Ok to leave slave HD cable connected with power off?

  • Thread starter Thread starter Mark
  • Start date Start date
Mark said:
I want to use a slave hard drive occasionally as a Drive Image backup
disk. I don't like the idea of the backup hard drive always being powered
on
because hard drives are rated for a certain number of on/off cycles. Can
I just unplug the power cable and leave the flat ribbon cable connected
when I don't need to use the backup drive or will that create some
reliability issues with my master hard drive on the end of the cable?
Click to expand...

It won't always, but it _can_.

First DVD burner I ever had was an HP. At one point I had it installed in
the machine and had forgotten to attach the power cable. I ended up with
corrupted data on every disk in the machine, not only the slave but also
the ones on the other channel--the isolation there is not as great as one
might assume, or at least it wasn't with the Intel BX chipset.

This got me curious, so I reformatted everything and did some tests and
found out that with the DVD burner plugged into the signal cable but with
no power I did indeed get data corruption on the other attached drives, but
with it not plugged into the signal cable or with it connected to both the
signal and the power cables everything worked fine.

That's the only time I've ever had that happen or heard of it happening, so
it's unusual, but it _can_ happen and the only way to know for sure would
be to test.
 
Timothy Daniels said:
You are correct in that there should be a device
connected at the end of the cable to avoid the
reflections (and the timing of those reflections as
they reach the intermediate device), but of what
significance is there in the end device being
powered or not?
Click to expand...

Depends on what's there. Unpowered it may not supply what's needed to keep
the reflections down.
Implicit in your statement is
the assumption that the impedance looking into
the device from the connector is different
according to whether it was powered or not.
Click to expand...

Exactly and frequently that is the case and there could also be a bias
issue.
Why should the input impedance change with
power?
Click to expand...

Active devices may be involved.
 
Ron Reaugh said:
"Timothy Daniels" asked:

Depends on what's there. Unpowered it may not supply
what's needed to keep the reflections down.
Click to expand...


What might that be? That is, what active device might
there be that would function to suppress reflections from
itself or what it controlled?

*TimDaniels*
 
1) Is there a "state" when the circuits are unpowered?
2) How is the "state" of an unpowered circuit going to
affect the signal passing by?
Click to expand...

Going back a few years, ISTR that the manual for the TRS-80 Model I
stated that peripherals should all be powered up before (or at least
simultaneously with) the computer to avoid damage to unpowered
circuits. Don't ask me why this was recommended; but I *assumed* it
had something to do with signal current paths that were perhaps "open"
when they shouldn't have been, and thus more susceptible.
3) Why don't you try it and see?
Click to expand...


Cheers, Phred.
 
Timothy said:
What might that be? That is, what active device might
there be that would function to suppress reflections from
itself or what it controlled?

*TimDaniels*
Click to expand...

transistors
 
Timothy Daniels said:
Any idea how transistors would control their input impedance?
Click to expand...

Have you ever studied the impedance of a diode. What do you think the input
of a transistor is and then the possibilities get more complicated from
there. You're outa your depth.
 
Ron Reaugh said:
Have you ever studied the impedance of a diode. What do you
think the input of a transistor is and then the possibilities get more
complicated from there. You're outa your depth.
Click to expand...


You're saying the input side of a transistor is a diode?

*TimDaniels*
 
Going back a few years, ISTR that the manual for the TRS-80 Model
I stated that peripherals should all be powered up before (or at least
simultaneously with) the computer to avoid damage to unpowered circuits.
Click to expand...

It was always much more complicated than that.

RS232 was designed right from the start to not only
handle that fine, but to hand shorts in the cable fine too.

In spades with USB and firewire now too.
Don't ask me why this was recommended; but I *assumed* it
had something to do with signal current paths that were perhaps
"open" when they shouldn't have been, and thus more susceptible.
Click to expand...

To some extent. Not relevant now tho.
 
Timothy said:
Is a virtual ground how an IDE device controls signal reflections?

*TimDaniels*
Click to expand...

Not likely, although I haven't reviewed the input circuitry of every IDE
device and controller, so I can't rule out that possibility. It was
presented as one example of how a circuit might change impedance via
active circuitry, in response to your question (which had drifted away
from the original discussion, so maybe I shouldn't have bothered).

This wasn't just about impedance, or even reflections; at issue was
possible circuit damage to the unpowered drive.
 
CJT said:
This wasn't just about impedance, or even reflections;
at issue was possible circuit damage to the unpowered drive.
Click to expand...


Right. And circuit damage is even less likely than
data corruption. And before you dredge up some
arcane exception, let me restrict the following
comment to my own Maxtor DiamondMax Plus 9s:
It worked for me - no circuit damage!! I hooked 'em
up dead, first at the intermediate position, then at the
end position, and they both are still working.

*TimDaniels*
 
Timothy said:
Right. And circuit damage is even less likely than
data corruption. And before you dredge up some
arcane exception, let me restrict the following
comment to my own Maxtor DiamondMax Plus 9s:
It worked for me - no circuit damage!! I hooked 'em
up dead, first at the intermediate position, then at the
end position, and they both are still working.

*TimDaniels*
Click to expand...

Good for you. I'm going to refrain from that sort of thing,
except perhaps with a drive that's spec'd for hot plugging.
Call me overly cautious if you like, but my data is important
to me.
 
It was always much more complicated than that.
Click to expand...

It always is. :-)
RS232 was designed right from the start to not only
handle that fine, but to hand shorts in the cable fine too.
In spades with USB and firewire now too.


To some extent. Not relevant now tho.
Click to expand...

Getting back to the OP's original question then, I assume you're happy
to have a 2nd HDD unpowered but not unplugged? Or, perhaps more to
the point for convenience, which way would you do it -- disconnect the
power, or disconnect the interface? Or both?


Cheers, Phred.
 
Phred said:
Getting back to the OP's original question then,
I assume you're happy to have a 2nd HDD
unpowered but not unplugged? Or, perhaps more to
the point for convenience, which way would you do
it -- disconnect the power, or disconnect the interface?
Or both?
Click to expand...


For convenience, of course, put the power through
an externally operable switch. Arrange it so you can't
bump it accidentally and where you'd have to stop and
think before flipping it - maybe a rocker switch with
guards on either side of the rocker. Leave the data
connector permanently in place to avoid it getting loose
with wear and to avoid putting the signals through
another interface.

*TimDaniels*
 
It always is. :-)
Click to expand...

You're getting there, grasshopper |-)
Getting back to the OP's original question then,
Click to expand...

Tad radical.
I assume you're happy to have a 2nd HDD unpowered but not unplugged?
Click to expand...

Nope, I dont think THAT is a good idea.

It usually does work, but not always, you can see some
problems with the bios working out what drives are present etc.

Thats a different issue to the damage question tho.
Or, perhaps more to the point for convenience,
which way would you do it -- disconnect the
power, or disconnect the interface? Or both?
Click to expand...

I always disconnect both.

And I dont like the major kludge that removable drive bays
involve, I prefer to use a proper formal standard like USB
or firewire or SATA that defines that behaviour fully.
 
You're getting there, grasshopper |-)



Tad radical.


Nope, I dont think THAT is a good idea.

It usually does work, but not always, you can see some
problems with the bios working out what drives are present etc.

Thats a different issue to the damage question tho.


I always disconnect both.

And I dont like the major kludge that removable drive bays
involve, I prefer to use a proper formal standard like USB
or firewire or SATA that defines that behaviour fully.
Click to expand...

And you reckon *I'm* a tad radical! 8-)

Incidentally, does the SATA spec make such things simpler than the
older ATA?


Cheers, Phred.
 
CJT said:
Good for you. I'm going to refrain from that sort of thing,
except perhaps with a drive that's spec'd for hot plugging.
Call me overly cautious if you like, but my data is important
to me.
Click to expand...

Ok, all this talk of reflections on IDE cables has me confused - what should I
be avoiding to minimise reflections? having a drive attached without power?
having a single drive in the middle connector of an IDE cable with nothing at
the end?
 
Alex Hunsley said:
Ok, all this talk of reflections on IDE cables has me confused -
what should I be avoiding to minimise reflections? having a
drive attached without power? having a single drive in the
middle connector of an IDE cable with nothing at the end?
Click to expand...


Having an open connector (i.e. infinite and discontinuous
impedance) at the end of a transmission line (i.e. the data
cable) produces strong reflections. The open connector
at a mid-point is not as a drastic change in impedance since
the transmission line continues on past it. Since any connector
creates some degree of impedance discontinuity, the fewer
connectors in the signal path the better - which is an argument
against slide-in hard drive trays. The question is really how
far do you think your hardware will let you stray from IDE
specs before there's a data error problem. If you put a
removeable HD tray at the end of the cable with another HD
at the mid-point, the mid-point HD may have error problems
when the end tray is removed. If the removeable tray has the
only device on the channel, there probably won't be a problem.
If the removeable tray is at the mid-point, there probably won't
be a problem. But "probably" is a function of your hardware.
Personally, the only situation *I'd* avoid is the one with the
removeable tray at the end with another device at the mid-
point.

*TimDaniels*
 
Back
Top