Best "strategy" for ECC-corr. and re-reading with PC3000 Flash

[Markus Bauer]

Hello,

I have again a case on my PC3000 Flash where re-reading and error correction take a very long time.

I have selected the re-read method with the highest ranking and actually the re-reading has fixed a few data-blocks but on 60-70% of the blocks I just get a partial fix or fail. 

So re-reading is clearly making an improvement and I have seen chips where the re-reading or ECC correction start to work much better after a few tousend blocks.

How do you guys handle that?

Do you stop a re-reading after a while when it looks like it's not working great or do you let every pass run till the end before you try other options?

And how do you handle ECC corrections? Do you cancle ECC correction if it looks like it's not working great and hop to re-reading or do you let it always run till the end?

I am not sure if I maybe waisting to much time with re-reading passes and ECC corrections which don't work great. Would "method hopping" be better or maybe worse? 

[t...@desertdatarecovery.com]

Wow Markus that a lot to unpack.

 

1. Starting again is not normally good as you have to reread the bad ECC all over again.
2. You can select rereading of partially corrected sectors, those have a higher chance of being totally corrected.
3. You can use ‘intelligent reread mode’ which skips empty pages.
4. You can try ‘reread map generator’ and select full block only
5. You can try ‘reread map generator’ and see if the controller is listed/supported for reread, that can help.
6. Leaving the best until last, heating or cooling the chip can have significant effects on correcting ECC.

 

Tim Homer - Lead Engineer

Desert Data Recovery

t...@desertdatarecovery.com

www.desertdatarecovery.com

--
Data Recovery Certification Group / for issue with google group please email sc...@myharddrivedied.com
---
You received this message because you are subscribed to the Google Groups "DataRecoveryCertification" group.
To unsubscribe from this group and stop receiving emails from it, send an email to datarecoverycertif...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/datarecoverycertification/f1fc58b7-654f-480b-86a7-789d9729d39bn%40googlegroups.com.

[Markus Bauer]

Thanks Tim!

I am aware of coling and heating. I get it - being impatient and cancel things is not a good idea...

I have with version 8 the issue that the "intelligent reread mode" let the app crash. I will open a ticket if non of you may know why. But point 4 and 5 seems to be a good tip. I have to look into that functions. 

How often do you have luck with the "Autoselect of power supply" option? I have tryed that just 2 or 3 times and it was usually not very promissing and slower then the "Read Retry command". Should I consider that option more frequently or is that a option which don't work very often?  

[t...@desertdatarecovery.com]

Ace did have some issues with the app crashing while reading larger chips, but in 8.1 that should have been solved. However with flash recoveries, and especially correcting ECC you need more RAM that you think, might be the reason why it’s crashing?

 

Autoselect power supply is very slow as you say, so I very rarely use it. However if it’s a last resort I will choose a handful of chains from the middle of the map and try a wide voltage range setting on just those chains. You might get luck and find a voltage which works. Then stop the correction, set that voltage (or a smaller range around it) and reread. I am sure you know that already.   

To view this discussion on the web visit https://groups.google.com/d/msgid/datarecoverycertification/3f1deb42-92d6-4de7-9f78-2d295ac0ff8en%40googlegroups.com.