SATA cable suck
Posted: Tue Mar 30, 2004 11:16 pm
So I have some SATA 120g samsungs. Plugged them in to my AOPEN mobo and notice that the cables sure do wiggle around a lot, more on the drive than on the mobo. That cannot be good for connectivity.
Sure enough, I didnt figure it out before attempting a windows reinstall that go hosed so I had to start from scratch. it is only a matter of time before they will disconnect again and probably at a bad time since I am doing a RAID 1.
The cables simply do not hang on. Anyone have a solutions for this? I will also paste what I found out from the web. I would prefer to buy SATA drives that have a new connector as the current one is useless. I may try the bend-the-connectors method described below but these are new drives damnit!
Thanks for your wisdom once again.
This comes from http://www.lostcircuits.com forum.
======
especially if they are some of the originals that hit the market.
wuz having probs over the last 6 weeks, one drive was not getting recognized at boot. at first it was happening maybe once a week, but it just kept getting worse.
i changed out the Sata connectors but not the ribbons, still same problem, everytime it didn't boot a small wiggle on the connector/Sata Cable got her up and running with the next boot.
but, finally, tonight the connection just failed while surfing. big probs with the raid. thought it was lost.
long story short, put in a new Sata ribbon, the connection ends had more visable and higher contacts.
my my, after a few scarey moments and some disk checking, bootage was obtained and the raid array still lives. correctage of probs is wonderfully handled by the XP OS, big sigh.
the Sata ribbon i took off, although it looked exactly the same, internally it was not. inspection of the ends revealed very small and slightly recessed contacts, obviously erroding even more over time because of the poor contact.
wondering out loud if maybe newer ribbons have overcome this deficiency or if i just had a bad ribbon period.
baldy
By Yuri on Wednesday, November 05, 2003 - 05:37 am: Edit
There are quite a few issues with S-ATA cables as it stands. For example, a proposed erratum to the specification (one that is not in 1.0a) describes a connector that has a push-button latch to secure the connector to the device in order to prevent accidental disconnection, addressing one area of concern with current S-ATA connectors.
At T13, who are working to include the S-ATA spec into the ATA/ATAPI-7 standard, a lot of people have concerns about the quality of the connectors as it is. They say there are a lot of open issues still associated with them.
Not all S-ATA cables are of equal quality anyway, even if they adhere to the 1.0a spec completely. For example, cable size can be from 30 to 26 AWG (0.01003" to 0.01594" dia). Lower AWG means the cable is slightly thicker, but has less resistance and thus should provide a cleaner signal.
What is also not a requirement, but can help in noisy environments, is to have the cable include drain wires that are terminated to the ground pins in the S-ATA cable receptacle connectors. Another thing is that, according to the S-ATA spec, the choice of cable termination methods, such as crimping or soldering, is up to each connector vendor. One or the other is better, but I'm not sure which that would be in the case of these types of connectors. Suffice it to say though, that a properly crimped connector can be of better quality than one that was soldered with cheap, eh, solder.
One last thing that can determine signal quality is the cable length. P-ATA cables are allowed to be 18" maximally, S-ATA cables can be up to 1 metre (or 3' -- the spec isn't really clear on the exact length), but cables for S-ATA II can be of 18" maximum length again, the "Phase 1" edition anyway. That is, "Extensions to S-ATA" -- I don't know about the Phase 2 edition, it's still being developed.
My belief is simply: keep it as short as possible. If you can do with a 12" cable, why use a 3' one?
By MS on Wednesday, November 05, 2003 - 11:53 am: Edit
I know, I have been bitching at the SATA meetings about that problem myself, some of the connectors are so wobbly that they tend to fall off. The new proposals are with a latch or a little hook but it'll be awhile, until then, Ia m with Baldy, check your cables and, if necessary, use a rubber band to tie them to the drive. ( rarely had any problems on the board side of things)
By b on Friday, November 07, 2003 - 11:29 pm: Edit
further updateness regarding the attachments.
after a day or two of good boots, if happened again, but i DO HAVE the FIX!
the adaptor itself is to blame, i always wiggled the adaptor to get the boot, so i figured the hard drive ide pins were loose inside the Sata Adaptor.
took five minutes and bent all the pins slightly further apart on the hard drive IDE socket and reinserted the Sata Adaptor, Walla!
fixage this time for sure, not hearing the hard drive boinking in and out like it used to do, smooth as glass.
so there ya go.
finally.
baldy
Sure enough, I didnt figure it out before attempting a windows reinstall that go hosed so I had to start from scratch. it is only a matter of time before they will disconnect again and probably at a bad time since I am doing a RAID 1.
The cables simply do not hang on. Anyone have a solutions for this? I will also paste what I found out from the web. I would prefer to buy SATA drives that have a new connector as the current one is useless. I may try the bend-the-connectors method described below but these are new drives damnit!
Thanks for your wisdom once again.
This comes from http://www.lostcircuits.com forum.
======
especially if they are some of the originals that hit the market.
wuz having probs over the last 6 weeks, one drive was not getting recognized at boot. at first it was happening maybe once a week, but it just kept getting worse.
i changed out the Sata connectors but not the ribbons, still same problem, everytime it didn't boot a small wiggle on the connector/Sata Cable got her up and running with the next boot.
but, finally, tonight the connection just failed while surfing. big probs with the raid. thought it was lost.
long story short, put in a new Sata ribbon, the connection ends had more visable and higher contacts.
my my, after a few scarey moments and some disk checking, bootage was obtained and the raid array still lives. correctage of probs is wonderfully handled by the XP OS, big sigh.
the Sata ribbon i took off, although it looked exactly the same, internally it was not. inspection of the ends revealed very small and slightly recessed contacts, obviously erroding even more over time because of the poor contact.
wondering out loud if maybe newer ribbons have overcome this deficiency or if i just had a bad ribbon period.
baldy
By Yuri on Wednesday, November 05, 2003 - 05:37 am: Edit
There are quite a few issues with S-ATA cables as it stands. For example, a proposed erratum to the specification (one that is not in 1.0a) describes a connector that has a push-button latch to secure the connector to the device in order to prevent accidental disconnection, addressing one area of concern with current S-ATA connectors.
At T13, who are working to include the S-ATA spec into the ATA/ATAPI-7 standard, a lot of people have concerns about the quality of the connectors as it is. They say there are a lot of open issues still associated with them.
Not all S-ATA cables are of equal quality anyway, even if they adhere to the 1.0a spec completely. For example, cable size can be from 30 to 26 AWG (0.01003" to 0.01594" dia). Lower AWG means the cable is slightly thicker, but has less resistance and thus should provide a cleaner signal.
What is also not a requirement, but can help in noisy environments, is to have the cable include drain wires that are terminated to the ground pins in the S-ATA cable receptacle connectors. Another thing is that, according to the S-ATA spec, the choice of cable termination methods, such as crimping or soldering, is up to each connector vendor. One or the other is better, but I'm not sure which that would be in the case of these types of connectors. Suffice it to say though, that a properly crimped connector can be of better quality than one that was soldered with cheap, eh, solder.
One last thing that can determine signal quality is the cable length. P-ATA cables are allowed to be 18" maximally, S-ATA cables can be up to 1 metre (or 3' -- the spec isn't really clear on the exact length), but cables for S-ATA II can be of 18" maximum length again, the "Phase 1" edition anyway. That is, "Extensions to S-ATA" -- I don't know about the Phase 2 edition, it's still being developed.
My belief is simply: keep it as short as possible. If you can do with a 12" cable, why use a 3' one?
By MS on Wednesday, November 05, 2003 - 11:53 am: Edit
I know, I have been bitching at the SATA meetings about that problem myself, some of the connectors are so wobbly that they tend to fall off. The new proposals are with a latch or a little hook but it'll be awhile, until then, Ia m with Baldy, check your cables and, if necessary, use a rubber band to tie them to the drive. ( rarely had any problems on the board side of things)
By b on Friday, November 07, 2003 - 11:29 pm: Edit
further updateness regarding the attachments.
after a day or two of good boots, if happened again, but i DO HAVE the FIX!
the adaptor itself is to blame, i always wiggled the adaptor to get the boot, so i figured the hard drive ide pins were loose inside the Sata Adaptor.
took five minutes and bent all the pins slightly further apart on the hard drive IDE socket and reinserted the Sata Adaptor, Walla!
fixage this time for sure, not hearing the hard drive boinking in and out like it used to do, smooth as glass.
so there ya go.
finally.
baldy
So, I am supporting them with another Velcro strap -- and luckily on this motherboard, hot plugging is supported! 
It actually *snaps* into the motherboard socket and has *far* less play than either of the other two SATA cables I have on hand (Tripplite & ASRock supplied). The Tripplite is piss-poor and "fits" very loosely...