Dummy Load & 0 Readings

[dewbot6]

Hello,
I have the following Hue devices that I wants to measure:

• Perifo cylinder spot (929003617901)
• Festavia string lights (LCX015)
• Perifo cylinder pendant (929003618001)
• Hue Infuse ceiling (929003531702)
• Hue color downlight (LCD007)
• Hue Calla outdoor bollard (1741930V7)
  I am struggling getting good readings with my Shelly PM Mini Gen3. With no dummy load I get 0 watts at idle. If I use a dummy load (incandescent bulb) I am able to read the idle power, but the dummy load appears to change over time and gives me negative readings. So my question is: is there a better power sensor I can use? I attempted to use the Eve Plug, but it had the same issue. If the Shelly works well enough, is there a more consistent dummy load I can use?

[bramstroker]

It could help to warm up the incandescent bulb to prevent the resistance changing over time.

Also we have a PR going on which is in final stages of development: #2810.
This will also account for voltage fluctuations, and will also measure dummy load to see when it's stable enough before starting the measurement session.
But this is not merged and releases yet. Will probably within a few days.

Also connection multiple of the same light in parallel could be a solution.

[dewbot6]

@bramstroker perfect. I'll wait for the next release and preheat my bulb for a few hours before. Thanks!

[RubenKelevra]

Well, incandescent bulbs don't need much preheating time. You can give it a minute and then start measuring the dummy load. The script will then confirm that there's no further preheating necessary.

The suggestion by @bramstroker to preheat for multiple hours is required if you use ceramic resistors, which first need to heat up a lot of mass around them before their surface temperature stabilizes and thus the heat which they give up to the room stabilizes.

While testing I noticed that the script is even sensitive enough to pickup that I opened a window and thus the ceramic resistors lost some heat, which was quite fun to see.

Incandescent bulbs on the other hand have nearly no mass which is heated, as the most heat is lost directly by the IR radiation of the tungsten wire itself.

The heatloss the effect of the mass of the glass on the other hand is quite minimal. Therefore they stabilize much more quickly.

[dewbot6]

@bramstroker do you know when / which release #2810 will be included in?

[bramstroker]

@dewbot6, not sure yet. I have done some final rework last week. Asked @RubenKelevra if he can do a final test run before I merge.
Could be he is very busy atm, so he don't have any time to do so. Will merge it this weekend when no response before that.

[RubenKelevra]

Testing is done now, LGTM :)

[bramstroker]

It's merged now, thanks @RubenKelevra.
However have troubles / errors while building and publishing the docker image.

Got strange errors in the github action, no time to investigate further.

Tried deleting github caches now and retrigger new run manually now.

https://github.com/bramstroker/homeassistant-powercalc/actions/runs/13018030540

Fingers crossed this build will succeed.

[bramstroker]

Succeeded now. You can pull and run latest master:

https://hub.docker.com/layers/bramgerritsen/powercalc-measure/master/images/sha256-cdba90b2cdac75e3da4eb23142e50c9f5eaf997c567367375b54d27ed572c4ae

[dewbot6]

@bramstroker awesome. I am now getting 0.46W in standby. I'll start measuring my devices!

[RubenKelevra]

Great! Hope it works out okay now :)