MicroSD options, including MLC cards

You've definitely got something against Lexar mate as there are loads of sites in Korea advertising the 64GB as MLC which are far from eBay as you can see from that previous link I gave, this thread needs a sponsored by Transcend logo on it :)
 
I have nothing against Lexar and pointed out they're one of the best non-MLC options out there. But according to Lexar they're not MLC. Going by what Lexar says is not promoting others.
 
I have nothing against Lexar and pointed out they're one of the best non-MLC options out there. But according to Lexar they're not MLC. Going by what Lexar says is not promoting others.
Lexar used to say that the 633 cards are all MLC until they released the 128GB, after that they refused to say. Given the different access speeds for the 128GB it seems very likely that the 128GB is TLC and the rest are MLC. Unless you have some real evidence against that conclusion? In terms of lifetime in a dashcam, that means that the 128GB card will last aproximately the same length of time as the 64GB so is not worth paying much extra for.

Also, nobody else seems to have been able to make an MLC microSD card bigger than 64GB. For a long time now we have been stuck at 64GB for MLC and 128GB for TLC. Sandisk did manage to squeeze 200GB of something onto a microSD card, but their cards are all of questionable reliability.

Samsung now have a 256GB microSD card using 3D NAND and the rest of the manufacturers will probably follow this year or early next year. I think development has been delayed a bit because all the 3D NAND is being used for solid state disk drives where it is getting well tested and prooven to work. Presumably Samsung will release a larger card late this year and reduce the price of the 256GB to a more reasonable level, eg 2x the cost of their 128GB card. They already have SSD drives using 3D NAND with 50% higher stacks thus allowing 384GB in the same size chip as their current 256GB.
 
No love for Sandisk High Endurance micro sd? Packaging states they're meant for security cams and dash cams and I have a couple of them running in 2 different dash cams so far with no issue. Price on amazon for the sandisk is quite a bit lower compared to similar size Transcend high endurance 64GB ($27 vs. $39) and hopefully that's not due to lower quality component being used.
 
SanDisk high endurance cards are fine, the lower price currently is more likely due to when they were manufactured rather than component quality, flash price is way up at the moment
 
SanDisk high endurance cards are fine, the lower price currently is more likely due to when they were manufactured rather than component quality, flash price is way up at the moment
But in your opinion would you still go with Trascend over Sandisk? I also need to know as I have several security cams that can take micro sd as local backup and also need a memory card that's as reliable as possible.
 
But in your opinion would you still go with Trascend over Sandisk? I also need to know as I have several security cams that can take micro sd as local backup and also need a memory card that's as reliable as possible.

personally we use Transcend as we have a stable supply channel direct from the manufacturer, it's what our engineers use for testing and debug also, I think the SanDisk high endurance cards are probably fine though, Sandisk cards are generally consistent in quality but some models are not suitable for dashcam use at all and should be avoided (the red/grey ultra cards in particular)
 
Given the different access speeds for the 128GB it seems very likely that the 128GB is TLC and the rest are MLC. Unless you have some real evidence against that conclusion?
That's not how that works. If Lexar claims it, then that's good enough for me. They don't, and there is zero evidence that they're hiding MLC and selling it for TLC prices. And no, even if there were some speculation about why it would be so, they do not claim any difference in access speeds. But regardless, because they do not make the claim, any other conjecture is pretty much irrelevant. No one is stopping you from buying what you want based on beliefs rather than verifiable facts.

Samsung now have a 256GB microSD card using 3D NAND and the rest of the manufacturers will probably follow this year or early next year.
I thought Samsung had an exclusive patent on their version? I could easily be wrong about that, or they could be licensing the process out more broadly. Regardless it's the most exciting development in this field (at least for dash cams and the like) in a long time. Because there's so little damage done during the writing process compared to traditional NAND, it should in theory make MLC irrelevant, as long as the controller and resistance to the physical elements is sufficient. If they can show consistent reliability in the microSD form factor, then I see no reason why they wouldn't take completely over.

No love for Sandisk High Endurance micro sd?
It's definitely mentioned in the first post, in the list of manufacturers.
 
Last edited:
I thought Samsung had an exclusive patent on their version? I could easily be wrong about that, or they could be licensing the process out more broadly. Regardless it's the most exciting development in this field (at least for dash cams and the like) in a long time. Because there's so little damage done during the writing process compared to traditional NAND, it should in theory make MLC irrelevant, as long as the controller and resistance to the physical elements is sufficient. If they can show consistent reliability in the microSD form factor, then I see no reason why they wouldn't take completely over.
Well Intel's 3D NAND works completely differently, no charge trap technology there, so I don't expect there are any patent issues, you shouldn't be able to patent the idea of a stack!

"it should in theory make MLC irrelevant,"
Maybe you haven't realised that Samsung have two types of 3D NAND, their Evo range uses TLC cells and their Pro range uses MLC cells, just like their planar NAND chips ;)
 
I picked up one of the Adata cards from today. It was $15 USD for 32GB. Comes in a super shiny package and advertises MLC, Dash Cam use, Lifetime Warranty, 95 MB/s read, 50 MB/s write. Any suggestions on how to test a micro SD so I can write a proper review?

20170107_172500.jpg
 
Last edited:
I picked up one of the Adata cards from today. It was $15 USD for 32GB. Comes in a super shiny package and advertises MLC, Dash Cam use, Lifetime Warranty, 95 MB/s read, 50 MB/s write. Any suggestions on how to test a micro SD so I can write a proper review?

You can test speeds with crystalmark and compatibility by just running a full cycle through a dash camera. However you won't be able to do any meaningful reliability testing with just one card as it's just a single data point. You'd probably need 30+ cards from different batches being tested in the same camera in order to show anything meaningful. I'm making up those numbers as I don't know how much it would take to be statistically relevant but it's more than one.
 
Maybe you haven't realised that Samsung have two types of 3D NAND, their Evo range uses TLC cells and their Pro range uses MLC cells, just like their planar NAND chips ;)
Samsung claims no such thing. If you have new info, please link to the source. And Samsung's V-NAND is neither MLC nor TLC. It is completely different.
 
You can test speeds with crystalmark and compatibility by just running a full cycle through a dash camera. However you won't be able to do any meaningful reliability testing with just one card as it's just a single data point. You'd probably need 30+ cards from different batches being tested in the same camera in order to show anything meaningful. I'm making up those numbers as I don't know how much it would take to be statistically relevant but it's more than one.
I think that for a review, testing a single example is OK as long as it performs as expected and within it's specifications. Some things that do well in reviews do turn out to be unreliable, but you don't blame the review for that, everyone expects the review to be based on a single example. You can't test in all circumstances that the users will anyway so your reliability data will never be correct.

The only reason I can see for testing a large sample is if you are testing lifetime and there is a large variation in lifetime between identical cards, but I don't think that is the case, I guess they nearly all wear out within +-10% and it probably depends more on the manufacturing batch, which is something you can't test for.

When testing, it is good to do some real life tests, eg copy 64GB of dashcam video files from/to the card onto your computer using the normal Windows copy and paste and time how long it takes, and compare it to the same test for other cards. Also h2testw.exe is a good tool for testing writes since it writes in a similar way to how a dashcam will, although a lot faster and will cover the entire memory capacity. Remember that performance will depend on your computer, operating system and card reader.

"crystalmark" etc may be good for testing performance in all situations but the results they give are often meaningless for actual use, we always read and write our video files sequentially - a very simple test will tell us what we need to know.

Note that the cards do tend to slow down after a bit of use and the wear levelling starts to operate, so re-run the performance tests as the last thing you do.
 
Samsung claims no such thing. If you have new info, please link to the source. And Samsung's V-NAND is neither MLC nor TLC. It is completely different.
Presumably you read every document they have published and listened to every speech they have made before you wrote that? I don't see how you could possibly know that "fact" to be true!
 
I researched Samsung's documentation and release history and found no evidence of it. If you have even the slightest shred of evidence that your otherwise poorly informed claims are in any way true, then link to it or go be a troll somewhere else.
 
@DrekiTech that card boasts some pretty amazing specs for such a low price. If you have some other cards with similar specs then comparison is probably the most useful aspect, as there's no real way to tell with a particular test or another whether it's truly MLC or more importantly in general how well it holds up.
 
@DrekiTech that card boasts some pretty amazing specs for such a low price. If you have some other cards with similar specs then comparison is probably the most useful aspect, as there's no real way to tell with a particular test or another whether it's truly MLC or more importantly in general how well it holds up.

Thanks! Those are great ideas. I'll run some speed tests and throw it in my camera and hopefully it'll work for the next year or more. :)
 
Last edited:
This white paper from Samsung refers to V-NAND and MLC in the same context - ie this tested V-NAND MLC:

http://www.samsung.com/us/business/oem-solutions/pdfs/V-NAND_technology_WP.pdf

I've just seen the price of those 256GB Samsung microSD cards - jeez!
The price of the biggest cards is always way above the rest! Just have to wait for an even bigger one for the price to drop to sensible levels.

They normally refer to the MLC 3D NAND as "3-bit V-NAND":
SEOUL, Korea – August 11, 2015 – Samsung Electronics Co., Ltd., the world leader in advanced memory technology, today announced that it has begun mass producing the industry’s first 256-gigabit (Gb), three-dimensional (3D) Vertical NAND (V-NAND) flash memory based on 48 layers of 3-bit multi-level-cell (MLC) arrays for use in solid state drives (SSDs).
http://www.samsung.com/semiconductor/about-us/news/22659
 
The price of the biggest cards is always way above the rest! Just have to wait for an even bigger one for the price to drop to sensible levels.

They normally refer to the MLC 3D NAND as "3-bit V-NAND":

http://www.samsung.com/semiconductor/about-us/news/22659


Are they using this tech in micro SDs though or are the micro SDs just regular TLC chips? Everything I've read here has referred to SSDs unless I've missed something...
 
Back
Top