"V W Wall" wrote in message
...

I don't have any 80-wire ribbon cables at the moment where I can see
a tiny
(3mm) cutout section of a wire.
snip

It's definitely the wire which would connect to pin 34 on the MB. I
assume
that those cables without the "cut" do not connect the wire to pin 34
on
the motherboard.

No, the motherboard's pin 34 would have to be connected. Just because
it support UDMA-66 and up, it still needs to support the older drives
with the old ribbon cables. If the disconnect were on the mobo, even
the old 40-wire/40-pin ribbon cables would get "detected" (i.e., fixed)
to look like 80-wire/40-pin cables. That would mean an inappropriately
high ATA mode would get selected that the old ribbon cable and old drive
cannot support.

That's why my guess is that if you cannot see the cutout then it can
still exist but under the cap on the connector that presses the ribbon
cable against the forked teeth for the pins, or the tooth is broken so
the pin and forked tooth are not connected. I really don't want to
destroy my only 2 spare 80-wire/40-pin ribbon cables to find out but
curiousity might someday overcome my thriftiness.

The Seagate Barracuda Product Manual says: "This (ATA/100) cable uses
even-numbered conductors connected to the ground pins to improve
signal
integrety."

It also says: "The drive detects the 80-conductor by sensing a
capacitor at
the host side through the CBLID-signal." Don't see how it can do this
if
CBLID, (pin 34), is not even connected to the "host".

Yeah, that's a good one.

I have cables with and without the cutout.

And the cutout is in the wire for pin 34, right?

- You don't know if the signal pins use the odd- or even-numbered
wires.
You'll have to check by ohming out the pin to see if it is connected
to wire
#1 or #2. For my 2 samples, pin 1 went to wire #2 (so signals were
on the
even-numbered wires).

Does it really make any difference. As long as they are between the
signal
wires, they serve their purpose.

Yes, if you're trying to troubleshoot a dysfunctional machine. Could be
the ribbon cable got pinched around a sharp chassis corner and frayed or
cut into the outside wire. If it's a ground, no big deal, especially
since there is no signal wire on the other side in which it will reduce
crosstalk. You don't need to toss a perfectly good cable. If it's a
signal line (pin 1 for Reset) you do need to be concerned. Is the cable
compromised, can you simply apply a coating to prevent furture shorting
if it's a signal or not bother if it's a ground, and so on. Yeah, the
best answer is to replace the ribbon cable but when you're in the north
woods of Minnesota, some tiny town, or elsewhere in the field, your
supplies are limited and you might not have a handy fresh unused
replacement. No everything dealing with computers requires replacement
with brand new parts or happens in a big city with lots of retailers
available for immediate parts replacement.

As seen, whether the shield ground wires are the even- or odd-numbered
wires makes a difference in determining what a cutout was for. Count
using the wrong offset and you think the cutout was for a shield ground
instead of a signal.

You're right ... mostly! Although they are 80-wire/40-pin ribbon
cables and
NOT used at anytime for a cable select configuration, they appear to
be
setup to be usable for cable select (if you configure your drives
that way).
Pin 28 on the mobo connector (blue) shows continuity to pin 28 in
the far
end drive connector (black) but pin 28 in the middle connector
(grey) is not
connected to either of the other connectors.

I thought that's what I said! ;-)

No, you said that ALL of the ATA-compliant ribbon cables supported CSEL.
I disagree. ATA-1 and -2 specs indicate that at one time pin 28 was not
used (i.e., connected to all connectors but unused) in "normal" or
standard setups using master/slave configurations and a *special*
cable-select version of the ribbon cable allowed you to use CSEL. If
the middle connector were NOT unconnected (open), you couldn't use that
ribbon cable for CSEL on the drives. I'm not paying ANSI to get copies
of ATA-3 on up to see if the spec changed to make all ribbon cables CSEL
capable, or to see if it is still listed as an optional configuration
but maybe the cable and drive makers decided on making them all CSEL
capable to eliminate confusion and bad hardware setups since it didn't
affect master/slave setups.

As I said, my recollection was that at one time, maybe 5 or more years
ago, you had to be careful not to use the wrong ribbon cable type. If
you wanted a CSEL setup, you could not use a standard ribbon cable
(meant for master/slave configurations) that had pin 28 connected on the
middle connector. As a matter of fact, if you go to the parts store and
grabbed the connectors and ribbon cable and use a vise to squeeze the
connectors onto the ribbon cable, the resulting cable could NOT be used
for CSEL configurations (because pin 28 for the middle connector was
then connected to the other pin 28's on the other connectors). This
homemade cable works fine for master/slave but not for cable-select.
Knowing that pin 34 is supposed to disconnected at the mobo end (via
cutout) and if pin 28 is now supposed to not be connected to the middle
connector (again another cutout) are important if you want to make your
own custom length cables. Now with CSEL not being connected on the
middle connector, CBLID getting disconnected from the drive connectors
and shorted to ground inside the blue mobo connector, color coding, and
knowing which 80-wire connector to use so you match its odd- or
even-numbered positioning for the shielding ground wires, I haven't
bothered making homemade ribbon cables for quite awhile. The days of
just squeezing the connectors onto a ribbon cable are gone.

I think the spec that defined the 80 wire cables also defined the CSEL
connections.

The CSEL signal was defined way back in the ATA-1 spec (that required
"special" cabling to work). From
http://www.seagate.com/support/kb/di...ltraata66.html and
http://www.computerhope.com/help/ide.htm#02, 40-wire/40-pin cables were
specified up to ATA-4 (for xfer modes up to UDMA mode 2 for 33Mbps). It
was ATA-5 (for UDMA mode 4 for 66Mbps; UDMA mode 3 was 44Mbps but often
called UDMA-50) which introduced 80-wire/40-pin cables.

There is as much confusion about hard drives as there is about the
software
used with them! With SATA coming along, we'll probably see a new
round of
mis-information. ;-)

Yeah, we'll have to get used to calling the old ports as PATA (parallel
ATA) ports to differentiate them from SATA (serial ATA) ports. Calling
them IDE ports was wrong but fell into common usage. [E]IDE = enhanced
integrated drive electronics, and this controller is now on the drive,
but ATA (Advanced Technology Attachment) defines the interface spec. In
fact, EIDE/ATAPI (EIDE/ATA Packet Interface to include CD-type devices)
is the full spec to describe the interface but usage degenerated to just
IDE. So PATA and SATA seem to be the new buzzwords to use to
differentiate the port types. I haven't had the time to read up on SATA
(http://www.serialata.org/).