So.... we are to believe that because its not a cosine sensor that its somehow going to give readings on the order of ten times what your cosine meter would read at that distance (24 inches). Riiiiight.
Two words:
Prove. It.
You have the cosine sensor. 400 diodes. 22" x 22" frame. 10 strips (I'll even donate my two spares). Reflective walls 24" x24". Put up or shut up
It's easy to prove.
Hang a 48" strip at 36". Measure the distance from a sensor to an LED from the sensor to the LED 36" above and 24" to the side of the sensor (on the very end of the strip). Then measure the angle.
Now lower the strip to 12" and measure the distance and angle to the end LED.
While the LED above the sensor distance changed 3x, the end LED changed 1.6x.
ISL 36" to 12" = 9x
ISL 43.3" to 26.8" = 2.6x
Error: 346% (9
÷ 2.6) x 100
NOTE: This is the same error introduced by the LED on the other end of the strip.
These images are rotated and flipped with respect to one another due to size constraints of the app.
Now take a look at the angle of the photons emitted from the LED. The angles shown for the triangle where the LED is located must be adjusted to the LED's point of view. The 34° angle at 36" becomes 56° and the 27° at 12" becomes 63°. Looking at the Samsung or Bridgelux strip's datasheets the flux is reduced by 60% at 64° and 45% at 56° with respect to the center LED at 0°. So there is another 60% error factor which increases as the height gets lower.
_________________________________________________________
Wrong Tool for the Job
Fist off a lux meter is photometric meaning the radiometric watts are adjusted for humans NOT plants! Furthermore plants do not care about the energy of a photon so the watts need to be conversed by wavelength to quantum moles.
A lux meter must be designed to measure pulsed PWM light which the cheap meter are not. For example, the Konica Minolta T10-A is designed for PWM.
For error rates, an inexpensive Lux Meter is compared to CIE illuminate A which is a tungsten incandescent light source with a CCT of 2756K
From:
Colorimetry, 3rd ed., Publication 15:2004, CIE Central Bureau, Vienna
CIE standard illuminant A is intended to represent typical, domestic, tungsten-filament lighting. Its relative spectral power distribution is that of a Planckian radiator at a temperature of approximately 2856 K. CIE standard illuminant A. CIE standard illuminant A should be used in all applications of colorimetry involving the use of incandescent lighting, unless there are specific reasons for using a different illuminant.
Was a 2756K strip being measured? So then you must convert every wavelength of the CCT of the strip being measured to a 2756K standard.
Does this look anything like a 4000K spectrum?
More error from the mismatch in spectrum.
As the angles increase the sensors accuracy decreases. Below is the specs from a higher end Lux meter ≈ ($1,000-$2,000)
Notice when the angles get down to the 60°-80°, the error increases from 10-30%. and that is for a high end meter. Notice the 6% error over 3,000lux. But only ±1 digit. Look at the digit error on cheap lux meters.
This chart shows the sharp decrease in accuracy above 60°.
A $1,000 lux meter will conform to class A of JIS C1609-1 2006 (e.g. Yokogawa Model 51011: $1,155.00)
A $2,000 conform to class AA of JIS C1609-1 2006 and DIN Class B (e.g. Yokogawa Model 51021: $2,101.00)
Reference: https://www.instrumart.com/products/40519/yokogawa-510-series-lux-meters
