As this is a theme here, I wonder how private persons/buyers could backcheck on the binning of their racks generally?
Are diodes marked with something like a production stamp or comparable?
Hi Kassiopeija yes the bins are marked on the reels when you receive them. There will be a model identifier with the CCT (Kelvin temperature, including colour shift), voltage and luminous flux bins marked on the sealed package, as well as the reel itself. This is the case for all the LEDs we purchase from Nichia, Seoul Semi, LED Teknik and others. It is possible for a manufacturer to mislabel a reel and there is no way of knowing if the individual LEDs inside are what the manufacturer says they are, but you can find out through individual testing.
When we design a LED panel we will calculate on paper the total efficiency of the combined LEDs to get to where we think it should be. We will then produce and test a board to see if it matches our calculations. We will also swap LED types over and test those to see the difference in efficiency between individual diodes types. It is an exhausting process!
We love Nichia because it is one of the few companies that under-promises and over delivers on its LED efficiency bins. We always get what we pay for and sometimes are surprised that the LEDs are slightly more efficient than promised. I will go a bit further into this below.
Also, my understanding (which could be mistaken) is that they are tested for binning AFTER they are manufactured as part of the quality control process. At that point, depending on how much luminous flux is produced, they are binned accordingly. So perhaps they are not individually stamped.
LEDs are not individually stamped but they are individually binned after production. Each batch goes into a type of centrifuge which spins the LEDs around so that they come out one at a time on a conveyor where they move over an electrode which briefly pulses the LED and the subsequent voltage, flux and CCT are recorded. The individual LED is then directed to a plastic bin (which is why it is called "binning") depending on the requirements of the customer. Binning often happens after you place an order so that desired bins can be grouped together on reels to the customer's requirements. Leftover LEDs may be sold in batches to the retail market where they have wider tolerances or "steps". This is common when colour accuracy is required and the CCT will be binned in steps of how far away each LED is from its locus line or "ideal" colour for each Kelvin temperature.
When we purchase LEDs we always voltage match to within 0.1V and we also pay a premium for the highest flux bins available due to the extra steps of having to bin each reel to our requirements. Because we are not as interested in true colour rendition (there are only minor fluctuations in CCT anyway) as we are in efficiency and voltage, then this makes it easier to purchase "top bin" diodes.
There is a lot of confusion over what a "top bin" diode is. In LED manufacturing everything is based on yield. For example when a batch of 1 million LEDs is produced there will be many parameters under which each LED will fall due to slight variances in manufacturing tolerances. You may have three different voltages each 0.1-0.2V apart. Any LEDs that fall outside those three voltage bins may be discarded. Most of the LEDs will fall under one of those voltage bins, known as "typical voltage" and so will have the highest "yield". It is easier and more economical to order the most common voltage bins.
Efficiency bins are tricky. There is really no such thing as a "top efficiency bin" diode. You would expect the diodes with the highest flux to be the most efficient but this is not always the case. A diode that falls into the top flux bin may have a higher voltage, so it may not be as efficient as a diode that is in a lower flux bin but also has a lower voltage. Flux is a measurement of human eye sensitivity and not total number of photons, so it is possible to have two "top flux bin" diodes and the diode that has more red shift will produce more photons and have a higher umol/j efficiency.
Now we get to the interesting part. Out of 1 million diodes perhaps only 10% of those diodes are in the top flux bin. Of those perhaps 60% are in the typical voltage range, 30% are in the highest voltage range and 10% are in the lowest voltage range. Those last diodes that make up just 1% of the total yield are the most efficient (highest flux, lowest voltage). But even those will have some variance depending on colour shift. It may be that only 0.1% of those 1 million diodes (1000 in total) are "true" top bin diodes. These diodes will have the highest flux, the lowest voltage and the most red shift. But all these things are contradictory as higher voltages normally produce a higher flux and higher green and blue content also produces a higher flux.
So you can see how hard it is to buy "top bin" diodes based on yield alone. Now we have to compete with other buyers, so often you must pay a premium for what you want. Especially if everyone else wants them!
Everyone thinks Samsung diodes are the best but there is one big problem with that thinking: the more manufacturers who use Samsung diodes, the less "top bin" diodes there are to go around. I would suggest that companies that have a good relationship with Samsung, or those who pay the most money, would have first access to the most desirable LEDs. But it also depends on what LEDs are most in demand – typically 3000K CRI80 3030 LM301B/H etc. You just know that any manufacturer that uses these diodes is competing with all the other manufacturers and there simply is not enough yield to go around. Some have to settle on lower flux bins or mix their voltages.
Grow Lights Australia has a good relationship with Nichia (which are better than Samsung, despite what many people believe). We also use diodes that are not typically in demand. One of our favourite diodes is the 757 series 2700K CRI90. These diodes have a lot of red and far red in them and we always get the top flux bins because we are one of the few buyers (and because we also pay a little more for them). These diodes have been tested at 2.77 umol/j which is very high for a 2700K CRI90 diode. We also pay LED Teknik a premium for our 3535 series diodes which is why we have 660m monos that are just 2% off the very top "G Bin" Cree XP-G3 diodes. The problem with the Crees is that you cannot buy the "G" bins anyway. There is not enough yield to go around.
I know this is a long post but I hope it explains a little bit about how the industry works and why there are really very few manufacturers out there who have access to "top bin" diodes even though everyone claims they do. It is simple mathematics.