Pin 23 is defined as "
-DIOW: STOP". To me, this means we would have to hold the signal line high (+5v) to avoid the write of data into the drive.
Now, I'm assuming the IDE/ATA interface uses +5vc/0vdc for its signalling. I would need to confirm that before ever applying voltages of any kind to wires or pins on the cable.
But if it is simply a constant +5vdc on the line, then you have to face the challenge of breaking open the cable run somewhere along the line in order to inject +5vdc into it (into the drive ...
not into the controller).
IDE cables are cheap, and I suggest you use some 5h1tty old IDE drive that has no value ... many of the sub-1GB ones fit into this category. (I took a 130MB drive out of production the other day ... it's a perfect candidate for such experimentation.)
First figure out which cable line is pin 23. Chances are, with one side of the cable being marked in red for pin 1, then 23 will be the 23rd wire over from that.
Once found, take a razor and slice a section of the wire free from the cable. Cut that across the middle and note which end goes to the drive. Strip the insulation from that end (note that this wire is tiny and this will require exacting work to avoid cutting the conductor again), and then solder on a long red wire, like the kind used in the power lines coming out of your power supply. For info:
Take one of your spare power plugs and stick the other end of the red wire into it at the point of the plug's own red wire. To make the red wire fit into the plug snugly, strip a longer length than normal and then fold the exposed wire in half to make a double-wide wire end. Cram it in there good.
Viola!
You have now hard-wired your IDE drive to always be write-disabled. What happens now,
I really don't know. After all, the controller will still try to set -DIOW low and expect the hard disk to do the write operation. Rough signalling aside, the hard drive's controlling logic may respond to the IDE controller in some way that involves other signals, like along the data lines themselves ... and if that's not right, the IDE controller may consider the drive to have failed in some way.