What are the length limits for sata and esata cables?

Can someone point me to a source for PC adapter bracket that brings a
sata connector (not esata) out to the back of a PC.

Thanks.





--
Al Dykes
News is something someone wants to suppress, everything else is advertising.
- Lord Northcliffe, publisher of the Daily Mail

Al Dykes <adykes (AT) panix (DOT) com> wrote:

SATA: 1m. eSATA: 2m.

I got one with an ASUS mainboard. No idea whether it can be had
as acessory. Not so clean alternative is to use an 1m cable
and pull that out through some opening. I use such a set-up for
testing purposes, along with power (fused at 4A) pulled
out with a custom cable.

Arno
--
Arno Wagner, Dr. sc. techn., Dipl. Inform., CISSP -- Email: arno (AT) wagner (DOT) name
GnuPG: ID: 1E25338F FP: 0C30 5782 9D93 F785 E79C 0296 797F 6B50 1E25 338F
----
Cuddly UI's are the manifestation of wishful thinking. -- Dylan Evans

Arno wrote:
What are the limits for those eSATA cables that connect directly to a
SATA port in the motherboard through a simple converter?

Yousuf Khan

On 23 Feb 2010 07:36:21 -0500, adykes (AT) panix (DOT) com (Al Dykes) wrote:

Yousuf Khan <bbbl67 (AT) spammenot (DOT) yahoo.com> wrote:
1m total. True eSATA is electrically a bit (not a lot) different
to suppoer the increased cable lenght. However everything
going over (e)SATA is checksummed, so you should not get data
corruption with an occasional error. With a lot of interface
errors, you could also get silent corruption, since the checksums
have a probablility of not catching multiple errors in one
data package.

I would need to look up what checksums are used in detail
to estimate how likely that is. If it is CRC32, then the
probability of an uncaught error is very roughly one
in 2^32 for multiple (>= 32) random errors per data block.
That is pretty low and an SATA connection with that many
errors will be severely slowed down anyways, think > 100x
slower. In additition the OS (or RAID controller) will
decide that such a disk is unusable because of repeated
errors pretty fast.

Bottom line: It may fail or be slow but will in most cases
not silently corrupt your data, so you can try it out.
Just do not expet it to be really reliable, even if it
works.

Arno

--
Arno Wagner, Dr. sc. techn., Dipl. Inform., CISSP -- Email: arno (AT) wagner (DOT) name
GnuPG: ID: 1E25338F FP: 0C30 5782 9D93 F785 E79C 0296 797F 6B50 1E25 338F
----
Cuddly UI's are the manifestation of wishful thinking. -- Dylan Evans