How to measure nanoleaf

[jacobdonenfeld]

Has anyone thought of a solid way to calculate power from nanoleaf shapes? I thought about calculating it per "large triangle", but I use a variety of shapes so this wouldn't work exactly. Would making a LUT per type of shape work? What do you think the right approach would be?

[bramstroker]

Need to think about this some more, and also need to do some testing with a friend of mine. He also has nano leafs.
I think we need to measure 1 triangle. And than we can use the multiply_factor option to set the amount of triangles which make your shape.

However there are multiple issues with this approach.

• The consumption of the driver/controller will also be included when measuring one leaf, so that won't multiply when you have multiple leafs.
• Each leaf can have other colors / effects. As long as they can't have different brightness levels I think it's not a huge problem . But when 1 leaf is much brighter and is using 0.8 watt and the other 0.5 watt for example it is a problem.

Anyway stuff for thought ;-)

[jacobdonenfeld]

Makes sense regarding measuring one triangle. I have my nanoleafs set up with triangles, hexagons, and mini triangles, so I'd probably just have to multiply by each shape, and then add them together. Do you imagine this being done with grouped components in the yaml, or via some new configuration?

[jacobdonenfeld]

Any updates on this? Would love to help out or brainstorm with you.

[bramstroker]

I have some ideas how to integrate this in powercalc.
The configuration could become something like this:

- platform: powercalc
  entity_id: light.my_nanoleaf
  nanoleaf:
    type: Shapes
    leafs:
      - type: triangle
        count: 4
      - type: hexagon
        count: 2
      - type: mini-triangle
        count: 2

Than the component will lookup the LUT files for triangle, hexagon and mini-triangle and multiply by their respective count. And finally add them all together to get the full power of your complete nanoleaf configuration.

But first we need to know if the whole setup has one controller/driver or does each leaf has their own integrated driver?
Could you try measuring with 1, 2 and 4 triangles for example? Then we can see if it is as simple as multiplying by the amount of leafs.

Also could you have a look into config/.storage/core.device_registry to see what information about the nanoleaf is available. And also check the state attributes in developer tools? That information can be useful to determine what configuration needs to be supplied by the user and what information I can automatically retrieve from HA.

[bramstroker]

@jacobdonenfeld did you see my reply above?

[diamant-x]

Mind that nanoleaf could potentially be powered with more than one power supply, for very large setups.
https://helpdesk.nanoleaf.me/en-US/shapes-power-supply-15607

To be honest I would start with the approach from #702 (comment) we can identify the 'commercial kits' being sold and release an update with those, and then continue discussions with whoever has larger setups.

In fact, powersupply is mentions is 42V fixed output, so potentially they just behave like the WLED model. In fact in the following link Nanoleaf mentions: "Each standard 42W PSU supports up to 21 hexagons (2W per panel), 28 triangles, or 78 mini triangles." So we could ask the user number of units of each, then apply WLED model to each with the ratio of consumption from that (42divided21W hexa, 42/28 W per triangle, or 42/78 per mini)
https://nanoleaf.me/en-CA/products/nanoleaf-shapes/get-started/#psu-calculator

[jacobdonenfeld]

Yes, and there are two types of power supplies as well. However that shouldn't change the power draw per shape significantly.

[edwardtfn]

Ok, I understand the complexity for measuring the Nanoleaf Shapes, but that shouldn't apply for Nanoleaf Canvas or Nanoleaf Lines as, in my understanding, thos cannot mix different shapes.

I have two kits, one Nanoleaf Canvas kit with 1+16 panels, and one Nanoleaf Lines with 15 line bars and 1 Controller (which is not a bar), so for those models we could yse these setting you are proposing. It isn't solving the problem for Nanoleaf Shapes unfortunately.
I was thinking about a combination of measurements (this will require time), ideally reflecting the kits they have, so we will have the values for each canvas panel, each line bar and the respective controllers:

Nanoleaf Canvas:
a) 1 Controller + 16 panels
b) 1 Controller + 8 panels
c) 1 Controller + 3 panels

Nanoleaf Lines:
d) 1 Controller + 15 light bars
e) 1 Controller + 9 light bars
f) 1 Controller + 3 light bars

This means running the measurement utility 3 times per model and per color schema (both models supports color_temp and hs)... It would take time, but it looks feasible.

Any thoughts?

[bramstroker]

Thanks for sharing your thoughts. Seems a good idea to provide measurements for the different kits as most users have that configuration anyway.
We can use the sub profile feature for that, which is also used on some other lights which have different configurations / extensions. For example https://github.com/bramstroker/homeassistant-powercalc/tree/master/custom_components/powercalc/data/yeelight/YLDD05YL.

[jacobdonenfeld]

Curious if you gave this any more thought or any new developments support this more easily? @bramstroker

[edwardtfn]

A while ago I've measured the Nanoleaf lines (with 15 bars) the Nanoleaf Canvas (with 17 panels) that I have at home, then I've started the discussion here about how to handle the different possibilities and in the end I've never looked at this again.
I probably still have the files from my first measurements, and I can set the testing environment again and collect additional data, but I wouldn't like to do this before we have a conclusion in the best approach so I don't waste time with measurements that are not gonna be used.
Those measurements are quite painful and usually takes a couple of days (if very lucky) with the lights turning on and off all the time... 😩

[bramstroker]

Sub profiles are fully supported by powercalc, so to get the kits included (Lines and canvas) would be nice, as most people would use the standard kits anyway.
There were also some improvements in the measure script where each color mode only should take a few hours max.
Would be great if you can issue a PR with the measurements you have.

[edwardtfn]

I will do that.

[MirkoKas]

@edwardtfn Hey, just got into Powercalc and found this thread. Did you progress on the Nanoleaf Canvas? I started with 5,4W in standby and an estimate of 15 W when the light is on.

[edwardtfn]

Nope. In the end I've never collected that data... Sorry.