I would check IRQ sharing conflicts, but I'm not 100% sure if that would affect only one port or both.

I had some issues with my CF to IDE adapter that fit your description, so unless you try with a real IDE HDD, I wouldn't rule out your SATA-IDE adapter as the culprit. For example, Win 2000 setup on this particular machine immediately BSOD'd with the CF adapter, while a regular HDD worked perfectly fine.

If the issues persist, it can be a problem with the southbridge. Broken BGA solder joints maybe? Another explanation can be bad capacitors, so the voltage is not clean enough. Check if you can see any bulging or leaking of capacitors.

Hallo,

if you chipset is failing, there is unfortunately not much you can do. Or to be more precise not much that worth it in a case of PIII board.
There is a chance your BGA chip is loosing connection with your board hence you are experiencing random problems. There are two ways of possible repair:
- reflowing the chip - doable at home with some tool - but high risk of destroying the board and even me with some experience has pretty bad sucess rate
- reballing the chip - more permanent but more expensive and demanding
BUT - keep this as your last resort and exclude or the other possible fault.
---
1) You said just one USB port is working - it means that on the back of you IO shiled there are two ports and one is working? Do you have power in the non-working connector?
Burned fuse or coil ot whatever is in the 5V USB power line is a pretty common mode of failure. I have just found this on a salavaged s370 PIII board - all good - just burned coli and no power to the external USB - mos likely fixable.
In my case the rer port works flowlessly.
Do the followings - check in you have 5V power on the not working port. You gonna see 4 pins. Power, data plus, data minus and ground. You can check the power and gnd and than on a SWITCHED OFF board
use your multimeter and check the resistance of the data pins do ground and inbetween of them. First of all the those values should be the same on all the USB ports. I would expect approx. 15kOohm from data pin to GND.
I have one very old board where the USB controller is doomed and it shows in mega ohms.
Your controller is most likely OK, otherwise none of your ports would work. But there is stilla chance that - yes - your chipset has a connection issue and for exmaple one data line to your faulty port is disconected.

2) Press your chipset with a funger and check if it has any influence on the behavior of your board - if so - than we are back at faulty BGA chip and in such a case a well performed reflow is the way to try.

3)What makes me nervous the most is the IDE problem. Do both of the channels exhibiting the same behavior? Have you excluded cabling problems, drive problems, jumpering problems? Use jus one known good IDE device and a known good 80wire ribbon cable and test meticulously if the fault is common, if oyu can reproduce it and again - if moving the board and some presure on the chipset makes some difference.
--
Good luck.

MoltenEQ wrote on 2023-11-15, 10:41:

Hi! […]
Show full quote

Hi!

A couple years ago I got my hands on a Pentium III PC which I wanted to set up for retro gaming.
I managed to get it working mostly, but there are many things not working properly that are really annoying, pibe of them being that only one of the USB ports is working properly.

I tried investigating the issue but I just don’t have any idea what I could do about it. I suspect that something is bad with the south bridge on the mobo because IDE drives act weirdly and not always show up on post, and Win 98 likes to freeze or hang up with some pci errors.

It’s really like to fix this motherboard, because it still has an ISA slot and there are loads of old SCSI cards that I want to test eventually and I also want to get my hands on some retro sound cards for DOS gaming.

TLDR:
Can someone here assist me with investigating the and possibly fix this broken USB port (and the rest of the issues)?

(Specs:
P3 slot1 550 cpu
Fujitsu/Siemens s26361-d1107 Mobo
640 mb ram
Mx400 gpu
Some older generic psu
Kingston sata ssd w/ ide sata adapter
Cd drive / dvd drive (I tried with both)
Sb220 sound blaster live 5.1
A new PCI usb 2.0 card
Some Austrian or German PC case)

If you need any more info I’ll try to look it up.

Absolutely. I would not relate the IDE and USB problem at first.
Me personaly I have a very bad experience with the SATA - IDE contraptioins. Basically it works someties somewhere...and causing issues on many boards. I have put the thing aside.
He needs one known good drive to check. I would even remove the optical drives for the test!
I have even found 2 boards already that can not be used with my CF card adaptor.
I would suggest to exclude all those contraptions from a testing setup!
--

technokater wrote on 2023-11-15, 10:49:

I would check IRQ sharing conflicts, but I'm not 100% sure if that would affect only one port or both.

I had some issues with my CF to IDE adapter that fit your description, so unless you try with a real IDE HDD, I wouldn't rule out your SATA-IDE adapter as the culprit. For example, Win 2000 setup on this particular machine immediately BSOD'd with the CF adapter, while a regular HDD worked perfectly fine.

If the issues persist, it can be a problem with the southbridge. Broken BGA solder joints maybe? Another explanation can be bad capacitors, so the voltage is not clean enough. Check if you can see any bulging or leaking of capacitors.

...and as another person here mentiond - weird random issues can be cuased ba bad capcitors. I have seen boards where such a failure cused IDE drive problems and even data coruption. It is not the most common failure - mostly the CPU VRM caps fail and the PC become unstable or dead, but the possiblity is there.
Do all the suggested investigations and measurements before you plunge in to reflowing chips etc. This is your absolute last resrot. I have stupidly reflowed 2 GPUs and later on found the fault in a bad or missing component...
..

MoltenEQ wrote on 2023-11-15, 10:41:

Hi! […]
Show full quote

Hi!

A couple years ago I got my hands on a Pentium III PC which I wanted to set up for retro gaming.
I managed to get it working mostly, but there are many things not working properly that are really annoying, pibe of them being that only one of the USB ports is working properly.

I tried investigating the issue but I just don’t have any idea what I could do about it. I suspect that something is bad with the south bridge on the mobo because IDE drives act weirdly and not always show up on post, and Win 98 likes to freeze or hang up with some pci errors.

It’s really like to fix this motherboard, because it still has an ISA slot and there are loads of old SCSI cards that I want to test eventually and I also want to get my hands on some retro sound cards for DOS gaming.

TLDR:
Can someone here assist me with investigating the and possibly fix this broken USB port (and the rest of the issues)?

(Specs:
P3 slot1 550 cpu
Fujitsu/Siemens s26361-d1107 Mobo
640 mb ram
Mx400 gpu
Some older generic psu
Kingston sata ssd w/ ide sata adapter
Cd drive / dvd drive (I tried with both)
Sb220 sound blaster live 5.1
A new PCI usb 2.0 card
Some Austrian or German PC case)

If you need any more info I’ll try to look it up.

Hi, thanks for the replies, let me try to answer your questions.

Bad capacitors:
I visually inspected the board and did not see any sign of bulking or leaking. I did but check them with a multimeter though. How can I do that?

IDE controller:
I tried many options, including multiple cables, disc drives and the behaviour was weird. Both channels. Sometimes the disc drive is not detected in post but will work, sometimes it is detected but will just refuse to read discs.
I forgot to mention that even my floppy drive exhibits the same behaviour. At this point I don’t think the issue is the adapter (it’s a Logilink 2 way SATA- IDE, with a switch to toggle).
I’ve heard that the CF standard is quite similar to IDE and adaptors are very basic and should work most of the time. But I did not go that route because I already had spare sata drives, and I need to get 2 adapters (one for my main pc and one for the retro) and that the cards are quite expensive.

USB port behaviour:
I took some measurements a while ago (and I even posted them in another post).
The second port I think provides 5V normally (on offline and unplugged values rise - so I’m guessing some caps are getting filled up) but the resistances on the data lines are very weird. I either mesure nothing, or all over the place.
The bottom port works fine, I usually put a usb keyboard mouse combo on it. The bios has legacy USB keyboard support, and it works most of the time (unless the OS wants to load its own drivers); I use it because it’s much simpler to borrow it from my laptop, since I don’t have a lot of space. Both of the data lines measured 15 kohm, so I think the pull down (or up?) resistors are fine. I can also provide pics if needed.

I was thinking that maybe I should just donate this mobo to CPUGalaxy and maybe try to get my hands on another slot1 mobo with ISA.

Did the original desktop from which this mainboard came from have a front usb port ? If not you can disregard this.
I'm asking because on one of my mainboards, that came out of a desktop with a front usb port, there were two jumpers which needed to be set to be able to use both usb ports on the back of the board.
in the original case with the front usb, one of the back usb ports was routed to the front and thus didn't work. the original back shield also shielded the non-functional usb port.

I didn’t see any front USBs on this machine, I had to DIY a 5.25 usb front cover with USB extensions myself. (Note: I’ve texted the USB ports without the extensions) There are not any jumpers on this board that I’m aware of, there’s only a do switch for setting the CPU frequency for Cody without fixed multipliers.
This board is really weird. It appears to be some pseudo workstation/server board with some uncommon features, mostly PSU monitoring.
I’ve been handed it from a driving school where it was used to do exams. It originally came with XP, 128 mbytes of ram and an IDE hard drive, that died almost immediately.

Speaking about the USB - if you are measuring weirdo resistance and it differs a lot from the working port I would say you have found the problem. You can try to trace the line from the port to the chipset. I would expect some resistors in the way. There could be some potential damage. Or as you mentioned, there is a chipset problem - maybe just a bad connectoin of the BGA for one of the ports?
I would be much bothered by the USB. Installing a USB 2:0 card is a good option.
--
If I remember you have mentioned that even your FDD had been acting weirdly? I am asking becouse FDD controller is usually in the IO chip, not in the chipset. Not sure about your board, but many later chipsets do not feature FDD controller and there is a super IO chip featuring this function as well. If this is the case than the problem might be something else and not the chipset itslef.
--

MoltenEQ wrote on 2023-11-15, 17:19:

Hi, thanks for the replies, let me try to answer your questions. […]
Show full quote

Hi, thanks for the replies, let me try to answer your questions.

Bad capacitors:
I visually inspected the board and did not see any sign of bulking or leaking. I did but check them with a multimeter though. How can I do that?

IDE controller:
I tried many options, including multiple cables, disc drives and the behaviour was weird. Both channels. Sometimes the disc drive is not detected in post but will work, sometimes it is detected but will just refuse to read discs.
I forgot to mention that even my floppy drive exhibits the same behaviour. At this point I don’t think the issue is the adapter (it’s a Logilink 2 way SATA- IDE, with a switch to toggle).
I’ve heard that the CF standard is quite similar to IDE and adaptors are very basic and should work most of the time. But I did not go that route because I already had spare sata drives, and I need to get 2 adapters (one for my main pc and one for the retro) and that the cards are quite expensive.

USB port behaviour:
I took some measurements a while ago (and I even posted them in another post).
The second port I think provides 5V normally (on offline and unplugged values rise - so I’m guessing some caps are getting filled up) but the resistances on the data lines are very weird. I either mesure nothing, or all over the place.
The bottom port works fine, I usually put a usb keyboard mouse combo on it. The bios has legacy USB keyboard support, and it works most of the time (unless the OS wants to load its own drivers); I use it because it’s much simpler to borrow it from my laptop, since I don’t have a lot of space. Both of the data lines measured 15 kohm, so I think the pull down (or up?) resistors are fine. I can also provide pics if needed.

I was thinking that maybe I should just donate this mobo to CPUGalaxy and maybe try to get my hands on another slot1 mobo with ISA.

Okay, I've found the pinouts of the south bridge. The PC is completely assembled now, so it will take a bit of time to disassemble and test it.
I will report back about my findings.

I've found the pin out of the south bridge chip (440bx 82430FX)

Sadly, the other side of the mobo features only a small amount of vias, but luckily there were some test points for the USB, and it seems that I have broken traces? At least that's what I thought.
I checked where they go and... crap? It seems that the other USB port doesn't work, because it is connected to what would be a place for the chipcard reader? Also, it appears that the test pads near are connected to the usb port, bit not to the south bridge...
I also noticed that i had reserve switch on my dips set to on, while the manual says that it should be off (SKP)... I have have flipped it to off. Also, I have a reserve switch, I wonder what that could do. Also, a bios recovery switch, that supports bios flashing with floppies , even if something messed up? I think this would be fancy for this time.

How should I move forward? Can I rewire the chipcard reader stuff? Could the bad dip switch cause issues?

Ahhh
I think I have figured it out! It looks like this motherboard expects a chipcard reader to be installed. I think all I have to do is bridge the gaps, and it will work... (chipset connects to the second col, but the usb port connects to the third).
Can I mess this up? How should I connect them?

I tested it with by bridging the two pins like jumpers, and it works!
I can't believe it! Why wouldn't this be mentioned in the mobo docs? Or was this supposed to be an included accessory or somethig?
Anyway, I'll need some time to reassemble the PC. Maybe I'll also do a clean install of Windows 98 SE. Any tips are welcome.
(I'll see if the IDE and Floppy issues are fixed or not)
(I have also included the chipcard reader's pin out, I don't know if I should do something more with it)

Good to hear about your progess!
You have found the solution before I managed to answer. Yes, you did it right - some older boards need this bridging - sometimes you are choosing if the back ports or front usb are connected - I have seen a sticker on a case - with a warning about it. Most likely from the original box.
You have some fancy schmenzy board for cool people with some additional features - most likely from some nice prebuild.
--
Ok - USB fixed. I am close to think your chioset is OK
--
The wrong DIP switch settings could have possibly caused anything including your random IDE failures. Hard to tell A it is an undocummented factory setrings - maybe for internal testing. Seen that many times - maybe there are another rev. of the board with different parts and the manufacturers just change jumpers and are ready to go (?)
It could possibly change some timing, bus speed, what ever.
--
The bios recovery switch is a cool feature - saved my ass once in an Intel Socket 5 board. Good to have it. You can usually make a blind boot from a FDD and automaticly flash the bios in case you have no image and the bios is doomed. 😂 There is some spare block controlling the fdd. Nice to have!
Good luck!
--

MoltenEQ wrote on 2023-11-17, 10:16:

I tested it with by bridging the two pins like jumpers, and it works! I can't believe it! Why wouldn't this be mentioned in the […]
Show full quote

I tested it with by bridging the two pins like jumpers, and it works!
I can't believe it! Why wouldn't this be mentioned in the mobo docs? Or was this supposed to be an included accessory or somethig?
Anyway, I'll need some time to reassemble the PC. Maybe I'll also do a clean install of Windows 98 SE. Any tips are welcome.
(I'll see if the IDE and Floppy issues are fixed or not)
(I have also included the chipcard reader's pin out, I don't know if I should do something more with it)

Okay, I managed to reassemble the PC, and got PLOP boot to run on a floppy, so now I can also boot from USB, which I'M SO HAPPY ABOUT.
That said, I'm still facing issues with the Floppy and the IDE controllers. My CD drive and SSD W/ IDE converter is still randomly detected or not, while the floppy drive is sometimes detected, sometimes not, and sometimes post complains about bad floppy. At this point I'm curious if my cables are bad, that being said I tried loads for both and the results are the same...

I think I should ditch the CD drive, as it's really old, and doesn't really like to read discs well. I have a USB disc drive, that seems to work most of the time in Win98 anyway...

Hallo,

considered you have tested more cable sets I would not blame the ribbon cables then.
It is super weird you have problems with you FDD as well. Please check your board layout, but I am pretty positive your FDD is not in the same chip as your IDE. Most likely you have super IO that includes FDD controller.
Have you tried some PCI IDE controller card?
-
Anyway have you tried using only HDD, no CD drive on the cable? Do you use 40 or 80wire cabels?
What is your connection setup? Primary IDE for HDD and CD-DRIVE on secondary?
Please check again your jumpers - some drives need a different jumpering for stand-alnone and master.
Andd sometimes CD-ROM needs change from cable select to master, or vice versa.
Anyway try your setup without a CD-ROM - I have met a few of setups where a specific board,HDD and an optical drive faield to work together. Ant the symptoms where pretty close to your expereience.
-
Please try a known good 80wire ribbon, mind the orientation - blue end to the motherboard. Connect just the harddrive. (if possible try a different drive than you have as well). Remove the FDD and CD-rom. And give it a try if your problem is still there. I have seen many sneeky troubles - especially when any kind of IDE SATA, CF-IDE etc. contraption is used. But as mentioned above - even normal HDD and a CD-ROM can make a mess!
Good luck .

MoltenEQ wrote on 2023-11-17, 15:47:

Okay, I managed to reassemble the PC, and got PLOP boot to run on a floppy, so now I can also boot from USB, which I'M SO HAPPY ABOUT.
That said, I'm still facing issues with the Floppy and the IDE controllers. My CD drive and SSD W/ IDE converter is still randomly detected or not, while the floppy drive is sometimes detected, sometimes not, and sometimes post complains about bad floppy. At this point I'm curious if my cables are bad, that being said I tried loads for both and the results are the same...

I think I should ditch the CD drive, as it's really old, and doesn't really like to read discs well. I have a USB disc drive, that seems to work most of the time in Win98 anyway...