Velocity Reviews - Computer Hardware Reviews

Velocity Reviews > Newsgroups > Computing > Computer Information > Serial ATA

Reply
Thread Tools

Serial ATA

 
 
Richard Dower
Guest
Posts: n/a
 
      07-19-2003

"DeMoN LaG" <de_on-lag@co_cast.net> wrote in message
news:Xns93BD33B95D03AWobbly@216.168.3.30...
> Robert Baer <(E-Mail Removed)> wrote in
> news:(E-Mail Removed):
>
> > Convert it to raw serial data, without any added control bits =
> > 800Mbits/sec or almost 1Ghz data rate.
> > Then add in the control functions, which translates to 1Ghz or more..
> > And the claim it is faster than parallel IDE????????????????????

>
> See, the thing is right now SATA isn't really much faster (if any). The
> difference is that PATA is at the end of it's road, and has little future.
> SATA /can/ go faster when needed.


Indeed...SATA 300 is due mid 2004, and SATA 600 is due sometime 2006.


 
Reply With Quote
 
 
 
 
Robert Baer
Guest
Posts: n/a
 
      07-20-2003
DeMoN LaG wrote:
>
> Robert Baer <(E-Mail Removed)> wrote in
> news:(E-Mail Removed):
>
> > Convert it to raw serial data, without any added control bits =
> > 800Mbits/sec or almost 1Ghz data rate.
> > Then add in the control functions, which translates to 1Ghz or more..
> > And the claim it is faster than parallel IDE????????????????????

>
> See, the thing is right now SATA isn't really much faster (if any). The
> difference is that PATA is at the end of it's road, and has little future.
> SATA /can/ go faster when needed.


Only with rise/fall times in the 100 picosecond region on the basis of
what i stated.
Assuming one used 1Ghz signalling rate with 100 picosecond rise/fall
times, it is obvious that a parallel (byte-wide) protocol is *eight
times* faster than a serial (bit-wide) protocol.
If one used a fast signalling method like that to get greater speed,
then it is stupid to slow down the data transfer by a factor of eight
(at minimum), thereby wasting that greater speed.
If one thinks that 100Mbyte/sec is slow, certainly 1,000Mbytes/sec
would be fairly fast.
Obviously, the cable as-is cannot easily support such a data ratem so
differnt vabling would be in order - perhaps similar to whatever is
being contemplated for SATA (have not looked) ?
 
Reply With Quote
 
 
 
 
Tim Auton
Guest
Posts: n/a
 
      07-20-2003
Robert Baer <(E-Mail Removed)> wrote:
>Thor wrote:
>> "GB" <(E-Mail Removed)> wrote in message news:(E-Mail Removed)...
>> > Surely, then, the message is to avoid SATA for now. Whatever hardware is
>> > bought now will presumably not work with the new faster standard when it
>> > comes out?
>> >
>> > Is this just hype-ware?

>>
>> No, the intention is that future S-ATA modes will be backward compatible,
>> just as parallel ATA drives and controllers have been backward compatible
>> throughout it's lifetime. The key limitations you will likely see over time
>> will be in the BIOS just as before.

>
> A single cable with data and control signals cannot be physically or
>electronically "compatible" with a parallel cable having seperate data
>and signal lines.
> So anyone claiming direct "backwards compatibility" is a liar.


Thor meant future S-ATA modes will be backwards compatible with the
current S-ATA spec, not with PATA. In the same way as you can use an
ATA66 drive with an ATA133 controller or a USB1.1 device with a USB2.0
controller (or whatever they call USBx.x these days).


Tim
 
Reply With Quote
 
Thor
Guest
Posts: n/a
 
      07-20-2003

"Robert Baer" <(E-Mail Removed)> wrote in message
news:(E-Mail Removed)...
> Thor wrote:
> >
> > "GB" <(E-Mail Removed)> wrote in message news:(E-Mail Removed)...
> > >
> > > "DeMoN LaG" <de_on-lag@co_cast.net> wrote in message
> > > news:Xns93BD33B95D03AWobbly@216.168.3.30...
> > > > Robert Baer <(E-Mail Removed)> wrote in
> > > > news:(E-Mail Removed):
> > > >
> > > > > Convert it to raw serial data, without any added control bits =
> > > > > 800Mbits/sec or almost 1Ghz data rate.
> > > > > Then add in the control functions, which translates to 1Ghz or

> > more..
> > > > > And the claim it is faster than parallel IDE????????????????????
> > > >
> > > > See, the thing is right now SATA isn't really much faster (if any).

The
> > > > difference is that PATA is at the end of it's road, and has little

> > future.
> > > > SATA /can/ go faster when needed.
> > >
> > > Surely, then, the message is to avoid SATA for now. Whatever hardware

is
> > > bought now will presumably not work with the new faster standard when

it
> > > comes out?
> > >
> > > Is this just hype-ware?

> >
> > No, the intention is that future S-ATA modes will be backward

compatible,
> > just as parallel ATA drives and controllers have been backward

compatible
> > throughout it's lifetime. The key limitations you will likely see over

time
> > will be in the BIOS just as before.

>
> A single cable with data and control signals cannot be physically or
> electronically "compatible" with a parallel cable having seperate data
> and signal lines.
> So anyone claiming direct "backwards compatibility" is a liar.


First, I wasn't talking about direct backward compatibility between P-ATA
and S-ATA. I was referring to backward compatibility between different
generations of S-ATA. This is what the poster I repsonded to was asking.
whether S-ATA drives available now will work with the faster generations of
S-ATA controllers that are brought out in the future. There was backward
compatibility between different generations of P-ATA. S-ATA won't be any
different in that regard, thus no need to avoid using S-ATA. However, as to
the issue you mistakenly addressed, there are adapters available right now
that allow P-ATA drives to interface with S-ATA controllers and ATA drives
to interface with P-ATA controllers. No, it's not "direct" backward
compatibility, but it's damn close. Close enough that the compatibitliy
issues between the two standards are a moot point.



 
Reply With Quote
 
V W Wall
Guest
Posts: n/a
 
      07-20-2003
Robert Baer wrote:
>
> DeMoN LaG wrote:
> >
> > Robert Baer <(E-Mail Removed)> wrote in
> > news:(E-Mail Removed):
> >
> > > Convert it to raw serial data, without any added control bits =
> > > 800Mbits/sec or almost 1Ghz data rate.
> > > Then add in the control functions, which translates to 1Ghz or more..
> > > And the claim it is faster than parallel IDE????????????????????

> >
> > See, the thing is right now SATA isn't really much faster (if any). The
> > difference is that PATA is at the end of it's road, and has little future.
> > SATA /can/ go faster when needed.

>
> Only with rise/fall times in the 100 picosecond region on the basis of
> what i stated.
> Assuming one used 1Ghz signalling rate with 100 picosecond rise/fall
> times, it is obvious that a parallel (byte-wide) protocol is *eight
> times* faster than a serial (bit-wide) protocol.
> If one used a fast signalling method like that to get greater speed,
> then it is stupid to slow down the data transfer by a factor of eight
> (at minimum), thereby wasting that greater speed.
> If one thinks that 100Mbyte/sec is slow, certainly 1,000Mbytes/sec
> would be fairly fast.
> Obviously, the cable as-is cannot easily support such a data ratem so
> differnt vabling would be in order - perhaps similar to whatever is
> being contemplated for SATA (have not looked) ?


Thor explained the problems with increasing the data transfer rate with
the present parallel ATA standard(s). Also, remember that the ultimate
origin and destiny of the data is on the surface of a hard disk platter.
This is, of itself, a "serial" data train.

There is some work being done on "parallel" data recording. Multi layers
in optical data media are being investigated, but using magnetic media,
SATA looks like the best near-term solution.

Incidently, without some major change in optical recording, the data density
is limited by the wavelength of the recording "light". Magnetic recording
has no such theoretical limitation.

Virg Wall

--
Be not the first by whom the new are tried,
Nor yet the last to lay the old aside.
 
Reply With Quote
 
AD C
Guest
Posts: n/a
 
      07-20-2003
DeMoN LaG wrote:


> See, the thing is right now SATA isn't really much faster (if any). The
> difference is that PATA is at the end of it's road, and has little future.
> SATA /can/ go faster when needed.


That is true, but there is one problem with SATA and that is you can
only put one drive on each port, so to put four drive into a computer,
you need 4 SATA ports.

I will not doubt buy a SATA drive when I get a larger hard drive, the
prices are coming down already and as my board supports them, it leves
the PATAs free for CDwriter and other slow products.

 
Reply With Quote
 
AD C
Guest
Posts: n/a
 
      07-20-2003
GB wrote:


> Surely, then, the message is to avoid SATA for now. Whatever hardware is
> bought now will presumably not work with the new faster standard when it
> comes out?
>
> Is this just hype-ware?
>
>

It will be beack ward compatible, just like PATA and usb is now.

 
Reply With Quote
 
Thor
Guest
Posts: n/a
 
      07-20-2003

"AD C" <(E-Mail Removed)> wrote in message
news:bfekb0$b5q$(E-Mail Removed)...
> DeMoN LaG wrote:
>
>
> > See, the thing is right now SATA isn't really much faster (if any). The
> > difference is that PATA is at the end of it's road, and has little

future.
> > SATA /can/ go faster when needed.

>
> That is true, but there is one problem with SATA and that is you can
> only put one drive on each port, so to put four drive into a computer,
> you need 4 SATA ports.


True, but putting 4 S-ATA ports on the motherboard, or a controller card
isn't a problem. They are very small in comparison to parallel ATA
connections. Even with 4 S-ATA cables in the box, it is still easier to
route and less obstructive than having 2 big fat ribbon cables. Also, you
don't have to worry about having ribbon cables where the connectors can't
properly reach both drives at the same time. Each drive gets it's own cable.
Also no master/slave or CS jumpering to worry about. Just plug and go.

>
> I will not doubt buy a SATA drive when I get a larger hard drive, the
> prices are coming down already and as my board supports them, it leves
> the PATAs free for CDwriter and other slow products.
>



 
Reply With Quote
 
Robert Baer
Guest
Posts: n/a
 
      07-21-2003
V W Wall wrote:
>
> Robert Baer wrote:
> >
> > DeMoN LaG wrote:
> > >
> > > Robert Baer <(E-Mail Removed)> wrote in
> > > news:(E-Mail Removed):
> > >
> > > > Convert it to raw serial data, without any added control bits =
> > > > 800Mbits/sec or almost 1Ghz data rate.
> > > > Then add in the control functions, which translates to 1Ghz or more..
> > > > And the claim it is faster than parallel IDE????????????????????
> > >
> > > See, the thing is right now SATA isn't really much faster (if any). The
> > > difference is that PATA is at the end of it's road, and has little future.
> > > SATA /can/ go faster when needed.

> >
> > Only with rise/fall times in the 100 picosecond region on the basis of
> > what i stated.
> > Assuming one used 1Ghz signalling rate with 100 picosecond rise/fall
> > times, it is obvious that a parallel (byte-wide) protocol is *eight
> > times* faster than a serial (bit-wide) protocol.
> > If one used a fast signalling method like that to get greater speed,
> > then it is stupid to slow down the data transfer by a factor of eight
> > (at minimum), thereby wasting that greater speed.
> > If one thinks that 100Mbyte/sec is slow, certainly 1,000Mbytes/sec
> > would be fairly fast.
> > Obviously, the cable as-is cannot easily support such a data ratem so
> > differnt vabling would be in order - perhaps similar to whatever is
> > being contemplated for SATA (have not looked) ?

>
> Thor explained the problems with increasing the data transfer rate with
> the present parallel ATA standard(s). Also, remember that the ultimate
> origin and destiny of the data is on the surface of a hard disk platter.
> This is, of itself, a "serial" data train.
>
> There is some work being done on "parallel" data recording. Multi layers
> in optical data media are being investigated, but using magnetic media,
> SATA looks like the best near-term solution.
>
> Incidently, without some major change in optical recording, the data density
> is limited by the wavelength of the recording "light". Magnetic recording
> has no such theoretical limitation.
>
> Virg Wall
>
> --
> Be not the first by whom the new are tried,
> Nor yet the last to lay the old aside.


Ahhhh, you mentioned something i thought of in the mid 1980s..namely
that since all platters have heads, then *parallel* output is so easy
and obvious that it is a wonder how stupid HD makers have been (and for
so long).
Hell, with 8 platters (16 heads) and a different cable configuration,
true 16-bit computing could have been achieved at the outset of the
80386!
That was roughly the time that both IDE and SCSI first came out,
replacing the older MFM multi-cable systems; a lot of new schemes and
"space" for a (then) kick-ass truly parallel system.
Oh, you want to *double* the file access speed (on open and close,
anyway)?
Just put the FAT in the middle of the disk instead of having it at the
edge!
Not my idea; Tandy did that on 8 inch floppies, model 80 if i remember
correctly.
 
Reply With Quote
 
Thor
Guest
Posts: n/a
 
      07-21-2003

"AD C" <(E-Mail Removed)> wrote in message
news:bfgf6g$coj$(E-Mail Removed)...
> Thor wrote:
>
>
> >
> > True, but putting 4 S-ATA ports on the motherboard, or a controller card
> > isn't a problem. They are very small in comparison to parallel ATA
> > connections. Even with 4 S-ATA cables in the box, it is still easier to

>
> True, my old board got 2 PATA and two PATa raid and they take up a fair
> bit of space on the board, my new Abit, got two PAta and two S-ATA and
> it is difficult to find the S-ATA on the board, they are so small.
>
>
> > route and less obstructive than having 2 big fat ribbon cables. Also,

you
> > don't have to worry about having ribbon cables where the connectors

can't
> > properly reach both drives at the same time. Each drive gets it's own

cable.
>
> I hate that, I had a full size towercase and my old board had the
> connection low down on the motherboard, it was a pain to connect the CD
> writer and dvd drive. The smaller cables are nice, I must admit.
>
>
>
> > Also no master/slave or CS jumpering to worry about. Just plug and go.

>
> That is also nice as well.
>
> The only problem is that for a while P-ATA will still need to be on the
> board due to the fact there ar so many products that use them. CDrom,
> DVD rom,DVD writer,Cd writer, Zip drives and the LS drive. plus others.
>
>
> How do Sata drives get their power, do they use the normal molex

connector?

No. They use a new style small connector adjacent to the S-ATA Data
connector. However, current S-ATA drives are often coming with the old-style
molex for compatibility reasons. There are also adapters available for older
power supplies. The new ATX spec coming out will have the new S-ATA power
connector, so new power supplies should start showing up with it.


 
Reply With Quote
 
 
 
Reply

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
The best SIP ATA-1000 and SIP ATA-1100 with The lowest price!!! voxquick via HWKB.com VOIP 0 04-03-2006 12:24 PM
Comunications ATA 186 to ATA 186 (without any gatekeeper or CallManager) SoGo VOIP 1 03-12-2005 07:39 AM
Serial ATA HD And Ultra/ATA HD's In Same PC: O.K. ? Robert11 Computer Support 1 10-28-2004 10:14 PM
Ultra ATA/100 HDD on a Ultra ATA/33 motherboard Prinxce of Darkxness Computer Support 3 01-06-2004 05:36 PM
ATA 133 HD on a ATA 100 Board buffy Computer Support 2 12-23-2003 03:39 PM



Advertisments