VOGONS


First post, by auron

User metadata
Rank Oldbie
Rank
Oldbie

on my current batman's revenge setup, i am using a ~'92/93 UMC ISA i/o card because i don't have a cable with that rare pinout to fit the onboard serial header. it works and everything, but i've been wondering how does detection exactly work when there are now two SIO chips in the system. without the card plugged in all the ports are detected as FIFO capable in hwinfo, while COM1 is shown as a different type without FIFO when the card is plugged in. unlike later motherboards, there is no way to change the COM port addresses in BIOS, so wouldn't there be a conflict if both chips tried to use the same address? are performance differences between onboard and add-on SIOs conceivable?

this controller card has IDE as well and it can be jumpered on/off. however there is another jumper "HDC EXIST/NOT EXIST" and i'm assuming this stands for hard drive controller, but what does it do exactly? does it set the port to tertiary or quaternary as on sound cards?

Reply 1 of 2, by mpe

User metadata
Rank Oldbie
Rank
Oldbie

On Batman's Revenge there is a actually a setting for serial port:

In "PERIPHERAL MANAGEMENT SETUP"when you set "Programming option" to "manual", then you have "First Serial Port Address" and "Second Serial Port" with "Disabled, 2E8H, 3E8H, 2F8H, or 3F8H"values which allows you to reconfigure your ports as you like.

If you leave it on default "Auto" and use ISA card then the BIOS might disable on-board I/O to avoid conflict. However, with ISA some IRQ resource sharing is often possible if you don't use the other one. Just the I/O port need to be unique.

Not sure about the IDE setting, but I believe this might configure the on-card IDE as secondary port so that you can use the on-board one as primary. However, you shouldn't be using RZ1000 on the Batman board as it is known to corrupt data due to hw bug.

Blog|NexGen 586|S4

Reply 2 of 2, by auron

User metadata
Rank Oldbie
Rank
Oldbie

thanks, i hadn't paid enough attention to the settings and completely missed "programming option". putting it to manual and turning everything off, now only a 8250A/16450 UART is left in hwinfo, so everything as expected. although as you said maybe it was smart enough to disable onboard COM1 on auto anyway.

regarding the RZ1000, i'm aware of its notoriety and wanted to make a thread about it, but there's already been some discussion with links here. so far i have not noticed any corruption with using win95b/rebooting to DOS. my current BIOS version is 1.00.08.AF2, which i had to install via recovery after wiping the original BIOS by installing a SCSI card with OROM and not being aware of the flash write protection being off by default.