I regularly have the following effect: I change the position from a% to b%, but instead of going to b% position, the shutter opens fully. This happens roughly 30-50% of the time, and mostly when no commands have been sent for a while. And it happens on two shutter switches of the same model: I only have the two, but a hardware defect seems unlikely.
I only have 3 z-wave devices: gateway and the two shutter switches, all in the same room.
nodes: Technisat shutter switch with built-in Z-Wave: https://www.technisat.com/assets/medias/docus/169/BDA_Rollladenschalter_EN_FR_NL.pdf#page=22
Z-Wave gateway: Zooz ZST10 with Espressif Z-Wave 700/800 chip.
I checked the debug log from z-wave and what I see is that when the problem occurs, communication to the z-wave device seems to break down, it just stops.
The good case is: HA requests a new position, then the shutter switch reponds, informs how long it took, and confirms what the new position is, it even gives in-between updates on the position while it’s still changing.
In the bad case, the switch completely opens (instead of going to position b%), correctly informs it took an unexpectedly long time to complete the command (because it completely open the shutter instead of going to b%) and while in that last message “more updates follow” is set to true, no other message is sent by this node. Hence HA still thinks the shutter is now on b%.
If I then ask HA to move the shutter to position c%, I have roughly 30-50% chances that it will work fine, going from full open to the position c%.
In my testing it seems that the longer I changes position the less likely it is for the failure scenario to happen.
Can you think of anything in the message sent by the Zooz ZST10 that could throw off the shutter switch, knowing the technisat shutter switch is z-wave+ only (aka z-wave 500 afaik)?
I can also post the z-wave debug here if that helps.