40 vs. 80-pin IDE cables

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.

 

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

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

 

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.)

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

 

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

 

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

 

It seems to depend on how the "round" cables are made. No drive now

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/articles/200210/20021018WD2000xB_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/articles/200303/20030320WD360GD_2.html

 

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

 

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.

 

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

 

There's sure no reason to do so. ATA/100 is much easier to achieve
than real drive thru-put.

I just ran a test on a new Seagate 7200RPM Ultra ATA/100 drive
(ST380013A).
The specs show INTERNAL TRANSFER RATE (whatever that is) of up to
85.4MB/sec,
and a SUSTAINED TRANSFER RATE up to 58MB/sec. EXTERNAL TRANSFER RATE
up to 100MB/sec. That's where they get the Ultra ATA/100.

Test showed a max of 52MB/sec and a min of 42MB/sec. Looks like we're
even coming close to the hype! As you indicate, 10,000RPM drives can
even stress ATA/66. Note: All measurements are in mutliples of 10^3,
(1000), not 2^10, (1024). ;-)

SATA should end some of the marketing hype, though we'll probably be
seeing SATA2 or SATA3 soon, as if 150MB/sec were not enough.

Virg Wall

 

They don't make 80-pin IDE cables AFAIK.


Have a nice week...

Trent

Follow Joan Rivers' example --- get pre-embalmed!

 

Did it take you all the time from the OP on Tuesday, to count yours?

Virg Wall

 

You mean I'm CORRECT??!!! lol


Have a nice week...

Trent

Follow Joan Rivers' example --- get pre-embalmed!

 

Of course. There is no such thing as an 80 pin IDE cable. I believe the
only hard drives that use an 80 pin cable are SCSI drives with a SCA
connector. The OP (hopefully) meant 80 /conductor/ cable.

--
AIM: FrznFoodClerk (actually me)
email: [email protected]_cast.net (_ = m)
website: under construction
Need a technician in the south Jersey area?
email/IM for rates/services

 

I just thought it might be nice if somebody TOLD him that. lol


Have a nice week...

Trent

Follow Joan Rivers' example --- get pre-embalmed!

 

Of course the OP meant 80 _conductor_!