This is a picture of the sata data lines on the controller of a MYBOOK-1TB (WD10EACS green) drive. This drive had intermittently (mostly) refused to connect via USB to the host computer. The most glaring symptom would be erratic spin-up of the disk when connected via USB. It would either do so, or not. And some of the times it did spin up, nothing else happened aside from the calibration clicks and whirring motor. Windows just wouldn’t see the disk. Simple as that.
I found that sometimes bashing it upside the head would make it connect more or less, sometimes. This stunk the room up with the smell of a bad connection someplace. The mechanical noises (when the drive was working) sounded just fine. So that meant (to me) the problem was somehow electrical and intermittent-connection related.
To begin - I reflowed and cleaned some “important looking” parts & connectors on the SATA to USB bridgeboard. That didn’t help much. I also tested the drive with my trusty USB-to-IDE & SATA adapter I got at the dollar store. I was getting the same intermittent no connect behavior. That pretty much eliminated the MYBOOK housing, bridgeboard, WD suplied power supply, and WD supplied USB cable. I tested the same USB port with a KEYDRIVE and my Elements drive. The computer & USB port worked just fine. I even tested the nice spiffy carrying case and its shipping box, just to be sure! Nope not the problem. So I put all the **bleep** aside and focused on the bare metal drive itself now.
I then took off the disk’s controller card and looked around there. I noted that C2,C3,C5,C7 were pretty close to the edge of the PCB, close to a big and major connector block. These really tiny capacitors are in series with the TDR (time-domain-reflectometer) pulsers in the Marvel controller I.C. These capacitors are coupling caps in the LVDS (low-voltage-differential-signaling) lines.
Most any “of-this-vintage” drives will have two pairs of data lines running parallel like so. One pair sends, the other receives. The capacitors, connector, and lines are right in the middle of the picture.
Basically in layman’s terms, the data comes in off the red SATA cable we all know and love and into the drive and makes its way to the controller I.C. From there on after, a lot of magic happens and your data is saved to disk.
So these little capacitors are coupling capacitors, a crude form of signal conditioning if you will. They are right in the middle of the road going from the SATA connector to the Controller. If you take them out, then it would have the exact same effect as disconnecting the SATA data cable.
As you can see, these parts are pretty close to the SATA connector block. And the stress of being in a portable device with a rigid bridgeboard connector had probably weakened the connections, somehow. So I got out a small 20 watt soldering iron with a 1/25 inch tip and reflowed them. Be careful, you don’t need extra solder, just some flux. And be quick, because if you get these parts hot, they’ll just float right off the board and get buried in the blob of solder sitting on your soldering iron’s tip! Then you gonned and dund it but good!
Nice and slow, one side at a time, with plenty of flux.
I took this pic, quickly, right after testing my very hasty proof-of-concept repair - I was so elated that that was indeed the problem! I went back and reflowed it “correctly” and smoothly and cleaned up the flux for a nice professional touch. I also re-doed the 7 pins on the SATA data connector itself for good measure. I was thinking maybe it would speed up my drive and make it spin faster or something, but, alas, that was not the case. I figured if that didn’t do anything, then probably resoldering other randomly chosen parts and the power section of the SATA cable wouldn’t do anything more either!
So, I reassembled everything and tested the drive s’more. It connects 100% of the time and doesn’t “drop out of Windows” anymore. Testament to the error checking and correction protocol used in the SATA interface, there was no corrupt data. The drive either saved the data or it didn’t. It either read the data, or it didn’t. 
Here is some reference material if you want to smarten yourself up:
http://hddscan.com/doc/HDD_from_inside.html
http://en.wikipedia.org/wiki/SATA
http://en.wikipedia.org/wiki/LVDS
http://www.instructables.com/id/Disassembling-a-Western-Digital-My-Book/
