Jumper external HDD as master or as cable select?

[Knack]

Getting a little confused about this. I have an external IDE HDD that is to
be mounted within an external enclosure with a USB to IDE interface (it will
be USB cabled to a computer), should that HDD be jumpered as master or as
cable select? Does that jumpering matter whether the computer that it will
be connected to is a laptop having one internal HDD or is a cabinet PC
having a pair of master HDDs plus a pair of slave drives? Does the type of
cable (DMA33 vs. ATA 100) affect the jumpering?

 

[CJT]

Getting a little confused about this. I have an external IDE HDD that is to
be mounted within an external enclosure with a USB to IDE interface (it will
be USB cabled to a computer), should that HDD be jumpered as master or as
cable select? Does that jumpering matter whether the computer that it will
be connected to is a laptop having one internal HDD or is a cabinet PC
having a pair of master HDDs plus a pair of slave drives? Does the type of
cable (DMA33 vs. ATA 100) affect the jumpering?

Jumper it however the docs for the enclosure prescribe.

 

[Impmon]

Jumper it however the docs for the enclosure prescribe.

And if the doc are not readable or not present, assume master setting.

 

[Borg hater]

Master. Master/alone if a Western Digital.

 

[Rod Speed]

Getting a little confused about this. I have an external IDE HDD that is to be
mounted within an external enclosure with a USB to IDE interface (it will be
USB cabled to a computer), should that HDD be jumpered as master or as cable
select?

The jumpering is determined by what the bridge can handle.

Does that jumpering matter whether the computer that it will be connected to
is a laptop having one internal HDD or is a cabinet PC having a pair of master
HDDs plus a pair of slave drives?

No, what matters master/slave wise is just what
is connected to the bridge in the external enclosure.

Does the type of cable (DMA33 vs. ATA 100) affect the jumpering?

Nope.

 

[Knack]

And if the doc are not readable or not present, assume master setting.

The enclosure's installation guide doesn't mention the jumpering of the HDD.
However, the Seagate Ultra ATA HDD is labeled with a short installation
summary that states :

"Set the jumper to Cable Select if using an Ultra ATA cable. For older
cables, set one drive as a Master, and the other as a Slave."

The short length of cable within the enclosure is in fact an 80-conductor
cable. So the HDD's instructions state to set it as Cable Select, even
though my instincts (and everyone else's in this thread so far) think it
should be set as Master.

 

[CJT]

The enclosure's installation guide doesn't mention the jumpering of the HDD.
However, the Seagate Ultra ATA HDD is labeled with a short installation
summary that states :

"Set the jumper to Cable Select if using an Ultra ATA cable. For older
cables, set one drive as a Master, and the other as a Slave."

The short length of cable within the enclosure is in fact an 80-conductor
cable. So the HDD's instructions state to set it as Cable Select, even
though my instincts (and everyone else's in this thread so far) think it
should be set as Master.

I can't imagine any harm resulting from trying either way.

 

[Rod Speed]

The enclosure's installation guide doesn't mention the jumpering of the HDD.
However, the Seagate Ultra ATA HDD is labeled with a short installation
summary that states :

"Set the jumper to Cable Select if using an Ultra ATA cable. For older cables,
set one drive as a Master, and the other as a Slave."

That is seriously misleading. It isnt the cable that determins that,
its what is driving the cable at the non hard drive connector.

The short length of cable within the enclosure is in fact an
80-conductor cable. So the HDD's instructions state to set it as Cable Select,
even though my instincts (and everyone else's in this thread so far) think it
should be set as Master.

It isnt that black and white. The bridge can be designed to
do it either way and it makes some sense to allow both.

 

[Healthnut]

I can't imagine any harm resulting from trying either way.

Solved. Master is the correct way for this particular enclosure. Couple
nights ago I did a successful backup to that external HDD when it was
configured as Cable Select. Unfortunately since then I couldn't get reliable
spin-up of the HDD. I guess I was just lucky that night. I was beginning to
think the enclosure's adapter circuitry had somehow failed since then
because the HDD is very new, and I also tested another identicle HDD in the
same enclosure and it also was not spinning up. So since rejumpering the HDD
as Master I can now switch it off/on repeatedly and it spins up, and with OK
partition detection by the laptop's WinXP.

 

[Folkert Rienstra]

That is seriously misleading.

Nope. It expects that a competently designed IDE controller is used.

It isnt the cable that determins that,
its what is driving the cable at the non hard drive connector.

Which is a ground.
Only seriously flawed controllers don't have this signal
grounded in which case both drives end up as 'Slave'.

It isnt that black and white. The bridge can be designed to
do it either way and it makes some sense to allow both.

There is no such design other than have the C/S line grounded at the contoller end.

 

[Folkert Rienstra]

Master.

Master/alone if a Western Digital.

With most HD, actually. Just different wording used.

 

[Folkert Rienstra]

The enclosure's installation guide doesn't mention the jumpering of the HDD.
However, the Seagate Ultra ATA HDD is labeled with a short installation
summary that states :

"Set the jumper to Cable Select if using an Ultra ATA cable. For older
cables, set one drive as a Master, and the other as a Slave."

The short length of cable within the enclosure is in fact an 80-conductor
cable. So the HDD's instructions state to set it as Cable Select, even
though my instincts (and everyone else's in this thread so far) think it
should be set as Master.

What matters is the connector used at the HD end and the CS line
grounded at controller end, not the cable.
Chances are slim that a Slave connector is used if the controller
requires to see Master instead, so CS should actually work.

 

[Folkert Rienstra]

Solved. Master is the correct way for this particular enclosure. Couple
nights ago I did a successful backup to that external HDD when it was
configured as Cable Select.

Right, as it should.

Unfortunately since then I couldn't get reliable spin-up of the HDD.

Which should have nothing to do with CS vs M/S.

I guess I was just lucky that night.

Logic doesn't work that way.
(unless that particular model HD is seriously flawed)

I was beginning to
think the enclosure's adapter circuitry had somehow failed since then
because the HDD is very new, and I also tested another identicle HDD in the
same enclosure and it also was not spinning up. So since rejumpering the HDD
as Master I can now switch it off/on repeatedly and it spins up, and with OK
partition detection by the laptop's WinXP.

Weird.

Does the problem reappear if you change it to CS again.

 

[Rod Speed]

Yep.

It expects that a competently designed IDE controller is used.

Its more complicated than that, most obviously when
an 80 wire cable is used on say an ATA66 port.

Which is a ground.

Its more complicated than that too.

Only seriously flawed controllers don't have this signal
grounded in which case both drives end up as 'Slave'.

Its more complicated than that too.

There is no such design other than have
the C/S line grounded at the contoller end.

Wrong, as always.

 

[Rod Speed]

What matters is the connector used at the HD end and
the CS line grounded at controller end, not the cable.

Chances are slim that a Slave connector is used
if the controller requires to see Master instead,

Try that again in english. There is no 'slave connector'
with a very short cable with just one drive connector.

so CS should actually work.

Should and does are two different things, particularly if the
bridge has been designed for the drive jumpered as master
and CS hasnt been implemented at the bridge end.

 

[Rod Speed]

With most HD, actually. Just different wording used.

Wrong, as always. The WDs have a different jumper
config for master in a pair and single drive on the cable.

 

[Folkert Rienstra]

Wrong, as always. The WDs have a different jumper
config for master in a pair and single drive on the cable.

Nonsense.
Maxtor, IBM and Seagate have the same configs but are using different
(and both confusing) wording to explain them (or not as it turns out).

Maxtor, IBM and Seagate's 'Master' = WD's (Dual) Master
and
WD's 'single drive installation' = Maxtor, IBM and Seagate's 'Master with_slave_present'

Like I said, same concept, different wording.

WD's 'single drive installation' means that it just ignores that a slave is
present, which is what Maxtor's "Master_with_slave_present" and IBM's
"Device 0 with Device 1 Present" - contrary to their wording- mean also.

Seagate's "Master with non ATA-compatible slave" is the only one jumper
description that more or less adequately hints at what's going on behind
that confusing 'with ..... present' wording. You'll have to look in the
Maxtor and IBM manuals that it actually means "Non ATA-compatible
slave", "slave device that does not comply with the ATA specification" or
"Slave drive does not use the Drive Active/Slave Present (DASP– ) signal
to indicate its presence".
The last description is the biggest clue that the master will ignore the
slave's presence, hence will opererate as if it were a 'Single Drive'.

Looking forward to your imminent "Its more complex than that"
display of defeat. Or the more gracious "Wrong as always".

 

[Folkert Rienstra]

Try that again in english. There is no 'slave connector'
with a very short cable with just one drive connector.

Gosh Rodney, isn't that what I said.

Slave connector, you know, that grey one (not blue, not black)
that specifically has pin 28 (CSEL) disconnected/missing.

Should and does are two different things, particularly if the
bridge has been designed for the drive jumpered as master
and CS hasnt been implemented at the bridge end.

Yup, Rodney, that is what I said.
You are so perceptive that it's almost frightning.
Have you stuck your head in the Microwave again?

 

[Folkert Rienstra]

Its more complicated than that, most obviously when
an 80 wire cable is used on say an ATA66 port.



Its more complicated than that too.

Nope.
From the horses mouth: T13/1410D revision 3 (ATA6 spec)

Table 5 - - Driver types and required termination

Signal | Source | Driver-type | Host | Device | Notes
CSEL | Host | | Ground | 10k Pu | 4,6

Notes:
4 When used as CSEL, this line is grounded at the host
and a 10 kOhm pull-up is required at both devices.
6 Pull-up values are based on +5 V Vcc.

Active CSel is only used in PCMCIA, apparently.
Not really surprizing when Active CSel can only be used effectively
with a single device.

As per ATA spec.

Its more complicated than that too.

Thanks for so graciously showing defeat in your usual manner.
I know it took alot of you for doing so and I appreciate it.

Yup, and the cheapest way is passive CSEL, ie grounded.

Wrong, as always.

Nope, correct, as usual.

 

[Rod Speed]

Yep.

From the horses mouth: T13/1410D revision 3 (ATA6 spec)

Irrelevant to what the 'designer' of that system has chosen to do.

They've already chosen to flout that ATA spec with the
length of the 80 wire cable they have chosen to use.

Table 5 - - Driver types and required termination

Signal | Source | Driver-type | Host | Device | Notes
CSEL | Host | | Ground | 10k Pu | 4,6

Notes:
4 When used as CSEL, this line is grounded at the host
and a 10 kOhm pull-up is required at both devices.
6 Pull-up values are based on +5 V Vcc.

Active CSel is only used in PCMCIA, apparently.

So your original is just plain wrong and it is
indeed more complicated than you claimed.

Not really surprizing when Active CSel can
only be used effectively with a single device.

Which just happens to be the case with
the external hard drives being discussed.

Keep desperately digging, you'll be out in china any day now, again.

As per ATA spec.

See above.

Thanks for so graciously showing defeat in your usual manner.
I know it took alot of you for doing so and I appreciate it.

Even a terminal ****wit pseudokraut should be able to bullshit
its way out of its predicament better than that pathetic effort.

Obviously not. No wonder you're completely unemployable.

Yup, and the cheapest way is passive CSEL, ie grounded.

And it costs peanuts to do it the more bulletproof way so it
will work regardless of how the drive is jumpered, cretin.

Nope, correct, as usual.

Never once in a full decade. No wonder you're completely unemployable.