"Switch Out" Physical HDD (hard disk drive)

Franc Zabkar wrote in
:

On Thu, 14 Jun 2012 13:06:56 -0400, put finger
to keyboard and composed:


I would like to add a 2 pole switch in the wires that supply power
to the internal HDD so I can "switch out" the internal HDD (make it
"disappear") when I want to put a "Master" drive in the drawer.

I can't see why that shouldn't work. However, you might like to run
the jumper pins through a switch so that you could program the
internal drive as either master or slave. That way both drives would
be visible to the OS.


With that and two poles for the supply wires, it could get complex to the
point where MOSFET's might be better used to do the switching, partly to
prevent any change of state when power is applied, if logic gates were added
to prevent change when power was applied. Or is there a way to do any of this
safely during run time? As far I know, there isn't..

On Thu, 14 Jun 2012 20:52:02 -0500, Lostgallifreyan
put finger to keyboard and composed:

With that and two poles for the supply wires, it could get complex to the
point where MOSFET's might be better used to do the switching, partly to
prevent any change of state when power is applied, if logic gates were added
to prevent change when power was applied. Or is there a way to do any of this
safely during run time? As far I know, there isn't..

I wouldn't want to try that, but I don't have any technical basis for
saying so. I just wouldn't feel comfortable about it. In fact, it may
not even be advisable to have the motherboard connected to the IDE
interface in the absence of DC power. AIUI, latch-up used to be a
serious problem with early CMOS devices, but I don't know whether this
problem has any relevance to the OP's situation.

I still think the switchable master/slave jumpers would be the safest
solution. In fact I did a very similar thing in an old 286 machine
when I had to give the user the option of booting from either a 1.2MB
FDD or 1.44MB FDD. My solution was to wire the drive select signals to
a 4PDT switch. The user then changed the BIOS configuration as and
when required.

- Franc Zabkar
--
Please remove one 'i' from my address when replying by email.

Franc Zabkar wrote in
:

On Thu, 14 Jun 2012 20:52:02 -0500, Lostgallifreyan
put finger to keyboard and composed:

With that and two poles for the supply wires, it could get complex to
the point where MOSFET's might be better used to do the switching,
partly to prevent any change of state when power is applied, if logic
gates were added to prevent change when power was applied. Or is there a
way to do any of this safely during run time? As far I know, there
isn't..

I wouldn't want to try that, but I don't have any technical basis for
saying so. I just wouldn't feel comfortable about it. In fact, it may
not even be advisable to have the motherboard connected to the IDE
interface in the absence of DC power. AIUI, latch-up used to be a
serious problem with early CMOS devices, but I don't know whether this
problem has any relevance to the OP's situation.


I wouldn't want to either. But mainly because of complexity, logic gates
mean board design, etc.. I hadn't thought of latchup, but I had assumed that
the logic power would be drawn from the main 5V line, so either everything
gets power (except the drive), or nothing does. When seeing this condition in
various tests I've tried in times past, no damage resulted. The Lian-Li drive
trays I use have keyswitches that disconnect power while changing the
physical lock state, and while I don't leave them present but unlocked often,
it never caused bother. It just meant that the drive had to be locked in at
boot if it was going to get recognised.

I still think the switchable master/slave jumpers would be the safest
solution. In fact I did a very similar thing in an old 286 machine
when I had to give the user the option of booting from either a 1.2MB
FDD or 1.44MB FDD. My solution was to wire the drive select signals to
a 4PDT switch. The user then changed the BIOS configuration as and
when required.


While leaving both drives permanently connected to power lines? That would
work, and be simple too, if a tad hungry. FDD's are removable too, which
might mean a lot. I don't know if ATA can be induced to re-read all it needs
to, after boot, by any method.