no hint, just a suggestion. Clear CMOS ?

Some AMI BIOSes demand you enter BIOS by hitting F1 when RAM size has been changed.

Yeah, i would suggest to deliberately "mess something up" - disconnect FDD or HDD or something like that so that the bios would whine about it. And then check, is the keyboard properly registered on that alert? (It probably asks you to either continue or enter bios)

Tried to trigger this behavior with changing ram, etc.
No reaction.

Can you provide more details about the motherboard? Does it use a Dallas RTC? A picture or two could be useful, especially around the area of the chip where CMOS settings are stored (i.e. 82C206).

From your first post I understand that it used to work and suddenly stopped behaving this way? I would try following the traces that come out of the BIOS chip and see if there's any breaks.

Here is a pic that i think is good enough.
The board is in pretty good condition.
Inspected traces - everything checks out.
It feels like component(s) are starting to act.

The empty 24 pin socket below the keyboard controller seems to be where an RTC should go, probably one with an internal battery such as the Dallas 12887 as I don't see any barrel or external battery connector. Do you install something in that socket when using the board? I don't think it would work properly without an RTC, unless it's not a discrete chip but is integrated in one of the Symphony chips, not really familiar with those.

Also, have you tested continuity on the two thick traces between the top ISA slots? They look corroded.

yes, i plug in the empty socket a dallas module.
added the socket a while ago sine the original module was surface mounted.

you have a good point about the 2 corroded traces. i didnt check them since the no bios issue surfaced.
thanks for the reminder.

Regarding the Dallas RTC, I noticed some strange behavior when trying to use an Intel board when the battery is either flat or new: it just hangs when attempting to enter the BIOS, because the CMOS settings are in an unexpected/invalid state. But I can still boot off a floppy and run a program called GSETUP to initialize the CMOS with acceptable values. Not sure if these symptoms are similar to yours but it may be worth a try, there should be some thread about this on the forum.

I had the same issue with my 486SC-P (same chipset):

Ways to safely transplant / replace 160-pin QFP chips

Depending on the temperature of the 82C362 sometimes the system wouldn´t start at all, someteimes I could not enter BIOS, sometimes I could but the first BIOS screen would freeze accompanied by all possible beep codes imaginable.

These FIC mainboards are very reliable and durable otherwise.

Ouch.
Is this the death of Symphony Haydn season or what ?
Saying this because 3 weeks ago another Symphony board went silent on me.
It was a weak mobo so i didn't feel anything but the one in question here actually works with BL3 IODATA module which is very rare.
Will check those traces tomorrow.
Slap a Peltier element on the big chip - noticed that it gets quite hot quick.
The Peltier will chill it for good.

The 82C461 gets quite hot during operation - this is normal behaviour, the 82C362 however stays very cool. Try freezing it with some good coolant spray before turning on the system.
The 82C362 is absolutely crucial for the system since it controls all the periphery! When it fails anything is possible.
Still hope there´s some other culprit in your case...

Actually i have some other (late) Symphony Haydn boards where the 82C461 stays cool and ok.
The early (wide PCB) assemblies are flaky.

Maybe there were later improved revisions of the C461 or it was operating out of specs on early board designs. I have never seen a failed C461 though.

Yes, there are later production runs.
Their chips markings look real fancy.