From the headline I assumed that technicians were literally able to smell whether a chip is new, possibly by it releasing gases shortly after production. It reminded me of the distinct smell of new hard drives when you open the factory sealed pouch.
I was slightly disappointed when I found that the article describes software based testing instead.
Funnily enough, all diodes actually are light emitting, and light receiving, to some extent. It's why diodes are generally in opaque encapsulation. Whether a PN junction is an LED, PV cell or rectifier/switch is a matter of what the design optimises for. Try putting a voltmeter across a glass encased signal diode and shining a strong light on it to see it act as a PV cell. One wouldn't normally use a diode as a PV cell, but if the desired power was very small, a glass encased diode might be used as a cheap PV cell.
Of course chips don’t work if the smoke escapes, internally they operate on smoke. We tested this with a chip socket connected to 120VAC. All the chips that we connected released smoke, and none of them ever worked again.
From tinkering with electronics my brain has imprinted the smell of various components (resistors, capacitors), isolation. I often wonder what effect inhaling those fumes had on me, can't imagine it was a positive one.
It's, uh, a reference to "new car smell" -- the combination of fragrance and volatile outgassing from new plastics that only lasts for a few months after manufacture. Even if you missed the social reference you got the idea: "new hard drive smell" is surely a physically similar effect.
I understood that part of the reference, but "tech" is ambiguous. If it meant "technician", then it would imply to me that the "new memory smell" was both figurative and something that a human nose could detect, in a literal sense.
Reading the article makes it clear that "tech" refers to "technology", making it more likely that the "smell" is purely figurative.
One of the little-known fascinating facts about standard NAND flash is that it contains its own tiny microcontroller and internal oscillator, usually a custom 4-bit architecture, to manage the command interface and do the program/erase algorithms. If you're interested in this stuff, the book Inside NAND Flash Memories is highly recommended, and may be the only publicly accessible documentation of the instruction set of one such microcontroller along with some interesting snippets of the actual source code of the NAND flash firmware.
That book sounds really interesting. Shame about the crazy price (but I suppose they only expected a handful of readers...).
Curiously, if you connect a microSD card to an Arduino, there's more compute power in the card than the Arduino (typically an ARM micro controller or something of similar complexity to handle all the protocols and wear leveling that microSD cards support).
Seems really useful, but if my knowledge of flash memory is accurate, this testing process could highly degrade the memory you run it on. Due to quantum tunneling, erasures are very hard on flash memory, often the limiting endurance factor. That is one method mentioned for testing the chip in the article. Writing all 1's then erasing to all 0's multiple times to see what percentage are correct. On an already dated device this could be catastrophic!
A few dozen erases for testing are inconsequential. They are meant to be used.
Old chips are rated for 10k erases. Newer is 100k+. Only the most ancient parts have lower wear limits and nobody designs them into new products. The 10k parts are all going EOL now.
Yes, endurance and retention decreases exponentially with increasing bits per cell and is correlated with the capacity process size too --- higher density NAND made on a smaller process obviously has less area to store a charge in the first place, and the smaller gate oxide means less endurance. In the "good old days" SLC was typically 100K cycles and MLC 10K, now even good SLC is around 10K with MLC at 5K, and the newest TLC is ~1K or less. Manufacturers are, not surprisingly, increasingly secretive about the decreasing reliability of their products. It's only with complex error correction that such chips can even be used at all.
I doubt they're writing the whole chip—instead, a statistical sampling seems sufficient.
And wouldn't most controllers already handle accumulating errors, including any introduced by the testing process? I understood that a portion of flash storage is set aside for redundancy.
I don't think that is a safe assumption. While good controllers do wear levelling, there are many that have less capable wear levelling. Also, even with reserve blocks, having a constantly full drive can make inferior controllers cause abnormal wear.
This is probably less of a problem even with lower end controllers today, but let's not forget the early Sandforce controllers... those chips could be part of the recycling process.
Until I saw 'chip', I thought this is about flashbulb memory (https://en.wikipedia.org/wiki/Flashbulb_memory) and made a mental note to myself not to 'take' the test under any circumstances.
This is almost a shame. If the memory is well below it's expected life then it seems like a good thing to reduce electronic waste by recycling. It seems like detection would only discourage this.
On the other hand, this enables one to reliably grade the recycled chips. For example, you could use grade A for low cost servers, grade B for consumer computers, grade C for phones and tablets, etc. Before this, you'd have no indication of how much a chip has been used and maybe would be more reluctant to reuse it.
I have a problem buying a product marketed as new and it turns out to be recycled and inferior to new (lower expected life). I have no problem buying something knowing it's recycled and has tradeoffs, since I can buy it in the times the tradeoffs sync with my requirements.
It's a shame if a manufacturer uses recycled parts and sells the device as new, too. I might choose to buy a device marketed as containing some used parts, but I'd like to know what I'm buying.
otoh this may actually increase recycling rates, because of reliable indicators of age. like with cars, lots of people probably buy new just because it's easy to tell if something is used and not new, but not whether it is 5% used (like they claim) or actually 50% used.
It should discourage it. I don't want to buy new products and have them turn out to be recycled. Obviously, a used/recycled/refurbished product has unpredictable reliability.
This was already known in the flash industry for a long time. They characterize the NAND to a great degree and know how the behavior changes as the flash cells wear down.
> Older flash, even if it’s only gone through 1 percent of its lifetime, will reliably have more of these flipped bits than new flash will, his team discovered.
Creatively ironic use of "reliably" duly noted. :)
I was slightly disappointed when I found that the article describes software based testing instead.