The device in principle should not react to IR radiation, but it interprets it as visible. Watch the video again.
I am thinking maybe it is missing a filter to remove IR before the sensor.
Would be a good idea to add it, if the sensor is able to detect it unintentionally.
the chip we employed itself has responsivity range from 380nm to 1000nm even for F1~F8 channels which intend to only react to VIS spectrum, it does contain IR part, although it’s comparatively very small response but not zero.
seems we can do nothing to solve this as we are just chip integrator.
The easiest way to cut off the IR range is to put an IR filter. This is the easiest and cheapest way, but you need to calibrate the sensor already with the filter.
the chip has an IR channel, why adding IR filter to it?
and what do you mean by saying an IR filter? something like the white plastic thing(actually teflon) on top of the chip?
The spill of NIR into the other sensors (blue?) seems to be the least problem that this device currently has. At some point, you could indeed correcting the response when NIR is present in the spectrum. But before implementing something fancy, the LM4 should basically work.
Jurij raised IR issue.
how can you get the conclusion that LM4 doesn’t work? seems the problem lies in whether LM4 can measure flashlights correctly, I remember long ago people posted sentences like “LM4 seems OK with home lights”.
however, flashlights issue I still will try to fix it(not easy, who to procure flashlights is a first headache).
1 Thank
Just taking standard LEDs into account: The device should give figures for CCT/Ra/R9/CIE coordinates. Since it works not equally well across the set of available LEDs, the user would have to know the LED type to decide if the measurements are acceptable or not. But if he already knows this, there’s wouldn’t be any need for the LM4 in the first place.
So, it works somehow. It just cannot do what we (or I) expected.
flashlights are narrow light beam, will this lead to inaccuracy for LM4?
I have not studied the protocol for working with ams-AS7341. And especially the protocol of interaction with the app. (We have a war in our country and I am not up to it now) Maybe somewhere NIR and need to calculate, but certainly not when calculating the illuminance, CCT of ordinary LED lighting sources. I’m sorry, but this is definitely a f…p of your company’s engineers.
Usually IR filter is a bluish transparent plate.
Secondary optics usually change the CCT at different points in the light spot measurement. Total light loss is also present. It is impossible to tell everything in a few words.
I think I know what you’re up to. Yes, domed LEDs can show a tint shift, and the spectrum will not be the same everywhere in the beam. But does this really matter? Everywhere in the beam, the spectral distribution should be fitable to a basic LED spectrum. If one measures e.g. the spot, he’s interested in the characteristics of the spot.
This is pretty disapointing.
There is continued discussion about spectra, and i can understand that it isnt for halogen lights, but if it cannot measure the led in a flashlight then its highly questionaly it can measure a led in a home light. The variations in spectra across home lighting leds is just a smuch as that across flashlights.
Opple looks to be putting this to market as an OPPLE measuring device. Thats fine, but that should be stated. As mentioned previously, if its not reliable on flashlights imthen it wont be reliable on other manufactures home lights either. In fact the only area it would be accurate is if it is tuned specifically for the led chips opple puts in its products.
Bottom line, the LM3 has been shown to be accurate across a WIDE variety of products, and yet its sensor has fewer bands. The lm4, even with a superior sensor, is failing to measure light properly. If it isnt reliable on flashlight leds then it cannot be considered reliable on home lighting leds.
3 Thanks
LM3 does not measure R1~R14, let’s take Ra as an example, I think LM4 is more accurate than LM3 in measuring home lights, you can do experiments to verify that. I heard them saying phosphor formulas are diverse across even Led products, but not that diverse in terms of home lights(similar phosphor formulas).
LM3 appears to be a great product, that’s great. but is this because LM3 shows less metrics(especially R9, not easy to calculate accurately), so it gives you illusion that LM3 outperform LM4?
LM3 has four engineers working on it(except smartphone app), and the iteration cycle is every 2 years.
I think I can’t elaborate more.
No, saying the LM3 has less capability is correct. However the LM4 does not accomplish even as much as the LM3.
We are not comparing the LM3 to the LM4, we are comparing the LM4 to the LM3, and it is less accurate than the LM3.
These are not “flashlight” leds, they are not designed or sold for this purpose. These leds are used in a wide variety of uses, including home lighting. I dont think there is a led manufacture out there that designs or markets a LED just for flashlights.
1 Thank
without verification I partially agree with your points.
if I now possess 10 flashlights I will now go downstairs and do testing now.
however I don’t have flashlights, that’s a sad thing. I searched on e-commerce sites, it’s kind of pricey.
hope the war finish as soon as possible.
stay safe.
Unfortunatly standard cheap flashlgihts may not help much. Those typically use budget bin leds. The lights we use and discuss here generally use much higher quality led emitter chips (hence why they are for other uses). They are however a bit more expensive and imo start at $20 USD and go up in price from there (and not saying any $20 light would be the same).
You are talking lights like olight, raylight, emisar, fennix, convoy, and a whole bunch of others. As long as it uses cree, osram, nicha, samsung, and other “name brand” emitters and although there are fakes of those emitters those are probably of interest as well.
Point being, standard cheap lights likely have poor or wildly different specrums than cree xml, osram cslnm1, or sst-20, 219, 519a, etc. Those leds themselves also have various specteums, some of which have been analyzed here with results.
More connisour than consumer
Convoy S2+ flashlight with different LEDs:
Nichia 219B 3000K 90+ CRI
Nichia 219B 3500K 90+ CRI
Nichia 219B 4500K 90+ CRI
Luminus SST20 2700K
Luminus SST20 4000K
Luminus SST20 5000K
Luminus SST20 6500K
Luminus SST40 5000K <70 CRI
Luminus SST40 6500K <70 CRI
There are also S2 flashlights with different Cree and Osram emitters.
Any particular drivers with good regulated output?
1 Thank