why is eSATA enclosure with Oxford chipset slow?

[ananymous]

I have both 3.5 and 2.5 external enclosures, and with the 3.5 I am getting at least 5MB/s slower results on all of my speed measurements with HDtach.

The thing is, the 3.5 hdd is rated as much faster than the hdd in the 2.5 enclosure. I know the enclosure's controller/chipset matters too, but the 3.5 uses an Oxford chipset which I thought was one of the best. I'm not sure of the controller used in the 2.5 enclosure (see model below), but it was inexpensive and I never thought it would be better than Oxford which I've been hearing for years is superior.

I also tested another hard drive enclosure in the past...I didn't keep it, but I made note of the HDtach stats, and those results were also better than my current 3.5 setup. So that's the only two enclosures I've compared to it, and both beat the Oxford setup.

Can I conclude that the enclosure (or the Oxford 934DSA chipset) is not very good then? In tom's benchmarks the 2.5 hdd is much slower than the 3.5 I have. Here are model numbers of the components I'm comparing:

For the 2.5 I have a WD 3200BEKT-00F3T0 (320gb, 16mb cache) hdd in an Eagle ET-CS2XMESU2-BK enclosure

For the 3.5 I have a Samsung HD103SJ 1AJ1 (1tb, 32mb cache) hdd in a Bytecc ME-535SSU2FW-BK enclosure (with Oxford 934DSA chip)

One other thing I might mention is that I'm using esata through a pcmcia card on my laptop, so no setup will ever be able to realize full speeds they are capable of due to speed restrictions of card port on the laptop. Never seen any speed measurement (including burst) over 65.4 MB/s. However, I'm not trying to go faster than what the pcmcia card is able...I'm just trying to figure out why a faster drive and highly-rated chipset are performing worse.
So maybe the Oxford isn't the best at dealing with that card bottleneck?? Otherwise I don't understand what's going on.
Any ideas? Anyone know where I can see enclosure benchmarks?
Thanks!

 

[CurseTheSky]

5MB/s is really within the margin of error. If it was more like 15-20 MB/s I'd say you have a problem.

Are you consistently getting 5MB/s lower than the 2.5" hard drive, or did you just test once or twice? It could have to deal with fragmentation on the 3.5" hard drive.

 

[alexruiz]

I have both 3.5 and 2.5 external enclosures, and with the 3.5 I am getting at least 5MB/s slower results on all of my speed measurements with HDtach.

The thing is, the 3.5 hdd is rated as much faster than the hdd in the 2.5 enclosure. I know the enclosure's controller/chipset matters too, but the 3.5 uses an Oxford chipset which I thought was one of the best. I'm not sure of the controller used in the 2.5 enclosure (see model below), but it was inexpensive and I never thought it would be better than Oxford which I've been hearing for years is superior.

I also tested another hard drive enclosure in the past...I didn't keep it, but I made note of the HDtach stats, and those results were also better than my current 3.5 setup. So that's the only two enclosures I've compared to it, and both beat the Oxford setup.

Can I conclude that the enclosure (or the Oxford 934DSA chipset) is not very good then? In tom's benchmarks the 2.5 hdd is much slower than the 3.5 I have. Here are model numbers of the components I'm comparing:

For the 2.5 I have a WD 3200BEKT-00F3T0 (320gb, 16mb cache) hdd in an Eagle ET-CS2XMESU2-BK enclosure

For the 3.5 I have a Samsung HD103SJ 1AJ1 (1tb, 32mb cache) hdd in a Bytecc ME-535SSU2FW-BK enclosure (with Oxford 934DSA chip)

One other thing I might mention is that I'm using esata through a pcmcia card on my laptop, so no setup will ever be able to realize full speeds they are capable of due to speed restrictions of card port on the laptop. Never seen any speed measurement (including burst) over 65.4 MB/s. However, I'm not trying to go faster than what the pcmcia card is able...I'm just trying to figure out why a faster drive and highly-rated chipset are performing worse.
So maybe the Oxford isn't the best at dealing with that card bottleneck?? Otherwise I don't understand what's going on.
Any ideas? Anyone know where I can see enclosure benchmarks?
Thanks!

That is your solution. All my enclosures are Oxford 934DSA, and with the same drive as yours (HS103SJ) I easily get the 140 MB/s read that you get internally. I have eSATa ports in motherboard.

The OX934DSA is more stable that the competition, but as far as performance, they are very very close. I have had enclosures with initio and sunplus chipsets and they also read/write fast, which shouldn't be surprising as using eSATA the bridge chipsets are just going on pass-thru.

 

[ananymous]

5MB/s is really within the margin of error. If it was more like 15-20 MB/s I'd say you have a problem.

Are you consistently getting 5MB/s lower than the 2.5" hard drive, or did you just test once or twice? It could have to deal with fragmentation on the 3.5" hard drive.

thanks so much for responding.
I forgot to mention also that the test results were with the *unused* 3.5 drive, right after formatting. I tested multiple times, then and now (after adding files to the drive) but same results.
I just don't understand how the 3.5 could be slower than both "generic" enclosures (with slower hdd's) I compared it to.

btw, I just tested again while writing this, and this time it was more like 9MB difference.

surely someone does enclosure benchmarks somewhere?? Or compares the chipsets they use?

 

[ananymous]

That is your solution. All my enclosures are Oxford 934DSA, and with the same drive as yours (HS103SJ) I easily get the 140 MB/s read that you get internally. I have eSATa ports in motherboard.

The OX934DSA is more stable that the competition, but as far as performance, they are very very close. I have had enclosures with initio and sunplus chipsets and they also read/write fast, which shouldn't be surprising as using eSATA the bridge chipsets are just going on pass-thru.

thanks...maybe it is related to the capacity of the external hdd's and how they respond when they hit the pcmcia bottleneck...the 1TB drive being the largest is slower?
Mind if I ask what enclosure model are you using? I googled that particular Oxford chip and it seemed to always lead back to the Bytecc enclosure.

 

[Emulex]

i actually ran into a bridge USB TO ESATA recently.. i was like wtf this is sorta faster but taking up 100% CPU. talk about sucky.

does esata push full sata-II/III safely? or just sata-I speed?

 

[Old Hippie]

surely someone does enclosure benchmarks somewhere??

I've never seen an "enclosure benchmark" round-up.

Or compares the chipsets they use?

I've only ever seen this one but it's IDE/USB and from 2007. I didn't think that the enclosure's chipset actually processes the eSATA signal but just lets it pass-thru.

eSATA should give regular SATA speeds but sometimes/somehow the enclosure's chipset seem to slow it down a little.

You may be interested in the cardbus controllers comparison even though it's from 2008.

 

[alexruiz]

thanks...maybe it is related to the capacity of the external hdd's and how they respond when they hit the pcmcia bottleneck...the 1TB drive being the largest is slower?
Mind if I ask what enclosure model are you using? I googled that particular Oxford chip and it seemed to always lead back to the Bytecc enclosure.

Oh, I don't mind at all 🙂
The Samsung HD103SJ is a a beast of a drive 😀, hard drives matter little here, the interface is the problem

I have:

- Rocstor Rocport SA (2.5" eSATa, firewire and USB, Oxford 943DSA)
- Buslink CO-500-UFS2 (3.5, eSATa, firewire and USB, Oxford 943DSA) This one doesn't show in the specs. Buslink sells this enclosure with minimal tech information. I don't think even they know what it has inside 😛 I know if the 934 because I swapped drives. I got the buslink drives on fleabay from buslink ebay store. As you know, swapping drives invalidates the warranty
- Vantec Nexstar 360-UFS (3.5, eSATa, firewire and USB, Oxford 943DSA) This is a very nice enclosure, and if it wasn't for the proprietary power adapter, it might be perfect. I got it over 2 years ago, and as you can imagine, it wasn't cheap.


Alex


Edit: The Rocstor drive came with a 120GB drive inside. As I did with the buslink, I took the original drive out and replaced with a travelstar 7K500.