«

I just read that using a 40-pin cable with a ATA/66 or 100 drive
forces the drive into ATA/33 mode. Just wondering: how does the drive
know it's on a 40-pin cable, if the other 40 wires on an 80-pin cable
are just grounds?

Also, since the 40 ground wires were added to prevent bad EMI, surely
chopping up the cable and tying it together defeats their purpose to
some extent?

Thanks.

"mrT" <> wrote in message
news: om...
> I just read that using a 40-pin cable with a ATA/66 or 100 drive
> forces the drive into ATA/33 mode. Just wondering: how does the drive
> know it's on a 40-pin cable, if the other 40 wires on an 80-pin cable
> are just grounds?

According to the spec, the cable should have one line designated for the
purpose of identifying to the motherboard that it is 80 conductor.

>
> Also, since the 40 ground wires were added to prevent bad EMI, surely
> chopping up the cable and tying it together defeats their purpose to
> some extent?

I've wondered this as well, but have yet to see any reports of a problem
with cut and bundled round IDE cables.

mrT wrote:
>
> I just read that using a 40-pin cable with a ATA/66 or 100 drive
> forces the drive into ATA/33 mode. Just wondering: how does the drive
> know it's on a 40-pin cable, if the other 40 wires on an 80-pin cable
> are just grounds?

Pin 34 is used to detect the presence of the correct 80 conductor cable.
(Only on a drive and motherboard that are set up to use this
information.)
>
> Also, since the 40 ground wires were added to prevent bad EMI, surely
> chopping up the cable and tying it together defeats their purpose to
> some extent?

It seems to depend on how the "round" cables are made. No drive now
made saturates even the ATA100 interface. More problems can occur if
the middle connector is used and the end "stub" is left unconnected.
That's why the end connector is labled "master", although except for
pin 28, (CSEL), used for cable select, they are wired identically.

Virg Wall
--
A foolish consistency is the
hobgoblin of little minds,........
Ralph Waldo Emerson
(Microsoft programmer's manual.)

Thor wrote:
>
> "mrT" <> wrote in message
> news: om...
> > I just read that using a 40-pin cable with a ATA/66 or 100 drive
> > forces the drive into ATA/33 mode. Just wondering: how does the drive
> > know it's on a 40-pin cable, if the other 40 wires on an 80-pin cable
> > are just grounds?
>
> According to the spec, the cable should have one line designated for the
> purpose of identifying to the motherboard that it is 80 conductor.
>
> >
> > Also, since the 40 ground wires were added to prevent bad EMI, surely
> > chopping up the cable and tying it together defeats their purpose to
> > some extent?
>
> I've wondered this as well, but have yet to see any reports of a problem
> with cut and bundled round IDE cables.

Didn't see your response until after writing mine. Earthlink's news
server seems to be lagging behind. At least we said the same thing!

Virg Wall
--
A foolish consistency is the
hobgoblin of little minds,........
Ralph Waldo Emerson
(Microsoft programmer's manual.)

On Wed, 29 Oct 2003 01:06:13 GMT, V W Wall <> wrote:

>mrT wrote:
>>
>> I just read that using a 40-pin cable with a ATA/66 or 100 drive
>> forces the drive into ATA/33 mode. Just wondering: how does the drive
>> know it's on a 40-pin cable, if the other 40 wires on an 80-pin cable
>> are just grounds?
>
>Pin 34 is used to detect the presence of the correct 80 conductor cable.
>(Only on a drive and motherboard that are set up to use this
>information.)
>>
>> Also, since the 40 ground wires were added to prevent bad EMI, surely
>> chopping up the cable and tying it together defeats their purpose to
>> some extent?
>
>It seems to depend on how the "round" cables are made. No drive now
>made saturates even the ATA100 interface.

I don't know of any drive that can go as fast as ata33!!

> More problems can occur if
>the middle connector is used and the end "stub" is left unconnected.
>That's why the end connector is labled "master", although except for
>pin 28, (CSEL), used for cable select, they are wired identically.
>
>Virg Wall

> >It seems to depend on how the "round" cables are made. No drive now
> >made saturates even the ATA100 interface.
>
> I don't know of any drive that can go as fast as ata33!!

The've have surpassed that actually. For example, the western digital 200GB
7200RPM ATA drive has been rated to transfer about 56MB/sec on the outer
tracks, and 33MB/sec on the inner tracks, putting it's sustained average
transfer rate capability above ATA/33's limits.

http://www.storagereview.com/article...D2000xB_2.html

Then consider that the new Raptor 10K ATA drives managed to sequentially
(sustained) transfer a *minimum* of 45MB/sec and a maximum of 63MB/sec
putting them within spitting distance of the ATA/66 spec at least on outer
zone reads. Either way, it pretty much proves that you don't want to
bottleneck your newer ATA drives with an ATA/33 interface.


http://www.storagereview.com/article...WD360GD_2.html

derek / nul wrote:
>
> On Wed, 29 Oct 2003 01:06:13 GMT, V W Wall <> wrote:

> >It seems to depend on how the "round" cables are made. No drive now
> >made saturates even the ATA100 interface.
>
> I don't know of any drive that can go as fast as ata33!!
>
I agree! Just didn't want to start a debate with someone claming they
had one! Most of the higher ATA specs are pure hype. At least SATA
at 150 to start has a clear route to acheiving higher burst rates, if
somebody figures out how to get the drives to do it.

Any drive can dump the buffer at a higher rate than the drive latency
permits. I haven't seen many programs that take advantage of that.
My new 7200RPM has an 8MB buffer. It seems to defrag much faster
than the 7200/5400 would suggest. Does it use larger the buffer?

Being tied to "legacy" formats has limited drive performance. We can
hope the SATA interface is radical enough to allow some new drive
techniques.

Virg Wall
--
A foolish consistency is the
hobgoblin of little minds,........
Ralph Waldo Emerson
(Microsoft programmer's manual.)

V W Wall <> wrote in message news:<>...
> mrT wrote:
> > Also, since the 40 ground wires were added to prevent bad EMI, surely
> > chopping up the cable and tying it together defeats their purpose to
> > some extent?
>
> It seems to depend on how the "round" cables are made. No drive now
> made saturates even the ATA100 interface. More problems can occur if
> the middle connector is used and the end "stub" is left unconnected.
> That's why the end connector is labled "master", although except for
> pin 28, (CSEL), used for cable select, they are wired identically.
>
> Virg Wall

Thanks for the info (Thor too!)...

I suppose even if EMI did become a problem, most errors would be
corrected invisibly, right? But I guess that would slow down
transfers, right?

"mrT" <> wrote in message
news: om...
> V W Wall <> wrote in message
news:<>...
> > mrT wrote:
> > > Also, since the 40 ground wires were added to prevent bad EMI, surely
> > > chopping up the cable and tying it together defeats their purpose to
> > > some extent?
> >
> > It seems to depend on how the "round" cables are made. No drive now
> > made saturates even the ATA100 interface. More problems can occur if
> > the middle connector is used and the end "stub" is left unconnected.
> > That's why the end connector is labled "master", although except for
> > pin 28, (CSEL), used for cable select, they are wired identically.
> >
> > Virg Wall
>
> Thanks for the info (Thor too!)...
>
> I suppose even if EMI did become a problem, most errors would be
> corrected invisibly, right? But I guess that would slow down
> transfers, right?

There is built-in CRC checking of the data when it is transferred. If it
requires correction, a retransmit occurs, which obviously would slow down
overall throughput to a degree.

mrT wrote:
>
> V W Wall <> wrote in message news:<>...
> > mrT wrote:
> > > Also, since the 40 ground wires were added to prevent bad EMI, surely
> > > chopping up the cable and tying it together defeats their purpose to
> > > some extent?
> >
> > It seems to depend on how the "round" cables are made. No drive now
> > made saturates even the ATA100 interface. More problems can occur if
> > the middle connector is used and the end "stub" is left unconnected.
> > That's why the end connector is labled "master", although except for
> > pin 28, (CSEL), used for cable select, they are wired identically.
> >
> > Virg Wall
>
> Thanks for the info (Thor too!)...
>
> I suppose even if EMI did become a problem, most errors would be
> corrected invisibly, right? But I guess that would slow down
> transfers, right?

Depending on the nature of the EMI, and the error correcting capability
of the drive electronics, a slow dowm might be the only visable result.
But why tempt the computer gods? Just don't leave dangling un-connected
stubs. The round cables seem to be OK.

Virg Wall
--
A foolish consistency is the
hobgoblin of little minds,........
Ralph Waldo Emerson
(Microsoft programmer's manual.)