thanks Max.
i have another board with similar issue, so i may take you on the offer, but first i will try these,in order:
- bus mouse - if that works - chapter closed
- regardless of the above i will follow on the suggestions with connecting pins and echo i know it will probably not work, because the Term9x already is silent when moving mouse, but, still.

Do you have something assembled ?

---
weedee

i tried already all com, address, irq combinations.

not sure what you mean with the /M option comment. can you clarify please ?

Thanks for the picture. It matches the RetroWeb page for an Abit AB-AK3 / AB-AK4. https://theretroweb.com/motherboards/s/abit-ab-ak3#downloads
The MTL picture and jumpers match perfectly. The board silk screen matches very well too...

pshipkov wrote on 2023-06-18, 17:11:

- regardless of the above i will follow on the suggestions with connecting pins and echo i know it will probably not work, because the Term9x already is silent when moving mouse, but, still.

If you can wait a few days I can write a program that will echo anything sent via keyboard to screen via the serial port, pooling only (no interrupts). Since the voltatages look OK it might indeed be ISA interrupt issue. Perhaps that can be tested with SB card, anything that streams digitial audio would most likely hang or loop the sound. It might work if only IRQ3 and 4 are affected but worth trying. A few games and a MOD-player or two for testing.

Deunan wrote on 2023-06-18, 19:55:

pshipkov wrote on 2023-06-18, 17:11:

- regardless of the above i will follow on the suggestions with connecting pins and echo i know it will probably not work, because the Term9x already is silent when moving mouse, but, still.

If you can wait a few days I can write a program that will echo anything sent via keyboard to screen via the serial port, pooling only (no interrupts). Since the voltatages look OK it might indeed be ISA interrupt issue. Perhaps that can be tested with SB card, anything that streams digitial audio would most likely hang or loop the sound. It might work if only IRQ3 and 4 are affected but worth trying. A few games and a MOD-player or two for testing.

considering he says having tried all combinations, that would imply 3, 4 & 5 not working. Hoping he gave the correct parameters to ctmouse and didn't hope on some autodetection.

pshipkov wrote on 2023-06-18, 17:11:

weedee

i tried already all com, address, irq combinations.

not sure what you mean with the /M option comment. can you clarify please ?

The following.

pshipkov wrote on 2023-06-17, 22:55:

Noticed something about the CTMOUSE.EXE driver.
In the working system CTMOUSE /M reports "Installed at COM1 (03F8h/IRQ4) in Microsoft mouse mode". (or something like that)
But in the problematic mobo it reports "Installed at COM1 (03F8h/IRQ4) in Mouse systems mode".

FYI if you change the irq different from default, you will have to use ctmouse /S13 or /S15 when the card is set to COM1 and IRQ3 or IRQ5 respectively. I hope you did that.

yes, of course.
tried with the /S## flag all possible coonfigurations that match the controller card jumpers.

deunan, if you have an app like that i will take it.

Well I guess, if you have a jumper wire with some dupont connectors you could try the serial port at IRQ7 after figuring out which exact pins to use on the jumper block.
Should be fairly straightforward to figure out with a multimeter, or even just on sight.

Which other devices, hdd, floppy,... ? and cards, video (vga cga ega,mda), sound, ... do you have attached to the mainboard whilst trying to get the mouse to work ?
i find it a really weird problem you're having. it's either interrupts not getting through, but I guess floppy & hard drive work so that implies IRQ6 is working, and likely 14 & 15, which go via IRQ2. So to have 3 4 &5 not working, just odd.
or... ? I'd suspect ctmouse due to myself not having such great experiences with it but then again you're not getting any terminal feedback from the mouse so?
I dunno anymore.

FYI, the source for a polling read program from serial is here https://en.wikibooks.org/wiki/Serial_Programm … olling_the_UART
I think it's pascal or turbopascal .

yeah, it is such a weird problem.
everything else checks out, with video, sound and ide.
thats why reached to you guys for ideas.
all feedback is well appreciated.
there are some clear steps i will take to narrow it down further.
th bus mouse option sounds like the low-friction entry point tho.

never put much attention to inpit devices and their signal and programming models.
catching up on that now.
thanks for the link.
will give it a try.

You probably already answered but have you tried it with out the sound card and bare config.sys/autoexec.bat ?
What video card are you using ?

yes of course i tried the sound.
but testing is with vga and ide controllers only.

tried trident 8900cl, stb nitro (cl 5434), et4000ax.

Ok thanks. That leaves just a motherboard issue. Wondering if a BIOS upgrade or re-flash might help. DId notice your eprom has no cover on the window, it is possible certain parts of the code have become corrupt and that is part of the problem. Or not...
With everything you have tried, that and bad chipset is about all that is left imho. There are a lot of boards that used the UMC 491F, in your picture I see some possible corrosion near RN5 (but those go to cache) and some trace oddness that does lead to ISA slots from what it looks like.
Can you take a better picture of the UMC chip ?

Horun wrote on 2023-06-19, 02:32:

Ok thanks. That leaves just a motherboard issue. Wondering if a BIOS upgrade or re-flash might help. DId notice your eprom has no cover on the window, it is possible certain parts of the code have become corrupt and that is part of the problem. Or not...

I don't think that's the case, as the BIOS won't even POST if it is corrupt. One of the first actions of the startup process is verifying the BIOS ROM checksum, if that fails then you get no POST at all in most cases.

I would say an issue with the chipset is much more likely. Maybe running the "System board test" from CheckIt can point to any possible issues with the interrupt controller or some other thing.

The EPROM was freshly burned.
Checked these pins - they are all good.
Just some crusty bits, not that bad really - looks worse on the pic than actually is.

What's the voltage on ISA pin B25, B24, B23, B22, B21 ? (IRQ 3,4,5,6,7)
https://upload.wikimedia.org/wikipedia/common … SA_Bus_pins.svg

Forgot to respond to the CheckIt3 test status request.
Everything comes clean.
But the serial ports test looks like this.
Result is the same when run the test on computers where the serial ports work (with the same controller), so not sure how indicative this test is.

The voltage on these pins is ~4.8.
This is what the PSU passes down to the mobo.
In case you forgot me response above - i test with ATX+AT adapter and AT power supplies.
The AT power supplies provides ~5V, the ATX+AT provide 4.8V.
In both cases the outcome in terms of serial ports malfunction is the same.

pshipkov wrote on 2023-06-19, 20:28:

The voltage on these pins is ~4.8. This is what the PSU passes down to the mobo. In case you forgot me response above - i test w […]
Show full quote

The voltage on these pins is ~4.8.
This is what the PSU passes down to the mobo.
In case you forgot me response above - i test with ATX+AT adapter and AT power supplies.
The AT power supplies provides ~5V, the ATX+AT provide 4.8V.
In both cases the outcome in terms of serial ports malfunction is the same.

Those are the IRQ pins, so yeah they should all be the same.
But i'm not asking in regard to the voltage from the psu. I'm asking from a signal point of view.
It could've been one of them was stuck low or high and clearly they're not.

Now, I think, If you measure B22 while you're copying a file from/to the floppy drive, the voltage reading should be different, indicating that the IRQ is being triggered, since ISA interrupts are edge triggered.
It's more visual if you have an oscilloscope, but a voltage measurement should give you an average that's lower than the 4.8v while the system isn't doing anything.

edit: WRT the checkit test. Do you have a loopback plug attached ?

pshipkov wrote on 2023-06-19, 20:27:

Forgot to respond to the CheckIt3 test status request.

I think this test requires some loopback dongle to be connected at the serial port. I don't see how things like line control could be tested otherwise.
Did you try to connect pins 2 and 3, and running terminal program or this kind of test? BTW do pay attention to the pin numbering, no damage will occur from short-term shorts on any of the serial port pins but it's better to avoid them. I'm mentioning this because the voltage measurements you did only make sense if you tested the pins in this order: 5-4-3-2-1 and 9-8-7-6. If you are sure you went from pin 1 up then perhaps you are using some incompatible serial port socket+cable combo.

I provided a link to the loopback specs earlier in the thread.

CheckIt can do a more limited test of the COM/LPT ports without the loopback but you have to respond 'N' when it asks if you have the loopback connected before running the test.