How to quickly spot degraded P67 S-ATA

[NEOAethyr]

Better check with aida64 or speedfan, or something else.

If you now have these smart errors permanently on the drive, that's not good, even if the drive is still well.
It's gonna freak people out if that's what happened...

I smell a possible class action.
To early to tell though.

[dinos22]

its one of my 4-5 year old drives i use for benching lol. i just whacked it in an AMD board and it says the same thing, looks like this drive is a bit , i should probably use a newer one to check hehehe

[dinos22]

plugged in my new Corsair SSD drive and its showing no errors with it, looks like it was just my old arse drive thats seen its fair share of bluescreens from benching thats not happy

[Sparky]

here is the system log, there is quite a few of these errors scattered throughout the log. I havn't noticed any performance decrease though and my system is performing extremely well. So i don't know what to make of it!

http://i24.photobucket.com/albums/c1...7757/Log-1.jpg

Last time I saw that on someone's PC logs their SATA was on the way out (old well-used board, SIS chipset I think it was). Knew it wasn't the hard drive because it was tested OK in another system.

[some_one]

Better check with aida64 or speedfan, or something else.

Second that. Might also want to try different drivers and if the drive is using the latest firmware.

Here's a similar result with HDTune and SMART, except this is said to be using a SATA 6 port.

http://forums.anandtech.com/showpost...44&postcount=5

AFAIK the Intel problem is with the communications interface so rather than get crap data you would probably time out on CRC errors, with the disk possibly becoming unresponsive.

[shawnlut]

Something is wrong / conflict with the Intel Raid / Ahci driver version 10.1.0.1008 & HD Tune 4.60...
It will show similar error as the post #1...
Reverted back to version 9.6.0.1014 = NO Error with HD Tune 4.60...

[dinos22]

i have issues with new Intel driver on my main X58 rig actually and cannot go back to that v9.6, just not sure how to best do it without reinstalling the OS

[CoolZone]

Tested all 4 SATA2 ports on the ASRock Extreme6, along with the two SATA3 and I get the same results in HDTune, so I guess the ports are not degraded.

[tiborrr]

CoolZone: ... yet

[josh1980]

This thread is definitely interesting. But overall how are we going to know what's really wrong? Seriously? Every single SATA problem anyone has is now going to be blamed on Intel's p67 chipset. Hard drive works on my AMD but not my P67? Gotta be the chipset. CDROM won't work on port 3 but does on port 1? Gotta be the chipset. Smart errors? Gotta be the chipset. Nobody is going to try to defend an arguement that the drive isn't compatible, that the drivers are flawed, etc. Anyone who tries to make that arguement is going to be quickly squashed by the sudden onslaught of techies who are absolutely certain that p67s are unreliable. Now every reason for any device not working on a P67 is always going to be blamed on the same thing.... it's gotta be the chipset. Everyone is going to turn around and claim Intel has drastically underestimated the 5-15% failure rate over 3 years.

I'd heard through the grapevine that part of the reason Intel chose to go ahead with it's current plan of action was that Intel was anticipating that all SATA problems on P67s would be blamed on the chipset problem. Intel didn't want to have to deal with going to court and trying to prove a case when it would be next to impossible to prove their case. Imagine if Intel tested your failed p67 board and told you that your messed up P67 motherboard wasn't bad because of Intel's defect but because of something else. Who would really believe it when everyone already has the pre-conceived notion that their problem absolutely must be related to the chipset flaw and nothing else.

Based on some of the technical info I've read, I'd say the easiest way to determine if it REALLY is the chipset is to switch to another port that is still SATA 3Gbps. If all drives fail to be detected on, say, port 3, but work on port 4, then I might buy it. We're talking about a single transistor that is unique to each port. So one single port might fail, but the other ports would be unaffected unless they have been "used" about the same. (read here for the reason why I think a single port would fail but not all simultaneously) I think the only people that could confirm or deny this would be Intel employees that are "in the know".

Regardless, its funny how alot of people have bad SATA ports, but nobody was claiming they had a bunch of failed SATA ports here BEFORE Intel made the announcement. Now there's threads all over the place on many forums of people claiming to have a defective board.

Honestly, it's very disappointing that so many people think they have the answer to their problems. Look at all of the problems people have with RAID controllers and hard drives these days. Despite all the problems blamed on hard drive firmware, drivers, compatibility issues, few people have the actual answers. The companies aren't about to tell us the truth either, which only adds to the speculation and adds fuel to the fire of the conspiracy theorists. The rest of us just listen to everyone else and accept that you don't mix certain brand hard drives with certain brainds of RAID controllers.

[JMKS]

i just put my drive in all 3 different controllers (intel SATA2, Intel SATA3, Marvel SATA3) on my board and they all get the same results

there is a warning on that UltraDMA CRC Error Count lol so not sure that id use this software as gospel to find problems on the controllers exactly

I think that SMART is not doing this calculations on-the-fly, it checks that parameters while the disk is working and collects it. If errors occured, the "CRC error" parameter is updated and stored in SMART data for further reference.
That's why it shows errors regardless of the actual connection being good or bad.

[NEOAethyr]

I don't think it'll be that hard to track down some of this...

Don't run the drives in raid, that way you can use the generic ms driver (not really a big deal but in raid you might not have the smart data and temps, etc, whatever though).

If you have smart errors, check in the system logs.
If there's errors in the system logs, verify that the errors are real by copying files to the drive and checking the crc32's, if some of the crc32's don't match and the system is actually stable, then the controller is actually corrupting data.

If there are no real errors, then windows is getting mixed up on it, potential for causing slowdowns because it "thinks" there are errors when there is no real probs with the data it's self.

Problem is, not everyone knows what a crc32 is or how to check it.
Nor do they know if there drive is really any good or not.

Cables can cause corruption probs too, I'm sure someone around here at least knows that.

It's not gonna be easy for us to take a look at the data coming through to determain what is really happening.
I'm sure there's gonna be alot of false positives or instances where there is no solid proof one way or the other.
There's gonna be people that use the sata2 ports, that have some issues one way or another and then switch to the sata3 ports and have no probs.
And then apparently there's gonna be other types of probs to look through.

But in the end, when it comes down to it, rma the board's when the new ones are available.

Other then that, if it does end up that this problem is killing drives or messing them up, well, I guess it's all in the sake of messing around.
Not unless we start seeing a ton of dead drives that aren't say the ocz brand of ssd's lol.

What I'm interested in is if we see enough drives that "seem ok" (might not be verified but still) get bad smart data on the sata 2 controllers.
When moving them over to a diff controller they still have bad data (some values will stay that way regardless, but not all of them).
If we keep seeing something like that, even if the drives are still ok, it's still gonna be a big deal.
That's when people start freaking out thinking there drives have gone bad.
I'm not saying that is going to happen or anything.

I'm also interested in the stability of the south bridge as a whole (sata3 and other).
I know the press is saying all good.
But I'm still interested.

Edit:
Smart data is on the fly, within intervals...(timing, it's the programs used on that one(mostly, interface wise), most is within a sec or so)
Once it reports errors, it keeps track of them, it never gets refreshed to 0.
It happens regardless if you're using a program to check it or not.

[JMKS]

Smart data is on the fly, within intervals...(timing, it's the programs used on that one(mostly, interface wise), most is within a sec or so)
Once it reports errors, it keeps track of them, it never gets refreshed to 0.
It happens regardless if you're using a program to check it or not.

What I meant is that SMART is not generated only on the fly (which dinos kind-of suggested while saying it is strange that the error count is still there when connected to SATA3). Of course it is collected on the fly while disk is working, during normal usage.

[prava]

This seems like a ing competition:

"Lets see who blows the SATA2 first"

[jackbrennan2008]

This thread is definitely interesting. But overall how are we going to know what's really wrong? Seriously? Every single SATA problem anyone has is now going to be blamed on Intel's p67 chipset. Hard drive works on my AMD but not my P67? Gotta be the chipset. CDROM won't work on port 3 but does on port 1? Gotta be the chipset. Smart errors? Gotta be the chipset. ....snip...

That's a good post and it certainly is true. Although i do find my results above a little strange. The system log is reporting a controller error when i have my drives plugged into the SATA 2 ports, but not the SATA 3 ports. To be honest i didn't noticed any performance degredation and swapped over to the SATA 3 ports just to be on the safe side.

It's unfortunate that this has happened and i've already seen forum posts where people claim the P67 flaw has caused poor performance and many other things, but after using my system for 2 weeks with all my SATA 2 ports filled i didn't notice any performance loss at all dispite the system errors that were being logged.

I'm really not that annoyed and as long as i can swap my mobo out for a new one i'll be as happy as the day i got home with all my new components.

[cegras]

This thread is definitely interesting. But overall how are we going to know what's really wrong? Seriously? Every single SATA problem anyone has is now going to be blamed on Intel's p67 chipset. Hard drive works on my AMD but not my P67? Gotta be the chipset. CDROM won't work on port 3 but does on port 1? Gotta be the chipset. Smart errors? Gotta be the chipset. Nobody is going to try to defend an arguement that the drive isn't compatible, that the drivers are flawed, etc. Anyone who tries to make that arguement is going to be quickly squashed by the sudden onslaught of techies who are absolutely certain that p67s are unreliable. Now every reason for any device not working on a P67 is always going to be blamed on the same thing.... it's gotta be the chipset. Everyone is going to turn around and claim Intel has drastically underestimated the 5-15% failure rate over 3 years.

I'd heard through the grapevine that part of the reason Intel chose to go ahead with it's current plan of action was that Intel was anticipating that all SATA problems on P67s would be blamed on the chipset problem. Intel didn't want to have to deal with going to court and trying to prove a case when it would be next to impossible to prove their case. Imagine if Intel tested your failed p67 board and told you that your messed up P67 motherboard wasn't bad because of Intel's defect but because of something else. Who would really believe it when everyone already has the pre-conceived notion that their problem absolutely must be related to the chipset flaw and nothing else.

Based on some of the technical info I've read, I'd say the easiest way to determine if it REALLY is the chipset is to switch to another port that is still SATA 3Gbps. If all drives fail to be detected on, say, port 3, but work on port 4, then I might buy it. We're talking about a single transistor that is unique to each port. So one single port might fail, but the other ports would be unaffected unless they have been "used" about the same. (read here for the reason why I think a single port would fail but not all simultaneously) I think the only people that could confirm or deny this would be Intel employees that are "in the know".

Regardless, its funny how alot of people have bad SATA ports, but nobody was claiming they had a bunch of failed SATA ports here BEFORE Intel made the announcement. Now there's threads all over the place on many forums of people claiming to have a defective board.

Honestly, it's very disappointing that so many people think they have the answer to their problems. Look at all of the problems people have with RAID controllers and hard drives these days. Despite all the problems blamed on hard drive firmware, drivers, compatibility issues, few people have the actual answers. The companies aren't about to tell us the truth either, which only adds to the speculation and adds fuel to the fire of the conspiracy theorists. The rest of us just listen to everyone else and accept that you don't mix certain brand hard drives with certain brainds of RAID controllers.

I read:

Each set of ports requires its own PLL source.

...

The fact that the 3Gbps and 6Gbps circuits have their own independent clocking trees is what ensures that this problem is limited to only ports 2 - 5 off the controller.

Sounds like one transistor for all of 3Gbps to me.

Furthermore, do you know what I find a bit scary? There's NO mention that the transistor is turned off if nothing is plugged into the SATA 3 Gbps port. In fact, it could be running all the time. Which means that regardless of what you do, those ports will kill themselves over time.

Resale value = gone.

[Sam_oslo]

I read:



Sounds like one transistor for all of 3Gbps to me.

Furthermore, do you know what I find a bit scary? There's NO mention that the transistor is turned off if nothing is plugged into the SATA 3 Gbps port. In fact, it could be running all the time. Which means that regardless of what you do, those ports will kill themselves over time.

Resale value = gone.

Resale value? Who is going to sell a MB that will be refunded, or replaced soon?

EDIT:
With Intel's good move for a total recall. If you can get a refunds, then you have a 100% resale value. If you get it replaced, then you have a new MB that you can sell for higher than your used MB.

No matter how you look at it, now vi have got a MB with a higher resale value.!

[Boissez]

FWIW Gigabyte has put out a SATA-check utility for their boards:
http://gigabytedaily.blogspot.com/20...ata-check.html

EDIT: Actually all it does is checking which ports are populated and suggest to switch to using mainly Sata 3.0 ports. So kindda useless...

[SKYMTL]

This is a seriously misleading thread.

HDTune is about the worst utility one can use to check for this sort of thing since it exhibits a whole wack of problems when trying to read information from the the controller chip present on certain HDDs and SSDs. It also starts spouting errors in completely random situations.

Heck, I have a brand new Hitachi HDD that spouted a sea of red HDTune error messages from the moment it was plugged into my X58 board and yet never caused any type of performance issues. Samsung F1 drives have also been known to exhibit some issues as have certain Seagate and WD models. So as I said, HDTune is a poor tool for "how to quickly spot degraded P67" SATA ports.

Unfortunately, the OP linked us to a forum thread which contains more assumption rather than actual research into WHY the errors are being reported...

[initialised]

This seems like a ing competition:

"Lets see who blows the SATA2 first"

You say that like it's a bad idea.

[Leeghoofd]

You got a good tool to check the HD's Michael ? I'm running since monday 4 HD's (all different types) on the ports. One Western Digital is at 90% health in Harddisk Sentinel, HDD Scan and in the Western Digital lifeguard tools . I'm doing a daily Virus scan and a defrag... gonna hook up the bad drive to port 0 & 1 and see what gives...

EDIT: That particular drive is definetaly going down, nothing to do with the Intel chipset... time to backup :p

[halfwaythere]

HDtune is fine since the same drives work fine on the Sata3 ports and on other boards with different chipsets.

[SKYMTL]

You got a good tool to check the HD's Michael ?

Personally, I use HD Sentinel or WD's in-house tools. However, I would NEVER use any of them to diagnose potential PCH issues.

HDtune is fine since the same drives work fine on the Sata3 ports and on other boards with different chipsets.

Note necessarily. HDTune has the same issues with other controller / chipset combos. We have seen a few people in this thread mention oddities already.

[aytacaksel]

FWIW Gigabyte has put out a SATA-check utility for their boards:
http://gigabytedaily.blogspot.com/20...ata-check.html

EDIT: Actually all it does is checking which ports are populated and suggest to switch to using mainly Sata 3.0 ports. So kindda useless...

hmm now i know which port i used thx alot gb

[josh1980]

Sounds like one transistor for all of 3Gbps to me.

Just because they all share the same power source doesn't mean that they are all susceptible to the same failure. The ports may all be in a parallel configuration, which would give each set of port the same PLL voltage but they would still have independent transistors for each port.

From the anandtech article:

If you have a desktop system with six SATA ports driven off of P67/H67 chipset, there’s a chance (at least 5%) that during normal use some of the 3Gbps ports will stop working over the course of 3 years. The longer you use the ports, the higher that percentage will be. If you fall into this category, chances are your motherboard manufacturer will set up some sort of an exchange where you get a fixed board. The motherboard manufacturer could simply desolder your 6-series chipset and replace it with a newer stepping if it wanted to be frugal.

To me, that means that the ports can fail independent of each other. There was another article I read that provided a little more detail but I can't seem to find it right now.