should master drives be at the end of the cable?

[Aloke Prasad]

In modern disk controllers (Intel 875PBZ), using modern ATA devices (WD
250GB drive, Plextor DVD burners), on a cable with 2 devices,

1. does it matter if the hard drive is master and DVD writer slave or vice
versa?

2. is the master preferred to be at the end of the cable?

I thought all this didn't matter anymore, but the Plextor manual has many
pages devoted to this subject, with suggestions for a "best" way of
attaching these things...

 

[Lil' Dave]

If that's all you have, just make each master on separate channels.

I did personnally run across a WD60GB HD jumpered master/alone. Was on
middle position of ribbon cable connected to a Promise Ultra100. Ran like a
pio 2 device, although the Promise bios detected nothing wrong. Moved WD to
end of cable and things visibly speeded up substantially.
Dave

 

[Aloke Prasad]

I have more. I have:

1 SATA drive - on it's own connector
1 WD 250 gig - PATA
1 Plextor DVD-ROM - PATA
1 Plextor DVD-R/RW - PATA

I wanted to see if the location of drives was still an issue. Apparently,
it is.

Oddly, in all the examples in the Plextor manual, they had the master on the
_middle_ connector and slave at the end connector !!

I thought it should be the other way around...master (or single devices) at
the end of the cable.

I'm planning

PATA primary controller--------DVDROM(slave)-------WD250(master)
PATA secondary controller------empty------------DVDRW(master)

 

[chrisv]

Oddly, in all the examples in the Plextor manual, they had the master on the
_middle_ connector and slave at the end connector !!

I thought it should be the other way around...master (or single devices) at
the end of the cable.

I've recently begun to "lighten up" and let cable select do it's
thing. Of course, CS will have the master at the cable end, which is
likely the CDROM. Works fine. I like it.

 

[Timothy Daniels]

I have:

1 SATA drive - on it's own connector
1 WD 250 gig - PATA
1 Plextor DVD-ROM - PATA
1 Plextor DVD-R/RW - PATA

I wanted to see if the location of drives was still an issue.
Apparently, it is.

Oddly, in all the examples in the Plextor manual, they
had the master on the _middle_ connector and slave
at the end connector !!

I thought it should be the other way around...master
(or single devices) at the end of the cable.

I'm planning

PATA primary controller-------DVDROM(slave)-----WD250(master)
PATA secondary controller------empty--------DVDRW(master)

Unless Cable Select is set, it doesn't matter which connector has the
Master and which has the Slave unless there is only one device - in
which case the Master should go on the end connector to avoid any
possibly confusing signal reflections caused by the unterminated section
of cable that would otherwise be left dangling.

For what it may be worth, the SIIG UltraATA 133 PCI controller
card manual says:
"In order to achieve high-speed data transfer, a 40-pin/80-wire
Ultra ATA ribbon cable is recommended. Also, do not mix
Ultra ATA/133 hard disk with slower IDE or ATAPI devices
on the same channel."
It's not clear if that means, for your system's topology, that the
HD's transfer speed would suffer or whether just the transfer
between the DVD-ROM and the HD (via RAM) would suffer.
And opinions have been expressed daintily here that there
would be no affect whatsoever by mixing device speeds on
an ATA/133 channel.


*TimDaniels*

 

[Rod Speed]

In modern disk controllers (Intel 875PBZ), using modern ATA devices
(WD 250GB drive, Plextor DVD burners), on a cable with 2 devices,

1. does it matter if the hard drive is master
and DVD writer slave or vice versa?

Not usually, tho its more usual to make the hard drive master.

2. is the master preferred to be at the end of the cable?

Doesnt matter with two drives. You should
have a single drive on the end drive connector.

I thought all this didn't matter anymore,

In fact the ATA standard has changed to require
a single drive to be on the end drive connector now.
That wasnt specified with the older slower drives.

but the Plextor manual has many pages
devoted to this subject, with suggestions
for a "best" way of attaching these things...

That could make a difference with the older burners, but
generally doesnt anymore with the modern burnproof burners.

 

[Lil' Dave]

Looks okay from here. Just put the HD as master on the end of the ribbon
cable. Excellent name brand choices.
Dave

 

[Folkert Rienstra]

When/if

Cable Select is set, it doesn't matter which connector has the
Master and which has the Slave

Nor when set M/S.

unless there is only one device -

in which case the Master should go on the end connector to avoid any
possibly confusing signal reflections caused by the unterminated section
of cable that would otherwise be left dangling.

It doesn't need to be master (drive0).
Important is that it is at the end, terminating the cable.

For what it may be worth, the SIIG UltraATA 133 PCI controller
card manual says:
"In order to achieve high-speed data transfer, a 40-pin/80-wire
Ultra ATA ribbon cable is recommended.

Not recommended, crucial.

Also, do not mix Ultra ATA/133 hard disk with slower IDE or
ATAPI devices on the same channel."
It's not clear if that means, for your system's topology, that the
HD's transfer speed would suffer or whether just the transfer
between the DVD-ROM and the HD (via RAM) would suffer.

Basically it is just bare nonsense in that form.
It's a lot more complicated, lots of factors involved.

And opinions have been expressed daintily here that there
would be no affect whatsoever by mixing device speeds on
an ATA/133 channel.

Nope. If differs a lot when both are accessed simultaniously and data
goes from one to the other (read, then write) or that data is read from
both simultaniously, unrelated. In latter mode the HD is slowed down
considerably compared to when on it's own seperate channel where in the
former case the slowest drive sets the pace and nothing will change that.
It also matters if transfers are sequential or nonsequential and whether
reads or writes or reads_and_writes.

 

[Timothy Daniels]

When/if


Nor when set M/S.

According to the Maxtor ATA Hard Drive Installation Guide,

"Cable Select. This setting is an alternate method of hard drive
identification that can be used instead of the master and slave
settings. With cable select enabled, all hard drives positioned
on the black (master) connector of the ATA cable will identify
as master, and all hard drives postioned on the gray (middle)
connector will identify as slave."

Maybe this wording would have been clearer: When Cable
Select is enabled on both drives, the connector selects which
is Master and which is Slave. When Cable Select is NOT
enabled, the drives (not the connectors) select which is
Master and which is Slave.

It doesn't need to be master (drive0).
Important is that it is at the end, terminating the cable.

Agreed. My mistake. A single device on a channel can be
either a Master or a Slave.

Not recommended, crucial.

I take it you don't approve of "round" cables?

Basically it is just bare nonsense in that form.
It's a lot more complicated, lots of factors involved.


Nope. If differs a lot when both are accessed simultaniously and data
goes from one to the other (read, then write) or that data is read from
both simultaniously, unrelated. In latter mode the HD is slowed down
considerably compared to when on it's own seperate channel where in the
former case the slowest drive sets the pace and nothing will change that.
It also matters if transfers are sequential or nonsequential and whether
reads or writes or reads_and_writes.

Regarding mixing a hard drive and ATAPI device on the same channel,
Maxtor's installation manual only says:

"To avoid potential issues with hard drive detection and performance,
Maxtor does not recommend attaching a hard drive as a slave on the
same cable as an ATAPI device"

Maxtor doesn't seem to object to the ATAPI device being the slave
and the hard drive being the Master, though, as the original poster,
Aloke Prasad, had planned.


*TimDaniels*

 

[chrisv]

Regarding mixing a hard drive and ATAPI device on the same channel,
Maxtor's installation manual only says:

"To avoid potential issues with hard drive detection and performance,
Maxtor does not recommend attaching a hard drive as a slave on the
same cable as an ATAPI device"

Maxtor doesn't seem to object to the ATAPI device being the slave
and the hard drive being the Master, though, as the original poster,
Aloke Prasad, had planned.

Well shit, it's almost impossible to connect the CDROM to the middle
connector, then loop the cable back down to the HD, so it can CS to
master. I suppose one could mount the HD above the CDROM, but that's
not the "normal" location... Of course, we always have the option of
jumpering to master and slave, but I don't think it's worth worrying
about - just let the HD CS to slave.

 

[Rod Speed]

Well shit, it's almost impossible to connect the
CDROM to the middle connector, then loop the
cable back down to the HD, so it can CS to master.

Nope, not if the hard drive is in the 5" bay stack.

I suppose one could mount the HD above the CDROM,

Or below it, in the 5" bay stack.

but that's not the "normal" location...

Completely irrelevant whether it is or not.

Some choose to put hard drives in the 5" bay
stack just so they dont get as hot as they would
if you have the hard drives adjacent in the 3.5"
bay stack with drives that get quite warm anyway.

 

[Folkert Rienstra]

I take it you don't approve of "round" cables?

No, my comment was about "80-wire" cable being crucial.
"40 wire" cable definately won't achieve high(est) speed
data transfers on a ATA133 channel.

Regarding mixing a hard drive and ATAPI device on the same channel,
Maxtor's installation manual only says:

"To avoid potential issues with hard drive detection and performance,
Maxtor does not recommend attaching a hard drive as a slave on the
same cable as an ATAPI device"

This may have to do with a master checking for a slave to get ready and
giving the all clear sign when it is. Perhaps it is their opinion that some
ATAPI drives don't do this well.