My guess is that your mouse is the issue. When the P2B-F came out there were no Optical mice (they were first released a few years later), it could be drawing to much current if it is a newer Optical mouse and slowly ruining the port.
Have you tried a basic "non-Optical" PS/2 mouse on the P2B-F ? I see basic PS/2 mice for cheap at all the thrift stores ($4 or less) every day...
What brand and model is you mouse ?

Search the forum for "asus c151" and you'll find plenty of threads. It's a capacitor array that sometimes go short-circuit thus killing the PS2 ports. Apparently the solution is to remove it.
P3B-F PS/2 ports stopped working

quicknick wrote on 2021-05-18, 10:07:

Search the forum for "asus c151" and you'll find plenty of threads. It's a capacitor array that sometimes go short-circuit thus killing the PS2 ports. Apparently the solution is to remove it.
P3B-F PS/2 ports stopped working

I saw that, but I'm not sure its the same issue. I got power to the mouse, and my USB keyboard work fine.
One thing I noticed is when plugging the keyboard to the PS/2 port I could hear clicks from the pc speaker, as if the keyboard is pressed and the buffer fills (as stated in that topic).

Is there any potential drawbacks by removing this chip?

PEBKAC error, PS/2 ain't supposed to be hot plugged. Power off before changing peripherals.

CesarDRK wrote on 2021-05-18, 18:28:

I saw that, but I'm not sure its the same issue. I got power to the mouse, and my USB keyboard work fine.

I think someone stated that the capacitor is shorting the +5v to ground, but IMO it's more likely that either Clock or Data are affected (shorted or pulled to GND or +5V, or even between themselves).
Of course, there's no guarantee that your problem is caused by C151, and if you decide to remove it please don't use anything other than proper (de)soldering tools.

BitWrangler wrote on 2021-05-18, 20:55:

PEBKAC error, PS/2 ain't supposed to be hot plugged. Power off before changing peripherals.

It was NEVER hot plugged, i dont know where you get that idea from.

quicknick wrote on 2021-05-18, 23:12:

I think someone stated that the capacitor is shorting the +5v to ground, but IMO it's more likely that either Clock or Data are affected (shorted or pulled to GND or +5V, or even between themselves).
Of course, there's no guarantee that your problem is caused by C151, and if you decide to remove it please don't use anything other than proper (de)soldering tools.

I´ll check for a short.

Seemed like the board was live if the speaker was clicking "when you plugged it in" .... either that or you've got a massive static problem.

BitWrangler wrote on 2021-05-18, 23:58:

Seemed like the board was live if the speaker was clicking "when you plugged it in" .... either that or you've got a massive static problem.

No, i meant while the POST was running (with the keyboard plugged in before booting the machine) if i keep pressing random keys the pc speaker would emit a beep like when the buffer is full.

Oh I see. How old is the keyboard? Some of the older "pre windows key" keyboards do not support typematic rate setting, which may need to be disabled in CMOS setup. Or they misbehave totally on faster systems, post 96ish or in the ballpark. I don't know if they expected a breather now and again and newer stuff polls them too fast or what. You occasionally get the odd "early windows key" board or design of board, start messing up on Ghz plus systems as well, but they seem to work fine on era appropriate boxes.

However, if it's making stuck key noises, you might want to investigate for a stuck key, as unless ignore keyboard errors is on a lot of BIOSes will fault the keyboard and turn it off. Putting it over you knee and giving it a good spank is often effective.

quicknick wrote on 2021-05-18, 23:12:

I think someone stated that the capacitor is shorting the +5v to ground, but IMO it's more likely that either Clock or Data are affected (shorted or pulled to GND or +5V, or even between themselves).
Of course, there's no guarantee that your problem is caused by C151, and if you decide to remove it please don't use anything other than proper (de)soldering tools.

Think you are right !
On the P2B-F there are five separate surface mount caps right behind the PS/2 port. One of them could be shorting. The silkscreen near them says C9-C13, the four very tiny ones (c10-c13) are the same as the 4-section cap marked C151 used on later P3B-F, Cubx, Cusl, etc boards that is known to short. The larger cap C9 most likely not be the issue but never know. Because the P2B and P2B-F use separate caps instead of a "quad pack" the odds of a failure is more rare but not unheard of.
Because they are very tiny you need a very good needle soldering iron to remove them, I would do one then test then another, etc. You might get lucky and find only one or two need removed OR after removing all four the PS/2 still acts up.

BitWrangler wrote on 2021-05-19, 01:14:

However, if it's making stuck key noises, you might want to investigate for a stuck key, as unless ignore keyboard errors is on a lot of BIOSes will fault the keyboard and turn it off. Putting it over you knee and giving it a good spank is often effective.

Yes ! But he also said the PS/2 mouse is not working proper. Everyone should have at least one or more spare PS/2 kb and mouse but sounds like OP does not (they go for $3 at the thrift stores 🤣)

The keyboard is not very old, it does have the windows key.
I´ve checked for shorts on both the PS/2 port pins and couldn´t find nothing, also couldn´t find the "C151" chip like the P3B-F.

Anyway.. I had a breakthrough at least, the MOUSE started working again all of the sudden (the keyboard stayed dead, so i went with a USB keyboard).
I used it for like 60 minutes, both in DOS, Windows 98 and Windows XP. Then, i reboot one last time to DOS (from Win98) and it stopped working again, reboot back to Win98 and still dead.

So, at least the PS/2 mouse it´s not completely DEAD as i tought, but i´m not sure what the problem is.

Here is a picture near the PS/2 ports of the motherboard:

This may be really inconvenient, but it might be worth removing the board and examining it closely to see if the PS/2 connector is still well soldered on every pin. Also try to assess the connector itself. These things can sometimes get broken pins or cracked solder joints due to their life of being flexed around.

To go even further, you could use a meter to check continuity through the connector. Stick a pin into one of the 6 contacts, and check for continuity to those caps on the motherboard. Wiggle the connector a little and make sure it's not intermittent. Do the same for all 6 contacts.
Instead of sticking pins into the connector, you could also open up your keyboard and measure continuity from there.

I don't know if intermittent contact would cause phantom keypresses though.

Unrelated side note: Your board has a Coppermine capable voltage regulator (visible in your photo) so it will go down to 1.30V Vcore. Just in case you don't already know.

Following on from Horun's post, FWIW, my guess is that +5 for the PS/2 ports (from +5VSB?) comes in via that green Fuse1 (1.1A PTC fuse I think), through L7, its noise is decoupled through C9, then out to the PS/2 devices, then the ground comes back via the inductor just under where the silkscreen says Fuse1. Data and clocks then come in from the ports, are decoupled through C10-13 and also go through L8-11 before heading off across the motherboard.

If you've got a meter then it might be worth checking that the resistance across the inductors and fuse is 0, and the resistance across the caps is high. Also with the board on and mouse and keyboard plugged in, check the voltage on L7 relative to ground (if it's reads as 0V, check the other big inductor beneath C13). Given that both mouse and keyboard are affected, it's possible that there's a high resistance point on the +5 supply near the port. Maybe the mouse is more tolerant of a low voltage / draws less current so is borderline working, and the keyboard just can't get enough power. I think PS/2 keyboards are supposed to send a signal to the host after they've powered up, so perhaps it gets part way through that, fails, resets and tries again, and that's what causes the 'stuck key' effect?

shamino wrote on 2021-05-19, 04:56:

This may be really inconvenient, but it might be worth removing the board and examining it closely to see if the PS/2 connector […]
Show full quote

This may be really inconvenient, but it might be worth removing the board and examining it closely to see if the PS/2 connector is still well soldered on every pin. Also try to assess the connector itself. These things can sometimes get broken pins or cracked solder joints due to their life of being flexed around.

To go even further, you could use a meter to check continuity through the connector. Stick a pin into one of the 6 contacts, and check for continuity to those caps on the motherboard. Wiggle the connector a little and make sure it's not intermittent. Do the same for all 6 contacts.
Instead of sticking pins into the connector, you could also open up your keyboard and measure continuity from there.

I don't know if intermittent contact would cause phantom keypresses though.

Yeah, this will be a little time consuming. But i´ll see if i can do this on until the end of this week.

shamino wrote on 2021-05-19, 05:04:

Unrelated side note: Your board has a Coppermine capable voltage regulator (visible in your photo) so it will go down to 1.30V Vcore. Just in case you don't already know.

Humm. I´m already using a Coppermine P3 750 with a generic slocket adapter. I wish i could use a Tualatin.

snufkin wrote on 2021-05-19, 09:56:

PS2_Caps.jpg

Following on from Horun's post, FWIW, my guess is that +5 for the PS/2 ports (from +5VSB?) comes in via that green Fuse1 (1.1A PTC fuse I think), through L7, its noise is decoupled through C9, then out to the PS/2 devices, then the ground comes back via the inductor just under where the silkscreen says Fuse1. Data and clocks then come in from the ports, are decoupled through C10-13 and also go through L8-11 before heading off across the motherboard.

If you've got a meter then it might be worth checking that the resistance across the inductors and fuse is 0, and the resistance across the caps is high. Also with the board on and mouse and keyboard plugged in, check the voltage on L7 relative to ground (if it's reads as 0V, check the other big inductor beneath C13). Given that both mouse and keyboard are affected, it's possible that there's a high resistance point on the +5 supply near the port. Maybe the mouse is more tolerant of a low voltage / draws less current so is borderline working, and the keyboard just can't get enough power. I think PS/2 keyboards are supposed to send a signal to the host after they've powered up, so perhaps it gets part way through that, fails, resets and tries again, and that's what causes the 'stuck key' effect?

Wow! Thanks for the detailed tips. I´ll check this and also shamino tips when i have some time later this week.