Jim wrote:
> It could just be that the PSU cannot support all the drives
> starting-up at
> the same time.
Good point. Though the difficulty seems to occur after the
CMOS BIOS and the various controller BIOS's have run and
init'ed the drives. Not out of the question that if the
drives are then asked to spin up at the same time there
might be an extra surge that puts it over the top. However,
we don't see problems when the system is running, so I kind
of doubt it is the PSU (which is probably a little light at
400W).
>
> And ... the required fans to cool all that - I do hope your caddies
> incorporate fans
Yup, and a couple have temp. alarms. <g>
> It seems to me that you already have 6, or perhaps 8 P-ATA
> drives and now 2
> more S-ATA drives.
We have 6 drives (4 PATA + 2 SATA).
>
>
> Re the memory usage, I'd have thought that was unlikely -
I've personally encountered a conflict between the BIOS of
a Promise Fast Track RAID contoller card and the builtin SATA.
Confirmed by ASUS support, and Promise -- with no fix, because
there are no jumpers on that Promise board. Thus, I don't
think it is a big leap to believe the Promise Ultra/ATA card
has problems too.
> I would expect that either they are engineered/designed to
> co-exist in the
> system, or they would always require the same fixed memory locations
> Perhaps an IRQ conflict - but the manufacturers should be
> able to confirm,
> or deny that or the memory problem.
see above. Note that there is no IRQ conflict or any other conflict
after XP has booted.
>
> If they don't seem interested, then that would be something to post to
> people like Toms Hardware, and the Microsoft hardware
> certification section
> (The controllers are Microsoft XP certified I hope)
Beats me. I think these disk controller cards are recognized
by standard Windows drivers. I believe that technically they
(SATA and Ultra/ATA) both appear as SCSI controllers.
> Do the Systems/add-in boards/drives have a delayed startup
> facility/option.
No but the BIOS might. Again, I think the observed facts
don't indicate that power is the problem.
>
> Also - you may find that a PCI-x controller may speedup the system -
> remembering that the base PCI connections have a single
> memory access path
> that means the drive controllers get memory access
> asynchronously, while the
> new master/x PCI connections run at about 4 x the old PCI
> connection speed,
> and can run concurrently, with a large number of
> 'intelligent' drives that
> may double the actual system throughput.
True. We tend to use these large drives on a project
by project basis (hence the use of caddies), so typically
we're working on only one/two large drives at a time,
and perhaps copying stuff back/forth. So, the need for
a lot of overlapped I/O isn't high.
I appreciate your comments/suggestions though, and will definitely
take them into consideration. thanks, - Gary
--
----------------------------------------
The WIN-HOME mailing list is powered by L-Soft's renowned
LISTSERV(R) list management software. For more information, go to:
http://www.lsoft.com/LISTSERV-powered.html
