ZEN15 returning ridiculous kWh readings and dropping power

Hardware
,
1

My new ZEN15 is doing some odd things. It’s reportedly running V2.10 FW. My other ZEN15 is running 1.6 and I have not had a bit of trouble with it in the past two years I’ve had it–but it’s used for something else, so… Both report “up to date” in HA. (???) I got it to monitor energy use by my new freeze drier. It periodically reports 30+ kWh jumps EVEN WITH THE OUTPUT OFF. This makes monitoring energy use impossible. When the drier is running it has periodically dumped the output power at a load approximately 1/2 the device’s rating. When it restores the output (all by itself) I get a characteristic motor inrush amperage when the compressor and vacuum pump start back up, and the drier is reporting power outages so I KNOW it is dumping the output power and it’s not just an indication issue. There are no automations controlling the device, so it should not turn off unless I command it manually, which I do between drier runs, which is why I know the ridiculous jumps happen with the power output turned off.
Reached out to Zooz, they claim “After performing extensive tests last winter, we did not find any issues in our firmware and the only issues are reported by HA users. We recommend reaching out to the Z-Wave JS team or we can set up a return.”
So I’m reaching out (as best I can since “Z-Wave JS” is not available as a tag) to see if anyone has any ideas prior to setting up the return.

2

The 1.6 version is the original 500/700-series model. The 2.10 version is the newer 800-series LR model.

Well yeah, they are up-to-date. You can confirm for yourself by going to the source:

1.6 is the latest for the 500/700-series, and 2.10 the latest for the 800-series.

However, the HA update entities are based on the Z-Wave JS firmware update service. Someone needs to provide this information, and it relies mostly on manufacturers providing their updates to ZJS, so may not be always accurate as either many manufacturers don’t bother, or they are slow to provide them. Here are the available Zooz firmware updates for example: https://github.com/zwave-js/firmware-updates/tree/main/firmwares/zooz

One thing about Zooz you should know, they will always blame HA or Z-Wave JS for any problem, and won’t investigate further without insistence. Sometimes it is a HA/ZJS problem, but there are at least an equal or more number of issues in their devices, and often ZJS has to workaround them. As a Zooz customer myself, I have encountered small paper-cut issues with their devices. Also, just take a look at their change logs (BTW it’s great they provide them), it’s not like they don’t fix any device bugs. For instance, despite performing “extensive tests last winter”, they still had to release a fix for the ZEN15 800 in May 2024 with v2.10, which sounds like it had a pretty serious bug (doesn’t sound like HA’s fault):

Fixed a bug when the Power Switch could turn off in edge cases when a large appliance like a dryer was initially starting. The Power Switch will no longer turn off on the initial start up unless overload protection is triggered.

One of my 800 series devices has the exact same bug they claimed to fix in a previous 500/700 model iteration, clearly the big still exists but their solution was to just factory reset it (which worked, but why?). So while I’m satisfied overall with their products, sometimes the support responses can be frustrating (at least they reply).

So I’m reaching out (as best I can since “Z-Wave JS” is not available as a tag) to see if anyone has any ideas prior to setting up the return.

Well, if you want to reach out to Z-Wave JS that would be at https://github.com/zwave-js/node-zwave-js/issues/new/choose. Whether you want further help here or there, you’re going to need to provide more information about your installation (how you are using Z-Wave JS), and driver debug logs to either demonstrate the problem or prove it exists elsewhere (e.g. driver vs. HA).

3

Oh, also just found that someone had reported a similar sounding problem in the past. Not sure the user ever followed through with providing more information as it’s still unresolved.