HDD GURU FORUMS

Hi,

I got this no-name USB drive for repair with an not branded UP21 48-pin controller, whose connector was violently ripped off, with tracks coming together. The flash drive was not in use when the accident occured.

I am now awaiting an identical flash drive that my client could find recently, so that I will have all necessary parts for the repair or a chip transfer. So, my question is just out of curiosity, to confirm that my diagnosis is correct.

I did a first repair with the original connector, which was unsucesseful.

So, I desoldered it, took a photo of the tracks and put a new connector into place.

The 1.5Ω (1R5) resistor near the positive pin had to be replaced by a resistor of the same value because the metal was no longer attached to the ceramic. The replacement resistor was taken from a dead USB drive from my trash of electronic boards, and consequently is possibly defective.

I bridged what could be, and re-tined the pads of the controller and NAND chip.
The flash drive is still not detected: no sound when connecting it in Windows and not visible in the disk management.

On the positive pin, the voltage is well around 5.07V, and is 2.76V after the 1.5Ω resistor.
So, the controller is powered with 2.76V.

The UP21 controller of the flash drive has :
- VCC on pin 47
- D- on pin 42
- D+ on pin 43
- GND on pin 41

This is similar to the pinout of the Phison PS2251-67 controller.

So, I assume 5V being expected on pin 47 and the 1.5Ω resistor being defective.

Just wanted to confirm my hypothesis that there currently is a problem of insufficient VCC to the controller.

Does the resistor measure 1.5 ohms on your meter?

Using Ohm's Law ...

Current drawn by flash drive = (5.07 - 2.76) volts / 1.5 ohms = 1.54 amperes

Power dissipated by resistor = I^2 x R = 1.54 ^2 x 1.5 = 3.6 watts

Power dissipated by controller and NAND = 2.76 volts x 1.54 amps = 4.25 watts

ISTM that the resistor should be well and truly cooked by now.

What resistance do you measure between ground and the controller end of the resistor?

Before doing anything else I'd pull the nand and dump it before you send it up in smoke trying to fix the pcb. You might also want to spend some time working on your soldering and get some sensible sized enamel magnet wire for repairs.

The "donor" flash drive arrived to me today, but although from the same "lot" it uses a different PCB and controller. So, unusable for this repair.

The original resistor has correct voltage.
I observed that the original desoldered 1.5 Ω resistor still has some thin metal that flush its top black coated surface.
This extremely small and thing pad, allowed me to mesure the resistor with my multimeter.
Measured resitance was exactly 1.5 Ω.
So, apart from its missing metal pad on one side, this resistor is perfectly healthy.

I tried soldered back this original healthy resistor and tried restoring a connection from the thin metal coating on the resistor surface, but despite good flux, the solder does not take at this location.

Yes, the donor resistor was defective.
The useful maths as suggested by fzabkar also convinced me that there was some problem with the donor resistor (which was taken from a defective USB flash drive). Unsoldered it and measured a resistance around 140 Ω (~ 100 times the one required).

Seeking another 1.5 Ω resistor.
I searched among boards that I have if I could find a 1.5 Ω resistor, but this resitance value seems not being broadly used.
1Ω (1R1) and 2.2Ω (2R2) are much more common. Maybe could I try with a 2.2Ω resistor and add a 0.75Ω resistor in parallell if the circuit is underpowered.

Before doing anything else I'd pull the nand and dump it before you send it up in smoke trying to fix the pcb.

This flash drive simply had a ripped off connector and was not in use when the accident occured.
Under the microsocope is no other damage to the board.
So far, the problem clearly comes from an underpowered controller, due to excessive resistance at the 5V power making not enough current flow.
I don't see risk the PCB going in smoke in such a situation ...

You might also want to spend some time working on your soldering and get some sensible sized enamel magnet wire for repairs.

I practice weldering for years, and have done many prettier solders, including tiny ones like repairing USB-C tracks.

Sometimes however, you are in a rush and the only thing that matters is restoring the electrical connectivity. In this context, aestetic doesn't really matter. The solders were carefully examined under the macroscope and are all ok. Also, the wires are not touching other components than they should. All the extra wires visible on the picture were to ensure proper connections in case the PCB tracks would have had a crack.

About supplies

Thank you for mentioning the enamel magnet for repairs. This is the kind of supplies I've been dreaming of for long time.

For thin wires, I initially started the hard way, using strands from telephone wire. The PVC coating was the major problem as lacking flexibility.
Then, I found some silicon coated wires. The red one visible on the picture is a 30 AVG, from which I kept here only 3 out of 11 strands.
This cable is very flexible, but the thickness of the silicon coating is the most painful thing for tiny repairs.

Reworking existing cables has always been ok, but for sure, the cables that you mentioned for repairs will be very helpful and bring a much-appreciated comfort.

I don't see risk the PCB going in smoke in such a situation ...

Im sure you don't, which is why I suggested you make a dump before doing anything else.

I can see shorts on the data lines, the power rail and who knows what's going on with the ground.

So far, the problem clearly comes from an underpowered controller, due to excessive resistance at the 5V power making not enough current flow.

I don't think it's clear at all. Pull up resistors are usually measured in k, 1.5 Ohm is essentially a wire and your reading a drop of 2.3v across it, it should burn your fingers. Save the clients data, worry about the electronics of it later.

Data successfully recovered!

I softly stripped the resistor's black coating to see the extremely thin metal layer.
I could then attach a small drop a solder to it, at relatively low temperature.
Then, grew the size of the solder drop, with additional solder.

I can see shorts on the data lines, the power rail and who knows what's going on with the ground.

@Lardman: Each situation is unique and the history of what arrived to the device really matters.

Of course, sometimes a client can give fake informations (e.g. if they say a dropped hard drive was not).
There are also some risks (although small) doing a NAND dump.

In this case, I was confident about the history of the flash device and the tear-off the connectors was sharp.

PCB was not bent. Tracks were visible and obviously in good condition.
Furthermore, the PCB had been slightly scrapped near the controller and NAND legs, and those had received fresh tin, to make sure the contact was good where the legs connect to the PCB.

The data lines to the controller were in good visual condition and had been doubled with insulated cable.
(one visible on the photo).

I would have considered things differently for a multilayered PCB, or if the cause of the failure was unknown, but apart the ripped-off connector, after careful inspection under 10x magnification, the circuit board revealed as being in good condition.

The resistor must have been bad. If the flash drive was otherwise OK, you could have replaced the resistor with a short or a fuse.