Binary sensor status reporting

[Roland declerck 0]

I am trying to connect my HC2 to my alarm installation (jablotron 100) by means of the jablotron module JA118N and the fibaro binairy sensor

based on attached documentation i found on the credex website.

 

It works but it is unrealiable. Status do not get updated all the time.

I first believed that there was an issue with the communication between sensor and HC2. Just to make sure i added a FGS223 in the same room

to take advantage of the mesh network. I did a network reconfiguartion for the sensor.

All seemed to work ok for a short time but then i got the same issue again: when i put the alarm on 

the fibaro sensor status was not updated.  NO reaction on switching jablotron alarm on or off.  In jablotron the PG status is responding correctly.

When i did a "read configuration" for the module it picked up the correct status.

 

I would like to understand what is causing this issue and i do not believe it is a signal weakness issue.

Has somebody an idea what is wrong here (polling?), anything to test .....

 

Please login or register to see this attachment.

Please login or register to see this attachment.

[jakub.jezek 353]

Hi @Roland declerck,

 

Does Jablotron sends impulse or it gets just turned on till alarm ends? I found out, that impulse with duration < 2 seconds in unreliable and Fibaro UBS has problem to detect it.

[Roland declerck 0]

Thanks,

it is a valid point.

But  as far as i can see  it is on/off:      position 2-3 function is on/off

Please login or register to see this attachment.

 

Is there anyway i can for 100% sure exclude the weak signal.
I believe it is not the issue but how can you be 100% sure? 

What happens if de binairy component sends a signal which is not captured by HC2.   

Will the binairy component try  to resend untill succesfull or will it stop after the first failed trial?

 

regards

 

[cag014 632]

It might be an issue with the power consumption of the UBS (Universal Binary Sensor). Please try to use external power supply.

According to jablotron module JA118N specifications "A current of up to 100mA can be drawn from the +U terminals. If the current consumption is higher (overload), the voltage is disconnected. If the current consumption drops, the voltage is automatically reconnected."

UBS total power consumption is unknown,  but the suggestion is to use power supply of 9-12V - 1A. 

May be Fibaro team has the exact information about the current consumption.

 

 

Edited November 29, 2017 by cag014

[Roland declerck 0]

Thanks,

It is a valid observation, whether it is the explanation i don't know.

Not sure what to do to be sure this is or is not the problem?

 

I would say on 6 to 7 times out of 10 the UBS respond correctly to the changes and reports a correct status.
If status is not correct i can sometimes correct it by arm and immediatly disarm the alarm system or doing a reread of the parameter is also solving the problems. 

 If what you explained  is the cause should i see this kind of performance?

If i look to the logbook of jablotron status changes are correclty reported and there is no message of overconsumption or what ever.
 

Fact is that you can not build on the UBS status  with this degree of realiablility.

 

 

 

 

[petergebruers 1,265]

Interesting suggestion by @cag014. From the top of my memory: the UBS needs 60 mA to start (at around 6 V, so that is below spec)), but uses less than 30 mA @ 12 V when it is running. It has a switched mode power supply, so current decreases as voltage increases. I don't power is the issue. This is not a Z-Wave Plus module, so it has less range than your FGS-2x3 switches (very crude estimate: 30% less). I would try to move the UBS and keep the antenna away from (grounded) metal objects. I do not think it supports explorer frames, so if you change something please do a mesh reconfigure on that module...

[Roland declerck 0]

Peter,

The JA118N module and the FBS are mounted in the Jablotron centrale. The cover of the security centale is from plastic.

The FGS 223 is about 1 meter away from de security centrale and about  12 m away from HC2 (with 2 walls in between).

In principle this should work i think. I have never seen that HC2 could not contact the FBS which you would expect to see

from time to time. 

Will try to reposition the fgs223 but . . ., lets try

 

A bit fustrating that it is trial and error.  

 

 

 

 

 

 

[petergebruers 1,265]

15 hours ago, Roland declerck said:

The JA118N module and the FBS are mounted in the Jablotron centrale. The cover of the security centale is from plastic.

OK, but the PCB is copper and grounded...

The UBS might pick up noise from the PCB

 

15 hours ago, Roland declerck said:

The FGS 223 is about 1 meter away from de security centrale and about  12 m away from HC2 (with 2 walls in between)

 

So you expect the UBS to pick this FGS as a router. It does not work that way. The UBS stores a list of neigbours it can talk to. An undisclosed algorithm first tries to send to the controller, then a neighbour. It  does not try "the best" and it does not try indefinitely either. Going by my personal experience with these older devices, it is better to have "no direct connection to the controller" than "a weak connection to the controller". I am talking about modules with the same chip and about the same Z-Wave firmware, like the older FGS.

 

The route chosen by the controller, to the device, is based on different information than the one chosen by the device to the controller. So they might be different. But to make things a bit more interesting... The network is source routed, meaning if A sends to B via C then B acknowledges via C to A. This is not like routing IP packets, it is like routing packets on a "token ring" network (anyone still remembers those?). How a Z-Wave chip uses all this information to determine the best neigbour is unknown to me. I've never seen anything substantial published...

 

At the moment, Fibaro does not have a Z-Wave Plus module with DC input (and better range, more recent firmware). The RGBW and D/W sensor Gen 1 are not Z-Wave Plus.

 

15 hours ago, Roland declerck said:

I have never seen that HC2 could not contact the FBS which you would expect to see

from time to time

 

Because of the differences in firmware and routing information, this is possible. I observed this with some of my older modules and it has been reported by other users too. So controller -> device is OK 99,9% but device -> controller = sometimes not OK is quite possible (though not very common).

 

15 hours ago, Roland declerck said:

Will try to reposition the fgs223 but . . ., lets try

 

You can try, but I'm suggesting you try to move the UBS (you might even try only 10 cm or 30 cm, away from the alarm system), not the FGS... Then mesh reconfigure the UBS, not the FGS.

Edited December 1, 2017 by petergebruers

[Roland declerck 0]

Interesting stuff.
It sound a little bit like black magic.

 

Good suggestion to move the UBS. I have one problem: this is inside the alarm central. If i open this i will get a tamper alarm which i can not switch off.

I will need to talk to the installer of the alarm system to help me out here. may take a bit of time.

 

 

[StefanS 1]

I'm having some kind of same problem with my UBS.

I'm opening my door via RFID normal open relais but the UBS didn't get every status change and keeps hanging on the last status.

I also set the relais of the RFID to 4 seconds but it does not make it functioning 100%

[Roland declerck 0]

This issue drive me nuts. From time to time the UBS reported status reflecting the jablotron alarm status (switched on or off) is incorrect. Some times it is ok for 2 days and then

all of a sudden it is not picking up the changes.  Mostly when i switch the alarm on and off 2 or 3 times in a row it works ok again.

Jablotron PG outputs are always refelecting the correct status. Soo it is a problem between UBS and HC2.

I tried a couple of things without success except my last attempt which looks promishing 
I have set up a polling time for the UBS:

 

I have some hopes that this is the solution to the problem. It is now working ok for 4 days (never achieved this before).

Normally the UBS should not be polled but it seem to make a difference.  

Soo this looks really promishing but time will tell .....

 

I would appreciate some documentation/explanation on how exactly this polling parameter works! What happens?
The HC2 takes contact with the device and get status back?  

Any idea where to find more info?

 

 

 

 

Please login or register to see this attachment.

[petergebruers 1,265]

8 minutes ago, Roland declerck said:

Normally the UBS should not be polled but it seem to make a difference.  

Soo this looks really promishing but time will tell .....

 

I would appreciate some documentation/explanation on how exactly this polling parameter works! What happens?
The HC2 takes contact with the device and get status back?

 

Yes, the HC ask the device to send status...

 

I can recommend this old topic started by me:

 

[Roland declerck 0]

Peter, great will read tommorrow.