I think I've solved my Spin Retry Count problem

[Yousuf Khan]

I have stated previously that I have an older drive that seems to get a
lot of Spin Retry Counts in SMART, but other than that it's rock solid:
it has outlasted a whole bunch of newer drives that were previously
giving no error messages whatsoever in SMART. Well, it looks like I've
now also solved its Spin Retries too.

Here's the older thread for reference:

comp.sys.ibm.pc.hardware.storage | Google Groups
http://groups.google.com/group/comp...han+spin+retry&&q=bbbl67+OR+yjkhan+spin+retry

Well anyway, what I did recently was that I converted two of those old
IDE drives into SATA drives using an IDE-to-SATA converter board that
fits right onto the back of the drives themselves. I started noticing
that the Spin Retries were gone about two weeks ago, but that was only a
week after I installed them, so I didn't want to say anything until I
saw a bit more time pass. Well, it's now the three week mark, and I
haven't seen any additional spin retries! I've also rebooted the system
quite a bit more often than I usually do in that time. HD Sentinel is
also now starting to upgrade its Health rating automatically, it had
gone down to 40% at its lowest and today it's at 52%.

I had previously suspected the IDE cables to be the culprits, so I did
change them, but that didn't help any at that time. So at that point I
began suspecting the drive itself was getting old (which it is). Looks
like now that it's going through SATA cables instead of IDE ones, that
the cables were the culprit, even with the newer IDE cables. I'm not
sure if this is endemic to IDE cables in general, or if it was just two
bad IDE cables in a row.

Yousuf Khan

 

[Daniel Prince]

I had previously suspected the IDE cables to be the culprits, so I did
change them, but that didn't help any at that time. So at that point I
began suspecting the drive itself was getting old (which it is). Looks
like now that it's going through SATA cables instead of IDE ones, that
the cables were the culprit, even with the newer IDE cables. I'm not
sure if this is endemic to IDE cables in general, or if it was just two
bad IDE cables in a row.

How long were your IDE cables? Were they flat or round?

 

[Rod Speed]

Yousuf Khan wrote

I have stated previously that I have an older drive that seems to get a lot of Spin Retry Counts in SMART, but other
than that it's rock solid: it has outlasted a whole bunch of newer drives that were previously giving no error
messages whatsoever in SMART. Well, it looks like I've now also solved its Spin Retries too.

Here's the older thread for reference:

comp.sys.ibm.pc.hardware.storage | Google Groups
http://groups.google.com/group/comp...han+spin+retry&&q=bbbl67+OR+yjkhan+spin+retry

Well anyway, what I did recently was that I converted two of those old IDE drives into SATA drives using an
IDE-to-SATA converter board that fits right onto the back of the drives themselves. I started noticing that the Spin
Retries were gone about two weeks ago,

Gone from what report ? They shouldnt go from the Everest SMART
report, the raw value is a total and should just stop increasing.

but that was only a week after I installed them, so I didn't want to say anything until I saw a bit more time pass.
Well, it's now the three week mark, and I haven't seen any additional spin retries!

Presumably thats what you meant and the previous is just poorly worded.

I've also rebooted the system quite a bit more often than I usually do in that time. HD Sentinel is also now starting
to upgrade its Health rating automatically, it had gone down to 40% at its lowest and today it's at 52%.

I've never considered those numbers are very useful, what matters is the raw value.

I had previously suspected the IDE cables to be the culprits, so I did
change them, but that didn't help any at that time. So at that point I
began suspecting the drive itself was getting old (which it is). Looks
like now that it's going through SATA cables instead of IDE ones, that
the cables were the culprit, even with the newer IDE cables. I'm not
sure if this is endemic to IDE cables in general, or if it was just
two bad IDE cables in a row.

Its much more likely to be a lot more complicated than that.

What else did you change when you moved from IDE to SATA ?

Did you change the motherboard or power supply ?

The old thread did say you upgraded the power supply.
Did you see the spin retry count keep increasing after that ?

The only raw value you did post wasnt that high, just 20.

 

[Yousuf Khan]

How long were your IDE cables? Were they flat or round?

They were around 20 inches long, from the furthest ends.

Now, they were rounded cables in both cases. I had thought of using flat
cables, but in the crowded conditions in my case it was pretty difficult
to use them without moving a lot of stuff around. Plus the rounded
cables gave better cooling.

Yousuf Khan

 

[Yousuf Khan]

Yousuf Khan wrote

I've never considered those numbers are very useful, what matters is the raw value.

Well, they seem to be based on the same criteria that we'd use to judge
quality: how many errors are there, and how often do they occur.
Recently they've been occurring much less frequently, so it's been
upgrading its rating for the drive.

Its much more likely to be a lot more complicated than that.

What else did you change when you moved from IDE to SATA ?

Did you change the motherboard or power supply ?

Neither were changed. Power supply was obviously changed in the previous
round of troubleshooting, so it didn't need to be changed again.
Motherboard and processor remained what they were too; as a matter of
fact the reason this change was done was to eventually upgrade the
motherboard and processor down the road -- it's extremely difficult to
find new motherboards with more than one IDE channel anymore, so I am
preparing for the next generation. It was difficult to find a
motherboard with two IDE channels (i.e. 4 connectors) the last time too.

There were 4 SATA connections, and 4 IDE connections. Two SATAs were
previously unused, so I just moved two IDEs over to SATA. The only
things left on IDE are the optical drives.

The old thread did say you upgraded the power supply.
Did you see the spin retry count keep increasing after that ?

Yes, it did keep increasing after that. But the PS upgrade fixed many
other unrelated problems so I wasn't unhappy about that.

The only raw value you did post wasnt that high, just 20.

Well, it's gone upto 32 since then -- and holding.

Yousuf Khan

 

[Michael Cecil]

They were around 20 inches long, from the furthest ends.

Now, they were rounded cables in both cases. I had thought of using flat
cables, but in the crowded conditions in my case it was pretty difficult
to use them without moving a lot of stuff around. Plus the rounded
cables gave better cooling.

Yousuf Khan

I thought of using certified life vests, but the cardboard ones were much
cheaper. Plus the cardboard ones were easier to store.

 

[Arno]

I had previously suspected the IDE cables to be the culprits, so I did
change them, but that didn't help any at that time. So at that point I
began suspecting the drive itself was getting old (which it is). Looks
like now that it's going through SATA cables instead of IDE ones, that
the cables were the culprit, even with the newer IDE cables. I'm not
sure if this is endemic to IDE cables in general, or if it was just two
bad IDE cables in a row.

Yousuf Khan

Well, it is nice that the problem went away, but I frankly cannot
imagine any way an IDE cable would be responsible. Did you change or
at least unplug/replug the power connections? Bad power
connections can cause spin problems as especially old drives
draw a lot of power on spindle startup.

Arno

 

[Rod Speed]

Yousuf Khan wrote

Rod Speed wrote

Well, they seem to be based on the same criteria that we'd use to
judge quality: how many errors are there, and how often do they occur.

The raw numbers for that particular drive dont support those 40% and 52%
claims, particularly with a parameter like spin retry count thats much more
likely to be a problem external to the drive than internal to the drive.

Recently they've been occurring much less frequently, so it's been upgrading its rating for the drive.

It shouldnt have ever given it a 40% with the raw values you posted.

Neither were changed. Power supply was obviously changed in the
previous round of troubleshooting, so it didn't need to be changed
again. Motherboard and processor remained what they were too; as a matter of fact the reason this change was done was
to eventually upgrade the motherboard and processor down the road -- it's extremely difficult to find new motherboards
with more than one IDE channel anymore, so I am preparing for the next generation. It was difficult to find a
motherboard with two IDE channels (i.e. 4 connectors) the last time too.

Sure, I wondered whether you had changed the motherboard already.

There were 4 SATA connections, and 4 IDE connections. Two SATAs were previously unused, so I just moved two IDEs over
to SATA. The only things left on IDE are the optical drives.

Yes, it did keep increasing after that. But the PS upgrade fixed many other unrelated problems so I wasn't unhappy
about that.

Well, it's gone upto 32 since then -- and holding.

I still find it hard to believe its a cable problem before the change to SATA.

 

[YKhan]

Well, it is nice that the problem went away, but I frankly cannot
imagine any way an IDE cable would be responsible. Did you change or
at least unplug/replug the power connections? Bad power
connections can cause spin problems as especially old drives
draw a lot of power on spindle startup.

Arno

The power connections were all brand new when the new power supply was
previously installed. The exact same power cables are still attached,
but now the data cables are SATA.

Yousuf Khan

 

[Wad Medani]

Well, it is nice that the problem went away, but I frankly cannot
imagine any way an IDE cable would be responsible. Did you change or
at least unplug/replug the power connections? Bad power
connections can cause spin problems as especially old drives
draw a lot of power on spindle startup.

Arno

His cables were round instead of falt ribbon and over the 18" length
spec limit so it could have very well been his cables. Round cables
are for case mod doodz and not computer savvy minded people.

*----------------------------------------------------------------------*
"A good neighbor is one who stays on their side of the ****ING fence."
*----------------------------------------------------------------------*

 

[Mike Tomlinson]

Well anyway, what I did recently was that I converted two of those old
IDE drives into SATA drives using an IDE-to-SATA converter board that
fits right onto the back of the drives themselves. I started noticing
that the Spin Retries were gone about two weeks ago, but that was only a
week after I installed them

I think the change in behaviour won't have anything to do with the
cables. A dodgy IDE cable is going to affect a lot more than the Spin
Retry count.

It's probably more to do with the fact hat you have now converted the
drive to SATA. Those plug-in converters are notoriously unreliable
(especially if you buy the ebay cheapo ones). I've used several
different ones with mixed results. How do you know the SMART data is
getting through from the drive correctly, if at all?

 

[Franc Zabkar]

Well, it is nice that the problem went away, but I frankly cannot
imagine any way an IDE cable would be responsible. Did you change or
at least unplug/replug the power connections? Bad power
connections can cause spin problems as especially old drives
draw a lot of power on spindle startup.

Arno

I agree. There seems to be no plausible causal relationship between a
faulty data cable and a spin retry issue. It would be a bit like
blaming a flat tyre on a flat battery.

I'd be watching and comparing other SMART attributes, eg ...

201 C9 Soft Read Error Rate (Number of off-track errors)

207 CF Spin High Current (Amount of surge current used to spin up the
drive)

208 D0 Spin Buzz (Number of buzz routines needed to spin up the drive
due to insufficient power)

209 D1 Offline Seek Performance (Drive’s seek performance during its
internal tests)

This article attempts to explain the SMART attributes:
http://en.wikipedia.org/wiki/S.M.A.R.T.#Known_ATA_S.M.A.R.T._attributes

Here are the OP's previous data:

C9  <vendor-specific>                    0    253  252           1
CF  <vendor-specific>                    0    224  207          23
D0  <vendor-specific>                    0    253  252           0
D1  <vendor-specific>                    0    241  241         154

If the OP is handy with a multimeter, he may be able to measure the
actual startup current by locating the current sense resistors for the
spindle motor and measuring the voltage across them.

In Seagate drives that employ an ST Microelectonics "SMOOTH" spindle
motor and voice coil motor (VCM) combo controller chip, there will be
five parallel connected 1R00 (1.00 ohm) resistors. The four 1R00
resistors sense the VCM current.

Here is a datasheet for an L7250 controller:
http://wandrew.regruppa.ru/PCInfo/TechDoc/L7250(Smooth).pdf

I believe this may be the OP's PCB:
http://www.tux-tech.ca/images/logo/pcb.jpg

The spindle motor current sensing resistors appear to be R512, R511,
R510, ...


- Franc Zabkar

 

[YKhan]

I agree. There seems to be no plausible causal relationship between a
faulty data cable and a spin retry issue. It would be a bit like
blaming a flat tyre on a flat battery.

There is one possibility that might explain it that I've been thinking
of recently. When the drive was connected via IDE, it was one of the
first drives detected by the BIOS, whereas now that it's connected via
SATA it's one of the last drives. It's perhaps possible that the
additional resting time in between gives it more time to get ready for
the BIOS when the detection procedure comes around to it.


Yousuf Khan

 

[YKhan]

It's probably more to do with the fact hat you have now converted the
drive to SATA.  Those plug-in converters are notoriously unreliable
(especially if you buy the ebay cheapo ones). I've used several
different ones with mixed results.  How do you know the SMART data is
getting through from the drive correctly, if at all?

As a matter of fact, I did get them on Ebay. They are not perfect, I
found them to be pretty finicky about being pushed in tightly, but
once they started working without flaw.

How do I know that the SMART data is getting through? Well because HD
Sentinel displays them, quite simply. HDS always polls that data, it's
not saved data.

Yousuf Khan

 

[YKhan]

His cables were round instead of falt ribbon and over the 18" length
spec limit so it could have very well been his cables. Round cables
are for case mod doodz and not computer savvy minded people.

There were two drives connected through that same cable. It was always
just the one drive that had any problems, the other one didn't. There
are good reasons for non-case-mod-doodz to use rounded cables too:
they are more flexible, and they allow better airflow.

Yousuf Khan

 

[Rod Speed]

There is one possibility that might explain it that I've been thinking
of recently. When the drive was connected via IDE, it was one of the
first drives detected by the BIOS, whereas now that it's connected via
SATA it's one of the last drives. It's perhaps possible that the
additional resting time in between gives it more time to get ready for
the BIOS when the detection procedure comes around to it.

That wouldnt produce spin retry counts in the SMART stats.

 

[Franc Zabkar]

There is one possibility that might explain it that I've been thinking
of recently. When the drive was connected via IDE, it was one of the
first drives detected by the BIOS, whereas now that it's connected via
SATA it's one of the last drives. It's perhaps possible that the
additional resting time in between gives it more time to get ready for
the BIOS when the detection procedure comes around to it.

AIUI, all drives spin up as soon as power is applied, unless you have
enabled PUIS (power up in standby), in which case the BIOS or the OS
could stagger the spinup. But I still can't see how either scenario
could affect the spin retry count.

Maybe warmer weather has made the fluid bearing a little less viscous
???

- Franc Zabkar

 

[Arno]

The power connections were all brand new when the new power supply was
previously installed. The exact same power cables are still attached,
but now the data cables are SATA.

Very strange indeed.

Arno

 

[Arno]

His cables were round instead of falt ribbon and over the 18" length
spec limit so it could have very well been his cables. Round cables
are for case mod doodz and not computer savvy minded people.

For interface errors and data corruption, yes. But for
Spin-Retry? How would that work?

Incidentially I have made good experiences with rounded cables
as long as they were twisted pair, it was at least ATA66
(has mandatory checksums) and cable length was at or below 60cm.
The only problem I was with a CD burner that only used ATA33,
apparently without the checksums that are optional below ATA66.

Arno

 

[Arno]

There is one possibility that might explain it that I've been thinking
of recently. When the drive was connected via IDE, it was one of the
first drives detected by the BIOS, whereas now that it's connected via
SATA it's one of the last drives. It's perhaps possible that the
additional resting time in between gives it more time to get ready for
the BIOS when the detection procedure comes around to it.

Actually that makes sense. Or that it just has more time to spin-up
before it is detected and needs to answer commands.

Arno