with 5 min delay. Crazy thing.
By looking a the group monitor, I see a flood of messages. I've caused a network storm on the bus. There are 2 GA talking all the time, it's related to the same actuator, Berker 8 Fach 75318004.
By removing the status GA, bus activity is normal again, ouf.

Bizarre ! Any comment ?

[SebastianFey]

seems like you did build a loop in your HS-logic?

again: don't touch the actuator its working fine, the problem is the switch.
Here is a little drawing

under I) you see the normal solution which you cant use:
- switch Transmits on the on/off-GO
- actuator receives on its on/off-GO
- and Writes to the internal state, which has the effect that the relais is switched.
- status info is processed similar but in other direction

at II) the only thing you got to do is also set the W flag on the switchs on/off GO and make sure that GA2 (status) is behind GA1 (NOT sending).

but this solution does only work if the application program cares about writes to the on/off GO, this may not be the case.

I)

(T)switch--on/off|>----GA1---->|on/off--actuator(W)
                   
(W)switch--status|<----GA2----<|status--actuator(T)

II)

(T,W)switch--on/off|>----GA1---->|on/off--actuator(W)
                   |<----GA2----<|status--actuator(T)

[Warichet]

seems like you did build a loop in your HS-logic?

Don't think so, the only thing I've modified is adding the status GA behind the switch GA of a Berker actuator.
By removing the association, the bus went back to normal activity. Didn't touch the HS.

the problem is the switch.
Here is a little drawing

Thank you for the drawing, I think I got it now (better late then never )

at II) the only thing you got to do is also set the W flag on the switchs on/off GO and make sure that GA2 (status) is behind GA1 (NOT sending).

Here is my mistake, I was putting the GA2 (status) behind GA1 .... on the actuator (instead of behind GA1 on the switch. I've been told to do it that way to avoid the double push.

If you don't mind, I've still questions:
- the trick you explained above, is that valid only in the particular case of "split" switch (I mean 2 separated toggle PB instead of 1 pair of switch PB) ?
- does it hold true for a standard binary input, i.e. US/U ?

The whole problematic started when I've implemented the "All-Off" function and wanted to avoid the double push after an "All-Off" and still have the PB's LED reflect the actual status.
Is this the same problematic as above ? or am I mixing things ?

Thank you for your time and patience.

[SebastianFey]

Don't think so, the only thing I've modified is adding the status GA behind the switch GA of a Berker actuator.
By removing the association, the bus went back to normal activity. Didn't touch the HS.

Problem could be in the HS anyway, could you provide a log fram the groupmonitor?

- the trick you explained above, is that valid only in the particular case of "split" switch (I mean 2 separated toggle PB instead of 1 pair of switch PB) ?

no

- does it hold true for a standard binary input, i.e. US/U ?

iirc the US/U provides an explicit status object for each physical input. Use th status object if possible

The whole problematic started when I've implemented the "All-Off" function and wanted to avoid the double push after an "All-Off" and still have the PB's LED reflect the actual status.
Is this the same problematic as above ? or am I mixing things ?

the root of the problem is the same: if the PB does not know that status of an actuator channel, a toggle cant work and also the LED can be in the wrong state.

[Warichet]

Hello,

I came across an actuator (Merten 6478 29) which has no status KO.
Is it possible to "fake" a status KO or generate it in any way, in order to use it with the above procedure ?

I guess I could do it at the HS level, by reading the actual status of the actuator, but at the ETS level ?

[SebastianFey]

I guess I could do it at the HS level, by reading the actual status of the actuator, but at the ETS level ?

Yes, the HS has a built-in mechanism for this. You can't do this at the ETS-Level (without extra hardware). Best thing would be to get rid of the actuator, it'll cause more trouble than a switch actuator is worth today.

[Warichet]

@Sebastian

I've streamlined my installation and it works as expected.
I wanted to thank you once more for your kind help.

I've still another aspect of the same issue pending, which I didn't dare to tackle (yet). Still not convinced if the benefits are worth the troubles.