Comment by nolok
1 day ago
When it comes to industrial manufacturing, a think of lot of people are not realizing (by lack of education on the matter in general knowledge or schooling) the difference between levels of manufacturing, the precision required for some things, and how the hard part is having the full chain (making the tool that can make the tool that can make the tool that ...) because you can't jump from nothing to milimeter precision.
Also known as "why did China who already owned world manufacturing insisted and struggled on making ballpoint pen until 2017", "why are car manufacturers not making random cheap cars that have the curbs of beloved sports cars", "why are barely 5-6 countries able to make decent jet engines" and all that.
Manufacturing is hard. It's built upon layers and layers of deep knowledge and abilities. And when don't have it or you lose it, just knowing how to make the last layer is not enough, you need to rebuild the entire stack.
Which in this case becomes "painting something black is easy, making a fan black is easy, making a high quality high precision fan black from the starting point of the same fan in another color is an industrial challenge".
We are so used to high quality high precision manufacturing, we have a bazillion factories pumping out millions of very high tech things for random usages or tools now and we stopped noticing it ... And then someone makes a small mistake and you get a "Samsung Note 7 explodes randomly" because of a margin of error small than what our brain can easily comprehend.
(I did a couple months of industrial engineering in university and while it wasn't for me, I loved what I learned about the field)
This analysis is missing price though.
A lot of times it's cheaper to just full send it than produce a full run at a given quality with a low rejection rate.
The "old" way of making a black fan is you just QC check them, send the good ones to Noctua, send the crappy ones to someone who DGAF because they're putting them some sort of industrial appliance that needs airflow through the box.
Everyone "wins" this way because Noctua gets their fan to spec cheaper and the people building plasma cutters or control units for chemical washers or ATMs get a fan that's "fundamentally good" if sloppily executed and the manufacturer gets less waste. Ain't no different than how the pork belly that doesn't become your bacon becomes dog food and die lubricant.
I suspect this is where a lot of the "X compatible" power tool stuff on Amazon comes from. That and/or the repurposing of "worn out" dies.
Yes you provide a great example of binning and market separation. Though I think in this case there's some limiting factors that make it infeasible to bin these fancy Noctua fan rotors including: 1) tooling have limited lifetime and will get sloppier and worse yields as time goes on. It's inefficient to use precious cycles of a precise tool and die on producing lower grade parts. 2) the material itself is likely more expensive than what industrial/lower grade use cases require. Why use reject Noctua when you can get regular crappy plastic for 1/500th the cost? 3) I expect Noctua stuff to be a much lower volume than lower cost/quality vendors so the volume of Noctua rejects is likely too low for a company to dedicate a product line using it. 4) brand/marketing reasons
Another obvious use case of binning is for microchips where the same die can be "wounded" to create multiple product variants that target different market segments, and also yield improvement from being able to isolate and disable an area of the die that are defective. However improving the manufacturability and yield itself is still fundamentally important