Fast forward a few weeks later... If the disk runs on an old chipset by emulating the PATA standard, I would really like to integrate it into a recent system with fellows. I shook the idea that it is connected to its USB Bridge: -first he worked on a system of PATA emulating SATA, -After I tried it on another USB/SATA Bridge based on an ASMedia chipset, it worked flawlessly -then I changed-completely, completely-the embedded software of the original USB/SATA Bridge based on JMicron and it still works with it -Finally, using WD-Marvell 4 beta and its (new?) AHCI addressing, it works by plugging the drive into the SATA port H270 while it was connected to the USB and identified without losing power. My conclusion was: this drive is not a well-conforming SATA drive and something in the SATA initialization sequence after the connection does not work well; some careless SATA controllers may pass it on, but others do not. So I dug a bit more and using Linux debugging and a few DOS routines, I discovered that the disk is discovered, on SATA & AHCI non-working system, as a 48bit blocks of 512 physical bytes (the drive is 512e AF , 4096 physical), present a signature of a SEMB (SATA enclosure management bridge) instead of a SATA disk. To be continued...
|