Or worse than normal Li-ion numbers. If you’re going to foist off lying batteries, you might as well make them cheap Temu lying batteries.
There are even techniques now to suss out most of the secrets in a microchip these days, and I would say with near certainty, every significant new piece of silicon is analyzed almost to the atomic level. Too much money in it to not.
-
- Decapsulation and Imaging: Specialized labs remove the chip’s epoxy packaging (decapsulation) and use high-powered electron microscopes to image the die.
- Delayering and Circuit Extraction: Technicians remove the metal and dielectric layers of the chip one by one to reveal the circuit design (transistor layout) at each level.
- FIB Cross-Sectioning: Focused Ion Beam (FIB) tools are used to cut precise cross-sections of the chip to analyze 3D structures, such as 4nm process nodes or 16-metal layer stacks.
My ears were burning. It is almost impossible to reverse engineer a chip from silicon, but you wouldn’t need to. You could get most of the information you need by looking at cache sizes, number of cores, and the instruction set.
There have been lots of papers on how to tell if someone put a Trojan horse into a chip - hardware added to allow a bad guy to take it over. They all assume someone at the fab will do it, and as far as I know, no one has ever demonstrated a case of someone doing this. (Putting them into your own chips is something else.) What has been done is putting a new chip on a circuit board.
There was a big legal case like 40 years ago about one company cloning a chip from another company by copying and studying the microcode. Now when someone like AMD makes a copy of an Intel chip it is done in a clean room, with the designers only being allowed to see publicly available information, like the instruction set and the spec sheet.
Battery teardowns (among similar chemistries) aren’t exactly uncommon… e.g. see:
They use CT scanners to look at the construction inside each battery. That same company also does a lot scans for other electronics: Scan of the Month—you've never seen anything like this.
And ifixit does plenty of teardowns: Gadget Teardowns - iFixit
Plenty of third parties also try their best to test batteries in various ways, whether it’s a teardown or running them through accelerated use charge cycles to try to get an idea of their longevity.
The company doesn’t “allow” it; once it hits the market, people are going to do it regardless of what the company wants.
The Verge motorcycle battery supports fast charging and can maintain limited self-discharge over a period of a few days. OK cool. That’s great, that’s a step. But it’s not really addressing the cycle life doubts. They could easily submit entire sample packs for independent testing if they wanted to do that.
From Donut Lab shows solid-state battery pack charging at 100 kW in Verge motorcycle | Electrek :
Today’s test is another step forward, but it still falls short of what skeptics need. A pack-level charging demo proves the cells work together and can sustain high power, but it does not address the most extraordinary claims around cycle life and energy density. Independent third-party testing of a complete pack — not just cells — under standardized conditions would carry far more weight.
Shrug. It’s just not worth jumping on the hype train for. Battery techs routinely get announced and only very rarely make real-world waves. Small incremental improvements year after year from many companies around the world with big research groups, sure, but that’s very different than one dubious company with one untested product making grandiose claims.
If and when this becomes widely available and the testing data supports the claim that this is revolutionary, well, very cool, hopefully it’ll take over like li-ion did. Until then, just… why believe the marketing? It’s just marketing.
This is the sort of claim that, if true, sells itself and the company is an instant mega-success. That they need to go out of their way to sell https://idonutbelieve.com/ t-shirts and such just makes it seem even less trustworthy and more marketing…
How long do you expect it to be until various of these third parties do these tests and release the data, now that the product is reportedly shipping?
Given the successful hype around the product I gotta think they are chomping at the bit to get testing?
I have no idea. If it were the latest Apple release, a thousand YouTubers would be racing for it (eg Macbook Neo teardown or the hundred similar videos). Battery techs are exciting to us nerds and EV geeks, but doesn’t have the same broad consumer appeal (and thus publicity/ad dollars).
It’s probably also grandiose enough of a claim that if any reputable lab were to either try to affirm or debunk it, they’d want to be sure to do a thorough and repeatable job at it so as to not hurt their own reputation.
Hopefully sooner rather than later. I’d love to see a battery leapfrog as much as the next person, I just don’t wait for it with bated breath.
For comparison, LiFePo batteries were invented, in what, 1996? In 2016, an ex of mine was writing an academic paper on them in between D&D games. It took until the 2020s for them to really hit the mainstream, and even today they aren’t exactly ubiquitous yet. Stuff like that takes a while to productize, mass-manufacture, and become cheap enough to become commoditized (and even then, many cells today are fake or exaggerated; see above video).
Of course, they don’t have to do that now. Even the hardware is software based. Stuff like a Rasberry Pi, PC motherboard, router or thumb drive, runs a hidden OS or CPU loader, that can be patched or subverted.
That has nothing to do with revealing proprietary info such as solid state battery chemistry.
Keyword discoverable
In the case of say an Apple M5 motherboard …whatever do you think you can discover about it? The hardware and software is tightly bound and VERY proprietary. Apple goes after anyone leaking a hardware product prelaunch very hard. They will reveal only what they want us to know.
-
Allegations of Theft: Apple alleges that Prosser and an associate plotted to gain unauthorized access to a development iPhone in the possession of an Apple employee, Ethan Lipnik.
-
Method of Leak: The lawsuit claims that Prosser’s associate tracked the employee’s schedule, stole the device, and initiated a FaceTime call to allow Prosser to see and record the device’s contents.
-
Trade Secrets: The phone allegedly contained significant trade secrets, including unreleased software, which were subsequently leaked and turned into re-created renders for Prosser’s videos.
-
Employee Termination: Apple terminated the employee involved for failing to report the security breach.
Order a Verge and have at it.
The truth about Donut’s new batteries’ capabilities is eminently discoverable. I could do it in my garage given a motorcycle to (ab)use.
How they perform the chemical and spatial magic to make those batteries will be a harder problem for folks who are real battery scientists and engineers with real laboratories.
The truth will out in a few weeks at most. Until then we’re all (including me) mostly throwing stones blindly from our fortresses built mostly of bias and innuendo.
What? No. Let’s take this in two parts.
First, battery chemistry is eminently discoverable, as pointed out by @LSLGuy already. Any reasonably advanced battery manufacturer or testing lab has the ability to run every kind of performance measuring tests to fully define the battery operational parameters before moving to destructive testing. They’ll do CT scans, X-rays, they’ll cut them open and do every kind of chemical analysis and ultimately do electron photomicroscopy of the physical structures.
It’s not obfuscation or low observability that protects a manufacturer of a novel product, it’s the patents. And converse to your point, patents are also highly observable, because they have to contain detailed descriptions of the physical properties of the protected devices as well as the operational steps to make them
So, after the research facility has torn these things down to atoms, assuming they are what the Donut folks say they are, the competitors are limited in their ability to create a new product by the patent protection unless as part of their investigation they find alternate, non-protected formulations and processes that result in a similar performing product.
Second, your Mac M5 example doesn’t hold water. Apple went after Prosser for pre-production leaks and theft, not post-release reverse engineering. Once the product is in the market, there’s virtually nothing Apple can do.
As pointed out up thread, scads of YouTubers will immediately do tear downs, with the committed sites building out full spec bills-of-material. And the deep pocket competitors will take that M5 chip for example, delid it, and take electron microscope photos of every layer as well as doing the fancy cross-sectioning and other things I dimly understand.
If a tech product is sold to the public, it’s discoverable. And if it’s interesting enough at all, it will be discovered.
Well, you can’t do a test on a reported shipment.
I expect at least a couple of months of “We’re sorry that we couldn’t ship your product, because we were so unexpectedly overwhelmed with demand that we had to ship to a bunch of people before you”, followed by people getting crappy Temu batteries that fail every test, followed by “Well it’s not our fault, you broke it by doing that, we told you specifically not to do anything that could falsify our claims and that voided the warrantee”, followed by at least a couple of years of litigation.
You only need to worry about that after a battery passes the performance tests. Which it won’t. To date, there have been zero truly independent tests of anything about this battery (a test that only tests the exact things that the company tells them to test, without even verifying fundamental basics like size and weight, is not at all an independent test).
Understand and agree. As I said (or at least tried to say), I can do the debunking tests myself with nothing but a crescent wrench and a suitable battery charger.
If, by some miracle I totally do not expect, the darn thing performs as advertised, then the real battery industry experts will be needed to carry the detailed analysis work forward. If.
Verge is only asking the agency to test very specific data. They are not conducting a thorough analysis of the product.
Actually, this is maybe the biggest red flag. Where are the patents? They probably have a few, but not the biggest ones, because there are still too many things that we don’t know, that we would if there were a patent. Theoretically, not getting a patent could be a legitimate business decision… if you think that your tech is so far advanced beyond the state of the art that you think that nobody will be able to duplicate it for more than 14 years (even with the extremely strong incentives that will be out there for your tech people to leak to other companies). But since that’s pretty much impossible, the only reason to not get patents is because you’re running a scam.
There’s a certain amount of tactical decision-making on timing of patent filings. To late is disastrous, while too early can piss away a lot of first-mover advantage (read “IPO value”).
But for damned sure they need to be filed before the first example of your product is in public hands. Which date is very, very close to now if they’re actually shipping anything but smoke & mirrors. And for something like a battery that’s a component of a larger product, like a motorcycle, you really ought to have those patents filed (not necessarily granted yet) before the first battery leaves your factory for the motorcycle factory.
Moderating:
This really really reads like an unattributed AI output. The SDMB requires that you attribute any text you didn’t write. For AI,a good standard is to say what AI you asked (they have different strengths, weaknesses, and biases, so it matters) and saying what the prompt was.
Please reply to this moderation with that information, unless you actually wrote this yourself, in which case, FSM help you.
FPGAs are chips whose hardware structure is programmed when the board is booted. They usually get loaded from a ROM on board. The programming port is encrypted, but that is another potential weakness.
You don’t want to primarily measure the output of the charger but rather the output of the battery.
So “battery exit” is the gate to watch, not “battery entrance”.
As it could convert a lot of the Chargers output into unproductive heat during charging .
Yes and no.
I agree 100% that the way to know how much energy is in a fully charged battery is to watch what you can get out of it, not what you put into it. And so part of my garage testing mechanism is charge it fully, then go ride the bike until it’s dead. The bike either rode much, much farther than any other comparable battery motorcycle, or the charge capacity claims of the battery are BS.
But one of the key magic claims of this particular battery is that you can stuff electricity into it at heretofore impossible rates. And to test that, we need to have a charger that can deliver insane volts & amps, make that available, and watch whether the battery happily slurps up all that power, doesn’t slurp up much, or cheerfully catches fire. Only the first of those possibilities supports the grandiose claims about charge rate.
I wrote it myself with the exception of 4 paragraphs on the Apple theft.