• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Do all current TLC planar SSD suffer like the 840 EVO did?

taisingera

Golden Member
I am considering an SSD, something lower cost to upgrade an old laptop, and am concerned about TLC. I already owned a TLC Samsung 840 Plain, that I sold 2 years ago. Do all the current planar TLC drives like in this thread, http://forums.anandtech.com/showthread.php?t=2444453 have the degraded read speed issue? Are some of the newer controllers in some of these designed to work better with TLC?
 
pretty much. but most of them (s10 and sm2256 for sure) use rewriting techniques to mitigate this.

so as long as u use the drive, it should be fine.
 
Actually I'm conducting tests on a number of different SSDs to see if they suffer from read speed slowdowns.

There are two combinations of NAND and controller that I've found to suffer from this issue (at least in certain conditions).

And that would be the SM2256 paired with 16nm TLC NAND from either Micron or SK Hynix.

Even with those two combinations read speeds should be ok as long as you do a full read of the SSD now and then to let the LDPC adjust for conditions and improve read speeds.

And while I don't have enough information to say for certain I think it seems that litography isn't as much of a factor regarding drift voltage as it used to be.
 
  • Like
Reactions: cbn
Interesting, so Samsung right now is the only company producing an SSD that does not suffer from this issue?
 
Actually I'm conducting tests on a number of different SSDs to see if they suffer from read speed slowdowns.

There are two combinations of NAND and controller that I've found to suffer from this issue (at least in certain conditions).

And that would be the SM2256 paired with 16nm TLC NAND from either Micron or SK Hynix.

Just wondering if you happened to test a SSD (like Patriot Blast) with Phison S10 controller coupled to Micron 16nm 128Gbit TLC NAND?
 
Last edited:
Just wondering if you happened to test a SSD (like Patriot Blast) with Phison S10 controller coupled to Micron 16nm 128Gbit TLC NAND?

It is a combination I'm definitely interested in testing to find out if it would do better or worse than the BX200.

However so far the relatively high price for a Patriot Blast here coupled with a lot of reports of them failing within weeks has prevented me from getting one.

Should prices go down then it is most likely I will buy one to run tests on but prices have (unfortunately) remained pretty steady for the Patriot Blast.
 
pretty much. but most of them (s10 and sm2256 for sure) use rewriting techniques to mitigate this.

so as long as u use the drive, it should be fine.

My tests may be incomplete but:

So far Trion 100 seems pretty resistant to any read speed slowdowns.

The Trion 150 hasn't been tested as long as the Trion 100 but hasn't displayed any issues so far about 1,5 months in.

And while I do intend to check more thoroughly for rewritten files (primarily on the BX200) I think the ones using SM2256 as a controller tend to adjust their read algorithms rather than rewrite files.
 
  • Like
Reactions: cbn
Interesting, so Samsung right now is the only company producing an SSD that does not suffer from this issue?
3D NAND (or V-NAND) is also available from a couple other companies, but Samsung was first-to-market by a pretty long period of time.

I think the Crucial MX300 uses Micron 3D NAND, and it JUST came out. Otherwise I don't know of any, but... coming any minute. 😀
 
My tests may be incomplete but:

So far Trion 100 seems pretty resistant to any read speed slowdowns.

The Trion 150 hasn't been tested as long as the Trion 100 but hasn't displayed any issues so far about 1,5 months in.

And while I do intend to check more thoroughly for rewritten files (primarily on the BX200) I think the ones using SM2256 as a controller tend to adjust their read algorithms rather than rewrite files.

are drives powerred down or in active use ?
 
My tests may be incomplete but:

So far Trion 100 seems pretty resistant to any read speed slowdowns.

The Trion 150 hasn't been tested as long as the Trion 100 but hasn't displayed any issues so far about 1,5 months in.

And while I do intend to check more thoroughly for rewritten files (primarily on the BX200) I think the ones using SM2256 as a controller tend to adjust their read algorithms rather than rewrite files.

Any updates?
 
Trion 100 and Trion 150.


Think it would be accurate to say that both adjust for it good enough that if they do have any more than minor issues with read speeds dropping it is not really evident in the tests I've run.
Read speeds have some individual variance but there are few drops that are even approaching something major.


Now on to whether they maintain those read speeds through rewrites.


The Trion 100 has no SMART-value for amount written to it except for percentages of write cycles used which means that any amount of rewrites will be concealed.
The pattern it has when it comes to read speeds I think suggests that it does not rely very much on rewrites however.

The Trion 150 does have a SMART-value for amount written to it however and I've seen that it has gone up between tests.
However without knowing what those increases mean in terms of amount of data written it is uncertain whether it is doing rewrites or possibly just performing some GC.



I'd say that unless they have a strong aversion to certain temperatures (or a strong affinity to those temperatures they've experienced in my tests) any read speed degradation should not be noticed when the drive is powered on and would still be pretty well controlled after it has been powered down (depending on time, wear and temperature of course).

Which is why I intend to test drives with a folder of files written to it at low temperatures to reduce the possibility that they do well because they like heat (when writing).
How to do this seems difficult however since they heat up fairly quickly and then there is also the question of getting a credible temperature (especially difficult for those that lack a working temp sensor).
 
  • Like
Reactions: cbn
Back
Top