Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Apple Said to Be Stopping Use of TLC NAND Flash in iPhone 6 and 6 Plus After Reported Issues
- Thread starter MacRumors
- Start date
- Sort by reaction score
That BusinessKorea report is a fraud. It had been proven by TechRepublic's Bill Detwiler (with clear screenshot) that the storage chip inside iPhone 6+ 128G is SK Hynix's SmartNAND/E2NAND (part no. H2JTFG8YD1BMS), not the Toshiba TLC as implied by that Korean media. In fact, Toshiba does NOT have such product as SmartNAND+TLC at all.
Besides, the iOS stability problem (with 700+ Apps installed) is a known issue prior to iPhone 6 launch. It can be tracked back to iOS 6 and iPad 4 128G. Although the real reason is unknown, but definitely has ZERO relation to TLC. That's way too ridiculous.
Fraud stop at the wise.
Besides, the iOS stability problem (with 700+ Apps installed) is a known issue prior to iPhone 6 launch. It can be tracked back to iOS 6 and iPad 4 128G. Although the real reason is unknown, but definitely has ZERO relation to TLC. That's way too ridiculous.
Fraud stop at the wise.
Have you tried restoring them to factory defaults and without any non-Apple apps reinstalled? If the phones are still unstable in stock configuration without any apps that didn't come standard with the phones then you could be fairly certain of a hardware defect.
That cannot be correct: H2JTFG8YD1BMS is a 512Gb (=64GB) NAND. This could be inside the iPhone 6 (or 6+) 64GB, not 128GB.
https://www.skhynix.com/inc/pdfDownload.jsp?path=/datasheet/Databook/Databook_Q4'2014_NAND.pdf
64GB is the largest capacity of E2NAND/SmartNAND category. It's possible that Apple had combined two memory modules inside.
So.. since constantly having Siri on requires more battery, the battery case is essentially for Siri?
I kid, it probably doesn't use that much, but cool to know.
I highly doubt about that.
I have 2 iPhone 6 with me now, 64Gb and a 16GB.
Its more than 10 seconds slower on boot for 64GB to 16GB. I first notice it when I switch to iPhone 6 64GB from a 5S. The installation process for apps are much longer, and the apps loading too. Then when this news hit, I did a comparison between the 2 iPhone. 16GB is noticeably faster like iPhone 5S.
This is just way too ridiculous how Apple is trying to cut cost.
Then the chip would have a different number.
Hmm. I think all the teardowns so far have been of 16GB and 64GB iPhones.
Has anyone even taken apart a 128GB iPhone yet? If so, does anyone have a link?
Thanks!
The slower boot time is well known. But for those of us whose phones are stable it's a non-issue since reboots are seldom needed. I haven't rebooted can mine in weeks. I had a 5s before and have not observed any performance problems either. Did you try what I suggested and restore to factory defaults as a new phone and don't restore your settings or apps from backup. Try just using it in stock config for a bit and see if it's stable. Then add apps one by one.
When we're talking electronic components, the term "slower" is in the scale of microsecond (0.000001). It's not something normal humanity can "feel". If the target is flash memory, the key of performance is write speed per page (a small block of memory, typically 2k bytes,) since the difference of read speed is not significant. Flash memory need to be erased page by page (clear all bits from 0 to 1) and then reprogram (mark bits from 1 to 0), and it takes much much longer. For some flash technology, it can be 300% time slower than others. Even though, the overall performance difference is still hard to detect for normal users.
I don't know the detailed procedure of iPhone boot-up, but IMO it's quite normal if the device needs to check the data integrity of storage file system first. And the kernel also needs to load the services (notifications, privilege authentications, share sheets, configuration panels in system settings, and iCloud drive usage), saved sessions, and preserved memory status prior to shutdown for every Apps installed on the device.
So I guess installed Apps will make a greater impact to the booting performance. And it may also be the real reason behind the 700+ Apps crash: those background services registered by Apps eat up all the RAM space.
Unless you have exactly the same Apps with same configuration installed on both devices, the comparison does not imply anything.
So.. since constantly having Siri on requires more battery, the battery case is essentially for Siri?
Siri always has to listen and analyze what your saying, so yes it uses A LOT more battery. That's probably why they didn't want people to use it that way all the time since people would undoubtably blame Apple for their really crappy battery life.
In the source I mentioned above ( http://www.techrepublic.com/pictures/cracking-open-the-apple-iphone-6/23/ ), Bill Detwiler had noted it as "128G memory module."
Even if Bill had mistaken the model, there is no reason to argue that 128G model may use different memory type.
E2NAND and TLC NAND are two different flash categories. You can't simple swap the E2NAND with TLC chips. You'd need to place extra passive components and controller chips for TLC on the board. E2NAND is the most compacted commercial memory module. Since now the board is designed to host E2NAND module(s), it's not possible to fit TLC (with extra components) into the same pin layout as E2NAND.
If Apple intend to use different memory type for each model, the only thing they can do is to reserve spaces for extra components that host TLC modules. Which means, these component slots are left blank in smaller capacity models. However, we did not find such white space in iFixit 16G tear-down.
In other words, unless Apple had two versions of boards, there is no way to host different memory type in different models. For some obvious reason, it's ultimately stupid to maintain two versions of board.
It's just as bad to use rumor to call something fake, as it was to use rumor in the first place. (It's only a rumor that Apple "insiders" deny TLC as the problem.)
For better or worse, Apple rarely speaks up. So all we do know, is that some people are experiencing a reboot, and nobody yet knows why.
Yep, he corrected it last night.
Actually, there is. Remember the circuit diagram leaks, such as the one that let us figure out the existence of the barometric sensor? There was also this leak, which showed various chip suppliers:
Notice that the 64GB model could be MLC or TLC.
The diagram also showed a 128GB TLC PPN (Apple's Perfect Page New) chip in place:
This jibes with other rumors that the problems can occur on 64GB models as well, if they contained Sandisk TLC NAND instead of Hynix MLC NAND. (Assuming that TLC is the reboot problem in the first place. It very well might not be.)
That's true if we assume that the TLC NAND is raw, without an error correcting controller. We do not know that, because we haven't seen the part numbers for the 128GB phones yet.
Right, if it was raw memory, then there'd need to be two versions of boards. Which is not as "stupid" as it sounds. Apple has made up to eight versions of each iPhone now to accommodate various radio needs around the world. And that's without taking into consideration color and memory variations.
However, because of the schematic above, it seems most likely that any TLC chips being used are pin and interface compatible with the MLC chips.
Is there a way to find out which kind of NAND we have through benchmark apps?
I've tried AnTuTu but it gives only a single "storage I/O" value.
If a benchmark exists with detailed stats (sequential/random read/write), we should be able to understand if there really are differences between different sizes. MLC and TLC should definitely give different performance!
I've tried AnTuTu but it gives only a single "storage I/O" value.
If a benchmark exists with detailed stats (sequential/random read/write), we should be able to understand if there really are differences between different sizes. MLC and TLC should definitely give different performance!
That was another part of the overall rumor.
It was claimed that you could figure out what kind of NAND your 64GB/128GB iPhone had via benchmarks. Which makes sense, as TLC is normally slower than MLC.
I looked for a while, but couldn't find any benchmarks done between the different storage variations.
--
Again, I would note that even if TLC is used, it could very well have nothing to do with the reboots. Nevertheless, it's of interest to find out if TLC and MLC chips were mixed in production, and if so, are the phones with TLC actually slower.
Readers are cautioned not to jump to conclusions.
Sounds like backboard is restarting. Check your Settings->Privacy->Diag&Use Data. Look for JetsamEvent with backboard as the Largest Process or a LatestCrash with backboard.
Pretty sure E2NAND is a Hynix name for their MLC NAND with ECC. That is, an ECC controller, buffer, DSP, etc. wafer with NAND wafer(s) on either side. Their maximum spec is 64GB. The Hynix chip can be found in the iP 5s.
Apple have their higher-level controller for ECC enabled NAND in the CPU. It is in the digital domain, handling wear leveling, FTL and so on. So when the ECC NAND is based on TLC, one can presume extra considerations (read signal processing patents owned by Apple via Anobit) are in the ECC controller wafer of the NAND chip. Noting that the last letter of the iP 5s Hynix 64GB part is different than the part in the Hynix catalog, is it possible it uses an Anobit enhanced version with better signal processing?
We could do it ourselves by comparing our devices.
I'm attaching results from 3 benchmark apps; left to right:
- AnTuTu (storage: 2580)
- PerformanceTest Mobile by Passmark (Write 174, Read 363)
- PhoneDoctor (paid app) (write 280, read 763)
It's a 128GB iPhone 6, black, space grey, serial C39.
Can you provide your results?
iP6 Gold 64GB, Serial
1st try:
- PerformanceTest Mobile by Passmark (Write 35.6, Read 212)
Quit apps:
- PerformanceTest Mobile by Passmark (Write 53.8, Read 337)
Reboot:
- PerformanceTest Mobile by Passmark (Write 46.1, Read 326)
Rut Roh.
Just WOW!
I just retested several times (with open apps, closing apps, after reboot, etc), and my write performance was ALWAYS is the 140-180 range
The Korean site was reporting TLC NAND chips in iP6 64GB and iP6+ 128GB. TLC is probably not the cause for crashes... but for sure your iPhone is almost 4 time slower in writing performance. That would seem consistent with the rumors...
-----
Originally Posted by kemal View Post
iP6 Gold 64GB, Serial F73N9... pre-order, 19SEP Delivery USA Verizon
- PerformanceTest Mobile by Passmark (Write 35.6, Read 212)
- PerformanceTest Mobile by Passmark (Write 53.8, Read 337)
- PerformanceTest Mobile by Passmark (Write 46.1, Read 326)
-----
However, TLC is supposed to have slowness issues with reading, not writing. Perhaps I just have the slow brand of flash.