Philio PAN04 multichannel problem

Tags: #<Tag:0x00007f326a8b6c00> #<Tag:0x00007f326a8b6b10> #<Tag:0x00007f326a8b6930>

[email protected],

my z-wave environment with 9 nodes and many other non-zwave things is working very well. The last node I’m implemented - the philio PAN04 dual relay switch - didn’t work well. The inclusion process was working fine and in ozwcp and HA everything shows up with right values (3 values of every sensor/ switch - 2 for any switch and 1 with overall values). After a time - I think after the next sync/ update - the values are not ok anymore. One of the 2 switch-sensores was overwritten with the overall values and the switches (on/off) seems to be turning around, too.

Searching around the web and googling through z-wave world shows, that it seems to be a problem with the multichannel endpoints.

Is here anybody who runs in the same situation or have a working philip PAN04 implementation?

thanks a lot,
mpoi

UPDATE:
the log of the update/sync:

2017-04-05 12:11:17.611 Detail, Node009,   Received: 0x01, 0x14, 0x00, 0x04, 0x00, 0x09, 0x0e, 0x32, 0x02, 0x21, 0x44, 0x00, 0x00, 0x05, 0x52, 0x0c, 0x22, 0x00, 0x00, 0x05, 0x52, 0x93
2017-04-05 12:11:17.612 Detail, Node009, Refreshed Value: old value=false, new value=false, type=bool
2017-04-05 12:11:17.613 Detail, Node009, Changes to this value are not verified
2017-04-05 12:11:17.613 Info, Node009, Received Meter report from node 9: Energy=13.62kWh
2017-04-05 12:11:17.614 Detail, Node009, Refreshed Value: old value=7.23, new value=13.62, type=decimal
2017-04-05 12:11:17.614 Detail, Node009, Changes to this value are not verified
2017-04-05 12:11:17.614 Info, Node009,     Previous value was 13.62kWh, received 3106 seconds ago.
2017-04-05 12:11:17.615 Detail, Node009, Refreshed Value: old value=7.23, new value=13.62, type=decimal
2017-04-05 12:11:17.615 Detail, Node009, Changes to this value are not verified
2017-04-05 12:11:17.615 Detail, Node009, Refreshed Value: old value=494, new value=3106, type=int
2017-04-05 12:11:17.616 Detail, Node009, Changes to this value are not verified
2017-04-05 12:11:17.617 Detail, Node009, Notification: ValueChanged
2017-04-05 12:11:17.631 Detail, Node009, Notification: ValueChanged
2017-04-05 12:11:17.647 Detail, Node009, Notification: ValueChanged
2017-04-05 12:11:17.662 Detail, Node009, Notification: ValueChanged

If you are comfortable editing code, you can try to edit https://github.com/home-assistant/home-assistant/blob/dev/homeassistant/components/zwave/workaround.py to apply WORKAROUND_REFRESH_NODE_ON_UPDATE which is enabled for PAN07 also for PAN04

thanks a lot - I give it a try… already implemented so far. Lets keep a eye on it :smiley:
I will give (a hopefully successful) response later…

Update: Doesn’t work :frowning:

Update #2
It seems that it only occurs on manual switching by the switch itself.
On periodically update of the switch the values are fine again (until the next manual switching).

The workaroud of andrey isn’t working (I added support fpr PAN04).
Any other idea?

thanks to everyone!
Marco

What my workaround does - it makes a full refresh whenever the device sends any update. Maybe PAN04 doesn’t send an update at all?

Could you check on OZW log if any message is sent when you manually switch the switch? If if sends anything - please post it here. I’ll check why the workaround isn’t triggered.

OK, when I switch the light off the following log entries were written:

2017-04-12 22:34:13.851 Detail, Node009,   Received: 0x01, 0x09, 0x00, 0x04, 0x00, 0x09, 0x03, 0x25, 0x03, 0x00, 0xde
2017-04-12 22:34:13.852 Detail,
2017-04-12 22:34:13.852 Info, Node009, Received SwitchBinary report from node 9: level=Off
2017-04-12 22:34:13.852 Detail, Node009, Refreshed Value: old value=true, new value=false, type=bool
2017-04-12 22:34:13.852 Detail, Node009, Changes to this value are not verified
2017-04-12 22:34:13.852 Detail, Node009, Notification: ValueChanged
2017-04-12 22:34:14.903 Detail, Node009,   Received: 0x01, 0x14, 0x00, 0x04, 0x00, 0x09, 0x0e, 0x32, 0x02, 0x21, 0x34, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xcc
2017-04-12 22:34:14.903 Detail,
2017-04-12 22:34:14.903 Detail, Node009, Refreshed Value: old value=false, new value=false, type=bool
2017-04-12 22:34:14.903 Detail, Node009, Changes to this value are not verified
2017-04-12 22:34:14.904 Info, Node009, Received Meter report from node 9: Power=0.1W
2017-04-12 22:34:14.904 Detail, Node009, Refreshed Value: old value=418.3, new value=0.1, type=decimal
2017-04-12 22:34:14.904 Detail, Node009, Changes to this value are not verified
2017-04-12 22:34:14.904 Detail, Node009, Notification: ValueChanged
2017-04-12 22:34:14.910 Detail, Node009, Notification: ValueChanged

So the report says that the switch is off and that the power consumption changed form 418.3W to 0.1W
What is missing?

OK, let me explain on another value, that was gives a better look at the main error. The switch has three endpoints:

0: switch1
1: switch2
2: overall

On initial state everything shows up correctly like this:

0_previous_reading: 1.0 kWh
1_previous_reading: 0.9 kWh
2_previous_reading: 1.9 kWh

Or in this case:

0_power: 193.1 W
1_power: 225.2 W
2_power: 418.3 W

On manual switching the light off (like in this case) the values are changing like this:

0_power: 0.1 W (seems the same as endpoint 2 but could be right)
1_power: 0.2 W
2_power: 0.1 W

and

0_previous_reading: 1.9 kWh (<- this is not ok anymore!! overwritten with value of endpoint 2 by manual switching light off)
1_previous_reading: 0.9 kWh
2_previous_reading: 1.9 kWh

On restarting homeassistant the values are OK again. Otherwise the values are false near almost the time HA is running. Sometimes (I don’t know why) the values are again correct (until the next update).

Switching lights on again results in this output:

0_power: 418.3 W (<- this is not ok, too…)
1_power: 225.2 W
2_power: 418.3 W

0_previous_reading: 1.9 kWh (<- this is not ok!!)
1_previous_reading: 0.9 kWh
2_previous_reading: 1.9 kWh

thx,
zheim

OK - another update:

Not only the manual switch results in this behavior, although a update from the device itself causes this problem :frowning:
Only a restart of HA is resolving it…

Nobody else having the same issue?

Sorry, got me a while to get back to this…

Are confusing power (in W) with energy (in kWh)?

When the device sends an energy update it replaces previous_reading with energy reading and updates interval with the time between readings.

Are @zHeim and @mpoi the same user?

What @mpoi posted seems like a report without the workaround applied.

I would expect, whatever the problem is to be fixed by applying the workaround.

Could you post to hastebin a full OZW report of what happens when a switch is switched manually? Both with and without the workaround aplied?

Note that after starting the HA / ozwcp you should wait till all awake nodes are queried (OZW log stops getting messages)

Hi Andrey,

yes - the login account wasn’t right - so the posting account was mpoi for the answer.

Now I’m really confused :expressionless:
The workaround is already applied.

The behavior - how it looks today:

0_previous_reading: 4.87 kWh
1_previous_reading: 4.21 kWh
2_previous_reading: 9.08 kWh

what looks right.
After update - it seems that it is not only by manual switching the light switch:

0_previous_reading: 9.08 kWh
1_previous_reading: 4.21 kWh
2_previous_reading: 9.08 kWh

OK - like I wrote a time ago. Now I keep the false values for a while and this happens:

0_previous_reading: 12.76 kWh
1_previous_reading: 6.25 kWh
2_previous_reading: 12.58 kWh

Does the update relative value modifications? It seems like a value is added to the sensor value.

Confusing. I give a log as soon I am able to do…

thank and kind regards,
zheim

Nobody with same problem here?

Hi Andrey,

I think - like you - that the workaround doesn’t get hit. Is there a debug info or everything else I can confirm the workaround runs? I don’t know if every parameter (product_id) I implementatet is the right one :expressionless:

thanks! kind regards, Marco

I have problem as well with the PAN-04 and Endpoints using Home Assistant- it leads to wrong states for the switches , as well the sensors getting mixed up.
It looks like the configuration parameter 3 “Selected end point” is wrongly defined in HA: 1: Relay 1 and Relay 2. 2: Relay 1. 3: Relay 2
It should be 1: Relay 1 2: Relay 2 3: Relay 1 and Relay 2
Additionally the configuration parameters 7 to 10 is missing.

hi don, is there a solution already? how did you “work” with the switch?

Well , there seems to be a mixup , probably based on some old documentation on the device.
(link to the old documentation teldus PAN4 where it states Value instance 1 is the combined:
http://www.telldus.com/wp-content/uploads/312352-PAN04-1-2-3-manual-20130530.pdf)
(link to new documentation for PAN4 https://www.z-wave.pl/pl/p/file/4db1bb56907dd365ee1ec6c431c1923c/PAN04-1-2-3B-manual )

The binary switch instance 1 and 2 is on the device switch 1 and switch 2. Value instance 3 is the “combined” switch state “on” whenever switch 1 or 2 or both is on.
However in setting states Value instance 1 is used as the “combined” switch. That is why this get wrong.
Doing polling on the device as a workaround this get the correct states on the switch correct in HA
However the Energy - in your case get wrong numbers then for switch one reading values for the combined data.
So a permanent fix should be done for the new version of the device.
Look at case in GitHub https://github.com/home-assistant/home-assistant/pull/15203

With help from philio - done correct installation, now pan04 works as they should :slight_smile:

More here:

Found a way to get this working right. In Openzwave UI go to the problem PAN04 node and then go to Config Values and set Primitive frame to 2 and check the message view and make sure it changed then that’s it, they work without switching the other on in the UI.