Hello Zbigniew,

  ZB> It somewhat unusual situation, because it seems to not
    > be dependent on the connection "point", neither on
    > jumpering.

That seems to support my suspicion that at least one of the
drives is jumpered for "cable select", something that I have
always avoided.

  ZB> I'm using a board, which has one "ordinary" PC-IDE
    > controller, and one HPT370. My main HDD is connected
    > to that faster HPT370 - and, when it's just one HDD in
    > the system, it's seen by NetBSD as wd0. To the
    > secondary channel of the slower (primary) controller
    > there is CD-R and ZIP-drive connected, the primary
    > channel is disabled.

I don't know what an HPT370 is, but I imagine it's a faster
ATA host bus adaptor than the one built into your mainboard.

  ZB> I made several tries:
    > - disconnected the CD-R from secondary channel of the
    >   IDE-controller (that "ordinary" one), and connecting
    >   there that additional HDD

That might have worked if it were not for the Zip drive,
which is probably ATAPI.

  ZB> - disconnecting ZIP-drive from the primary channel of
    > the same controller, and replacing [..]

The same applies.

  ZB> - turning on the primary channel of that controller,
    > and connecting HDD there

I don't know for sure, but that may cause a conflict with
your HPT370 host bus adaptor.

You still haven't answered my questions regarding how the
drives are jumpered.

- Andy Ball