Reseller,
No, the Atapi Zip wasn't the problem. At one point I thought it was, at another point I was sure it was the fault of a cheap CDROM changer. I thought this because if either Atapi device was on the failing port, and I disconnected it (and rejumper the Master accordingly), the problem went away.
But only one port driver failed at a time, never both, and it failed for either Atapi device, and which port failed also depended on which version of the Intel UDMA driver was installed! It was maddening.
I don't know for sure what it was, but 3 possibilities occur to me:
1) Mobo bug. I downloaded and installed IWILL's latest BIOS, but no change. I think it is more likely the mobo is the culprit, but have no proof.
2) HD bug. But it seems so unlikely both could have the same problem. On the other hand, both HD's, irritatingly enough, required rejumpering when going from Master-without-slave-present to with-slave-present. Both of the Atapi devices have only one Master setting. So I suppose it is possible the HD's might not have been recognizing an Atapi device as a slave. For instance, when I tried the two HD's on the primary port, and the two Atapi's on the secondary, the problem went away. But I didn't want that configuration, because I wanted the fastest throughput when backing up whole partitions - and that means putting them on separate ports. So apparently, the UDMA HD's get along well with their own kind, but not when Master to an ATAPI device slave. But contradicting that is the fact that only one of the two port drivers failed at a time, though I could control which one it was, as I mentioned above, by changing Intel drivers. Finally, when I substituted either UDMA HD with a non-UDMA HD (a 1 year old 2.5GB WDC) no port drivers ever failed.
3) Software bug. Whenever I re-installed either Win95 (I've done this too many times already for one lifetime) or either of the Intel drivers, the problem always went away for one re-boot. Upon the second reboot, it always reappeared and stayed.
Summing up, with the IWILL P55XB2 mobo, I can have at most one UDMA HD in the master position with an Atapi slave. Now why can't they write that simple rule in the mobo manual? <g> |
