First post has no model number of the motherboard (if integrated) or picture of IO card 🙁
might be just blown 1488 1489 line drivers/receivers https://www.onsemi.com/pdf/datasheet/mc1488-d.pdf https://www.onsemi.com/pdf/datasheet/mc1489-d.pdf

Hm. I haven't thought about that. So the V.24 itself might be okay ?

PS: How about connecting some LEDs to the output pins?
In combination with the VB5 program for Win95 I mentioned earlier, the individual pins can be toggled easily.

rasz_pl wrote on 2023-06-25, 00:56:

First post has no model number of the motherboard (if integrated) or picture of IO card 🙁
might be just blown 1488 1489 line drivers/receivers https://www.onsemi.com/pdf/datasheet/mc1488-d.pdf https://www.onsemi.com/pdf/datasheet/mc1489-d.pdf

Later post does have a photo of the cards
and blown 1488 is out of the question since card works in another computer and also not the issue since it's an integrated one chip does most solution kinda card.
Which would make one think it's likely a voltage issue, yet voltages are apparently all there.
+12,+5;-5,-12, according to OP

edit : link to post with photo of card Re: serial mouse not detected on a 386 motherboard

easiest way to try is to do a loopback skipping 1488/1489
12V might be there, but extremely noisy for example
best would be oscilloscope on the output

.. And an ISA POST card with individual LEDs for +-12v and +-5v.
Just because the PSU has these voltages doesn't mean the ISA slot has, as well.
Maybe there's a battery damage on the motherboard (trace eaten by acid), making an open circuit.

Remember that we went over this several times already. Voltages on all isa and com pins are as expected.
Razs_pl where you want me to check signal exactly ?

do you have an oscilloscope?
all voltage rails for ripple
serial TX pin while typing something into serial terminal, ideally with scope serial decoder enabled

pshipkov wrote on 2023-06-25, 06:05:

Remember that we went over this several times already. Voltages on all isa and com pins are as expected.
Razs_pl where you want me to check signal exactly ?

Remember when I told you to add a summary of the things tried to every reply you made from now on ? 😁
Or you could just add the summary as an edit to the first post.

pshipkov wrote on 2023-06-24, 16:39:

Btw, I tested this already with terminal. It does not echo.

I noticed that, you tested the output with a scope as well so we know the transmitter works (at least its output is not completly dead and that's the usual issue).
But pretty much all the software will attempt to use interrupts, which we suspect are not working right. That's why I want to try pooling-only approach, and with a twist as well in case the problem is reading the Interrupt Identification register.

EDIT: Here's another test, you will need to make a screenshot of the results or else there will be a lot of writing. For the last test to pass you also need to have pins 2 and 3 on the serial port wired together. Let's see what that does.

weedeewee wrote on 2023-06-25, 09:46:

Remember when I told you to add a summary of the things tried ....to the first post.

would be ideal 😀
where is the post with a scope, this?
Re: serial mouse not detected on a 386 motherboard
posting.php?mode=quote&f=46&p=1173298
>However Serial pin 2 is constant when outputting from terminal.
>Pin 3 is silent too.
>To confirm to your first question - the output is constant.

doesnt look like working output signals at all. Serial mouse uses some other pins for supply https://www.eltima.com/article/9-pin-serial-port.html Those should be settable (either with debug or loading ctmouse) even when interrupts dont work

additionally

weedeewee wrote on 2023-06-20, 16:49:

All voltages are present and accounted for +12,+5,-5,-12

I wouldnt claim that without a picture of those voltage rails on a scope, ideally while typing in serial terminal

Yeah not sure what the issue with not posting pictures is, every cell phone since about 2012 can do it. Would think a member since 2018 would know pictures speak a thousand words ;p

rasz_pl wrote on 2023-06-25, 22:32:

would be ideal :) where is the post with a scope, this? Re: serial mouse not detected on a 386 motherboard posting.php?mode=quo […]
Show full quote

weedeewee wrote on 2023-06-25, 09:46:

Remember when I told you to add a summary of the things tried ....to the first post.

would be ideal 😀
where is the post with a scope, this?
Re: serial mouse not detected on a 386 motherboard
posting.php?mode=quote&f=46&p=1173298
>However Serial pin 2 is constant when outputting from terminal.
>Pin 3 is silent too.
>To confirm to your first question - the output is constant.

doesnt look like working output signals at all. Serial mouse uses some other pins for supply https://www.eltima.com/article/9-pin-serial-port.html Those should be settable (either with debug or loading ctmouse) even when interrupts dont work

additionally

weedeewee wrote on 2023-06-20, 16:49:

All voltages are present and accounted for +12,+5,-5,-12

I wouldnt claim that without a picture of those voltage rails on a scope, ideally while typing in serial terminal

Fairly certain the wrong pins were measured since the next post with the loopback plug also mentions a changing output signal on the oscilloscope
See: Re: serial mouse not detected on a 386 motherboard

wrt the voltage rails... that's what op claims, several times.
I guess for best certainty, voltages should be measured right on pin 1 & 14 of the SN75188 chip

yay wrong pins measured 😀

>I see the signal changing [EDIT]in the oscilloscope[/EDIT] when typing something.

so now we need to see that signal, and as you say voltage on pins 14(1) of MC1488, ideally both at the same time while typing in terminal

rasz_pl wrote on 2023-06-26, 06:49:

yay wrong pins measured 😀

>I see the signal changing [EDIT]in the oscilloscope[/EDIT] when typing something.

so now we need to see that signal, and as you say voltage on pins 14(1) of MC1488, ideally both at the same time while typing in terminal

the signal on pin 14 and 1 of the MC1488, fyi it's an SN75188 on OP's cards, shouldn't change much aside from any possible interference or fluctuations.

I can't recall if it was mentioned that another psu has been tried ( see that's why a summary would've been so great ! ). I'm gonna assume he already did that.
and to re-itterate the card works in another computer, so that leaves the mainboard.

pshipkov, did you find and try a serial card or multi io card with more chips? ie not one component that does most.

mandatory header:
- all ISA pin voltages confirmed
- all COM port voltages confirmed
- all components work fine in other motherboards
- terminal outputting to the port does not echo anything back
- oscilloscope shows changing signal on B23 pin during data I/O between FDD/HDD
- tested with different PSUs
- tried different multi I/O controllers, including a dedicated COM/LPT port one with low-level of integration

---

Oh, wow, this thread swell further.
Replying for the sake of adding a mandatory header for clarity.

How about setting up another computer connected over null modem? If the problem is in the motherboard logic and not voltage levels you should be able to TX from this computer and receive on another one no problem. This will leave interrupts after filling FIFO as the only possible culprit.

Already allocated much more time on this than i anticipated.
It is a curious problem, so i will probably get back to it at some point not far from now.
If you have other ideas that don't require more elaborate setup - i will be happy to try them.
Thanks rasz_pl.

This means this motherboard is the issue with serial function in particular.

Could be in chipset itself or bios firmware not good. Only way is replacement after confirming the bios firmware is good.

I use diode reading using black probe on the pins and red probe on ground to make sure the transistor junctions are good but I really think the chipset itself is bad.

Cheers,

pshipkov wrote on 2023-06-26, 22:30:

If you have other ideas that don't require more elaborate setup - i will be happy to try them.

I've updated my previous post with another test program, please try it.

Things I can't see confirmed... elimination of memory managers (EMM386 and QEMM in particular can screw with things in some configs) cleanest of clean boots on a retail MSDOS version (No OEM compatibility hacks screwing things up) ... just try things out of lower 640k first, if it works, then we deal with corruption or errors in HMA.. BIOS error might not be BIOS error if the shadow is bad.