| HDD GURU FORUMS http://forum.hddguru.com/ |
|
| Samsung 840EVO 500GB SSD http://forum.hddguru.com/viewtopic.php?f=10&t=36564 |
Page 1 of 2 |
| Author: | MasterT [ March 15th, 2018, 14:30 ] | |||
| Post subject: | Samsung 840EVO 500GB SSD | |||
I need some help with this drive to pinpoint the problem. Drive does Not come ready. I checked for short all capacitors, none came short. I checked all resistors and none is blown. If i power drive and let it powered for a while it looks like cpu is heating up, a sign that it is alive. If anyone can help me move forward with this one i attach hi res pictures for both sides. Thanks.
|
||||
| Author: | fzabkar [ March 15th, 2018, 15:03 ] |
| Post subject: | Re: Samsung 840EVO 500GB SSD |
For the 250GB version ... http://www2.futureware.at/~philipp/ssd/TheMissingManual.pdf http://www2.futureware.at/~philipp/ssd/SamsungEVO840Voltages.pdf |
|
| Author: | MasterT [ March 16th, 2018, 4:46 ] |
| Post subject: | Re: Samsung 840EVO 500GB SSD |
Even if the pcb described in the pdf is a diffrent type, voltages measured are kind of the same. Seems to me more like a firmware /software issue. Question? If the content of the internal flash memory of 128kb, used to start the "safe mode", as described in pdf, is damaged, is there a way to write it back with a good file to get ssd back to life? And next question? Has anybody dumped a good file out of a working ssd of this type? Thanks. |
|
| Author: | lcoughey [ March 16th, 2018, 6:39 ] |
| Post subject: | Re: Samsung 840EVO 500GB SSD |
This series of Samsung SSD is not great. The NAND degrades and loses data fast. You should consider getting it to a professional lab with PC3000-SSD sooner than later. I can't speak for others, but my lab generally charges $900 CAD for this type of recovery. |
|
| Author: | MasterT [ March 16th, 2018, 7:49 ] |
| Post subject: | Re: Samsung 840EVO 500GB SSD |
I am the professional lab, i own pc-3000 ssd, and i charge a lot less for this kind of recovery. I also know from working with lots of types of controllers that when a controller does not have en external eprom to store the boot code but instead has an internal flash, just for the case when the boot code stored in the internal flash got damaged there is also a rescue procedure to be able to get it back alive. What i was asking is if anyone knows it, as it seems to me boot code got corrupted. |
|
| Author: | MasterT [ March 16th, 2018, 8:04 ] |
| Post subject: | Re: Samsung 840EVO 500GB SSD |
On the subject please. |
|
| Author: | lcoughey [ March 16th, 2018, 13:30 ] |
| Post subject: | Re: Samsung 840EVO 500GB SSD |
http://forum.acelaboratory.com/viewtopic.php?t=8735 The NAND is failing. You must not have PC3000-SSD or you would have already jumpered to safe mode, loaded the virtual translator and started to image with lots of read errors. |
|
| Author: | fzabkar [ March 16th, 2018, 14:59 ] |
| Post subject: | Re: Samsung 840EVO 500GB SSD |
Roman_TS wrote: I spoke with our ACE Lab developers and did you know what we found? Everything is very simple - NAND memory in Samsung 840 EVO is TLC based, and it is VERY VERY Bad. After some time when you write the data on memory, the charge inside NAND cells is flow away. The voltage inside cells become worse and worse, and finally - flow away forever. It means that if you will not use your 840 EVO drive for some time, you will find that your SSD is EMPTY. You ask me - are the guyes from Samsung stupid bastards? Why they did not find the way to fix this bug?! And I will answer - they release new FW, which make very simple thing - in background, when you working with your laptop or with PC, it just rewrite customer data again and again, again and gain by cycle, every time with the only goal - to keep the data and charges inside cells "fresh". When drive is not connected to power source, charge from cells is flowing away. Of course this bug fix is not solve the main problem - bad quality of NAND memory. Wow! Can Samsung expect a class action? |
|
| Author: | MasterT [ March 16th, 2018, 15:04 ] |
| Post subject: | Re: Samsung 840EVO 500GB SSD |
I am the professional lab, i have pc-3000 ssd, i have tried to get into safe mode with no success, so there is no way to load virtual translator and go on. If cannot get into safe mode there may be 2 problems: 1. pcb component failure problem 2. boot code is damaged I suspect the second. From my experience, when a controller does not have external eprom it ALWAYS have a backdoor procedure to be able to overwrite damaged boot code from internal flash to be able to recover. I was asking if anyone knows how can it be done and if anyone has a dump from a working ssd of this type. Thanks. |
|
| Author: | fzabkar [ March 17th, 2018, 0:25 ] | ||||||||
| Post subject: | Re: Samsung 840EVO 500GB SSD | ||||||||
Did you check the crystal? References: TPS62130A-Q1, Texas Instruments, 3V to 17V, 3A adjustable Step-Down Converter with DCS-Control, marking PA6IQ, VQFN 16-pin: http://www.ti.com/lit/ds/symlink/tps62130a-q1.pdf This (expired) URL suggests that GUILL = TPS62130D2: https://chipworks.secure.force.com/cata ... =TEX-GUILL STEF4S, STMicroelectronics, electronic fuse for 3.3V and 5V lines: http://www.st.com/web/en/resource/techn ... 095426.pdf
|
|||||||||
| Author: | sourcerer [ March 17th, 2018, 4:37 ] |
| Post subject: | Re: Samsung 840EVO 500GB SSD |
I developed a tool for exactly this kind of start problem: https://github.com/thesourcerer8/SSDdiag Please provide the reports it generates, then I can tell you more about the status of it. If you need any help to get it running, let me know. |
|
| Author: | HaQue [ March 17th, 2018, 5:19 ] |
| Post subject: | Re: Samsung 840EVO 500GB SSD |
fzabkar wrote: Roman_TS wrote: I spoke with our ACE Lab developers and did you know what we found? Everything is very simple - NAND memory in Samsung 840 EVO is TLC based, and it is VERY VERY Bad. After some time when you write the data on memory, the charge inside NAND cells is flow away. The voltage inside cells become worse and worse, and finally - flow away forever. It means that if you will not use your 840 EVO drive for some time, you will find that your SSD is EMPTY. You ask me - are the guyes from Samsung stupid bastards? Why they did not find the way to fix this bug?! And I will answer - they release new FW, which make very simple thing - in background, when you working with your laptop or with PC, it just rewrite customer data again and again, again and gain by cycle, every time with the only goal - to keep the data and charges inside cells "fresh". When drive is not connected to power source, charge from cells is flowing away. Of course this bug fix is not solve the main problem - bad quality of NAND memory. Wow! Can Samsung expect a class action? I expect not, as I believe this is deduced not from testing, but using generally accepted "reasons" for why an SSD has failed. such as it has TLC so... it must be that that failed. I would be interested to see the testing report that showed NAND cell voltage flowed away. They make MILLIONS of SSDs.... so why are we seeing only 10's to Hundreds fail worldwide? That's a pretty good % in my opinion. In many cases the cause of SSD failure is not found, but guessed at. Actually yet to see a case study that proves a cause of a failure. Who knows, it may be true.. what I am saying is there is as yet no proof on causes of failure in 99% of cases. |
|
| Author: | MasterT [ March 17th, 2018, 6:23 ] | ||
| Post subject: | Re: Samsung 840EVO 500GB SSD | ||
I made the measurements.
|
|||
| Author: | sourcerer [ March 17th, 2018, 6:55 ] |
| Post subject: | Re: Samsung 840EVO 500GB SSD |
Regarding the following question: If the content of the internal flash memory of 128kb, used to start the "safe mode", as described in pdf, is damaged, is there a way to write it back with a good file to get ssd back to life? What usually gets damaged is the normal bootloader on the NAND flash chips, or the SA structures on the NAND flash chips, but not the safe mode bootloader. The usual damage can be corrected by booting it in SAFE mode and then running a firmware update (if it affects the firmware part), or by fixing the SA structures (given that you understand them, which I don't yet). The internal memory contains only the safe mode boot loader, which is only used when you bridge the 2 safe mode pins during powerup. Perhaps there is a third bootloader inside the controller which checks the SAFE mode pins and boots SAFE mode or loads the normal firmware from the NAND chips, but I haven't seen any hints for the existance of that third bootloader yet, I only guess that it somehow must exist, because the NAND flash interface to load the firmware looks too complex to be implemented in hardware. I have not found any way to change the internal memory yet. And I am not sure, whether it is actually changeable, perhaps it is really just a ROM. If is not ROM, I would expect it to be SPI Flash, but I haven't found that SPI interface yet. If the internal memory with the SAFE mode firmware is really broken and that really concerns you, I would suggest to unsolder the controller and replace it with a working controller from a donor. But I wouldn't expect a broken SSD to be fixable with that. I guess that the controller could even start normally from the NAND flash chips with a broken SAFE mode bootloader. If you think that the controller might be at fault or the PCB might be at fault, and you are good with BGA soldering, I would suggest that you take the NAND flash chips off from the patient and solder them in the same place (and orientation!) onto a donor PCB. There is no configuration data stored inside the main controller, all configuration is on the NAND flash chips. If you aren't sure, whether the internal flash memory is working correctly, just plug the SSD into a SATA port, bridge the SAFE mode pins and power the SSD up, then you should see a ~500 MB storage device on the SATA interface, if the internal flash memory works. If that doesn't work, then try to connect with JTAG and see whether the 3 ARM cores are there. But as I said, the internal flash memory of the controller is most likely not the reason for why the SSD doesn't work anymore. |
|
| Author: | fzabkar [ March 17th, 2018, 16:36 ] |
| Post subject: | Re: Samsung 840EVO 500GB SSD |
HaQue wrote: I would be interested to see the testing report that showed NAND cell voltage flowed away. They make MILLIONS of SSDs.... so why are we seeing only 10's to Hundreds fail worldwide? That's a pretty good % in my opinion. The "charge decay" problem is real: https://www.extremetech.com/computing/190746-samsung-840-evo-has-serious-performance-bug-fix-on-the-way https://www.anandtech.com/show/9196/samsung-releases-second-840-evo-fix http://www.overclock.net/t/1507897/samsung-840-evo-read-speed-drops-on-old-written-data-in-the-drive |
|
| Author: | fzabkar [ March 17th, 2018, 16:50 ] |
| Post subject: | Re: Samsung 840EVO 500GB SSD |
MasterT wrote: I made the measurements. As you have already stated, they match sourcerer's results. |
|
| Author: | fzabkar [ March 17th, 2018, 17:22 ] | |||
| Post subject: | Re: Samsung 840EVO 500GB SSD | |||
Would it make sense to wire each of the R/B pins to ground and then see how the drive behaves?
|
||||
| Author: | MasterT [ March 19th, 2018, 5:54 ] | ||
| Post subject: | Re: Samsung 840EVO 500GB SSD | ||
I also checked the crystal.
|
|||
| Author: | outbreak198x [ January 16th, 2019, 12:34 ] |
| Post subject: | Re: Samsung 840EVO 500GB SSD |
have you ever found a solution for this, have similar drive here (PM851) - no bsy mode, nothing, CPU RAM NAND getting warm - but nothing happens, no safe mode and so on. i think the only way is to resolder chips to PCB board... |
|
| Author: | MasterT [ January 16th, 2019, 15:17 ] |
| Post subject: | Re: Samsung 840EVO 500GB SSD |
Hi outbreak198x, I had no solution to my case. All ports were dead (SATA, JTAG, Serial). Most probably boot code was damaged that's why no interface came up, and there is no way to fix boot code yet. Customer was not willing to go for changing chips to a working donor board. If your JTAG is not dead you could try this: https://github.com/thesourcerer8/SSDdiag sourcerer may help you as he is the expert in this matter. Good luck. |
|
| Page 1 of 2 | All times are UTC - 5 hours [ DST ] |
| Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group http://www.phpbb.com/ |
|