Well, you're basically echoing what I had previously said in
my first reply.

In your other emails you stated in this case to a member's
post (I think it was "Doug"); the HD should be the slave and
the CDROM drive the master on his SEC IDE port.  I was
pointing out this should not be the case.  He wanted to put
his new HD on the SEC port, and if he did, it should be the
master with the CDROM drive as the slave.

The rotational speed is not going to be any different for a
7200rpm HD if slaved to a 5400rpm drive.  That is the PLATTER
speed and it of course is governed by the motor and the
+5v/+12v supplied to it, and not the position of an IDE
cable.  RPM tests show this.  What might change and
usually does, is the data transfer rate in MB/sec due to
being a slave device....which is what I mentioned in the
other emails on this subject about the desired preference
to use a controller card so HD's will not be slaved.

Also you may be referring to instead of an ATA66 and
ATA100 HD on the same cable.  There is no *real* way
to test it, but the ATA100 may 'operate' @ ATA66.
Like I said above, a shared IDE controller does not
offer the same performance as a stand alone device
and the *performance* is diminished (but not the RPM
rating).  Whether or not it *TECHNICALLY* reverts
back to ATA66 is hard to show since the ATA standard
(whether 33, 66, 100 or 133mb/sec) is not anywhere
near it's rating in actual tests in the first place.
-Clint

God Bless Us All
Clint Hamilton, Owner
http://OrpheusComputing.com �

----- Original Message -----
From: <[EMAIL PROTECTED]>

Date: 09.10.02
Sender: OrpheusComputing.com
Time: 09:26

> Where did you hear that?  In most typical cases, hard
drives
> should always be the master, albeit PRI or SEC IDE port.

Did I write something else?

> Making a 2nd HD a slave to a CDROM drive on the SEC IDE
port;
> benchmarks show decreased performance.
>
> If it was a CDRW drive; sometimes CDRW drives must be the
> master on the SEC IDE port, sometimes they don't need to
be.
> Depends on the CDRW drive.

Rule of thumb: CD-RW master, CD-ROM slave (on the same
controller).


> This is exactly why the HD (horse) should be the master and
> the CDROM drive (chicken) be the slave.

No. I mean, not just this.

This is also why the HD (horse) should be on the same
controller with the
HD (horse) and not with the CDROM (chicken).

I was not referring to the master/slave thing; I wanted to
point out that
having different devices (HD and CDROM) on the same
controller is not
good.


> This is also why one device per channel (port) is
desirable, and a
> controller card is best to use if a mobo only has 2 IDE
ports and you
> wish to add another HD if you already have an existing HD
and CDROM
> drive.  IDE devices always work best being a 'stand alone'
on an IDE
> port.  If they are not, even regardless of how they are
'jumpered',
> performance decreases.

I guess I mistook your point; I just wanted to say that the
"canonical"
connection would be like:

Primary Master: HD 1
Primary Slave: HD 2

Sec. Master: CD-RW
Sec. Slave: CD-ROM


And if HD2 rotates max at 5400rpm while HD1 rotates max at
7200rpm, then
under this connection HD1 will rotate at 5400rpm.


> Doug, since it's no big deal to change them around, test it
> yourself.

There's also a small executable raptest.exe (or something)
that is used
mainly for testing the overlay capabilities of the graphics
card. Apart
from this, it has a built-in HD test utility.


> "i.e. if there is a CD-ROM drive on the cable
> you are putting it on, you will want to change
> the CD-ROM drive to slave and set the new
> hard drive as master."

That's true. But connecting a HD on the same controller with
a slower
device (CD-ROM/CD-RW/DVD/super-chicken) does not let HD run
full speed.
============= PCWorks Mailing List =================
Don't see your post? Check our posting guidelines &
make sure you've followed proper posting procedures,
http://pcworkers.com/rules.htm
Contact list owner <[EMAIL PROTECTED]>
Unsubscribing and other changes: http://pcworkers.com
=====================================================

Reply via email to