Western Digital My Passport Essential SE 750 GB USB3.0 Hard Drive Repair

Good day to all.

I warn you right away that all the actions described in the article are carried out at your own peril and risk, I, as the author of the article, are not responsible for the loss of information or damage to your device.

I somehow bought an external USB 3.0 hard drive Western Digital My Passport 750 GB. He liked me because it has the ability to connect to a USB 3.0 port and its compact size. I used it a little more than one year, and on one terrible day, it ceased to be determined by the system (both Windows and Linux). This happened due to a power failure. Then the voltage was applied, as a result of which, most likely, the hard drive failed. It was connected to an external USB hub, which is powered by an additional 5 V power supply. I’m not sure, of course, but from the part, it was probably his failure. But I do not want to talk about this, but about how I was able to repair it.

This HDD has a single board and all the elements and a SATA controller and a SATA to USB 3.0 converter are built into it, which is why it actually turned out so compact. The problem was that this hard drive was defined in the system as a certain device BAD PCB USB Device, and the system does not see the drive itself and any file system. He began to search the Internet for what kind of problem this was and how to solve it. As I found out from similar problems on the forums, the main problem in such hard drives (associated with the BAD PCB USB Device) is the failure of the converter (or you can call it the controller) SATA to USB 3.0, and WD has external hard drives with such dimensions as above, but to other volumes. It turned out that the hard drives similar in design to WD have different circuitry and configuration on the board. I did not find detailed instructions for revitalizing my disk, so I had to carry out repairs, as they say, in the image and likeness, based on the recommendations of other users of such disks or masters. Since the spindle in my drive, when connected to the USB jack through the cable, began to rotate, and the heads did not cramp, I concluded for myself that, most likely, there is still a point in the repair, and it’s worth a try.

Judging by the forums, the malfunction was due to the failure of the SATA to USB 3.0 converter chip. The repair consisted in the fact that the hard drive did not completely lose its functions, which means that it can work if it is connected directly to the SATA interface or to the new SATA to USB 2.0 / 3.0 converter, as follows from these discussions on the forum:
IXBT .

After reading the above forum and other sites proposed on it, I had a hope, for which I am very grateful to the authors of articles and comments on the forums.

I will give a few pictures of the printed circuit boards from the hard drives at the following links:

image
Fig. 1

image
Fig. 2

image
Fig. 3

The pictures show that the conductors that lead to the SATA connector are soldered to the pads on the board. It is not difficult to guess that the pads on the board are connected to the SATA interface circuits, namely, they are connected to the SATA to USB converter.
I unscrewed the board from my HDD, and it turned out that the circuitry on this board was different from those on the above sites and on the above pictures, but the principle became clear to me, and the direction I saw where to dig.

I immediately drew attention to the microcircuits that were used on my motherboard, and the main microcircuits: the SATA controller and the SATA to USB converter were different from those given in the links. Then, I also drew attention to the tracks (see position 1. in Figure 4 below) leading from the obvious SATA controller to the SATA to USB converter. Each of the tracks on its way passed through the contact pads with vias on the outside of the board (pos. 2, Fig. 4). These tracks approached the capacitors C90, C91, C92, C93 (item 3, Fig. 4), and then the signal flowing along the tracks, passing through the capacitors, was fed to the SATA to USB converter chip to which the connector is connected, for connection USB 3.0 cord.

image
Fig. 4

Now you need to make sure that these tracks are indeed the SATA interface, and if so, find out the polarity and purpose of these tracks. For these purposes, I began to look for a description of the applied microcircuits: 88i9146-TFJ2 (SATA controller), and SW6316-3VB14 (SATA to USB 3.0 converter). To my regret, I could not find anything, for some reason these microcircuits are kept secret, or I didn’t look there, I usually use the site www.alldatasheet.com to search for reference data on microcircuits (this is not for advertising purposes).
Then I started looking at the links that the search engine returned for the request for the 88i9146-TFJ2 chip, and found a photo of the board on this site: www.storagereview.com/western_digital_caviar_green_3tb_review_wd30ezrsdtlon which this chip was applied, and on this board there was a connector for connecting the SATA interface:

image
Fig. 5

From this picture, I was convinced that this chip is a SATA controller and that the tracks that I immediately noticed on my motherboard are exactly the SATA interface, as Figure 5 shows that the 88i9146-TFJ2 IC outputs that are connected to the “unusual” tracks on my board, lead to the SATA connector in fig. 5. After which, after looking at the pinout of the SATA connector, I determined the polarity and purpose of the SATA interface tracks on my board, which means which contact pads of the vias (Fig. 4, pos. 2) are responsible for which channel of the SATA interface (see. Fig. 6 )

Now the task was to carefully solder the conductors to these tracks, which should be output to the SATA interface or to the new SATA to USB converter. For myself, I bought a 2.5ʺ AGE Star hard drive box with a built-in SATA to USB 3.0 converter. Since my hard drive did not have a direct SATA interface, but only had a USB 3.0 micro-B connector, my hard drive did not fit in this box. In this regard, it was necessary to unsolder the USB 3.0 type micro-B connector from the hard drive board, observing all safety precautions in terms of preventing mechanical stress on the board and temperature overheating. The USB connector in the hard drive board was sealed securely, and in order not to overheat the board, I decided to file it, namely to saw off a microdrill, into which a detachable disk (like a “grinder”) used in dentistry was inserted on a special axis. With a “grinder”, I sawed the side racks of the connector along the lateral edges, after which it became easier to remove the connector “in parts”. On the converter board of the new box, I also unsoldered the SATA + SATA power connector, which would also interfere with the packaging of my HDD in the box.

image
Fig. 6

Well, then things are small, you need to remove the SATA interface conductors from the HDD board, and solder the +5 V powered wires from the converter box of the new box.

First, I unsoldered the capacitors (pos. 3, fig. 4), since they lead to the microcircuit, which is out of order. I do not know the components of the microcircuit that failed, and in order not to risk, the capacitors should be removed, because through them, the useful signal of the SATA interface can be bypassed in the SW6316-3VB14 chip if the inputs of this chip are broken or connected in some way to a common wire.

The HDD board has very thin tracks and extremely small contact pads of vias, especially those holes that I needed for soldering (see. Fig. 4, item 2). In my opinion, the diameter of these holes, taking into account metallization, is approximately 0.3 mm, and the contact area has a diameter of, at best, 0.6 mm, which is very unreliable in terms of mechanical strength. Those. if I start soldering conductors to these paths, then with one wrong movement along the wire forming, I can easily disrupt this contact pad. I did the following:

1) carefully, so as not to touch the neighboring tracks, I cleaned the soldering mask from the solder mask (green color) on both sides of the contact pads of the vias indicated in Fig. 4 poses 2 and fig. 6;
2) inserted into these holes thin tinned copper wires, approximately 0.2 mm in diameter;
3) soldered these wires on both sides, carefully so that the solder does not float on adjacent tracks;
4) bit off the excessively sticking ends of the wires with side cutters.

That's it, now the mechanical strength of the contact pads is much more reliable, and you can now solder the mounting wire to these pads.

As a mounting wire, I took the MGTF-0.07 wire, it is quite thin in diameter, has fairly thin wires, which gives it flexibility, and most importantly has a heat-resistant fluoroplastic insulation, which is very important so that the insulation does not melt when soldering and didn’t slide down the bare wires, which could cause a short circuit. I carefully cleaned and irradiated the ends of the wires, and then soldered them to the contact pads, so that there were no extra influx of solder and that the droplets of solder had the lowest possible height above the board. This is done so that the drops do not rest against the housing of the new box, due to which it is possible to break the contact pads with wires or to pierce the insulation of the box lid (in the case when the lid is metal and insulated with a film, as in my case, a short circuit may occur). I also pulled out the 5th GND wire (the conventionally accepted common wire for the SATA interface) for the interface, and passed it along with 4 interface wires under the cable clamp from the insulated wire, which I soldered in the place where the USB socket had previously been 3.0 micro-B. The clamp can not be soldered, but it is recommended to do this for a more reliable fastening of the wires on the board so that they do not come off “with the roots”. It is also recommended that, at a distance of about 5 mm from the soldering points, the cable loop be glued to the board with glue based on ethyl cyanoacrylate (the so-called superglue), which will securely and quickly fix the wires to the board. but it is recommended that this be done to more securely attach the wires to the board so that they do not come off “with the roots”. It is also recommended that, at a distance of about 5 mm from the soldering points, the cable loop be glued to the board with glue based on ethyl cyanoacrylate (the so-called superglue), which will securely and quickly fix the wires to the board. but it is recommended to do this for more reliable fastening of wires on the board so that they do not come off “with roots”. It is also recommended that, at a distance of about 5 mm from the soldering points, the cable loop be glued to the board with glue based on ethyl cyanoacrylate (the so-called superglue), which will securely and quickly fix the wires to the board.

I connected the power wires to the hard drive board with the service connector, which in all likelihood was on it, which is made in the form of an end two-row plug with 12 contacts (like PLD2-12-R). I determined the power contacts by a dial tone, first on the board I connected the multimeter probe to a common wire, and then I found a common wire on the connector pins. I also detected a +5 V power contact by a call, but now I connected the probe on the board to the soldering point of the power contact No. 1 of the USB 3.0 micro-B connector, and with the second probe I found the power contact on the end connector. I soldered the supply wires directly to the pins of the connector and insulated with a PVC pipe. All this can be seen in Figure 7.

image
Fig. 7

After all the above operations, I finally soldered the wires to the SATA to USB 3.0 converter board of the new box, which I will not talk about and describe here, because each one can have a different one, depending on which box you purchase.

My hard drive worked, no information was lost from it, and all information is read without problems, i.e. no encryption "on the fly", which I read on different forums, was not provided by the manufacturer in this model of a hard disk.

I hope that my article will be useful to someone.

Also popular now: