A Disgusting and Humiliating Discovery

[BonzaiDuck]

I'll try and make this simple and short.

2007, July: After extensively modding a pre-ATX Compaq ProLiant-Server case, I built my first LGA-775 system quad-core with 680i chipset and a 3Ware 9650SE hardware RAID controller. I chose the Seagate 7200.10 320GB drives, and installed four in RAID5 on that controller. And I THOUGHT that these things were running at 3GB/sec or full SATA2 potential.

Late 2007, early 2008: I picked up some spare Seagate drives of the same size, model and spec -- for spares in case a drive in the RAID5 went bad. Two of these, I used to build a VISTA64 system with RAID0 this year.

December, 2008: One of those drives went "south" -- went "on the fritz" -- and I ordered WD Caviar Black 500's to replace both of them.

Now I have the Seagates from the VISTA system, and bought a "BlacX" ThermalTake docking station (eSATA and USB2) so that I could test them and use the good one for backup.

I'm getting old -- over 60 -- and my eyes aren't what they used to be. I needed a magnifying glass to review the diagram and small print on the manufacturer's label for the Seagates.

NOW I discover that the default tiny gray shunt or jumper on the drives limits their operation to 1.5 GB/sec operation!! Yeah! I've been running that RAID5 array on the other system as they were shipped -- with the jumper installed.

I could KICK myself!! I started working with this technology in 1982, and this is something I should've caught. I just can't believe it.

Now -- I'm guessing that I can remove the shunts/jumpers on all four drives in the RAID5 array, and performance will increase by leaps and bounds.

Does anyone know if I'll run into problems -- having created the array for drives running at a slower speed? I've gone all this time without capturing the full potential of either the controller or the array of drives attached to it.

 

[Zap]

You shouldn't run into any problems, but don't expect performance to be much better.

 

[tcsenter]

Just double-confirm the jumper indeed is for SATA mode rather than other features such as capacity limit or spread spectrum. A lot of drives use jumper blocks for those features, as well.

 

[taltamir]

it is most likely because the drive is very unlikely to ever go FASTER than 1.5gbps. They are selling it with a 3gbps connector because people for some reason prefer SATA2 drives over SATA1 drives even if the SATA1 drive is much faster.

There should be ABSOLUTELY NO PROBLEM with you removing that jumper and just enjoying better speed... but to play it safe, turn off the computer, do it for ONE drive. turn it on... if it DIDN'T work you would get a degraded array (which is fixable because its just one drive), but I see absoltely no reason why it shouldn't work...

Intels SSD is the first drive ever to go over 1.5gbps (1.5 giga BITS per second... aka 187.5 mega BYTES per second)... There is NO regular platter drive, AT ALL, that can get faster than SATA1 speed. MAYBE they have some instantanous peak speed going over that on occasion, but I seriously doubt that! Most are ~ 50MB/s or 70 MB/s

 

[Denithor]

taltamir +1

3Gbps = (3,000,000,000 bit/sec)/(8 bit/byte) = (375,000,000 byte/sec)/(1,000,000 byte/MB) = 375 MB/s

1.5Gbps = 187.5 MB/s

As talt mentioned most drives have <100 MB/s sustained transfer rates so there is unlikely to be much difference observed by removing the jumpers & therefore the speed limit.

Even the newest SSD drives from Intel offer transfer rates in the 200-250 MB/s range.

 

[Shmee]

you make a good point about MOST drives not being fast enough to fill the SATA I cap. But what about a velociraptor? Or a 15K seagate cheetah?

 

[taltamir]

the velociraptor gets 120MB/s compared to the 50 or 70 on most drives, well below the 187.5MB/s limit of SATA1.

Oh, and those 32MB of cache? that is to prevent peak speeds from mattering... the drive reads into cache, while the SATA protocol transfers the data from the cache to the ram and CPU.

the 15k cheeta and the like get 110-140MB/s, still under SATA1 speed... and they are typically SCSI drives, not SATA (of any number).

 

[thermite88]

Originally posted by: BonzaiDuck
2007, July: After extensively modding a pre-ATX Compaq ProLiant-Server case, I built my first LGA-775 system quad-core with 680i chipset and a 3Ware 9650SE hardware RAID controller. I chose the Seagate 7200.10 320GB drives, and installed four in RAID5 on that controller. And I THOUGHT that these things were running at 3GB/sec or full SATA2 potential.

Late 2007, early 2008: I picked up some spare Seagate drives of the same size, model and spec -- for spares in case a drive in the RAID5 went bad. Two of these, I used to build a VISTA64 system with RAID0 this year.

December, 2008: One of those drives went "south" -- went "on the fritz" -- and I ordered WD Caviar Black 500's to replace both of them.

Now I have the Seagates from the VISTA system, and bought a "BlacX" ThermalTake docking station (eSATA and USB2) so that I could test them and use the good one for backup.

I'm getting old -- over 60 -- and my eyes aren't what they used to be. I needed a magnifying glass to review the diagram and small print on the manufacturer's label for the Seagates.

NOW I discover that the default tiny gray shunt or jumper on the drives limits their operation to 1.5 GB/sec operation!! Yeah! I've been running that RAID5 array on the other system as they were shipped -- with the jumper installed.

I could KICK myself!! I started working with this technology in 1982, and this is something I should've caught. I just can't believe it.

Now -- I'm guessing that I can remove the shunts/jumpers on all four drives in the RAID5 array, and performance will increase by leaps and bounds.

Does anyone know if I'll run into problems -- having created the array for drives running at a slower speed? I've gone all this time without capturing the full potential of either the controller or the array of drives attached to it.

Reading your post is like reading my own script except the outcome. I ordered my first Seagate 7200.10 320GB drive in May 2007 and eventually added up to 4 drives for a RAID5 based on Intel ICH8R southbridge. Never get it to work.

http://forums.anandtech.com/me...=y&keyword1=thermite88

18 months and 4 failed drives later, I gave up on the RAID and tried the Seagates in single drive mode using AHCI. Making a little progress that the Vista 32 completed installation and all the upgrades except Service Pack 1 and any of the language packages. I still constantly get blue screen of death. It happened whenever the disk access light came on for a long time and, then, another memory dump and reboot. BTW, I did remove the jumper from the first day on. Tried putting them back on and it did not improve the system stability.

I still have two of the Seagates in somewhat working condition. One is lost by USPS in return to Seagate who decline to help. Another one is a DOA replacement from Seagate. I have been on the phone half a donzen times to get a pre-paid mailing label. Seagate always said that they will email it, but never did. The 5-years warranty is good, but no substitute for reliable drives.

I am in the process of deciding on a replacement set of either the Samsung Spinpoint F1 RAID or WD Caviar Black. The 500 GB drives are priced about the same from different vendors. Both are slightly less than what I paid for my Seagates in 2007.

 

[Zap]

Originally posted by: thermite88
One is lost by USPS in return to Seagate who decline to help.

Seagate Return Packing and Shipping Information

You must pay postage charges for all shipments from you to Seagate.

Send your product to Seagate using the carrier of your choice. Please use a method with ability to trace shipment. Seagate is not responsible for shipments for which the carrier cannot provide proof of delivery.

Seagate is not responsible for loss or damage during transit. We suggest that you purchase freight insurance on your shipments.

USPS has delivery confirmation, not tracking. Those are two different things.

Did you purchase insurance? If not and USPS did not deliver, you have zero recourse.

 

[BonzaiDuck]

Originally posted by: thermite88

Originally posted by: BonzaiDuck
2007, July: After extensively modding a pre-ATX Compaq ProLiant-Server case, I built my first LGA-775 system quad-core with 680i chipset and a 3Ware 9650SE hardware RAID controller. I chose the Seagate 7200.10 320GB drives, and installed four in RAID5 on that controller. And I THOUGHT that these things were running at 3GB/sec or full SATA2 potential.

Late 2007, early 2008: I picked up some spare Seagate drives of the same size, model and spec -- for spares in case a drive in the RAID5 went bad. Two of these, I used to build a VISTA64 system with RAID0 this year.

December, 2008: One of those drives went "south" -- went "on the fritz" -- and I ordered WD Caviar Black 500's to replace both of them.

Now I have the Seagates from the VISTA system, and bought a "BlacX" ThermalTake docking station (eSATA and USB2) so that I could test them and use the good one for backup.

I'm getting old -- over 60 -- and my eyes aren't what they used to be. I needed a magnifying glass to review the diagram and small print on the manufacturer's label for the Seagates.

NOW I discover that the default tiny gray shunt or jumper on the drives limits their operation to 1.5 GB/sec operation!! Yeah! I've been running that RAID5 array on the other system as they were shipped -- with the jumper installed.

I could KICK myself!! I started working with this technology in 1982, and this is something I should've caught. I just can't believe it.

Now -- I'm guessing that I can remove the shunts/jumpers on all four drives in the RAID5 array, and performance will increase by leaps and bounds.

Does anyone know if I'll run into problems -- having created the array for drives running at a slower speed? I've gone all this time without capturing the full potential of either the controller or the array of drives attached to it.

Reading your post is like reading my own script except the outcome. I ordered my first Seagate 7200.10 320GB drive in May 2007 and eventually added up to 4 drives for a RAID5 based on Intel ICH8R southbridge. Never get it to work.

http://forums.anandtech.com/me...=y&keyword1=thermite88

18 months and 4 failed drives later, I gave up on the RAID and tried the Seagates in single drive mode using AHCI. Making a little progress that the Vista 32 completed installation and all the upgrades except Service Pack 1 and any of the language packages. I still constantly get blue screen of death. It happened whenever the disk access light came on for a long time and, then, another memory dump and reboot. BTW, I did remove the jumper from the first day on. Tried putting them back on and it did not improve the system stability.

I still have two of the Seagates in somewhat working condition. One is lost by USPS in return to Seagate who decline to help. Another one is a DOA replacement from Seagate. I have been on the phone half a donzen times to get a pre-paid mailing label. Seagate always said that they will email it, but never did. The 5-years warranty is good, but no substitute for reliable drives.

I am in the process of deciding on a replacement set of either the Samsung Spinpoint F1 RAID or WD Caviar Black. The 500 GB drives are priced about the same from different vendors. Both are slightly less than what I paid for my Seagates in 2007.

Just to share latest developments: as I mentioned, I've begun to notice at age 61 what the computer said to spaceman "Dave" in the movie "2001 Space Odyssey:" My mind . . . is going . . . . I can feel it. . . . I can feel it . . . "

Apparently, I DID read the drive label on the Seagates when I installed the 4-drive RAID5 in 2007. The shunts had been removed from all of them. But even for the fact that I designed and built slide-out latching cages to make the array more accessible, it was still a P-I-T-A pulling it out just to check, and then find that I'd done everything properly when I first installed them.

 

[VirtualLarry]

I had the opposite experience with the shunts. I ordered four WD 320GB SATA2 HDs, and I was using them with a SI3114 PCI 4-port SATA RAID controller card, that only supported SATA1. I kept losing drives out of teh array, for no good reason.

So I jumpered all of the drives for 1.5 operation. Still didn't help. Never got my situation resolved, so I parted out the drives from the RAID5 array. I finally gave it another shot with WD 500GB SATA drives, and this time it seems to be working.

 

[BonzaiDuck]

Originally posted by: VirtualLarry
I had the opposite experience with the shunts. I ordered four WD 320GB SATA2 HDs, and I was using them with a SI3114 PCI 4-port SATA RAID controller card, that only supported SATA1. I kept losing drives out of teh array, for no good reason.

So I jumpered all of the drives for 1.5 operation. Still didn't help. Never got my situation resolved, so I parted out the drives from the RAID5 array. I finally gave it another shot with WD 500GB SATA drives, and this time it seems to be working.

The Caviar Black WD 500's seem to have consistent and rave reviews, given their price. As I say in another thread, I'm getting ready to run the VISTA performance-index program to see if they added something over the crippled (7200.10's set to SATA1 speed) RAID0.

$70 each and no shipping is a good price for those WDs at NewEgg.

 

[Denithor]

The 640GB drives are slightly better, higher platter density = more throughput.

 

[EvilBob]

I second Denithor on the WD 640s - I dropped a pair of SE16 640s in my last build and love 'em. The newer Caviar Black 640s are comparable, and both are typically in the top tier in the reviews.

 

[Woofmeister]

Don't feel bad, I did the exact same thing with my data drive. I was excited when I discovered the issue and eagerly configured the jumper for 3GB. It made no difference in performance whatsoever.

Still, it's nice to know I have that theoretical 3.0 GB transfer rate. :roll: