Weird. Maybe there's a short circuit somewhere in there.

That is weird.
If you have a multimeter, I guess what I'd try would be to sanity check all the pins and that the case is grounded.
Check if that spot on the case is grounded relative to the PSU casing, or maybe test it against a ground pin at one of the power connectors.

With the serial port plugged into the board but not mounted, check the resistance of each pin vs the outer shell (which normally makes contact with the case). Figure out which pins are in electrical contact with the shell.
Compare with a diagram of the serial port pinout, and see if any of those pins in contact with the shell are not supposed to be connected to ground. If you find such a problem, then measure it again with it unplugged from the board.

here we go - the case is grounded to the psu ground pin.
and so is pin 9 (ring indicator) instead of pin 5 (ground)! what the hell?
here's a pinout: http://www.db9-pinout.com/

none of the pins is connected to the shell. I've tried and looked at different db9 connectors with attached cables, none of them had one of the pins connected to the shell.

why on earth is pin 9 grounded when connected to the mainboard?

edit:
it seems that the motherboard has another pinout. however I don't understand why it is still working when not grounded.
this mainboard (486-GIO-VT) seems to be wired from left to right, like this (looking from top to the onboard connector):
6 7 8 9
1 2 3 4 5

normally i guess it is like this:
2 4 6 8
1 3 5 7 9

that would explain why 9 is grounded and 5 isn't on my board. do you know of any boards which were wired like that?

edit 2:
installed a multi i/o card with serial port, disabled everything but the serial port and parallel port, disabled onboard serial and parallel ports -> BOOM everything working.
besides: when the connector didn't have contact to the case and thus wasn't grounded, the mouse only was working within DOS and using ctmouse. as soon as win 3.11 started, the mouse stopped working. now, with the i/o card, it is working even in windows.
that was a hell of a wtf moment.

The headers were never standardized. Strange luck that it would work as well as it did.

firage wrote:

The headers were never standardized. Strange luck that it would work as well as it did.

yuck...

seems that the layout

2 4 6 8
1 3 5 7 9

was more common, as all my headers and motherboards (as far as i know) are like this. only that FIC board seems to do it differently...

IMHO with a little screwdriver you should be able to take the contacts out of the connector and reorganize them

alexanrs wrote:

IMHO with a little screwdriver you should be able to take the contacts out of the connector and reorganize them

yeah, been there, destroyed that 🤣

OMG ahhaahahahahah
I managed to do that with a bunch of USB cables, but it does take a somewhat steady hand and a bit of precision.