Background: New to Inovelli. Just installed and started playing with settings. Tried to set the color to predefined “Orange” but noticed it looked really yellow instead. So, I manually picked a hue value that looked orange to my eyes. This ended up being 18º. It looked great… until I turned the light off and the LED strip dimmed to the default level and it turned red.
Findings: I then dug around in the Hubitat driver code and experimented with different settings. On my dimmer anyway, if the hue setting is less than 23º (Parameter13 < 17) and the intensity is at 10% (Parameter15 = 1), then it turns red. It seems when the led strip is dimmed this low, the green LED completely shuts off and we’re left with just the red LED glowing dimly and so it suddenly turns from yellow/orange to red.
Doesn’t seem like a hardware limitation, but maybe something in firmware miscalculating how to drive each of the R, G, and B LEDs??? I read that these are firmware update-able, but how do I know what firmware version is on there now? And are OTA updates even possible with my Hubitat?
Can anyone else replicate my findings so I will know it’s not just my particular dimmer?
I can duplicate your results. If I set Parameter 13 a value less than 17, and set the brightness to 10% or 20%, the led appears red.
I am not sure if it is a firmware issue or a hardware limitation though. It may just be that the green LED is not capable of dimming low enough enough to create an orang-ish color at that brightness level.
I agree it could be a hardware limitation, not because the diode itself cannot dim low enough, but because the LED drivers may not be able to dim low enough.
Seems like the brightness quantization is too linear at the low-end of the dimming curve (or maybe not linear enough – I haven’t really thought through the logic of which way it is). This could explain why at higher brightness levels when the green LED is not off, the “Orange” setting (properly defined at 21/255 ≈ 30°) looks very much like yellow (Hue 60°) and the “Yellow” setting (defined properly as 42/255 ≈ 60°) still looks like yellow – Because the Green LED is ramping up/down in perceived brightness too quickly (too large of steps) at the low-end of the dimming curve. My perception of orange at a setting of 18° (looks like 30°) is also consistent. I guess only Inovelli engineers can answer this.
I’m hoping the problem is on the firmware-side where this is defined, and not on the hardware-side where it could be that there is a small number of bits used to quantize each of the colors and the scaling is done in hardware.
I’ll continue to play with the color settings and see if low-values of the other colors (to the right of green and blue) gives any further clues, but I don’t have a color meter (maybe re-purpose a Spyder screen calibration device?) and photoreceptors in human eyes are most sensitive to the orange part of the spectrum (between the red & green).
Is this with the latest FW? There has been a couple of FW updates to the LZW31-SN since my experimentation. But with COVID work from home and three kids, I haven’t had time to update the FW nor have I had any time to play with LED color settings.
I have just run into this as well. I noticed the that the color distinctly changes at the values 17, 34, 53, 69, … I also test this out on the LZW30 (Black On/Off) and it does not have this issue.
Edit: I tested on the LZW31-SN with firmware version 1.48 & 1.57.