Half/quarter speed beeps would make me suspect the deturbo function may be active.

Hi Doornkaat,
mmh I don't know what a deturbo function is...
But the machine has no turbo button and also no jumper on the board marked anything like turbo.

There is a jumper pair tough that says 33/40/50MHz. Only 33/40 are with pins soldered in, and if I change it from 33 to 40 the beeps and the floppy seek goes even slower. Like the floppy stepper is literally moving slower.
It seems so crazy to me...
I think it is a 486 SX but I'm 100% not sure, there is a round tall heatsink stuck on the top so I can't identify it. But it is soldered to the board has a coprocessor socket next to it so I'm pretty confident about it being SX

Some machines are like this with a stuck key if you've turned the keyboard check off. Some machines go like this when one of the oscillators breaks and goes way off frequency. Some machines go like this when there's excessive power ripple from either the main PSU or the onboard voltage converters.

Stepping speed is up to FDC, it's programmable. 8272A, which is the father of all PC floppy controllers (and itself based on 765 I belive) can have it set between 1 and 16 ms, with 1ms resolution. This timing comes from internal timer and that in turn depends on the input clock (which should be 8Mhz). Seek command should run entirely on FDC internal timer and turbo, keys and other factors should not affect it.

Does this system still read floppies? If the input clock is wrong it should not be able to. And in that case I would suspect the motherboard clock generator since all of the system seems slow. It can be something as trivial as dust buildup between the crystal oscillator leads, if that dust gets damp it can easily load the circuit down or even stop it.

8272A is Intel designed and is not pin compatiable. The 765 floppy controller was NEC initial design long ago and others cloned it, but not intel's, WD has their own design, not commonly seen outside of WD-designed floppy and hard drive controller cards.

Cheers,

Deunan wrote on 2022-02-14, 23:30:

It can be something as trivial as dust buildup between the crystal oscillator leads, if that dust gets damp it can easily load the circuit down or even stop it.

I recently noticed that dust that looks like fibrous fine sawdust or cardboard or box dust can go like that, it gets blown in the system dry, then a touch of humidity and it swells and packs itself between pins, and becomes fairly conductive in it's damp state. I mean normally you could put damp card over the top of a 9V battery and it wouldn't discharge it much, probably thousands of ohms resistance but when it packs in between pins and is compressed from the moisture it seems to drop to mere hundreds or high tens even with close pin spacing. Had a motherboard that I surmise must have come from a packing/box factory and had to "pick it's teeth" in detail because even at 40% humidity it was wigging out in various ways.

Hi,

after ages of waiting I finally got in bios and was able to change the settings.
The culprit was BL-Clock it was set to one... when setting it to 2 or 3 it works fine.
I changed it to 2 and also the jumper on the board form 25 to 33 mhz now it seems to be ideal and boots windows 95 nicely 😀
Thank you 😀

*edit*
A weird problem is still there though...
The audio playback in windows 95 is very choppy and seems way too fast...
Could it have something to do with that situation?
It doesn't matter if I change it to 25 or 33MHz and also doesn't change if I set BL2 or BL3
*edit2*
And windows 3.11 for workgroups doesn't get past the splash screen...
I have the feeling something really weird is going on.

I still think it's something with the mobo clock generator, did you try reading from floppy? If the ISA clock is also too slow that would explain soundcard issues, and too slow system timer might cause WfW 3.11 to trigger bugs that only manifest on much faster CPUs. There's a known issue, which was present also on Win95 I think, and was related to the way internal delay loops in networking stack were calibrated. DOS runs it's internal clock on system timer so just boot the system, check the time and let it run for a few minutes, then check DOS time again. If it's way off, you do have mobo clock issues.

Ah it is all fixed now 😀
As it turns out winodws 95 is not well suited for that machine and for windows 3.11 FW it was that problem:
Windows 3.1 OK? 3.11 locking up? Here is the fix!
That seemed to do the trick
Also Audio working flawless with the Aztech Sound Galaxy NX Pro.