Strange behavior new disk -- explain?

Can anyone explain the following strange behavior I encountered installing a new WD Black 2TB SATA disk?

Using a PCI Promise TX4200 controller in JBOD mode.

1. I physically installed the disk, and powered up. Windows did not believe the disk existed. Disk Manager and Device Manager showed nothing. Using the Promise manager, the controller reported the card was installed, had the right size, and was functioning properly. No change with a couple of power cycles. I already had a Seagate 2TB disk that had been working flawlessly on this controller for ages.

2. I decided to see if the problem was Windows or hardware, so I booted into a live Linux distro (Slax 7.0.8). Using parted, it said that disk had an unrecognizable label, and gave no info for the disk, but was happy with all my other disks (13 disks in total, including external USB). I figured maybe the disk was configured with a new label type designed for >2GB disks, even though this was a 2GB disk. So, I used parted to write an "msdos" label, and this seemed to succeed. parted now reported the right size and type of disk, with 512 byte sectors. This persuaded me it was not a hardware problem.

3. Powered back up, but Windows still did not acknowledge existence of the disk -- exact same results as before.

4. I decided to see what would happen with Win 7, so I installed the disk in an external USB enclosure and attached it to a win 7 machine. No problems -- I was able to format the disk with NTFS (non-GPT), and successfully wrote a small file.

5. Reinstalled the disk back into the XP machine. No change whatsoever -- not recognized, no indication of existence in Device Manager, etc.

6. I realized that at the last install, I had forgotten to plug back in an IDE disk on another PCI controller (which still had the master plugged in). So, I powered down, plugged that one in, and powered back up.

7. Miracle: now Windows recognizes the new WD disk just fine, as if nothing had ever been wrong.

What I can't understand is why the behavior should be different between the last 2 power-ups?

--peter