Beiträge von Dale Phurrough

Oh dear. Shelly Inc. writes this is one of their official support methods and links to it. That means someone is untrue. But it doesn't really matter. The bug exists and is being tracked in github, here, and in two closed issues internal to Shelly Inc. Shelly is unfortunately unwilling to admit their faulty design. Which they can fix. But are chosing not to. Firmware can be changed by Shelly inc. They do it often.

Last, "dangerous" is a general English word meaning: likely to cause problems or to have adverse consequences.

There are many "adverse consequence" of their firmware bug to not update voltage as I describe.

Please keep the discussion between paid customers and Shelly a discussion professional. In English culture, writing "What is your point here" is aggressive. the defensiveness from your support team is clearly here in your writing as well as both closed issues I have linked above.

I have reported in four locations a design bug on voltage notifications. Clear reproduction, the source of the bug, and the opportunity for Shelly to fix the firmware.

As the firmware is today, Shelly devices with the PM1 component (perhaps others for voltage) send NO notifications for voltage changes for anything except:

- total ON, total OFF ==> because they are more than 5%

- dangerous overvoltages and dangerous undervoltages 218.5v or higher than 241.5v == > because they are more that 5% of 230volts

As I wrote, the design bug is limited the the notification channels. When the Shelly is directly force polled, it will return voltage.

Modern systems, like Home Assistant, do not poll. You want gen2 and gen3 to be modern...right?

Therefore, voltage updates from Shelly do not occur except for dangerous situations.

The 5% rule is a bug for voltage. And if any other values trigger in a similar errant way (power, energy, etc.), then they are also errant.

Meanwhile, shelly support ticket carnival on this firmware issue is ongoing

https://support.shelly.cloud/en/support/tickets/92886

https://support.shelly.cloud/en/support/tickets/93111

naturally, only Shelly manufacturer (or me) can open those links

The two test devices I have are both Shelly PM mini gen3

Please notice I wrote COMPONENT. "PM1" is the name of the COMPONENT within the Shelly firmware architecture

json diag of:

```

"deviceInfo": {

"name": "ShellyPMMiniG3-1",

"id": "shellypmminig3-84fce63c4960",

"mac": "84FCE63C4960",

"slot": 0,

"key": "redacted",

"batch": "2347-Broadwell",

"fw_sbits": "00",

"model": "S3PM-001PCEU16",

"gen": 3,

"fw_id": "20240223-141904/1.2.2-g7c39781",

"ver": "1.2.2",

"app": "MiniPMG3",

"auth_en": false,

"auth_domain": null

}

```

While debugging everything Shelly, I traced that voltage changes are not being sent from the Shelly to HA on ws notifications. The logs show only aenergy and ret_aenergy. The Shelly documentation does not clearly define what measured variables are sent in a NotifyStatus notification (as compared to the full detailed NotifyFullStatus. I did find https://shelly-api-docs.shelly.cloud/gen2/Component…#webhook-events which writes webhook events only fire when the voltage changes at least 1V and at least 5% from the last reported value.

But... this is voltage....230 volts. 5% is +/- 11.5v. So if Shelly's internal logic is to only send webhooks (and likely notifies) if the voltage goes lower than 218.5v or higher than 241.5v ... that's just too large a range in which to send no voltage notifications.

Similar rules apply to current; at least 0.05a and 5% from previous value. However, this might be more reasonable for amps.

I see the Shelly integration system option for polling was already set on install to be TRUE. But I don't see evidence of polling in the debug logs (attached below)

Manually curl to the shelly correctly returns voltage

Home Assistant Shelly integration related bug at https://github.com/home-assistant/core/issues/111598