Errors after updating to MQTTController build 25139

[toggledbits]

Also, did you restart Reactor after adding the node?

[gwp1]

@toggledbits Answering in reverse order: yes on restart of MSR (multiple times). Also reloaded the HA ZWaveJS plugin as well as restarted HA as a whole.

The the first ask:
647.yaml

manufacturer_id: 647
manufacturer_name: "HAB Home Intelligence, LLC"
version: 23093
revision: 1
format: 1

and I do not have a ext/ZWaveJSController/zwave_devices.yaml but do have a ext/ZWaveJSController/zwavejs_devices.yaml and it shows

version: 23052
revision: 1
format: 1

Looking in my files I see the latest version was 23093 so I'm grabbing 23094 from your repo and installing now.

[toggledbits]

Looks like your Hass devices aren't actually faring much better... nulls there too for some values, and that's a straight pass from Hass itself, not likely something MSR is doing (or not doing).

Something off about this device. I'm going to PM you instructions to upload a data file from your system.

[gwp1]

@toggledbits Ok, release deployed.

ext/ZWaveJSController/zwavejs_devices.yaml now shows

version: 23194
revision: 1
format: 1

647.yaml still shows

manufacturer_id: 647
manufacturer_name: "HAB Home Intelligence, LLC"
version: 23093
revision: 1
format: 1

Node attributes:

battery_power.since=1694192043559
cover.state=null
position.value=null
x_debug.dt={"entity_class":"Cover","match":"deviceClass.generic.key=17;deviceClass.specific.key=7","capabilities":["cover","toggle","position"],"primary_attribute":"cover.state","description":"Window Blind Controller","model":"iblinds V3","default_name":"Window Blind Controller","values":[{"valueId":"38:currentValue:","capabilities":{"cover":{"attributes":{"state":{"expr":"!(value == 0 || value >= 99)"}},"actions":{"open":{"valueId":"38:targetValue:","value":50},"close":{"valueId":"38:targetValue:","value":99},"stop":false}},"position":{"attributes":{"value":{"expr":"round( value / 99, 2 )"}},"actions":{"set":{"valueId":"38:targetValue","value":{"expr":"max( 0, min( 99, round( 100 * ( parameters.value ?# 0 ), 0 ) ) )"}},"increase":{"valueId":"38:targetValue","value":{"expr":"max( 0, min( 99, round( 100 * ( ( entity.attributes.position?.value ?# 0 ) + ( parameters.amount ?# 0.10 ) ), 0 ) ) )"}},"decrease":{"valueId":"38:targetValue","value":{"expr":"max( 0, min( 99, round( 100 * ( ( entity.attributes.position?.value ?# 0 ) - ( parameters.amount ?# 0.10 ) ), 0 ) ) )"}},"relative":{"valueId":"38:targetValue","value":{"expr":"max( 0, min( 99, round( 100 * ( ( entity.attributes.position?.value ?# 0 ) + ( parameters.amount ?# 0.10 ) ), 0 ) ) )"}}}},"toggle":{"actions":{"toggle":{"valueId":"38:targetValue:","value":{"expr":"entity.attributes.cover.state ? 99 : 50"}}}}}}],"manufacturerId":647,"productType":4,"productId":113}
x_zwave_values.Battery_isLow=false
x_zwave_values.Battery_level=100
x_zwave_values.Configuration_Close_Interval=15
x_zwave_values.Configuration_Default_ON_Value=50
x_zwave_values.Configuration_MC=1
x_zwave_values.Configuration_Maximum_Tilt_Level=99
x_zwave_values.Configuration_Minimum_Tilt_Level=0
x_zwave_values.Configuration_Movement_Duration=0
x_zwave_values.Configuration_Override_Response_to_ON_Command=null
x_zwave_values.Configuration_Remote_Calibration=1
x_zwave_values.Configuration_Reset_Button=0
x_zwave_values.Configuration_Reverse_Direction=1
x_zwave_values.Configuration_Send_Reports=0
x_zwave_values.Indicator_Node_Identify_3=0
x_zwave_values.Indicator_Node_Identify_4=0
x_zwave_values.Indicator_Node_Identify_5=0
x_zwave_values.Indicator_identify=true
x_zwave_values.Indicator_timeout=null
x_zwave_values.Indicator_value=null
x_zwave_values.Manufacturer_Specific_manufacturerId=647
x_zwave_values.Manufacturer_Specific_productId=114
x_zwave_values.Manufacturer_Specific_productType=4
x_zwave_values.Version_applicationBuildNumber=43707
x_zwave_values.Version_applicationFrameworkAPIVersion="3.12.1"
x_zwave_values.Version_applicationFrameworkBuildNumber=35
x_zwave_values.Version_applicationVersion="3.12.1"
x_zwave_values.Version_firmwareVersions=["3.12"]
x_zwave_values.Version_hardwareVersion=2
x_zwave_values.Version_hostInterfaceBuildNumber=0
x_zwave_values.Version_hostInterfaceVersion="unused"
x_zwave_values.Version_libraryType=3
x_zwave_values.Version_protocolVersion="7.12"
x_zwave_values.Version_sdkVersion="7.12.2"
x_zwave_values.Version_zWaveProtocolBuildNumber=35
x_zwave_values.Version_zWaveProtocolVersion="7.12.2"
x_zwave_values.Window_Covering_currentValue_23=0
x_zwave_values.Window_Covering_duration_23={"value":0,"unit":"seconds"}
x_zwave_values.Window_Covering_levelChangeDown_23=false
x_zwave_values.Window_Covering_levelChangeUp_23=true
x_zwave_values.Window_Covering_targetValue_23=0
zwave_device.capabilities=[106,112,114,128,134,135]
zwave_device.endpoint=0
zwave_device.failed=false
zwave_device.generic_class="Multilevel Switch"
zwave_device.impl_sig="23194:1:23093:1"
zwave_device.is_beaming=false
zwave_device.is_listening=false
zwave_device.is_routing=true
zwave_device.is_secure=false
zwave_device.manufacturer_info=[647,4,114]
zwave_device.max_data_rate=null
zwave_device.node_id=20
zwave_device.specific_class="Motor Control Class C"
zwave_device.status=4
zwave_device.status_text="alive"
zwave_device.version_info=[null,"3.12.1"]

(Did a manual restart via Tools>Restart after the deploy as well.

[toggledbits]

Yeah, this device is off. It's reporting Multilevel Switch for its generic class, and Motor Control Class C as the more specific, and usually this device class reports command class 38, Multilevel Switch. That's how your other iBlinds units that identify the same way (same 647 manufacturer ID, product type 4, product ID 114) work. But this particular unit does not appear to be advertising command class 38 Multilevel Switch, and is instead reporting command class 106 Window Covering, which is relatively new.

Once you upload the data I PM'd you about, I'll be curious to see how the reporting firmware compares. But this new unit is definitely working differently from everything else you have, and it doesn't look like HA actually likes it any better -- I think that's why it's reporting the slats angle in HA as null; I bet it shows the same in their dashboard.

I'll know more/confirm once I see you data. Probably won't get to that until Sunday or Monday at this point, though.

[gwp1]

@toggledbits file has been uploaded. This blind is in a guest room that's not in use currently - no rush.

Do you think excluding/including with HA/ZWjs would have any effect? Could something have gone sideways during pairing that a complete redo might clear?

[toggledbits]

ZWaveJSController version 23254 has been released. It should handle your oddball iBlinds unit. I'd love to know (not from you, from the manufacturer) what triggers the difference between reporting status in command class 38 versus command class 106 -- only this most recent unit of yours reports/controls via 106; all of the others use 38. And there's no difference in reported model or firmware version. No obvious configuration parameter that might trigger such a difference (e.g. some others, like some Aeotec devices, can change the reporting CC based on a config parameter). Just goes to show you that despite a fairly strict standard and specification, manufacturers can continue to make subtle choices that create headaches.

[gwp1]

@toggledbits this was deployed and tests as working. Really appreciate it - I bought two new units and have held off installing the second until it was determined not to be a firmware problem (broken vs them just randomly changing something).

Thank you!

[gwp1]

@toggledbits I contacted iblinds Support. This was their response:

*We've had similar reports. I believe it is related to a recent update with Home Assistant supporting the Z-Wave Window Covering Command Class now. So when you add your new devices they are seen with the new command class. I would imagine if you excluded and included one of the older motors they would now be Window Covering devices as opposed to multi-level switches.

We initially worked with the ZWaveJS developer when they added the Window Covering CC Support. But unfortunately, our HA in our lab is no longer working so we have been unable to test locally.

We have been waiting to get a replacement RPi so hopefully we can do some testing on our end soon.

I hope this helps.

Best Regards,
Eric B.
The iblinds Team*

Further, I just discovered if I do a reinterview on an existing blind via ZWaveJS in HA it switches to this new Window Covering mode.

[tamorgen]

@gwp1 said in [SOLVED] New iblind and zwaveJScontroller/MSR not communicating:

Further, I just discovered if I do a reinterview on an existing blind via ZWaveJS in HA it switches to this new Window Covering mode.

For what it's worth, I have both v2 and v3.1 iBlinds in my home. Reinterviewing only changes the type from multilevel switches to window covering on v3.1 blinds. Version 2 blinds remain multilevel switches after reinterviewing.

The Window Covering class is supposedly better. The MLS class is designed for rolling blinds, so on/off will have the bllinds all the way open, or all the way off. With horizonal blinds, the 0/100 values are just 180 degree opposites of a closed blind. With Window Covering class, I believe there are values to tell the Z-Wave system what type of blind it is, and specify what is open versus closed. I dealt with iBlinds support maybe 4 or 5 months ago, and at the time, no Z-Wave system supported Window Covering. Zwave JS implemented the change shortly thereafter.

[gwp1]

@tamorgen I'm poking at it now. It seems for the iblinds I can now use cover.close and the [0] NO (default) setting for direction vs having it set to [1] YES to make it tilt the way I want it to when closed.

I was using power_switch.off previously to close my v3.1's.

[gwp1]

UPDATE: after many lost hours last night/this morning (hello, 0230 ET, looking at you) I found cover.x and power_switch.x to be woefully unreliable after reinterviewing the blinds.

What has proven reliable is position.set for which I use 0 for closed, 50 for open and have the blinds (as they were previously) set to [1] YES for Reverse Direction.

[tamorgen]

I've been using the ID hass>cover_blind_horizontal_slats_angle and x_hass_cover.open_cover_tilt or x_haas_cover.close_cover_tilt and it works flawlessly for open and close. If I want to set the position for partial close when the sun is glarring into the room, I use x_hass_cover.set_cover_tilt_position with the same ID, and set the tilt_position to whatever I want to reduce the glare.

My only problem now is that I use mushroom for my dashboard in Home Assistant, and the templates for covers hasn't caught up with the new Window Covering class. It makes adjusting the blinds more problematic from HA.

[toggledbits]

@gwp1 said in [SOLVED] New iblind and zwaveJScontroller/MSR not communicating:

UPDATE: after many lost hours last night/this morning (hello, 0230 ET, looking at you) I found cover.x and power_switch.x to be woefully unreliable after reinterviewing the blinds.

Can you please be more descriptive here? I don't have these devices, so I can't test them, but that doesn't mean I'm not interested in getting them working reliably. But the description given isn't enough for me to do any troubleshooting at all.

[gwp1]

@toggledbits of course, no problem! I cannot seem to upload images into this response. I've uploaded to your Dropbox shared folder a pair of images.

Using the iblind I sent you logs and such for previously as the example, I have a blind called Beach Blind (two guest rooms, one is named Ocean, the other Beach). I use the commands from the ID Beach Blind from zwavejs. As @tamorgen notes, there is also a new one (with the changes HA made) from my HASS controller called Beach Blind Horizontal Slates Angle which has a few less commands including those mentioned starting with x_hass_*.

I use the ZwaveJSController commands.

cover.open
cover.close
power_switch.on
power_switch.off
position.set
zwave_device.refresh
zwave_device.set_config

These are the ones I most commonly use to open, close, and tilt the blinds. refresh is used to force HA to update dasheboards with any changes and set_config I use less often, weekly, to recalibrate the blinds as they do wander sometimes and "open" becomes "open-ish".

I'd been using power_switch.on to open and cover.close to close. The blinds have been happiest that way. With these new changes from HA, these commands have become hit/miss. In the HA ZWaveJS plugin the status would reflect 0/closed but the blind was clearly wide open.

Last night/this morning I opted to use position.set and it's been pretty solid from what I've seen throughout the day.
position.set = 0 nets me a closed blind, position.set = .5 nets me a wide open/"slates level with the ground" blind. I used position.set = .2 or .3 to tilt the sun away during specific dayparts based on sun azimuth.

I did not grab MSR logs during those tests - apologies, it was 0200 ET and I just wanted the blinds closed for the night. I have amassed 13 of these iblinds over the past couple years.

If you wish, I'll set some time tonight or tomorrow night and generate some MSR logs.

[tamorgen]

@gwp1 said in [SOLVED] New iblind and zwaveJScontroller/MSR not communicating:

With these new changes from HA, these commands have become hit/miss.

To be precise, it's not so much a change from HA, but from ZWaveJS. They're recognizing a new command class that the developers at iBlinds actually recommend.

I don't believe the old power switch commands work with the new command class. When I go into a specific device in ZwaveJS, the switch type devices are greyed out.

if you go into the switch, it shows it as unavailable.

I would guess the power on/off cover.close / cover.open are remnants and probably need to be removed.

[gwp1]

@tamorgen said in [SOLVED] New iblind and zwaveJScontroller/MSR not communicating:

I don't believe the old power switch commands work with the new command class. When I go into a specific device in ZwaveJS, the switch type devices are greyed out.

I noticed this as well. I've also noticed that tapping into the dialogue you're showing (how did you copy/paste images, I keep getting a parsing server error...) nets a slider in HA that does nothing. It literally says UNKNOWN at the top where a percentage would be display of openness or closure.

[tamorgen]

@gwp1
It seems to me that something may be missing from this command class integration. I know iBlinds was working with the ZWaveJS developers. I'm curious is this is what they expected.

When I look at the entities in MSR, there are a lot of entries that have null or unknown. I'm not sure if this is coming from MSR expecting entries in a specific command class or if ZWaveJs is providing it to MSR that way.

Of all the actions in that window, the cover.xxx don't seem to work for me. The only ones that seem to work are:

x_hass_cover.close_cover_tilt
x_hass_cover.open_cover_tilt
x_hass_cover.set_cover_tilt_position

Others may work, but these are the one's I know specifically work and from an automation perspective meet my needs.

[gwp1]

@tamorgen Yep. The ZWaveJSController from @toggledbits was doing the same and that's what started this thread. He reviewed data from my system and updated his ZWaveJSController to work with those command classes.

[tamorgen]

Okay, gotcha. I did verify that the cover.open using the zwavejscontroller entries does work on my system with the updated release. Curious that the cover.open command does not work using the equivalent haas entity.