[N8VEM: 16955] N8-2312 problem with on-board PPIDE and RomWBW 2.5.2

[Arlen]

Another N8-2312 problem I finally debugged late last night was an
unresponsive on-board PPIDE interface.

I could not get the PPIDE drive to work in RomWBW 2.5.2. After checking
all the upstream hardware, I could not understand why the 8255's chip
select was staying stubbornly inactive.

But when I ran my own tiny test loop I could see the chip select waggling
properly on the scope. At this point I began to suspect that the BIOS code,
unlike my 5-byte test loop,wasn't actually talking to the 8255 parallel
port. In other words, a software glitch, not hardware (for a change :)

So I went back to config_n8_2312.asm in the RomWBW 2.5.2 distribution and
discovered a weirdness in the port number used by the software to set up
the PPIDE code. In the file I changed

PPIDEIOB .EQU $60 ; PPIDE IOBASE
to
PPIDEIOB .EQU $80 ; PPIDE IOBASE

which matches the 8255's base I/O port number of $80 -- as indicated by the
address decoding circuit on the N8-2312 schematic. After re-building the
ROM with this one change, PPIDE now works perfectly with both hard drives
and CF adapters.

Has anyone else encountered this? It had me running in circles for a while.

Arlen Michaels

[Wayne Warthen]

Hi Arlen,

Nice work tracking that down. Not sure how it got screwed up. The N8 2511
config file has it correct. The N8 2312 has the value set to what would be
used for an ECB SBC board. I have already corrected this and it will be
included in the next release of RomWBW.

I'm afraid that the number of hardware combinations I am trying to support
makes it virtually impossible to do a complete regression test when I
release a new version. So, I do depend on these reports of problems.

Thanks,

Wayne

Post by Arlen
Another N8-2312 problem I finally debugged late last night was an
unresponsive on-board PPIDE interface.
I could not get the PPIDE drive to work in RomWBW 2.5.2. After checking
all the upstream hardware, I could not understand why the 8255's chip
select was staying stubbornly inactive.
But when I ran my own tiny test loop I could see the chip select waggling
properly on the scope. At this point I began to suspect that the BIOS code,
unlike my 5-byte test loop,wasn't actually talking to the 8255 parallel
port. In other words, a software glitch, not hardware (for a change :)
So I went back to config_n8_2312.asm in the RomWBW 2.5.2 distribution and
discovered a weirdness in the port number used by the software to set up
the PPIDE code. In the file I changed
PPIDEIOB .EQU $60 ; PPIDE IOBASE
to
PPIDEIOB .EQU $80 ; PPIDE IOBASE
which matches the 8255's base I/O port number of $80 -- as indicated by
the address decoding circuit on the N8-2312 schematic. After re-building
the ROM with this one change, PPIDE now works perfectly with both hard
drives and CF adapters.
Has anyone else encountered this? It had me running in circles for a while.
Arlen Michaels

[Arlen]

Ah, so $60 was carried over from the original ECB SBC port assignment.
That explains it nicely These things are bound to happen when the software
bravely supports so many different hardware configurations.

Thank you for RomWBW. It's a superb piece of work!

Arlen Michaels

Post by Wayne Warthen
Hi Arlen,
Nice work tracking that down. Not sure how it got screwed up. The N8
2511 config file has it correct. The N8 2312 has the value set to what
would be used for an ECB SBC board. I have already corrected this and it
will be included in the next release of RomWBW.
I'm afraid that the number of hardware combinations I am trying to support
makes it virtually impossible to do a complete regression test when I
release a new version. So, I do depend on these reports of problems.
Thanks,
Wayne

Post by Arlen
Another N8-2312 problem I finally debugged late last night was an
unresponsive on-board PPIDE interface.
I could not get the PPIDE drive to work in RomWBW 2.5.2. After checking
all the upstream hardware, I could not understand why the 8255's chip
select was staying stubbornly inactive.
But when I ran my own tiny test loop I could see the chip select waggling
properly on the scope. At this point I began to suspect that the BIOS code,
unlike my 5-byte test loop,wasn't actually talking to the 8255 parallel
port. In other words, a software glitch, not hardware (for a change :)
So I went back to config_n8_2312.asm in the RomWBW 2.5.2 distribution and
discovered a weirdness in the port number used by the software to set up
the PPIDE code. In the file I changed
PPIDEIOB .EQU $60 ; PPIDE IOBASE
to
PPIDEIOB .EQU $80 ; PPIDE IOBASE
which matches the 8255's base I/O port number of $80 -- as indicated by
the address decoding circuit on the N8-2312 schematic. After re-building
the ROM with this one change, PPIDE now works perfectly with both hard
drives and CF adapters.
Has anyone else encountered this? It had me running in circles for a while.
Arlen Michaels

[Wayne Warthen]

Exactly! And I'm glad it is working well for you.

--Wayne

Post by Arlen
Ah, so $60 was carried over from the original ECB SBC port assignment.
That explains it nicely These things are bound to happen when the software
bravely supports so many different hardware configurations.
Thank you for RomWBW. It's a superb piece of work!
Arlen Michaels

Post by Wayne Warthen
Hi Arlen,
Nice work tracking that down. Not sure how it got screwed up. The N8
2511 config file has it correct. The N8 2312 has the value set to what
would be used for an ECB SBC board. I have already corrected this and it
will be included in the next release of RomWBW.
I'm afraid that the number of hardware combinations I am trying to
support makes it virtually impossible to do a complete regression test when
I release a new version. So, I do depend on these reports of problems.
Thanks,
Wayne

Post by Arlen
Another N8-2312 problem I finally debugged late last night was an
unresponsive on-board PPIDE interface.
I could not get the PPIDE drive to work in RomWBW 2.5.2. After checking
all the upstream hardware, I could not understand why the 8255's chip
select was staying stubbornly inactive.
But when I ran my own tiny test loop I could see the chip select
waggling properly on the scope. At this point I began to suspect that the
BIOS code, unlike my 5-byte test loop,wasn't actually talking to the 8255
parallel port. In other words, a software glitch, not hardware (for a
change :)
So I went back to config_n8_2312.asm in the RomWBW 2.5.2 distribution
and discovered a weirdness in the port number used by the software to set
up the PPIDE code. In the file I changed
PPIDEIOB .EQU $60 ; PPIDE IOBASE
to
PPIDEIOB .EQU $80 ; PPIDE IOBASE
which matches the 8255's base I/O port number of $80 -- as indicated by
the address decoding circuit on the N8-2312 schematic. After re-building
the ROM with this one change, PPIDE now works perfectly with both hard
drives and CF adapters.
Has anyone else encountered this? It had me running in circles for a while.
Arlen Michaels