IDE Wiring

[tokeer.port5.com]

Hi,

I've got an hp8580c, it has 3 optical bays.
In the top bay i've got a dvd burner, in the 2nd and 3rd bays I've got
removable IDE trays. the internal hard disk resides in the internal 4th
bay.

I've had this working perfectly for many years, recently i upgraded my
ram, and unplugged all the optical devices. when I went to plug them
back in I couldn't remember which IDE cable went to which device.

I plugged the Master of the primary IDE channel to the dvd burner, the
slave to the 2nd bay.
The master of the secondary channel went to the 3rd bay and the slave
went to the internal hard drive. This WAS not the original setup
because I had put a pin to opne the cable for the removable hdd. When I
boot up the PC the master on the secondary was not recognized on the
bios.

So i tried switching the cable on the secondary channel, this alos
cannot be the original setup because I had to twist the head of teh
master around the 3rd bay and plug it into master. The bios does
recognize all 4 drives, but the hdd in the 3rd bay posts approx 30
seconds after the 1st three post.

The question, what is the ideal way to wire 4 drives?

Regards,
t.p5.com

 

[Rod Speed]

I've got an hp8580c, it has 3 optical bays.

In the top bay i've got a dvd burner, in the 2nd and
3rd bays I've got removable IDE trays. the internal
hard disk resides in the internal 4th bay.

I've had this working perfectly for many years, recently i upgraded my
ram, and unplugged all the optical devices. when I went to plug them
back in I couldn't remember which IDE cable went to which device.

I plugged the Master of the primary IDE channel
to the dvd burner, the slave to the 2nd bay.

You'd normally have the boot drive as the master on the primary IDE.

The master of the secondary channel went to the 3rd bay and the
slave went to the internal hard drive. This WAS not the original setup
because I had put a pin to opne the cable for the removable hdd.

Not clear what you mean in that last bit.

When I boot up the PC the master on the
secondary was not recognized on the bios.

So i tried switching the cable on the secondary channel,

Presumably you mean the internal drive became the master.

this alos cannot be the original setup because I had to twist the
head of teh master around the 3rd bay and plug it into master.

The bios does recognize all 4 drives, but the hdd in the
3rd bay posts approx 30 seconds after the 1st three post.

You can get that effect when that drive doesnt like the other drive
on the same cable, or its not jumpered properly for the config.
Some manufacturer's drives are more fussy than others.

The question, what is the ideal way to wire 4 drives?

No such animal.

Its more usual to have the boot drive as the master
on the primary IDE channel, and in your config one
of the other drives obviously has to be the slave.
It shouldnt matter much which, tho it might be better
to have the DVD burner on the same cable as it,
just because these two dont get affected by what
is happening with the removable drive bays.

This cable is the most problematic length wise,
it may not be feasible to have some of the drives
in the 'optical drive' stack on that cable just due to
the amount of cable between the drive connectors.

Then the removable drive bays are on the same cable.
It may or may not matter which is master depending
on what drives you have in the removable drive bays.
Some drives like the WDs can be a problem in removable
drive bays because they have a different jumper config
when they are the only drive on a cable and one of a pair.
That can be a problem if both removable drive bays dont
have the drive inserted.

 

[tokeer.port5.com]

Thanx again for your help.

The way the drives are stacked makes it difficult to try all possible
scenarios.
I am convinced that the internal boot drive was not the master on
either the primary or secondary channel in the original setup.
Will try to play around with other combinations over the weekend to see
if I can get it back to normal.

t.p5.com

 

[Isis]

Thanx again for your help.

The way the drives are stacked makes it difficult to try all possible
scenarios.
I am convinced that the internal boot drive was not the master on
either the primary or secondary channel in the original setup.
Will try to play around with other combinations over the weekend to see
if I can get it back to normal.

t.p5.com

Should be quite starightforward if you take the drives (CD, DVD etc) out of
the PC. On the back of the drives are the Master|Slave selector plugs.

Each IDE cable has two connectors - you must attach one Master and one
Slave to each cable - you can change each device to either master or Slave
using the plugs on the back.

The hard drive 'should' be plugged onto the 'end' of the 'primary' IDE
cable - the cable that is plugged into IDE connector 1 on the motherboard.

In fact the machine 'should' boot from the hard drive whichever IDE (1 or
2) you plug it into if you are running 2000 or XP. Just make sure that
there is a Master and Slave on each or just a Master in the case of your
hard drive.

Only plug in the CD, DVDS and hard drive to start until you get it up and
running.

DVD 'should' be set as Masters on cables with CD devices attached - Hard
Drives and CD/DVD's should NOT be attached to the same cable - it slows
down the hard drives.

HTH

 

[Folkert Rienstra]

Utterly clueless.

Should be quite starightforward if you take the drives (CD, DVD etc) out of
the PC. On the back of the drives are the Master|Slave selector plugs.

Each IDE cable has two connectors - you must attach one Master and one
Slave to each cable - you can change each device to either master or Slave
using the plugs on the back.

The hard drive 'should' be plugged onto the 'end' of the 'primary' IDE
cable - the cable that is plugged into IDE connector 1 on the motherboard.

In fact the machine 'should' boot from the hard drive whichever IDE (1 or
2) you plug it into if you are running 2000 or XP. Just make sure that
there is a Master and Slave on each or just a Master in the case of your
hard drive.

Only plug in the CD, DVDS and hard drive to start until you get it up and
running.

DVD 'should' be set as Masters on cables with CD devices attached - Hard
Drives and CD/DVD's should NOT be attached to the same cable - it slows
down the hard drives.

HTH

 

[Arno Wagner]

Should be quite starightforward if you take the drives (CD, DVD etc) out of
the PC. On the back of the drives are the Master|Slave selector plugs.

These are called "jumpers". Description on how to set them for different
effects are to be found often on the dirve label.

Each IDE cable has two connectors - you must attach one Master and one
Slave to each cable - you can change each device to either master or Slave
using the plugs on the back.

Not quite: A single drive should be master (some set-ups also work
with only a slave). If it is two drives one they must be set differently,
one to master and one to slave.

The hard drive 'should' be plugged onto the 'end' of the 'primary' IDE
cable - the cable that is plugged into IDE connector 1 on the motherboard.

Does not matter. IDE is not terminated you can do a middle-bus tap, i.e.
connect a single drive to the connector in the middle of the cable.

In fact the machine 'should' boot from the hard drive whichever IDE (1 or
2) you plug it into if you are running 2000 or XP. Just make sure that
there is a Master and Slave on each or just a Master in the case of your
hard drive.

In my experience MS stuff is very picky about having it exactly its way.
I think the disk with the boot partition has to be master on the first
controller port. It may have gotten more flexible by now, I would not
notice, since I boot Windows with Grub.

Only plug in the CD, DVDS and hard drive to start until you get it up and
running.

Actually Only use the HDD and graphics card first and if that works
start adding stuff until it breaks. But depends on the scenario.

DVD 'should' be set as Masters on cables with CD devices attached

Huh? And why ist that?

- Hard Drives and CD/DVD's should NOT be attached to the same cable
- it slows down the hard drives.

Depends. May or may not happen. If yiu have two fast HDDs they can slow
each other down as well, I made bad experience with a RAID1 and
two HDDs on one cable. However since CD/DVD drives have lower data
rates they can share a cable more easily.

Bottom line: You need to try it out but my advice is to not
have more than one HDD per IDE channel (i.e. cable).

Arno

 

[Folkert Rienstra]

And clueless number two.

These are called "jumpers". Description on how to set them for different
effects are to be found often on the

dirve label.

No kidding.

Not quite: A single drive should be master (some set-ups also work
with only a slave). If it is two drives one they must be set differently,
one to master and one to slave.


Does not matter. IDE is not terminated you can do a middle-bus tap, i.e.
connect a single drive to the connector in the middle of the cable.


In my experience MS stuff is very picky about having it exactly its way.

But you have no experience.

I think the disk with the boot partition has to be master on the first
controller port. It may have gotten more flexible by now, I would not
notice, since I boot Windows with Grub.


Actually Only use the HDD and graphics card first and if that works
start adding stuff until it breaks. But depends on the scenario.


Huh? And why ist that?


Depends. May or may not happen. If yiu have two fast HDDs they can slow
each other down as well, I made bad experience with a RAID1 and
two HDDs on one cable. However since CD/DVD drives have lower data
rates they can share a cable more easily.

Bottom line: You need to try it out but my advice

Like you should be the one to give advice.

 

[latmu]

These are called "jumpers". Description on how to set them for
different effects are to be found often on the dirve label.

Not quite: A single drive should be master (some set-ups also work
with only a slave). If it is two drives one they must be set
differently, one to master and one to slave.

That assumes that cable select isnt being used. It often is with modern systems.

Does not matter.

It does actually, particularly with just one drive on the cable.

IDE is not terminated you can do a middle-bus tap, i.e. connect
a single drive to the connector in the middle of the cable.

The ATA standard now requires that a single drive be on the end
connector, basically because the unterminated stub that is there
if the middle connector is used can be a problem at higher speeds
due to reflections off the end of the unterminated stub.

In my experience MS stuff is very picky about having it exactly its way.

Nope, tho the NT/2K/XP family can get rather confused if you boot a
clone with the original visible on the first boot after the clone is made.

I think the disk with the boot partition has
to be master on the first controller port.

Nope, that has never been true.

It may have gotten more flexible by now,

It never was as rigid as you claim.

I would not notice, since I boot Windows with Grub.

Actually Only use the HDD and graphics card first and if that works
start adding stuff until it breaks. But depends on the scenario.

Huh? And why ist that?

Depends. May or may not happen. If yiu have two fast
HDDs they can slow each other down as well, I made bad
experience with a RAID1 and two HDDs on one cable.
However since CD/DVD drives have lower data
rates they can share a cable more easily.

You're both mangling that story considerably.

Bottom line: You need to try it out but my advice is to not
have more than one HDD per IDE channel (i.e. cable).

Works fine and is often the best config purely because its
mechanically more convenient to have one cable for the 3.5"
drives and the other for the optical drives in the other bay stack.

 

[Arno Wagner]

That assumes that cable select isnt being used. It often is with
modern systems.

True. Personally I never use cable select, I like more control.

It does actually, particularly with just one drive on the cable.

The ATA standard now requires that a single drive be on the end
connector, basically because the unterminated stub that is there
if the middle connector is used can be a problem at higher speeds
due to reflections off the end of the unterminated stub.

And how do the reflections go away if there is another
unterminated drive at the end of the cable? Or did ATA recently
discover that terminating signal lines is a good idea?

[...]

Arno

 

[latmu]

True. Personally I never use cable select, I like more control.

You have just as much control, it just controls the master and slave
relationship
by the connector used for the drive instead of with jumpers on the drive.

And how do the reflections go away if there is
another unterminated drive at the end of the cable?

The other drive is much better reflections wise than no drive at all.

Or did ATA recently discover that terminating signal lines is a good idea?

Nope, just realised that a drive is better reflections wise than no drive.

That isnt a problem with the middle connector with no drive connected.

 

[Arno Wagner]

The other drive is much better reflections wise than no drive at all.

So? And how does that work with a high-input resistance digital
CMOS input? Unless the input specification was changed to include
some form of mandatory load or termination the second drive
is _not_ visible electrically while the first is active. Your
argument sound very much like unfunded hot air to me.

Care to give some _technical_ reference?

Arno

 

[Eric Gisin]

Arnie the clueless. All the ATA specs specify termination in host and devices.

 

[latmu]

So?

So that is a better config for higher speed ops where reflections are more
important.

And how does that work with a high-input resistance digital CMOS input?

It isnt that high, because none of the lines that matter are purely inputs.

And its always going to be lower than with nothing plugged
into that drive connector at all with the end connector.

Unless the input specification was changed
to include some form of mandatory load

No change, that has always been the case, all the lines
that matter are always loaded by the digital drivers etc.

or termination the second drive is _not_
visible electrically while the first is active.

Thats just plain wrong with reflections of the signals on the ribbon
cable. All that matter with them is the electrical load on the line.

Your argument sound very much like unfunded hot air to me.

Your problem.

Care to give some _technical_ reference?

The ATA spec clearly says that the drive should be on the
end connector when there is just one drive on the cable.
From memory it also says that thats to minimise reflections,
tho that last may have been outside the standard itself, forget.