As with so many others, I've had the same problem of my eSATA drives disconnecting at random times.
PC: is a PowerSpec G208 with 16GB RAM
VG: nVidia graphics
HD: 20TB across 11 drives (3 internal SATA, 2 eSATA, 2 USB3, 4 USB2).
OS: Win 7 Ultimate; Win 10 Pro.
HD: (2) 2TB LaCie Quadras (USB2, eSata, FW) connected with a Rocket eSATA controller card with independent CPUs.
Whit that scenario, I had random disconnects with the eSATA Quadras every so many days, weeks, or months. I also had Power setting to NEVER power off drives which didn't help.
When they disconnected, I'd usually reboot the entire system to get them to be recognized, again.
On rare occasion, unplugging then plugging back the drives' power and eSATA cables along with hitting "Scan for New Hardware" in Windows Device Manager would seduce my system into letting them play. Since my PowerSpec G208 also has a seemingly undefeatable 3-4 min pre-BIOS cursor delay, a reboot's total round trip is a painful 5+ minutes to final disk quietude.
The research into my own eSATA case suggests it may be overly long periods of drive inactivity - not short periods or disk writing as some found in their cases.
Since my eSATA Quadras seem to are run continuously as they serve 10 SyncBackPro backup procedures executed from several machines in my SOHO network, it is rare they would be quiet for more than an hour or two - but it happens. E.g., if a remote machine that is the source or target of a backup procedure is offline, then the backup procedure will fail and wait for the next procedure's time slot. This fits my scenario wherein the time lapse between disconnects was in the range of days, weeks, or months.
My Quadras have a 3-position power switch (Off, ON, ON+Power Saving) but neither ON position stopped the disconnects.
I ultimately resorted to the trusty method of writing/overwriting a tiny text file to the root of each drive Quadra every so often. I started with a cycle of 15 minutes, decreasing it until the disconnects stopped. To my surprise, it required a fairly frequent 8 minute cycle to stop the disconnects - cpntrary to my previous conjecture. Ugh! So, an 8 minute prod *solved* my Quadra disconnects on THAT system ... read on.
== Embarrassing Oversight ===
To finish this post, I rechecked my set up (system, ports, scheduler, batch files, etc). To my chagrin, I then realized that when I switched to my 2nd OS partition which I upgraded from Win 7 Ultimate to Win 10 Pro on Sep 30, 2015, I did NOT enable disk power off NOR did I implement the frequent writing scripts for the Quadras.
To my surprise, in the Win 10 system, both Quadras have been running without disconnects for about 2.5 months (since Sept 30, 2015) - which is as long or longer than the inter-disconnect times I had since the problem started. However, I have reboot a few times in that period. So, time will tell if they really do continue without disconnecting again.
Assuming the Quadra's are not reading my this post these, my problem may have evaporated with the upgrade to Win 10 or perhaps something else changed in Windows 10's HAL layer or drivers because my drives are NOT set to power down and my recurring write to root batch file is NOT running. Or, perhaps, it has something to do with some Win 7 updates that I had NOT installed BEFORE I upgraded partition 2 to Windows 10. Too many changes to finger the exact reason. I'll post an update it they disconnect again in this new environment.
Before I discovered my inadvertent solution of using my 2nd partition's Win 7 upgrade to Win 10, I was going to end this post with a recommendation that those with eSATA disconnects simply rip out the naked drives and install them in a USB 3.0 enclosure which would likely fix the disconnects while also increasing disk read speed though with a generally reported 10-30% decrease in write speed with common USB 3 controllers. I was going to do just that until I discovered I may no longer have the problem.
FYI: If you have a top notch USB 3.0 card, you can approach eSATA write speeds from USB 3.0. I installed a higher end Rocket 4-port USB 3.0 card which has independent controllers for each port. As such, my USB 3.0 drives (two new 6TB WD My Books with HW encryption) are running faster than my eSATA Quadras - probably due to the dedicated controllers for each port which completely avoids the sharing of bandwidth and duplicitous overlapping protocol overheads.
So, for now, I remain using eSATA for the two 2TB Quadras.
I am using 2 of the Rocket USB 3.0 ports dedicated to my two new 6TB WD's. The other 2 Rocket USB 3.0 ports are feeding two Anker 7-Port USB hubs (14 in all). One of those is, in turn, feeding a Tripp-Lite 4-port USB 3.0 hub. That makes a total of 17 USB 3.0 ports running off the 4 Rockect USB 3.0 ports. Collectively, off the other 2 Rocket ports, I have about 12 devices connected: mouse, keyboard, display, older USB drives, film scanner, flatbed scanner, Epson R2800 photo printer, phone, Spyder calibrator, Olympus LS-11 Field Mic/Recorder, Bluetooth chip, USB 3.5" floppy drive, etc.
I hope my prosaic deluge provides some helpful ideas.
Happy now - though still scratching my head!
XOR_42
Facebook
Twitter
Instagram