SATA connectors

  • Thread starter Thread starter Son4ta
  • Start date Start date
S

Son4ta

Here is the text of a letter I am about to send to the Serial ATA
International Organization, any thoughts or comments in the NG's are
welcome.

Hello,
I have been using SATA drives (3 at the moment) since last year, and
have been anxiously awaiting SATA II ever since. As much for NCQ/TCQ
as for the 300mbs spec. Seagate has just come out with a drive that
supports NCQ, but at this time the only controller that supports
NCQ/TCQ is the Intel ICHR6 on the new 925 chipset mobo's. Why anyone
would buy one at this point in time is beyond me, I am waiting for
ASUS to come out with a board that supports DDR2, PCI Express, and
SATA II, and even then, I will wait until the dust settles a bit on
the implementation of all these new specs and the peripherals to go
with them.
The reason I am writing however, is that I would like to know what
DUMBASS (S) designed the SATA drive and power connectors. The design
of these is stupid almost beyond comprehension. They are SO flimsy.
The connectors on the cables are fine, but why, on the drives and the
ports, is there no protective shroud that fits around the connector
(like PATA or 4 pin Molex) to securely hold it in place?
The way the design is now, the cables connect loosely and if you are
not EXTREMELY careful, just moving the connector up or down a little
bit will break off the flimsy little piece of plastic that 'orients'
the connector. Ditto for the SATA power connectors. They break off
with almost no effort. How could the designers be so stupid? How could
they look at what they had designed and let it go out as it is without
thinking 'we should design it so that the connection is solid and
secure'.
I just don't get it.
Western Digital at least had the foresight to include a 'legacy'
standard Molex power connector as well as the flimsy SATA one. They
also have a little kit that will secure the SATA drive connector, but
it also blocks the SATA power connector. More than likely, they
realized how flimsy the SATA connector design is, came up with their
own 'fix', and included a Molex power connector to make it all work.
Why doesn't SATA use the Molex/PATA 'shroud' design approach to both
the SATA data and power connectors? I just don't get it and I won't be
buying any new SATA drives (even though I am in the market for at
least 2), until SATA II is implemented, the connectors are fixed, and
there is a board (preferably ASUS), that has native SATA/SATA II
support, PCI Express, and DDR2.

end

---------------------------------------------------------------
A p.s. for the NG's -

And at that point it will also have to be 64 Bit as Linux has become
my personal OS of choice and can take good advantage of 64 Bit.
Currently I dual-boot Mandrake 10 and Windows XP Pro. There is no
consumer 64 Bit Windows at the moment, I know they're working on it,
but Windows is a crappy OS anyway and a 64 Bit version is likely to
only be twice as crappy and troublesome as 32 Bit Windows is. I paid
$300.00 each for Windows 2000 and XP Pro full retail and I don't think
I will ever buy another Microsoft OS. For Windows I'll just fall back
to Windows 2000... it's as good as Windows is ever gonna get.

Peace, out

end
-----------------------------------------
 
Son4ta said:
Here is the text of a letter I am about to send to the Serial ATA
International Organization, any thoughts or comments in the NG's are
welcome.

Hello,
I have been using SATA drives (3 at the moment) since last year, and
have been anxiously awaiting SATA II ever since. As much for NCQ/TCQ
as for the 300mbs spec. Seagate has just come out with a drive that
supports NCQ, but at this time the only controller that supports
NCQ/TCQ is the Intel ICHR6 on the new 925 chipset mobo's. Why anyone
would buy one at this point in time is beyond me, I am waiting for
ASUS to come out with a board that supports DDR2, PCI Express, and
SATA II, and even then, I will wait until the dust settles a bit on
the implementation of all these new specs and the peripherals to go
with them.
The reason I am writing however, is that I would like to know what
DUMBASS (S) designed the SATA drive and power connectors. The design
of these is stupid almost beyond comprehension. They are SO flimsy.
The connectors on the cables are fine, but why, on the drives and the
ports, is there no protective shroud that fits around the connector
(like PATA or 4 pin Molex) to securely hold it in place?
The way the design is now, the cables connect loosely and if you are
not EXTREMELY careful, just moving the connector up or down a little
bit will break off the flimsy little piece of plastic that 'orients'
the connector. Ditto for the SATA power connectors. They break off
with almost no effort. How could the designers be so stupid? How could
they look at what they had designed and let it go out as it is without
thinking 'we should design it so that the connection is solid and
secure'.
I just don't get it.
Western Digital at least had the foresight to include a 'legacy'
standard Molex power connector as well as the flimsy SATA one. They
also have a little kit that will secure the SATA drive connector, but
it also blocks the SATA power connector. More than likely, they
realized how flimsy the SATA connector design is, came up with their
own 'fix', and included a Molex power connector to make it all work.
Why doesn't SATA use the Molex/PATA 'shroud' design approach to both
the SATA data and power connectors? I just don't get it and I won't be
buying any new SATA drives (even though I am in the market for at
least 2), until SATA II is implemented, the connectors are fixed, and
there is a board (preferably ASUS), that has native SATA/SATA II
support, PCI Express, and DDR2.

end

---------------------------------------------------------------
A p.s. for the NG's -

And at that point it will also have to be 64 Bit as Linux has become
my personal OS of choice and can take good advantage of 64 Bit.
Currently I dual-boot Mandrake 10 and Windows XP Pro. There is no
consumer 64 Bit Windows at the moment, I know they're working on it,
but Windows is a crappy OS anyway and a 64 Bit version is likely to
only be twice as crappy and troublesome as 32 Bit Windows is. I paid
$300.00 each for Windows 2000 and XP Pro full retail and I don't think
I will ever buy another Microsoft OS. For Windows I'll just fall back
to Windows 2000... it's as good as Windows is ever gonna get.

Peace, out

end
Click to expand...
"America, the land of the FREE!" If you want to express your opinion,
then you certainly should!
 
Here is the text of a letter I am about to send
Click to expand...

<snip>

I predict that they will delete the email right about the
time they get to the "dumbass" part, if not before then when
you start out egocentric.

Skip the personal preferences and name-calling and just
focus on concerns about mechanical integrity of the
connector, short and sweet.
 
just out of curiosity do you spend your time playing with your SATA
connectors? If you do I'm sure there are MIL spec connectors somewhere you
just might not get them for 5 bucks
 
Son4ta said:
Here is the text of a letter I am about to send to the Serial ATA
International Organization, any thoughts or comments in the NG's are
welcome.

Hello,
I have been using SATA drives (3 at the moment) since last year, and
have been anxiously awaiting SATA II ever since. As much for NCQ/TCQ
as for the 300mbs spec. Seagate has just come out with a drive that
supports NCQ, but at this time the only controller that supports
NCQ/TCQ is the Intel ICHR6 on the new 925 chipset mobo's. Why anyone
would buy one at this point in time is beyond me, I am waiting for
ASUS to come out with a board that supports DDR2, PCI Express, and
SATA II, and even then, I will wait until the dust settles a bit on
the implementation of all these new specs and the peripherals to go
with them.
The reason I am writing however, is that I would like to know what
DUMBASS (S) designed the SATA drive and power connectors. The design
of these is stupid almost beyond comprehension. They are SO flimsy.
The connectors on the cables are fine, but why, on the drives and the
ports, is there no protective shroud that fits around the connector
(like PATA or 4 pin Molex) to securely hold it in place?
The way the design is now, the cables connect loosely and if you are
not EXTREMELY careful, just moving the connector up or down a little
bit will break off the flimsy little piece of plastic that 'orients'
the connector. Ditto for the SATA power connectors. They break off
with almost no effort. How could the designers be so stupid? How could
they look at what they had designed and let it go out as it is without
thinking 'we should design it so that the connection is solid and
secure'.
I just don't get it.
Western Digital at least had the foresight to include a 'legacy'
standard Molex power connector as well as the flimsy SATA one. They
also have a little kit that will secure the SATA drive connector, but
it also blocks the SATA power connector. More than likely, they
realized how flimsy the SATA connector design is, came up with their
own 'fix', and included a Molex power connector to make it all work.
Why doesn't SATA use the Molex/PATA 'shroud' design approach to both
the SATA data and power connectors? I just don't get it and I won't be
buying any new SATA drives (even though I am in the market for at
least 2), until SATA II is implemented, the connectors are fixed, and
there is a board (preferably ASUS), that has native SATA/SATA II
support, PCI Express, and DDR2.

end

---------------------------------------------------------------
A p.s. for the NG's -

And at that point it will also have to be 64 Bit as Linux has become
my personal OS of choice and can take good advantage of 64 Bit.
Currently I dual-boot Mandrake 10 and Windows XP Pro. There is no
consumer 64 Bit Windows at the moment, I know they're working on it,
but Windows is a crappy OS anyway and a 64 Bit version is likely to
only be twice as crappy and troublesome as 32 Bit Windows is. I paid
$300.00 each for Windows 2000 and XP Pro full retail and I don't think
I will ever buy another Microsoft OS. For Windows I'll just fall back
to Windows 2000... it's as good as Windows is ever gonna get.

Peace, out

end
-----------------------------------------
Click to expand...

Yeah, a letter like this should be polite and technical. And the
organization you send it to, may not be set up to handle input
from the public. It might be better to identify key individuals
in the organization, and copy them on the letter. At least that
will make you feel better :-)

I downloaded my copy of the spec a while ago. It was here:

http://www.serialata.org/collateral/zipdownloads/serialata10a.ZIP

There are multiple documents in there, but reading from
"Serial ATA 1.0 gold.pdf" page 38, it says:

"This section covers Serial ATA connectors and cable assemblies.
It defines the
* Connector mating interfaces
* Connector location on the device
* Electrical, mechanical and reliability requirements of the
connectors and cable assemblies
* Connector and cable testing procedures

It, however, does not define how the connector and cable assembly
should be implemented, such as
* The mounting feature of the connectors
* The cabling and cable terminations
* The methods on how the PCB connects to other components of
the system"

So, for better or worse, they chose not to constrain the whole design
of the connector. That is typically done, so no connector manufacturer
feels cheated, by a particular companies design being incorporated
into the standard. I agree with you, that they should have put info
on lock latches and other mechanisms in the spec, to protect against
shock, vibration, and strain on the cables.

Looking at how the connector is designed, there is a secret plan
at work here. The SATA people don't give a damn about the use of
SATA drives in personal computers. The SATA connector was designed
first and foremost, to be plugged into a backplane. This is a typical
mechanism for building large servers (like the SCA connector?). IOW,
the SATA committee wants to complete with the SCSI boys - they wanted
a drive that could be hot plugged into a working server, and they
wanted the economy of having a backplane to seat the drive against.
A backplane doesn't have the same stress and strain issues as does
discrete cabling. The move to SATA II will be an attempt to match
SCSI's 320MB/sec transfer rate. The command queueing is an attempt
to match the SCSI feature set in a server environment. NCQ/TCQ or
whatever, has nothing to do with the desktop, and everything to do
with competing in the server domain.

I would still send the letter. After all, desktop sales of SATA
drives is paying the bills, and I feel that desktop users deserve a
better quality solution than they are currently getting. Maybe the
standards boys could come up with a few alterations, without changing
the rest of the design too much, that would allow both server
and desktop to be satisfied. Maybe if they gave the design to Amp,
Amp could come up with a latch for the connector - Amp loves to
put latches on stuff :-)

Paul
 
Paul said:
Looking at how the connector is designed, there is a secret plan
at work here. The SATA people don't give a damn about the use of
SATA drives in personal computers. The SATA connector was designed
first and foremost, to be plugged into a backplane. This is a typical
mechanism for building large servers (like the SCA connector?). IOW,
the SATA committee wants to complete with the SCSI boys - they wanted
a drive that could be hot plugged into a working server, and they
wanted the economy of having a backplane to seat the drive against.
Click to expand...
It's not so much they want to compete with scsi, but more about being
compatible to SAS (Serial Attached SCSI). In fact, the backplanes for
SATA and SAS are universal, so you can plug in both sorts of drives.
This of course means SAS and SATA connectors must be both physically and
electrically compatible.
I'd agree though that this should not mean that desktop-users get flimsy
connectors (can't comment on that personally, I have a sata-capable
motherboard and even some sata cables (which sure look nice compared to
these 80pin flat-ribbon suckers!) but no sata drives...).

Roland
 
Son4ta said:
Here is the text of a letter I am about to send to the Serial ATA
International Organization, any thoughts or comments in the NG's are
welcome.
Click to expand...

[complaint about ridiculously engineered SATA connectors removed]

I believe they are addressing the situation because I saw some new
connectors on some seagate drives at a trade show. I too find it
hard to believe that they implemented what they did originally. That's
why being an early adopter is not good (especially in the marketing
driven, rather than quality driven, states). There's a lot of examples
of the seemingly mindless implementation of products (in this case
it's actually a standard specification!) . As long as it makes money,
the companies consider it a success. Engineering excellence is
secondary. It's almost like all they have are electrical engineers who
are trying to do the mechanical engineering and they are failing
miserably at it!

SATA II (at least at first) won't have 3GB/s transfer rate, btw. That's a
year or so away still (so I have heard). And you'll need motherboard
support for it anyway or an add-in card. NCQ doesn't mean much to
desktop users, but for servers could be significant.

T
 
There is a latching/external version of the SATA connector due,
under SATA-2 - go read the full specs. The design is poor largely
from a latching perspective - they can come loose in shipping.

However, the 4-pin Molex connector was hardly brilliant - very
sloppy in design, large, cumbersome, sometimes hard to remove.

There are better connectors out there, and whilst the SATA data
connector is quite compact, the power version is stupidly large.
Consider 2006+ sees 2.5" SATA becoming common, and then
we have 2008+ putting 1.8" SATA in laptops - often in pairs.

The connector design seems to have been designed in 3.5" era,
a form factor which lasted a surprisingly long time - although the
high-speed (15.3k-rpm) drives are actually a 2.5" platter inside.

SAS/SATA compatibility is the focus, as well as backplane use,
but a latching inline (cable) version need not prohibit compatibility.

So SATA is the product of the system that created it:
o A committee that designed a horse, producing a camel
 
Roland Scheidegger said:
It's not so much they want to compete with scsi, but more about being
compatible to SAS (Serial Attached SCSI). In fact, the backplanes for
SATA and SAS are universal, so you can plug in both sorts of drives.
This of course means SAS and SATA connectors must be both physically and
electrically compatible.
I'd agree though that this should not mean that desktop-users get flimsy
connectors (can't comment on that personally, I have a sata-capable
motherboard and even some sata cables (which sure look nice compared to
these 80pin flat-ribbon suckers!) but no sata drives...).

Roland
Click to expand...

I didn't even know there was a serial SCSI. So I guess it isn't a
competition after all. Seems like a lot of duplicated effort
though, but I guess that is the bread and butter of standards
committees.

Paul
 
Back
Top