eSATA card problem

tinpanalley

Golden Member
Jul 13, 2011
1,451
22
81
I've got these external enclosures with no USB 3.0 but with eSATA and while I could get new enclosures it seems ridiculous to replace these perfectly good metal ones with fans just because there's no USB 3.0.

Anyway, I have an eSATA card that used to work great and now doesn't. It's a Silicon Image SiI 3114 Softraid 5 that used to work as a regular eSATA input port and now doesn't read at all. The website for the manufacturer offers no help because they claim the card should work fine for eSATA connections out of the box and that the standard Windows driver is good enough and that you only download THEIR driver if you need RAID features. Frankly I don't even know what RAID is.


Can anyone recommend another eSATA card I could buy or a manufacturer or tell me if they've ever had a similar experience with an eSATA PCI card?

Thanks!
 

Burpo

Diamond Member
Sep 10, 2013
4,223
473
126
You don't say what O.S. you're using. If it's Win7, then check under optional updates for a new Silicon Image SiI 3114 driver. Moving the card to a dif PCI slot may also invoke a driver update from Windows update.
 

tinpanalley

Golden Member
Jul 13, 2011
1,451
22
81
You don't say what O.S. you're using. If it's Win7, then check under optional updates for a new Silicon Image SiI 3114 driver. Moving the card to a dif PCI slot may also invoke a driver update from Windows update.
I'm sorry. You're right. I'm on Windows 7. No optional updates at all. I've tried doing the PCI slot shift and it just grabs the same driver again and then won't recognise any connections. It's not the enclosure either because I've tried the four I have.
 

mikeymikec

Lifer
May 19, 2011
17,633
9,447
136
I realise this isn't of much direct assistance for your problem, but I tried out eSATA for a year or so and found it to be a lot less tolerant when connecting up drives that are in a more iffy state, so I migrated to USB 3.0. Admittedly most of the drives I connect via USB 3.0 are ones that I need to urgently back data up from or recover data so this is of importance to me (though perhaps not for you).

One other note is that when you say PCI do you mean PCIE? Because PCI will do an absolute maximum theoretical throughput of 133MB/sec IIRC, which many (most?) modern HDDs will max out when going full tilt. If it's PCIE, it's worth paying attention to what generation of PCIE your board can handle as well as the card (as well as whether the card is 1/2/4x etc). I've seen plain SATA AHCI PCIE cards that couldn't even handle the full potential of SATA 3Gbps.
 

tinpanalley

Golden Member
Jul 13, 2011
1,451
22
81
I realise this isn't of much direct assistance for your problem, but I tried out eSATA for a year or so and found it to be a lot less tolerant when connecting up drives that are in a more iffy state, so I migrated to USB 3.0. Admittedly most of the drives I connect via USB 3.0 are ones that I need to urgently back data up from or recover data so this is of importance to me (though perhaps not for you).

One other note is that when you say PCI do you mean PCIE? Because PCI will do an absolute maximum theoretical throughput of 133MB/sec IIRC, which many (most?) modern HDDs will max out when going full tilt. If it's PCIE, it's worth paying attention to what generation of PCIE your board can handle as well as the card (as well as whether the card is 1/2/4x etc). I've seen plain SATA AHCI PCIE cards that couldn't even handle the full potential of SATA 3Gbps.
Hmmm..ok. What can I do to get USB 3.0 speed transfers then if I decide NOT to use the eSATA enclosure? I have a dock that has 3.0, but I'm convinced the speed is WAY slower than it should be for 3.0. Sluggishly slow. And these are regular internal SATA HDDs at 7200rpm. Shouldn't be a problem. My 3.0 ports on my mobo are fast as well, 5Gbps so I know it's not them.
 
Last edited:

mikeymikec

Lifer
May 19, 2011
17,633
9,447
136
By having a USB 3.0 card (or rear 3.0 port) along with a USB 3.0 capable dock and USB 3.0 cable.

If you connect a USB 3.0 device to a USB 2.0 port, you're going to get 2.0 speeds.

If you need a USB 3.0 card, the same advice applies with regard to PCI/PCIE.

Alternatively, if your board's SATA implementation allows you to specify a particular SATA port to be hotplug-capable, then you could simply get a little adapter that provides an eSATA port on the back of your computer that has a cable hooked directly into the SATA port in question. I've never done this myself, but I've seen the adapter plenty of times.

If the enclosure really is 3.0 capable, then the USB connector into it should look like the connector that's either of the two in the bottom right corner of this image:
draft-lens19072204module157574320photo-1337779392.0-.png


The one that's labelled "USB 3.0 B plug pinout" is a little taller than the USB 2.0 B plug which is typically used by printers.
 
Last edited:

YBS1

Golden Member
May 14, 2000
1,945
129
106
I've frequently seen no notification when hooking up eSATA drives. Try opening up Storage Management as an admin, you may see the drive present and you will simply need to manually initialize them.
 

tinpanalley

Golden Member
Jul 13, 2011
1,451
22
81
Thanks. Looks pretty much like the card I have.
Does your computer notify you of anything untoward, like you've got a superspeed device that can perform faster?
If you pick a modern HDD and connect it up via USB 3.0 and benchmark it with this, what kind of results do you get?
https://www.attotech.com/disk-benchmark/
I will totally do that. I was looking for a program like that so I could post speeds with this thread but couldn't find one. Looking forward to checking the results. Thanks!!
I've frequently seen no notification when hooking up eSATA drives. Try opening up Storage Management as an admin, you may see the drive present and you will simply need to manually initialize them.
Yeah, tried that. They never show up with eSATA connections despite the fact that device manager shows me the card.
 

Burpo

Diamond Member
Sep 10, 2013
4,223
473
126
Does your card have a molex power connector & do you have it plugged in? The card I linked has a newer SI chipset and requires power supply lead .
 
Last edited:

mikeymikec

Lifer
May 19, 2011
17,633
9,447
136
I'm getting 38 MB/s Read, 45 MB/s Write. Too slow for USB 3.0 isn't it?

That's definitely in USB 3.0-ville. USB 2.0 tops out between 25-30MB/sec.

The most obvious device that could be slowing down throughput is the disk itself (or perhaps the disk being copied to). How old are the disks being copied from and to?
 

tinpanalley

Golden Member
Jul 13, 2011
1,451
22
81
That's definitely in USB 3.0-ville. USB 2.0 tops out between 25-30MB/sec.

The most obvious device that could be slowing down throughput is the disk itself (or perhaps the disk being copied to). How old are the disks being copied from and to?
Well the HDD is just a regular 7200rpm SATA HDD. Goes very quickly when connected via the mobo and when eSATA worked it went faster as well so I know it's capable. I could have gotten, I suppose, a crap dock that isn't really running at full speed.
The device copied from, a 2TB HDD is about 2 years old, the 1TB being copied to is about 4 years old.
 

tinpanalley

Golden Member
Jul 13, 2011
1,451
22
81
That's suspiciously close to USB 2.0 speeds. It seems like you might be having an issue where it's not linking up at USB 3.0. A bad connection or cable could cause this.
That's what I thought. I'll try to get another USB3.0 cable. Don't think I have another. I have to assume for now that the dock is fine because a cable replacement is a cheaper fix. Then I'll go from there if a new cable doesn't improve things. I know the mobo port is fine because it transfers at the proper speeds to flash drives.
 

mikeymikec

Lifer
May 19, 2011
17,633
9,447
136
Well the HDD is just a regular 7200rpm SATA HDD. Goes very quickly when connected via the mobo and when eSATA worked it went faster as well so I know it's capable. I could have gotten, I suppose, a crap dock that isn't really running at full speed.
The device copied from, a 2TB HDD is about 2 years old, the 1TB being copied to is about 4 years old.

It would be interesting to know what benchmark you get out of the drive when connected via SATA because there are enough variables to throw a data transfer speed test regardless of the interface used.

If we're talking 3.5" drives, the 7200.12 Seagate drive I previously used (bought in 2010) topped out at about 100MB/sec. A Seagate drive from around 2007 might do about 50-60MB/sec in my experience. 2.5" HDDs are a completely different kettle of fish, maybe half or 66% of the throughput for those eras. They may have caught up a bit more recently. My WD Black (2014) can do about 180MB/sec full tilt IIRC.

Having said that, I've seen my SSD sometimes only manage something like 170MB/sec when the benchmarks show it going full tilt at >500MB/sec, or my WD Black sometimes only doing about 100MB/sec or 130MB/sec.
 

tinpanalley

Golden Member
Jul 13, 2011
1,451
22
81
It would be interesting to know what benchmark you get out of the drive when connected via SATA because there are enough variables to throw a data transfer speed test regardless of the interface used.
That would be interesting, yeah. Just to make sure I'm doing this right, are ther instructions for that application anywhere?

Having said that, I've seen my SSD sometimes only manage something like 170MB/sec when the benchmarks show it going full tilt at >500MB/sec, or my WD Black sometimes only doing about 100MB/sec or 130MB/sec.
Confused.. if benchmarks say one thing and the actual transfer rate is something else, then what is the value of benchmarks? And are they different because the actual instance of copying specific files is different than the theoretical model tested in benchmarking software? Is that it?
 

mikeymikec

Lifer
May 19, 2011
17,633
9,447
136
That would be interesting, yeah. Just to make sure I'm doing this right, are ther instructions for that application anywhere?

I haven't followed any instructions or changed any settings. I just make sure that disk I/O has dropped to zero for say thirty seconds and then run it. If it's an SSD to be benchmarked, I've heard people recommend giving a computer half an hour of idle time before running the benchmark (so that garbage collection has likely done its thing).

Confused.. if benchmarks say one thing and the actual transfer rate is something else, then what is the value of benchmarks? And are they different because the actual instance of copying specific files is different than the theoretical model tested in benchmarking software? Is that it?

That's basically it, as well as other factors like what else is causing disk I/O in the background, state of fragmentation / garbage collection, AV, etc.

Benchmarks do have their uses though; you can be fairly sure that AV isn't messing with the benchmark, and if you've run a few benchmarks on similar drives you know what you ought to expect from the next drive, so if it doesn't hit in anywhere the same ball park, you know that something isn't quite right. ATTO helped me figure out an IDE driver problem once; Windows XP boot times were perfectly fine and the computer mostly seemed to perform "OK", but write performance was totally down the toilet. After I sorted the problem the computer did feel more responsive, but it wasn't "light and day" difference IMO.