The Dark Side of Buying Electronics: Adata's Bait and Switch Scandal Exposed
If you've ever purchased an electronic device from online marketplaces like eBay or Ali Express, you may have come across a product that doesn't quite live up to your expectations. You might think it's just a coincidence, but what if I told you that the same thing is happening on reputable websites like Newegg and Micro Center with trusted brands like Kingston, PNY, and Adata? Yes, you read that right. These well-known brands have been accused of releasing new products, only to silently downgrade them after the fact.
This practice, known as bait and switch, can be devastating for customers who are eager to upgrade their devices or try out a new product. The manufacturer sends out a wave of positive reviews from media and early enthusiast customers, making it seem like the product is performing flawlessly. However, for everyone who comes later, they have no way of getting the original product to compare, leaving them with only "mileage may vary" as their review. This approach can be misleading and unfair to customers.
In March this year, I put out a bounty for a particular drive, the Adata XPG SX8200 Pro, offering to pay our community members for testing and reviewing this device. We received several samples from different vendors, each claiming to be the original product. However, as we dug deeper into the packaging, specifications, and even the controller variants used in these drives, we discovered that something was amiss.
Out of the five drives tested, only two used the faster EN variant of the controller, while three others used the slower G variant. But what's more disturbing is that all five drives used not just different NAND flash but different NAND flash from completely different vendors. This raises questions about the performance and endurance of these drives. Without hundreds of samples, we can't perform any kind of meaningful endurance testing, but we can validate performance.
The results were surprising to say the least. The three drives using the slower G variant failed to meet the advertised sequential read and write speeds, while our EN variant samples came very close in CrystalDiskMark. However, when we ran PCMark, which is a benchmark that more closely simulates real-world drive use, the differences became even more apparent. The EN variants managed higher peak speeds for both reads and writes, but their overall performance was actually measurably lower.
What's even more astonishing is that when we reran the test with all of the drives, three-quarters filled with data, the EN variants took a more noticeable performance hit between seven and nine percent compared to less than one percent on the G drives. This shows that the controller swap has a significant impact on real-world performance.
Now, this is likely due to the SD card industry's approach to NAND flash procurement. The SX 82 pro launched right around the time that 96-layer flash was making its way onto the market. There's an RND cost associated with these transitions, but adding more layers generally lowers the cost per gigabyte, incentivizing this kind of development.
However, NAND pricing is subject to supply and demand pressure like any commodity. If a major player like Apple books a huge order for 96-layer NAND, it can drive up the price forcing smaller players like Adata to buy older 64-layer tech or whatever's available. This practice isn't new; Samsung uses both Snapdragon and Exynos on flagship phones, while Apple was sourcing modems from both Intel and Qualcomm simultaneously.
This approach can alleviate supply chain issues and force your vendor to stay cost-competitive if they know you can just buy it from someone else. AMD famously pulled a bait and switch with one of their products, the Ryzen 5 1600, by quietly moving it from 14 nanometers to 12 nanometers, giving customers a slightly better version of what's fundamentally the same product at a great price.
However, Adata doesn't appear to have followed either of these rules. In their statement to Tom's hardware, they said that the changes were "meant to improve performance and efficiency." But we found that this approach is misleading and unfair to customers. They didn't disclose the changes or ensure proper validation to distinguish between the two versions.
The consequences of such actions can be severe. It undermines trust in the industry, and customers may feel cheated out of their hard-earned money. As consumers, it's essential to be vigilant and do our research before making a purchase. We need to hold manufacturers accountable for their actions and ensure that they follow best practices when it comes to product development and transparency.
In conclusion, Adata's bait and switch scandal serves as a cautionary tale for the electronics industry. It highlights the importance of transparency, proper validation, and accountability in product development. As consumers, we need to demand more from our manufacturers and hold them responsible for their actions. Only then can we build trust back into the system and ensure that customers receive the products they pay for.
"WEBVTTKind: captionsLanguage: en- If you buy electronicson marketplaces like eBayor Ali Express, you caneasily end up with a productthat doesn't live up tothe performance or qualitythat you would expect.But what if I told you that the same thingis happening right nowon reputable websites,like Newegg and Micro Center,and with trusted brands likeKingston, PNY and Adata.That's right.All three of them have been accusedof releasing new products,then silently downgradingthem after the fact.It's the perfect crime, right?You get a wave of positive reviewsfrom media and earlyenthusiast customers, right?And for everyone who comes later,well, they have no wayof getting the originalproduct to compare.So you can just hand wave itaway as your mileage may vary.Right?Wrong.In March of this year, I put out a bountyfor a particular drive,the Adata XPG SX8200 Pro,offering to pay our communitymembers three times the MSRPif anyone could ship me aunique variant of this drive.I mean, how many could there be, right?Two, three, eight?Wow. That's a lot.And we've got five of themto answer three questionsonce and for all.Our manufacturers, bait andswitching their customers,why would they do that?And does it really matter anyway?Oh yeah. There's one more question.Who's our sponsor?Honey.Honey is the free to useshopping tool that helps searchfor some of the best promo codes on lotsof your favorite websites.Get it today@jointhoney.com/ltt.(low-pitched music)(upbeat electrifying music)On paper, the XPG SX8200 Pro is oneof the top performing PCIexpress GEN3M.2 drives.It's rated at up to 3500megabytes per second readsand up to 3000 megabytesper second writes.And in third-partyreviews, it managed to meetand sometimes even exceedthose lofty claims,whether sequential speedsare a meaningful measureof SSD performance isa separate discussion,but Adata used it as thestandard to measure this productand they delivered on it,except for one small problem.If you go to the store todayand pull an SX8200 Pro off the shelf,you might not actuallybe getting one of these.It ended up being a Reddit userby the name of Svarchimpans,who blew the whistle on this.They purchased three 88SX8200 Pro two terabyte drivesat different points in timeand found that a brand newdrive purchased in 2021was performing a boat aswell as a 94% full drivethat had been in use for over a year.Big yikes, as the kids say.Of course, just becausesomething was on Reddit,doesn't make it accurate.And in fact, Adata pointed outthat the user appearedto have connected oneof the SSDs directly toCPU controlled PCIE laneswhile the other was runningthrough the motherboards PCH.That actually could accountfor the observed performance difference,but I'm not fully satisfied.Let's go through all ofour different variantsthen and get them naked,starting with whatAdata sent to the media.There we go.Now there are threemain hardware componentsthat affect SSD performance,starting with the controller.This acts as both the brainand communications hubof the drive and better onesuse smarter storage algorithmsthat have a profound impacton both performance and endurance.Next, we've got the NAND flash,where all your data is actually stored.Fast high endurance NAND meansa fast high endurance drive.And then finally, we'vegot our DRAM cache.This holds data temporarilyso that it can be writtenin a way that, okay,this is becoming a bitof a pattern here,makes the drive faster and longer lasting.And this right here, thisis a good combination.Silicon motion SM 2262 ENG controller,4x256 gigs of IMFT 64layer TLC NAND flash.It's a total of one terabyte,and two packages of 512megabytes of Nanya DDR3L.Hold on a second.Given that speed and enduranceare the only things reallyseparating a premium SSDfrom a cheap one, it stands to reasonthat changing even one of thesecomponents has the potentialto turn this into an entirelydifferent class of product.So, what did Adata change?I'm so glad you asked?Everything.Thanks to a tool developed specificallyfor Silicon motion controllers,we were able to generate this tablewith the full actual specsof each of our five drivesand to summarize what'shappening here comparedto our original.Remember the one sent to the media,not a single drive had morethan one of these three componentsI just talked about that matches.Only this one purchased in July of 2019,even uses the same ENGvariant of the controllerwith the other three usingthe slower G variant.None of our samples used thesame DDR3 as our original,although I don't reallyexpect that to be a problem.And all five of them use notjust different NAND flash,but different NAND flash fromcompletely different vendors.That's right.Out of the top six NAND flashmanufacturers in the world,we have five of them represented here,which doesn't matter as longas they all meet the same barfor performance and endurance.But the question is, do they?Unfortunately, withouthundreds of samples,we can't perform any kind ofmeaningful endurance testing,but we can validate performance.And let's just say somediscrepancies were found.Right out of the gate,the three drives usingthe slower G variantof the controller failedto meet the up twosequential read and writespeeds advertised on the boxwhile both of our EN variant samplesat least came very closein CrystalDiskMark.But let's see if thesedifferences hold up in PCMark,which is a benchmarkthat more closely simulatesreal-world drive use.Now, this is interesting.Much like in the synthetic benchmark,the faster clocked EN variantsmanaged higher peak speedsfor both reads and writes,but their overall performancewas actually measurably lower.And when we rerun the testswith all of the drives,three quarters filled with data,the EN variants took a morenoticeable performance hitbetween seven and a half to 9%.That's compared to lessthan 1% on the G drives.So in the case of the controller swap,you're likely to end upwith better performancein the real world, butthat doesn't explain 88as loot box approach toNAND flash procurement.Now, this one is more likely downto pricing and market availability.The SX 82 pro launchedright around the timethat 96 layer flash was makingits way onto the market.There's an RND cost associatedwith these transitions,but adding more layersgenerally lowers the cost per gigabyte,incentivizing this kind of development,except that NAND pricing is subjectto supply and demandpressure like any commodity.So if apple, let's say booked a huge orderfor 96 layer NAND tightening the supply,that could drive up theprice forcing smaller playerslike Adata to buy older 64 layer techor whatever's available,all of which can be fine,just like you'll look finein that stealth hoodiefrom lltstore.com.The thing is sourcing componentsfrom two or more suppliersisn't a new strategy,and sometimes it's a necessary one.Samsung uses both Snapdragonand Exynoses on flagship phonesand Apple was sourcing modemsfrom both Intel andQualcomm simultaneouslyat one point while theymade the transition.This practice can alleviatesupply chain issuesand it can force your vendorsto stay cost competitiveif they know you can justbuy it from someone else.AMD famously pulled a bait and switchwith one of their products,the Ryzen 5 1600 a couple of years ago.But in that case, what theydid was quietly moved itfrom 14 nanometer to 12 nanometers,giving their customers aslightly better versionof what is fundamentally thesame product at a great price,no harm, no foul.It's just that the two rules to follow arethat proper validation is needed to ensurethat both parts areindistinguishable to the end-userand failing that thechanges must be disclosed.Adata doesn't appear to havedone either of those things.In their statement to Tom's hardware,they said, \"We do guaranteethat every SX8200 Prowill perform to spec withinthe regular variationssince the drives havebeen tested stringently.\"But while that may be technically correct,anything between zero and3500 megabytes per secondis technically up to 3500megabytes per second.It also has the potentialto be extremely misleading.There was a similar scandal back in 2014when Kingston and PNY were called outfor the same bait and switch tactic.Actually, as we werefinishing writing this video,new surface that PNY had allegedly quietlyreduced the accelerate CS3030's enduranceby nearly 80% on their500 gig one terabyteand two terabyte models yetthere was no revision numberor announcement and it's still labeledon shelves as the same drive.That means that customers areleft without the proper toolsto make an informed decision.Now these vendors might say,who cares about a 10%performance difference,or it's not our fault.It's down to the inconsistent supplyand it's unmanageableto make a new skill everytime there's a change.But my counter to that is A,people pay a lot of moneyfor 10% more performancein some cases.And B, if you can't bearsed to test and labeland get the specs rightfor your own products,maybe you need some oversight.I mean, this is whyorganizations like Vesa exist.Prior to Vesa's displayHDR certification program,HDR ready was casually thrownaround without any regardfor the level of experiencethat product would actually deliver.Now, the current systemain't perfect either.I mean, display HDR 400shouldn't even exist,but at least that label meansthat you can trust that productto meet a certain performance standard.Camera storage labels,like U3 and V 90 on SD cardsserve a similar purpose.They give consumers a quick and easy wayto ensure that their mediawill handle the speedstheir camera requires.Now, this hasn't really been an issuewith SSD boot drives up until nowbecause the main competitionfor them has beenfrom hard drives, but asgenuinely faster SSDs becomea requirement for gamerswho wanna take advantageof features like direct storage,it would be really niceto see someone hold theseguys accountable finallyfor playing so fast and loosewith their product specs.Oh, wait, I guess that's us.SSD manufacturers, stop it.We see you, all of us.You see that view counter down there,everyone knows about this nowand if you guys don't cut it out,we're gonna be stuck tellingeveryone to just stickto vendors like Samsung, whoare vertically integratedand don't randomly change suppliersfor literally every corecomponent of their product.Just like we tell everyone to stickto buying technology andcomputer hardware at Micro CenterMicro Center has 25 locationsacross the United Statesand you can check out MicroCenter's custom PC builderto spec out the best PC for your budget.They'll help you ensure allyour parts are compatible,find stock available at yournearest micro center location,and once you add it to your cart,you can arrange same day in store pickup.For a fee you can check the boxmarked same day pro assemblyand Micro Center's expert technicianswill assemble your PC for you.If you want help deciding what partsto put in your new custom gaming PC,join the online Micro Center community.It's a great place to discusstech with other enthusiasts.So don't wait, check out thelink in the video descriptionto learn more about Micro Centerand get a free pair ofwireless Bluetooth headphones,valid in store only,no purchase necessary.If you guys enjoyed this video,maybe check out the\"Linus was right\" video.It's not what it sounds like.It's the other line is,we're talking about ECC memory offer.(man laughing)\n"
