Corsair Memory Chip Help

[ddrecovery]

Just received this flash drive. One area I am not strong in is assessing what might be wrong electronically with a dead board. I can complete chip offs and repair broken connectors but know I need different skills. This flash drive is dead, does not mount and no LED so presume its a power issue. What are the steps I should be taking to find out whats wrong before completing a chip off?

Also if it does need a chip off, this drive has 'double decker chips' (see pic) and advice on removing and reading these vs normal TSOP chips? Never seen these before.

Attachments

[labtech]

Looks like flux residue on it. Did it arrive like that? If so, then asking customer questions can be very helpful in finding out in whose hands it has been.

AFAIR reading the NAND in reader, it is the same thing as regular NAND.

[fzabkar]

I would measure the voltage on either side of the 2R2 resistor (2.2 ohm). This will tell you whether +5V is getting to the drive.

Measure the voltage across the same resistor. This will tell you the current that is being drawn by the drive.

current = voltage / 2.2 ohms

Measure the voltages across the capacitors. You will probably find +3.3V for the NANDs. There may also be +1.8V and a Vcore supply (~ 1V ?).

Shake the PCB against your ear. If it rattles, then the 12MHz crystal (Y1) is probably shattered.

Test for continuity between the Data + and Data - pins and their respective resistors, R11 and R10. There should be a low resistance between the USB Data pins and the controller side of the resistors.

[ddrecovery]

Looks like flux residue on it. Did it arrive like that? If so, then asking customer questions can be very helpful in finding out in whose hands it has been.

AFAIR reading the NAND in reader, it is the same thing as regular NAND.

The residue is burnt glue. The drive was contained within a waterproof rubber enclosure and covered in glue presumably to keep it watertight. After spending an hour removing most of the glue I tried to reflow the board with no luck (hence the dark residue). I have tested the connector legs and there is power, so the issue is somewhere on the board (or perhaps controller failure).

Thanks Fzabkar. I will let you know how I get on.

[jeremyb]

Also if it does need a chip off, this drive has 'double decker chips' (see pic) and advice on removing and reading these vs normal TSOP chips? Never seen these before.

The metal leads are welded together, try not break the welds, you can use solder to reconnect them but it's a minor PITA, otherwise they show up as 4 Banks, nothing special.

DON'T SEPARATE THE TOP AND BOTTOM CHIPS

Silicone is holding the NAND chips to the PCB, in addition to heat slip a Xacto knife under the bottom chip to gently lift it from the PCB.

[Amarbir[CDR-Labs]]

Also if it does need a chip off, this drive has 'double decker chips' (see pic) and advice on removing and reading these vs normal TSOP chips? Never seen these before.

The metal leads are welded together, try not break the welds, you can use solder to reconnect them but it's a minor PITA, otherwise they show up as 4 Banks, nothing special.

DON'T SEPARATE THE TOP AND BOTTOM CHIPS

Silicone is holding the NAND chips to the PCB, in addition to heat slip a Xacto knife under the bottom chip to gently lift it from the PCB.

Jeremyb Sir ,
How About Using a Preheater To Heat Bottom To 100 Degree Celsius ? .This Should Melt The Silicon IMHO

[Amarbir[CDR-Labs]]

Looks like flux residue on it. Did it arrive like that? If so, then asking customer questions can be very helpful in finding out in whose hands it has been.

AFAIR reading the NAND in reader, it is the same thing as regular NAND.

The residue is burnt glue. The drive was contained within a waterproof rubber enclosure and covered in glue presumably to keep it watertight. After spending an hour removing most of the glue I tried to reflow the board with no luck (hence the dark residue). I have tested the connector legs and there is power, so the issue is somewhere on the board (or perhaps controller failure).

Thanks Fzabkar. I will let you know how I get on.

Hi Follow Franks Advice .
Do You Have aHigh End Multimeter At Hand Or a Scope of 20Mhz Or More ?

[northwind]

Also if it does need a chip off, this drive has 'double decker chips' (see pic) and advice on removing and reading these vs normal TSOP chips? Never seen these before.

The metal leads are welded together, try not break the welds, you can use solder to reconnect them but it's a minor PITA, otherwise they show up as 4 Banks, nothing special.

DON'T SEPARATE THE TOP AND BOTTOM CHIPS

Silicone is holding the NAND chips to the PCB, in addition to heat slip a Xacto knife under the bottom chip to gently lift it from the PCB.

+1

[jeremyb]

Jeremyb Sir ,
How About Using a Preheater To Heat Bottom To 100 Degree Celsius ? .This Should Melt The Silicon IMHO

Silicon is pretty good at dissipating heat (?right word?), you can get a little bit of play using heat (400C) but not a lot which is why I recommend using a spreader of some sort.

[Amarbir[CDR-Labs]]

Jeremyb Sir ,
How About Using a Preheater To Heat Bottom To 100 Degree Celsius ? .This Should Melt The Silicon IMHO

Silicon is pretty good at dissipating heat (?right word?), you can get a little bit of play using heat (400C) but not a lot which is why I recommend using a spreader of some sort.

Sir ,
When The Advise Comes From a Better Pro [ I Grab It ] ,Thank You

[ddrecovery]

I would measure the voltage on either side of the 2R2 resistor (2.2 ohm). This will tell you whether +5V is getting to the drive.

Measure the voltage across the same resistor. This will tell you the current that is being drawn by the drive.

current = voltage / 2.2 ohms

Measure the voltages across the capacitors. You will probably find +3.3V for the NANDs. There may also be +1.8V and a Vcore supply (~ 1V ?).

Shake the PCB against your ear. If it rattles, then the 12MHz crystal (Y1) is probably shattered.

Test for continuity between the Data + and Data - pins and their respective resistors, R11 and R10. There should be a low resistance between the USB Data pins and the controller side of the resistors.

Voltage either side of R2R is 6.6v which seems high
Voltage across R2R 0.1v
Most capacitors are 4.5v of 1.8v except:-
C5 0.0v
C3 6.6v
C4 0.3v
No noise from the crystal
Data+ 0.3
Data - 0.3

Hope that helps....

[fzabkar]

The voltage readings are too high by about 33%. Either your meter is bad, or perhaps it behaves that way when its battery is flat.

6.6 / 1.33 = 4.96V
4.5 / 1.33 = 3.38V
1.8 / 1.33 = 1.35V

[ddrecovery]

The voltage readings are too high by about 33%. Either your meter is bad, or perhaps it behaves that way when its battery is flat.

6.6 / 1.33 = 4.96V
4.5 / 1.33 = 3.38V
1.8 / 1.33 = 1.35V

Fzabkar you are indeed a genius. I had a new 9v battery, tested it and it read 12v. New battery now installed. You can take your new readings above as true. Do any of the other zero/low readings point to any particular issue?

[fzabkar]

The current through the 2.2 ohm resistor is about 45mA, so that would suggest that the board is drawing the right amount of current.

3.3V is what you would expect for NAND, but I'm not sure about the 1.35V supply. It could be Vcore, but the following circuit diagrams only show 3.3V and 1.8V. Your chip could be different, though.

ftp://helpedia.com/pub/temp/uploads/ala ... ematics.7z

The zero reading for C5 looks OK. I don't know the function of C4, so I can't advise there.

I would test the crystal with a scope. Other than that, I can't see what else you could do with a meter.

[ddrecovery]

Thanks for the advice. I am talking to a local PCB fabricator about getting one of their techs to give me a one-on-one for a day. Going through how to test these types of boards and what to look out for. Its been hard to find the right kind of class.

[HaQue]

bear in mind, the LED function is not a simple dumb "when there is power, the LED goes on" but the firmware in the controller is programmed to turn LED on undr certain conditions - usually data transfer. so if firmware is bunk - no led. not as simple as it must be a power issue.

You have to remember the flash drive is basically a complete computer with power supply motherboard, ram, storage, CPU enclosed in the controller chip, albeit likely only an 8-bit one.

[jeremyb]

Don't waste your time with this PCB.. just read the memory chips.

If you want to learn how to fix PCB's find a schematic for a controller (preferably a modern one). It explains every part, after a few you'll see they're pretty much all the same.