Samsung 840 EVO 512GB - burned - pls need advice

[danikd]

Hello,

great investigation, thank you!
Its seems to be the right IC. And is orderable via ebay and Mouser.
The pad under is really grounded, pin 5 is ground, but on SSD PCB is not connected, because IC is grounded via large pad - and pin 5 - as is in datasheet - is internally grounded.
Pin 6 is really grounded via cap on ssd PCB, so it is OK too.
Pin 7 - EN - is connected to 5V input on second side of the PCB, see photo, so its enabled right by 5V sata voltage.
Dimension are same, 3x3 mm
5A seems odd ... dont know ...


And yes, 5V Guill input is connected to 5V output of ther burned JS4NAM
C1 and C2 are OK, not shorted, resistance in MOhms, C1 has 63μF and C2 128μF.
I measured all caps on both sides, quick test, seems to be no short.

Thx

Daniel

[fzabkar]

If pin 5 is connected to the IC's pad, then you should have measured a short between pin 5 and PCB ground when the IC was installed on the PCB. But you stated that there was no ground connection to any IC pin. BTW, Farnell and Digikey also have the STEF4S IC ($3).

As for the 5A rating, the datasheet does in fact specify that this IC is targeted at SSD applications. Otherwise, could it be that Samsung's IC is a special order with OEM specs?

ISTM that the GUILL IC and its load are probably OK. You could confirm whether this buck regulator supplies the NAND array by testing for continuity between its output and the ceramic Vcc bypass capacitors adjacent to any of the NANDs.

The AKE4JR chip appears to comprise 5 buck regulators and some other circuitry which I don't recognise. I suspect that these regulators would provide Vcore and Vio for the flash controller (S4LN045X01-8030), plus the Vcc for the SDRAM (K4P4G324EQ-FGC2).

You could confirm that no regulator output is shorted to ground by testing the resistances of C1 - C5. L1 - L5 are the corresponding inductors.

I don't understand the function of the area consisting of L6 (?), C6A, C6B, and X. It might be worth checking whether any of these components are connected to ground or 5Vout. If X is a diode, then it could be that this circuit constitutes a boost regulator, but I can't understand what this would do.

Attachments

[danikd]

Any of pins PCB is grounded, only pad. Pin 5 on PCB is connected anywhere ( blind - dont know the right idiom - my English isnt good ) And in datasheet is : "5 GND Ground pin (can be left floating if TAB is connected to GND)" - so its OK.

And first page of datasheet :
Applications
• Hard disk drives
• Solid state drives (SSD)
• Hard disk and SSD arrays
• Set-top boxes
• DVD and blu-ray disc drivers


I just order 2pcs on Farnell, try to find another replacement and only this one has the right pinout.
Maybe same IC, but only OEM version for Samsung Corp.
Measurements tomorrow, today I am very tired and must sleep.

Thx master

Daniel

[Amarbir[CDR-Labs]]

Hello,

great investigation, thank you!
Its seems to be the right IC. And is orderable via ebay and Mouser.
The pad under is really grounded, pin 5 is ground, but on SSD PCB is not connected, because IC is grounded via large pad - and pin 5 - as is in datasheet - is internally grounded.
Pin 6 is really grounded via cap on ssd PCB, so it is OK too.
Pin 7 - EN - is connected to 5V input on second side of the PCB, see photo, so its enabled right by 5V sata voltage.
Dimension are same, 3x3 mm
5A seems odd ... dont know ...


And yes, 5V Guill input is connected to 5V output of ther burned JS4NAM
C1 and C2 are OK, not shorted, resistance in MOhms, C1 has 63μF and C2 128μF.
I measured all caps on both sides, quick test, seems to be no short.

PCB2.jpg

Thx

Daniel

Well,
I am Happy To Know The Ground Was As I Stated .But Frank did a Greater Job Then Me In Assisting You .The Caps You See Are Connected To Vin I think On One Side .If This Works Please Let Us All Know So That Someone Gets Assistance In Future In Similiar Cases

[fzabkar]

Amarbir, I have no idea what you are banging on about.

The location of ground is abundantly clear. I have always understood that the pad under the IC is grounded and I have identified it as "0V" in the photo clip. I did not need to ask the OP to verify this connection -- you did.

As for the grounded underside of the package ...

What has not been clear is whether the pad has an electrical connection to the IC or merely a thermal connection. It now appears that the pad performs both functions.

What is still not clear is why the OP's in-circuit measurements did not show continuity between the IC's pin 5 and ground. If the IC is an STEF4S, then pin 5 would be internally connected to the IC's pad, which is then grounded by the PCB. However, the OP indicated that "no pin [is] shorted to ground", so this is why I am questioning whether the IC is in fact and STEF4S.

That said, I am confident that an STEF4S will function as a replacement, although I'm wondering about the 5A trip current rating.

[danikd]

Hello,

sorry no measurement yet, no time Tommorow replacement will arrive from Farnell, i try it and let you know. Thx. Daniel

[fzabkar]

I notice that the 850 Evo 500GB has an unpopulated electronic fuse and a 5V TVS diode. The 5Vin and 5Vout are bridged with two components.

http://www.thessdreview.com/our-reviews ... lc-v-nand/
http://tssdrcdn.thessdreview1.netdna-cd ... B-Back.jpg

Attachments

5V_TVS_diode.jpg (12.87 KiB) Viewed 22260 times

e-fuse.jpg (32.7 KiB) Viewed 22260 times

[danikd]

Hello!

replaced with STEF4S and its working!!! So many thanx to all, especialy to fzabkar, you are really master!
Hope it will help to someone in future ...
Nice day to all!

Daniel

[fzabkar]

Congratulations. BTW, what are the markings on the STEF4S chip?

[danikd]

Heelo, see photo. Nice day. Daniel

Attachments

[fzabkar]

Thanks. I wish I had your soldering skill.

BTW I'm now seeing e-fuses everywhere.

Here is an example from an Intel S3700:

[danikd]

Hello

I thanks, wish I had your knowledge ...
Monday, maybe I will post another SSD problem, with Vertex 2. Solid green, but even after 200 restarts and connections, no detection ... I will start new thread.

Nice weekend.

Daniel

[gianni]

could you kindly confirm that the component is this one? thanks

http://uk.farnell.com/webapp/wcs/stores ... reId=10165

[danikd]

Hello, yes, the right one ... Daniel

[f-bone]

Hello guys,

I know thats an old topic but can anyone help me identify the AKE4JR chip? Anyone knows where I can source it from (except from donor)
Its on a samsung PM851 board and its completely blown.

TIA

[fzabkar]

If you can show us a photo with the chip removed, we could try to match it up with likely candidates.

Personally, I would measure the resistances between ground and each of the inductors, after removing the chip. This will tell us if any of the loads may have been damaged. Often in cases like this, the input voltage (+5V) punches through the PMIC and damages the other ICs.

[f-bone]

If you can show us a photo with the chip removed, we could try to match it up with likely candidates.

Personally, I would measure the resistances between ground and each of the inductors, after removing the chip. This will tell us if any of the loads may have been damaged. Often in cases like this, the input voltage (+5V) punches through the PMIC and damages the other ICs.

Thanks for your reply!

Yes Its pretty likely that other chips are damaged as well. its like 220v hit that chip as you can see on the pictures.

Attachments

[fzabkar]

I couldn't find anything in Texas Instruments' lineup. I know Samsung's products often have PMICs with Samsung's own part number and logo, so it might be an in-house part.

It could even be a DIY user programmable PMIC:
https://www.ti.com/power-management/multi-channel-ics-pmic/diy-pmic.html

These are 40-pin PMICs which are configured for various processors:
https://www.ti.com/design-resources/design-tools-simulation/processor-fpga-power/other-processors.html

Other IC makers (eg Active Semiconductor) also produce PMICs which are preconfigured for particular processors. You see them on Sandisk SSDs, but many are custom designs.

The bottom line is that I don't know where to look. :-(

In any case you may be wasting your time if any of the outputs are shorted to ground.

[f-bone]

I couldn't find anything in Texas Instruments' lineup. I know Samsung's products often have PMICs with Samsung's own part number and logo, so it might be an in-house part.

It could even be a DIY user programmable PMIC:
https://www.ti.com/power-management/multi-channel-ics-pmic/diy-pmic.html

These are 40-pin PMICs which are configured for various processors:
https://www.ti.com/design-resources/design-tools-simulation/processor-fpga-power/other-processors.html

Other IC makers (eg Active Semiconductor) also produce PMICs which are preconfigured for particular processors. You see them on Sandisk SSDs, but many are custom designs.

The bottom line is that I don't know where to look.

In any case you may be wasting your time if any of the outputs are shorted to ground.

I understand. Thanks anyway for your time. Much appreciated. I may have access to an identical disk so maybe I can source the chip or at least make some measurements. will keep you posted.

[sourcerer]

I have several of this AKE chip in my EVO840 collection:
AKE3NA
AKE4IP
AKE4QD
AKE4JJ
AKE3TK
AKE4GQ (twice)

AKE is the chip model, 3/4 is definitely the production year (2013/2014) , and the rest 2 characters are most likely a week code.

I have 2 of these AKE chips left from destructive research experiments, if you need them.