UDMA CRC error count

Do you have a question? Post it now! No Registration Necessary.  Now with pictures!

Threaded View
I'm thinking of buying a used 1TB HDD as a backup for
non-critical data. On request, the owner has provided a health
report by HD Tune. It has one line highlighted in yellow. The
line is "Ultra DMA CRC Error Count". The values are : Current -
200, Worst - 200, Threshold - 0, Data - 2, Status - warning

I know what DMA and CRC mean separately, but could anyone please
explain the significance of the report? I've heard that this
error is caused by a faulty cable (I'm assuming that this
includes poor contact at the connector). Is that all it is - that
a CRC error occurred twice in the past (probably due to the
cable) and normally does not indicate any problem with the drive

AFAICT, there doesn't seem to be any other abnormalities. No
reallocated sector, no spin retry. Some other values -
Start/Stop count - 206
Power on hours - 181

TIA for any helpful input.

Re: UDMA CRC error count

Pimpom wrote:
Quoted text here. Click to load it

As far as I've been able to determine so far, that CRC counter
doesn't get reset. And it's probably a count of cable
errors. If the drive itself had a defective SATA interface
chip, the counts could be caused by that, rather than a cable.
So the statistic doesn't identify what was broken. Just
that errors were detected. It could have been a defective
cable, but there is no way to be sure.



Re: UDMA CRC error count

Pimpom wrote:
Quoted text here. Click to load it

My only personal experience with this involves:

1) An internal Samsung that was slightly bumped during operation
(sitting horizontally, bumped on side).  1 UDMA error, no reallocated

2) an external vertical drive that fell on its side during operation.
Lots of UDMA errors, and so many reallocated sectors that it always
failed SMART testing right away.  After the drive was completely
overwritten, no visible defective or slow sectors were detected during
scans with MHDD.  I believe this drive is still in use as a backup.

3) Seagate, WD, Samsung, and Hitachi SATA drives plugged into a SATA-
PATA converter based on a SunPlus chip.  Each developed 1 UDMA error
after being scanned with HDAT2 and the program locked up with an error
that required reboot.  No reallocated sectors.  Apparently the new
version of this SunPlus chip doesn't have that this defect.

I've had lots of problems with SATA cables not being plugged in
straight, but whenever that happened the drive wasn't recognized at
all.  You can avoid that by using SATA data cables with metal clips on
their sockets, and Meritline.com often offers them for under $1,
delivered, when they run a coupon deal (coupon usually expires within
the first 3 hours after they open).

If you're buying used drives locally, try to test them with MHDD (self
boots from CD or USB stick)  because when indicates what sectors are
marginal and require multiple attempts to read them.   Apparently some
manufacturers considered it normal if a sector requires a dozen
retries to read because each such brand new drive still passed its
factory diagnostics.  A full scan of a 1TB drive takes roughly 3
hours, much more if it's connected through USB 2.0.   There's a
Windows version of MHDD, called HDDscan, but it's a lot more likely to
report slow sectors, probably because of stalls caused by Windows, so
I wouldn't trust it.   BTW these two programs don't seem to execute
SMART self diagnostics right with drives bigger than 1TB or 1.5TB.

Re: UDMA CRC error count

larry moe 'n curly wrote:
Quoted text here. Click to load it

Thanks LMC, I was looking for the name of that app just yesterday and
couldn't find it in my records.  I just built a system that's a little
sluggish and I was looking to remap any bad sectors to see if that cleared
it up.


Re: UDMA CRC error count

Pimpom wrote:
Quoted text here. Click to load it

Thanks for the replies and sorry about the late one from my side.
My internet connection has been acting up again. I decided to buy
the drive and am expecting it today or tomorrow. With help from
your inputs, I'm taking the chance that the CRC errors must have
been due to external causes.

Site Timeline