VOGONS


Reply 20 of 25, by JAKra85

User metadata
Rank Newbie
Rank
Newbie
the3dfxdude wrote on 2026-07-14, 20:06:
Since you weren't writing what you posted, then I'll just say the message is unclear what you did to solve your problem. […]
Show full quote
JAKra85 wrote on 2026-07-14, 18:50:
To me it seems that no one understands what I've wrote... Yeah, I was assisted by AI, however I think it's dead on. […]
Show full quote

To me it seems that no one understands what I've wrote...
Yeah, I was assisted by AI, however I think it's dead on.

This is not an IRQ conflict issue.
3C509B is set and was set to IRQ 2. I've tried different IO addresses 0x210, 0x280, 0x290, 0x300... all froze with the Xebec Controller.
This 3C509B also has a special ID address 0x110, it's hard coded not editable in any way.
Xebec Controller uses IO address 0x310 also by default not editable.
These two addresses get in conflict and there is no way to modify them.

Western Digital WD1002A-WX1 is set to use IRQ5 and works well on address 0x320.

Since you weren't writing what you posted, then I'll just say the message is unclear what you did to solve your problem.

You are saying there is an IO port conflict. I pulled up the driver source code from linux. There is a port *0x100 called ID_PORT, which is used to access the eeprom. That ID_PORT is used to find the actual IO port.
*The driver code technically scans 0x100-0x200, but all things considered, it's probably on 0x100. This means it is NOT hardcoded.

There is also the fact that I/O ports go up to 0x3FF on PC/XT.

If you got this adapter new, it'd likely default to IO port 0x300 and IRQ 10, which means you'd want to fix the IRQ. But who knows what you got initially. If it was using both 0x100 and 0x300 addresses I can't say why you think those addresses would overlap, likewise, I don't think anyone would have used any IO port above 0x2FF in the PC/XT era if it truncated/rolled over, but they did.

So there are issues with the information that has been shared.

This card below:

The attachment 20260713_193605.jpg is no longer available

Uses an ID port 0x110. There is no way to change this:

The attachment 20260715_124648.jpg is no longer available

Using network card configuration utility 3CCNFG.EXE I can change:
-IRQ (currently: 3)
-I/O address (currently: 0x210)
-PnP (OFF)
-Modem (OFF)
-Full-duplex or half-duplex (currently: Full-Duplex)
-Connection type (currently: RJ-45)

ID port address is not editable and I could not find anything on the net that could modify this ID port.

Last edited by JAKra85 on 2026-07-15, 11:27. Edited 1 time in total.

Retro PC Gallery:
https://docs.google.com/document/d/1ara32t4rV … nU/edit?tab=t.0

Reply 21 of 25, by JAKra85

User metadata
Rank Newbie
Rank
Newbie
Grzyb wrote on 2026-07-15, 08:03:

BTW: there's something special about slot 8 in IBM 5160 - some cards can't work correctly in there.

Is there anything in that slot?

ISA slot 8 is empty.

This is the machine I am working on:
IBM XT 5160

I was informed that the ISA8 is special:
Re: IBM XT 5160

Retro PC Gallery:
https://docs.google.com/document/d/1ara32t4rV … nU/edit?tab=t.0

Reply 22 of 25, by JAKra85

User metadata
Rank Newbie
Rank
Newbie

Further updates to this topic, important.

I am using a special (Nestor https://github.com/hackerb9/3C509B-nestor) packet driver that allows the use the 3Com EtherLink III 3C509B with a PC running a 8088 CPU.

Plus!
"The 3Com configuration program, 3C5X9CFG.EXE, would die claiming no card was in the machine. Predator99 fixed the problem so it now works on 8088 and 8086 machine. The resulting executable, now called 3ccfg.exe, has been compressed with PKLITE so that it fits on a single 360K floppy."
Opening network card configuration utility 3ccnfg.exe AFTER drivers loaded (3c509.com 0x60) still crashes the machine as previously.
The network card configuration utility 3ccnfg.exe runs fine without drivers loaded first.
I don't know if this is normal or not.

So, in the end the only issue I managed to fix by going from Xebec MFM to WD MFM is the copying of files trough FTP.
That works nicely now.

Retro PC Gallery:
https://docs.google.com/document/d/1ara32t4rV … nU/edit?tab=t.0

Reply 23 of 25, by the3dfxdude

User metadata
Rank Oldbie
Rank
Oldbie
JAKra85 wrote on 2026-07-15, 11:05:
Uses an ID port 0x110. There is no way to change this: […]
Show full quote

Uses an ID port 0x110. There is no way to change this:

The attachment 20260715_124648.jpg is no longer available

Using network card configuration utility 3CCNFG.EXE I can change:
-IRQ (currently: 3)
-I/O address (currently: 0x210)
-PnP (OFF)
-Modem (OFF)
-Full-duplex or half-duplex (currently: Full-Duplex)
-Connection type (currently: RJ-45)

ID port address is not editable and I could not find anything on the net that could modify this ID port.

Why would you want to change it?

While it's a PnP supporting card, it's clear that the designers wanted this card to work without PnP. The stock drivers would support 286, 386, and older 486 and pentium systems, which at the time still would be around. I would suspect the card, in order to support multiple of the same NIC or other devices in the 0x100-0x1FF range, has a way to arbitrate the location*. Because you would want to be able to access the eeprom information for each individually. So it's set up automatically and needs no user control.
*Not sure if this happens due to PNP support or it does it on it's own, but it's clear in the driver code it doesn't have to be fixed to a single location.

So based on how it works, I'd pretty much say, it's very unlikely that was causing the problem.

Reply 24 of 25, by rasz_pl

User metadata
Rank l33t
Rank
l33t

1 port 0x110 is definitely not the source of problems, it doesnt collide with anything of MFM HDD

2 its definitely moveable, as evidenced by /3C509B-nestor:

"Configuration & Diagnostic Port
(Do not change unless another device is on address 110) 110 /CONFIGPORT: 100-1E0, in increments of 10"

https://github.com/raszpl/sigrok-disk FM/MFM/RLL decoder
https://github.com/raszpl/FIC-486-GAC-2-Cache-Module (AT&T Globalyst)
https://github.com/raszpl/386RC-16 ram board
https://github.com/raszpl/Zenith_ZBIOS Zenith Z-386 MFM-300 ZBIOS disassembly

Reply 25 of 25, by JAKra85

User metadata
Rank Newbie
Rank
Newbie
rasz_pl wrote on 2026-07-15, 17:19:
1 port 0x110 is definitely not the source of problems, it doesnt collide with anything of MFM HDD […]
Show full quote

1 port 0x110 is definitely not the source of problems, it doesnt collide with anything of MFM HDD

2 its definitely moveable, as evidenced by /3C509B-nestor:

"Configuration & Diagnostic Port
(Do not change unless another device is on address 110) 110 /CONFIGPORT: 100-1E0, in increments of 10"

Hmm, good find!
I will set back the Xedec and do a retry, then I can rule out the AI suggested 0x110 and 0x310 conflict.

Any ideas on how to proceed, what to check if the system still crashes with the Xedec MFM when I initiate an FTP copy over the network?

Retro PC Gallery:
https://docs.google.com/document/d/1ara32t4rV … nU/edit?tab=t.0