Art said:
Timothy:
With the disclaimer that my comments pertain to PATA hard drives
and not SATA drives since I've had no experience with the latter
devices, here is my understanding of HD configuration. I would
appreciate your comments on the following...
1. If the HD is jumpered as Master or Cable Select (CS) and is the
only device on the IDE cable, it doesn't matter what position it
occupies on the IDE cable for it to boot -- it would be considered
as Master whether connected to the middle or end connector
of the IDE cable.
Being "considered as Master" only means 2 things:
1) the device would have a *default* position in the boot order
that is ahead of a device on the same cable/channel
that is "considered as Slave", and
2) the controller would differentiate it from a Slave.
Other than 1) and 2), "Master" and "Slave" have no meaning.
As you can see under 2), a single device on a cable/channel
doesn't need to be differentiated as Master or Slave, so
either setting would do, and the default boot order can be
changed in the BIOS and that change would affect all
subsequent bootups.
2. If the HD is a Western Digital (one that has a jumper setting of Single), it should be jumpered as Single or CS if it's the
only device on the IDE cable. It doesn't matter what position it occupies on the IDE cable for it to boot - it would be considered
as Master.
I have no experience with Western Digital hard drives, but since it
must conform to IDE/ATA specs, I would assume (until shown otherwise)
that being "considered as Master" has the same significance as for
other hard drives.
3. If the HD is installed at the only device on the IDE cable, it's considered good practice to connect it to the end connector
rather than the middle connector even though it *will* boot while
connected either way. I vaguely recall there was a ATA specific-
ation that cautioned against a "dangling end connector" but I've
been unable to find the reference. In any event there certainly
doesn't seem to be any negative implications to connecting a
single HD to the terminal end of the IDE cable.
Correct. The lengths specified for an IDE cable were (AFAIK)
for signal reflection tolerance, and a "dangling" end connector
(which would leave an open circuit at the end) was a no-no for
reasons of signal reflection. But being perverse, I tried that
booting a Maxtor Diamondmax Plus 9 (133MB/s) hard drive
that was a lone device at the middle connector of an IDE cable,
and it booted up fine and ran with no discernable problems.
I assumed from this that since the dimensional specs for IDE
cables didn't change with higher data rates as the rates climbed
from 33MB/s upwards to 133MB/s, that modern electronics can
tolerated signal reflectins from the open circuit end connector.
I would not run with an open end connector under normal operation,
though. But to suggest that the OP in this thread MUST put his
hard drive on the end connector to solve his problem was misleading.
4. The only time the position on the IDE cable is relevant is when the HD has been jumpered as CS and there are two devices on the
IDE cable.
Then the device connected to the middle connector is Slave (assuming
it's jumpered as Slave or CS) and the drive connected to the end
connector is Master (assuming it's jumpered as Master or CS). So if
inadvertently the HD was connected to the middle connector while
jumpered CS, it would *not* boot.
If there were 2 hard drives on the cable and Cable Select were the
common setting of both devices, and the 2 hard drives were swapped,
the boot order would be reversed and the other hard drive would be
given control (assuming it was bootable). If the BIOS's boot order
were still the default order, that means the hard drive put at the end
connector would get control because it would have the "Master" setting.
5. If the HD is jumpered as Slave, it will not boot, period! It's position on the IDE cable is immaterial.
Incorrect. A "Slave" can indeed boot (or more accurately, be given
control).
If there is another device on the cable jumpered as "Master",
and the boot order has been set to put the "Slave" ahead of the
"Master", the hard drive jumpered as "Slave" will be given control
if it is bootable by virtue of it being at the head of the boot order.
If the hard drive is jumpered as "Slave", it will be given control
because (assuming that it's on the Primary channel) it will be
at the head of the boot order. If it's on the Secondary channel
and there are no other devices on the Primary IDE channel,
it will be given control for the same reason - it's at the head of
the boot order. It's setting as "Slave" really has no other meaning
than the two I listed above:
1) default boot order, and
2) differentiating 2 devices on the same IDE channel.
6. If the HD is jumpered as Master, and there are two devices
on the IDE cable, and the HD is connected to the middle connector
of the cable, then that HD will *not* boot.
If the other device is jumpered as "Slave", i.e there are no conflicts,
a "Master" on the middle connector can indeed be given control.
All it needs is to be at the head of the boot order.
In my terminology, I differentiate between "given control" and "boot"
because although an "active" partition on a hard drive may have
control given to its boot manager (ntldr), its boot.ini file may
present OS options for OSes residing on other hard drives, and
I would call the loading of those OSes by ntldr to be simply "loading"
and not "booting". "Boot strap loading" is a process that begins
in the CPU and is taken over by firmware in the BIOS, then the
hard drive's Master Boot Record, then the "active" partition's
Boot Sector, and by the time ntldr gets control, it's just "loading"
and "starting".
*TimDaniels*
