FireWire 400: Master or Slave for ATA

[Martin Trautmann]

Hi all,

the docu of my external casing claims: Using it via USB I should jumper
the harddisk as Master, using it via FireWire I should jumper it as
slave.

Huh?

How does the USB/FW-ATA conversion know about a master or slave at all?
(using MacOSX 10.3.6).

Does it make any difference here?
(ICYBOX 350UE, Samsung SV1604N)

Thanks,
Martin

 

[Folkert Rienstra]

Hi all,

the docu of my external casing claims: Using it via USB I should jumper
the harddisk as Master, using it via FireWire I should jumper it as slave.

Huh?

How does the USB/FW-ATA conversion know about a master or slave at
all? (using MacOSX 10.3.6).

The same way maybe that an IDE controller does it, (or is that a too radical thought for you?).

 

[Martin Trautmann]

The same way maybe that an IDE controller does it, (or is that a too radical thought for you?).

That's why I made the single, one and only HD as Master. But why should
I jumper it as slave when using it via FireWire? I'm not very happy to
rejumper a harddisk whenever I decide to use the other port - can't
imagine that this is very reasonable.

I'll give it some speed tests later on, verifying the jumper influence.
But maybe someone knows the effekt of ATA Master on FW...

Thanks,
Martin

 

[Folkert Rienstra]

That's why I made the single, one and only HD as Master.

But why should I jumper it as slave when using it via FireWire?

Because they want you to. You have to ask *them* concerning why.

I'm not very happy to rejumper a harddisk whenever I decide
to use the other port -

From the manual. that is apparently how they designed it:
the USB firmware addresses the Master, the FW firmware
addresses the slave.

can't imagine that this is very reasonable.

Well, it at least makes it impossible to attach both interfaces at
once and have two computers indepently access the same harddrive.
Maybe they designed it so that you can use both interfaces to
address 2 different drives independently, but then, maybe not.

I'll give it some speed tests later on, verifying the jumper influence.

Should be huge. As in working vs not_working at all.

But maybe someone knows the effekt of ATA Master on FW...

None.
The bridge firmware has to address the hardrive using either master
or slave by setting the device0 or device1 bit in the commands to the
harddrive(s). It could find out what to use by doing a device scan,
just like a controller bios or a driver does, or they just limit use to
Master or Slave only and keep it simple.

 

[Martin Trautmann]

Should be huge. As in working vs not_working at all.

No, not that much: The drive ist jumpered as master, but used via
FireWire (which recommends slave) and rather good performance.

None.
The bridge firmware has to address the hardrive using either master
or slave by setting the device0 or device1 bit in the commands to the
harddrive(s). It could find out what to use by doing a device scan,
just like a controller bios or a driver does, or they just limit use to
Master or Slave only and keep it simple.

Up to now, fortunately, I do not observe any differences. Is there a
standard test suite for performance tests? Mine is not very efficient. I
suppose. I'll try to chain the FW drive with yet another FW drive for
further analysis...

 

[Folkert Rienstra]

No, not that much: The drive ist jumpered as master, but used via
FireWire (which recommends slave) and rather good performance.

Then their documentation is obviously false.

Up to now, fortunately, I do not observe any differences.
Right.

Is there a standard test suite for performance tests?

HD Tach, but Version 2 (Win9x etc) needs bios support to work.
Don't know about version 3 (W2k,XP)

 

[Martin Trautmann]

HD Tach, but Version 2 (Win9x etc) needs bios support to work.
Don't know about version 3 (W2k,XP)

sorry - although the group name states differently, I use Mac OSX or
Unix. On the other hand, IBM is out of PC business by now.

Any Open Source recommendation?