VOGONS


How to fix your dead on-board floppy controller

Topic actions

First post, by feipoa

User metadata
Rank l33t++
Rank
l33t++

I have a Biostar MB-8433UUD v3.1 motherboard which I am trying to fix the on-board floppy and serial controllers. I first thought the issue was the SuperI/O circuit that went bad. It has part number UM8663BF. I replaced it, but I still cannot access my floppy disk. I then thought maybe some of the DIP IC's might be bad (74LS14Nx1, 74F244PCx2, 74LS245Nx2, 74F08Nx1, UM9515-01), so I desoldered all of them, put in DIP sockets, and replace them with new chips. Still cannot access the floppy disk. I then recapped all the electrolytic capacitors (9x10uF, 2x47uF, 4x100uF) - still no floppy disk access.

The serial ports also don't work, which is curious, because the serial port IC is a seperate chip, UM8667. I didn't bother to replace this because I think whatever is causing the floppy controller not to work is likely the same thing which is affecting the serial IC.

If I use an ISA I/O card with an onboard floppy controller (UM8663AF), floppy access works fine as long as I leave "onboard FDC" enabled in the BIOS and set A to 1.44 MB 3.25" drive. So I know the issue is not with the floppy drive or cable. I suspect the issue may be with a certain section of the Southbridge (UM8886BF) IC. I say a "certain section" because the PS/2 mouse port, which is connected to the Southbridge, still works. Alternately, there may be some bad SMD resistors or capacitors, or a bad SIP resistor pack (sometimes called a "resistor network").

Has anyone had an SMD capacitor or resistor go bad? If so, how did you narrow down the failure?

This motherboard used to function as a server many years back and I specifically recall the on-board floppy working. After an uptime of approx. 5 months, the computer crashed, and which point the floppy drive stopped working entirely. If I used an ISA I/O card, everything works fine, however I'd still like to fix the on-board features.

Plan your life wisely, you'll be dead before you know it.

Reply 1 of 90, by Imperious

User metadata
Rank Oldbie
Rank
Oldbie

I was going to post about checking voltages, but floppy and serial ports are all signals only and ground pins.
There are two types of motherboard pinouts for the serial, but I guess You have checked that out, as well as the obvious not putting
the floppy cable the wrong way round.
I was just looking at a good photo of that board, there are 2 diodes and some resistor packs in that area. Check for any through hole corrosion
too. Replaced bios chip with a new eprom?

Atari 2600, TI994a, Vic20, c64, ZX Spectrum 128, Amstrad CPC464, Atari 65XE, Commodore Plus/4, Amiga 500
PC's from XT 8088, 486, Pentium MMX, K6, Athlon, P3, P4, 775, to current Ryzen 5600x.

Reply 2 of 90, by Errius

User metadata
Rank l33t
Rank
l33t

This touches a nerve. I recently killed the floppy controller on a nice old mobo after pulling out and reseating the memory modules. There was no need to do this, I just wanted to see if a particular module worked (it did). But just doing this killed the floppy controller, probably due to flexing of the board. (The original modules probably hadn't been removed for 20 years and took some pulling to get out.) A DOS machine without a floppy drive is useless so I threw out the board and got a replacement. Fortunately it's a common model so this wasn't too hard/expensive. Lesson learned though: if an old computer is running OK leave it alone.

Is this too much voodoo?

Reply 3 of 90, by feipoa

User metadata
Rank l33t++
Rank
l33t++

I did replace the BIOS chip some time ago and reflashed it. This did not help.

I suppose there could also be some bad traces, diodes, or solder joints, but nothing that obviously pops out. Anyone seen dead solid state diodes, resistors, or SMD caps on a motherboard before? I haven't. Seems highly unlikely. But there must have been people fixing motherboards back in the day with dead floppy controllers. What did they do?

Plan your life wisely, you'll be dead before you know it.

Reply 4 of 90, by Errius

User metadata
Rank l33t
Rank
l33t

That reminds me, one of the things I tried with my board was reflashing the BIOS, which had strange results: the floppy drive now worked in Windows NT, but still didn't work at boot or in DOS.

Is this too much voodoo?

Reply 5 of 90, by Aideka

User metadata
Rank Member
Rank
Member

I have seen dead resistors on motherboards before, a few were not resisting the current anymore, and a few didn't let any current through them. I would atleast probe around with a multimeter and use its continuity test on those passive components. If all pass the continuity test, then I would start measuring the resistance of the resistors.

8zszli-6.png

Reply 6 of 90, by feipoa

User metadata
Rank l33t++
Rank
l33t++

How did you determine the resistors were at fault?

Unfortunately, many SMD caps and resistors do not have values on them, so even if I were to test them out of circuit, I would not know the proper values. I will probe around anyway though.

Plan your life wisely, you'll be dead before you know it.

Reply 7 of 90, by Aideka

User metadata
Rank Member
Rank
Member

In my case the troubleshooting was pretty easy, since the MB started working after replacing the resistors that blocked all current, those had somehow decided to act as a kind of a fuse, and blown themselves open from the inside, there was no visual damage in any of them, but I managed to measure them in circuit luckily.

8zszli-6.png

Reply 8 of 90, by feipoa

User metadata
Rank l33t++
Rank
l33t++

I finished probing the board with my multimeter. I checked all resistors, inductors, diodes, and most resistor packs. I checked a few SMD capacitors. The only glaring item was this resistor pack with a chip on the corner. This chip changed the resistance for the 10th pin only. Instead of 4.7 K-ohms, it measured 8.22 K-ohms. I merely replaced the resistor pack such that all pins are 4.7 K-ohms. This did not solve the problem.

Resistor_Pack.jpg
Filename
Resistor_Pack.jpg
File size
58.03 KiB
Views
4503 views
File license
Fair use/fair dealing exception

I am attaching photos of my SuperI/O soldering job. The pins which look questionable, I confirmed that they are hitting the trace by ensuring connectivity with the other end of the trace.

SuperIO.jpg
Filename
SuperIO.jpg
File size
168.63 KiB
Views
4503 views
File license
Fair use/fair dealing exception

I'm also attaching a high-resolution photo of the current state of the motherboard. Maybe someone sees something I am missing. The RTC was removed for the photo only. All the electrolytic caps have been replaced as shown. For the 100 uF caps, I used aluminum polymer caps due to the low profile, the others are electrolytic. Also shown is how all the DIP IC's are socketed.

MB-8433UUD.jpg
Filename
MB-8433UUD.jpg
File size
2.38 MiB
Views
4503 views
File license
Fair use/fair dealing exception

Pending any other ideas, I will need to replace the southbridge chipset, which has a lot tighter spacing compared to the SuperI/O. I am doubtful that I can pull it off, especially due to the close proximity to the ISA, PCI, and DIP slots. I have soldered about 10 pins of a chipset with tight spacing like this in the past. It was a nightmare, but eventually worked out.

Southbridge.jpg
Filename
Southbridge.jpg
File size
340.91 KiB
Views
4503 views
File license
Fair use/fair dealing exception

Plan your life wisely, you'll be dead before you know it.

Reply 9 of 90, by Imperious

User metadata
Rank Oldbie
Rank
Oldbie

I'll be stuffed if I can find any datasheets for these chipset and IO ic's anywhere. I wonder if the 2 You have replaced have got a clock signal? 5v?
These motherboards are multi-layered so there could be a blown track between the top and bottom layer and You'll never see it.
I would forget about replacing that large IC, too difficult. The one You have replaced might look a bit better if You used flux paste, it's
impossible to recreate that factory solder job by hand wihout it.

Atari 2600, TI994a, Vic20, c64, ZX Spectrum 128, Amstrad CPC464, Atari 65XE, Commodore Plus/4, Amiga 500
PC's from XT 8088, 486, Pentium MMX, K6, Athlon, P3, P4, 775, to current Ryzen 5600x.

Reply 10 of 90, by h-a-l-9000

User metadata
Rank DOSBox Author
Rank
DOSBox Author

Soldering each pin individually? There are better ways 😉

So you say the serial ports are also not working?
Can you try the following in DOS debug? Assuming your serial port is at COM1.
o3ff 55
i3ff [it should print 55]
o3ff aa
i3ff [it should print AA]

This will show 'how good' the chip is still connected.

1+1=10

Reply 11 of 90, by luckybob

User metadata
Rank l33t
Rank
l33t

They make QFP 208 soldering nozzles. That would make the process quick and easy, if you happened to have a hot-air gun and the 208qfp nozzle.

and I now can't help but post this: https://www.youtube.com/watch?v=H8yQhXDquII

It is a mistake to think you can solve any major problems just with potatoes.

Reply 12 of 90, by Synoptic

User metadata
Rank Member
Rank
Member

Exact same issue with my version 2. I am out of the country right now and will do same kind of tests you guys suggest. If we can fix either feipoa's or mine, I think it will help the Biostar owners all over the planet.

Reply 13 of 90, by keenerb

User metadata
Rank Oldbie
Rank
Oldbie

Did the floppy/serial controller ever work?

If not, are you certain you've got the correct BIOS installed?

*edit*

I see you remember the floppy DID work prior to a server crash of some sort; I wouldn't be surprised to find that one of the troubleshooting steps was to upgrade the BIOS, potentially to a bad/incorrect version.

I'd definitely look into that before unsoldering any more chips...

Reply 14 of 90, by yawetaG

User metadata
Rank Oldbie
Rank
Oldbie
Errius wrote:

This touches a nerve. I recently killed the floppy controller on a nice old mobo after pulling out and reseating the memory modules. There was no need to do this, I just wanted to see if a particular module worked (it did). But just doing this killed the floppy controller, probably due to flexing of the board. (The original modules probably hadn't been removed for 20 years and took some pulling to get out.) A DOS machine without a floppy drive is useless so I threw out the board and got a replacement.

Er...why not just install a multi-I/O card with floppy and/or IDE connectors instead of throwing out the motherboard?

Reply 15 of 90, by feipoa

User metadata
Rank l33t++
Rank
l33t++
h-a-l-9000 wrote:
Soldering each pin individually? There are better ways ;) […]
Show full quote

Soldering each pin individually? There are better ways 😉

So you say the serial ports are also not working?
Can you try the following in DOS debug? Assuming your serial port is at COM1.
o3ff 55
i3ff [it should print 55]
o3ff aa
i3ff [it should print AA]

This will show 'how good' the chip is still connected.

Sorry for the delay - I am having issues with the local bylaw officer and got derailed. Who calls the bylaw officer? Neighborus of course!

When in DOS 7.1, I type DEBUG, then carriage return. Each new line is now proceeded with a dash -
-o3ff55 returns ^Error
-i3ff returns FF
-o3ff aa returns nothing (but the cursor goes to the next line)
-i3ff returns FF

Does this mean the serial is not working at all? I ran these commands with the serial mouse plugged into COM1.

keenerb wrote:

I see you remember the floppy DID work prior to a server crash of some sort; I wouldn't be surprised to find that one of the troubleshooting steps was to upgrade the BIOS, potentially to a bad/incorrect version.

I'd definitely look into that before unsoldering any more chips...

The computer crashed after 5 months of uptime. Immediately afterwards I rebooted to find a floppy error. I did not flash the BIOS between those two steps.

yawetaG wrote:

Er...why not just install a multi-I/O card with floppy and/or IDE connectors instead of throwing out the motherboard?

These motherboards, and hardware of this age in general, have becoming collectable items. As collectable items, having a defective onboard floppy/serial/IDE port severely weakens the collective bond. It is best to have the board fully functional.

Plan your life wisely, you'll be dead before you know it.

Reply 16 of 90, by yawetaG

User metadata
Rank Oldbie
Rank
Oldbie
feipoa wrote:
yawetaG wrote:

Er...why not just install a multi-I/O card with floppy and/or IDE connectors instead of throwing out the motherboard?

These motherboards, and hardware of this age in general, have becoming collectable items. As collectable items, having a defective onboard floppy/serial/IDE port severely weakens the collective bond. It is best to have the board fully functional.

Personally, I'd think having a partially broken board working with a multi-I/O card would be preferable to "throwing out a board". The latter implies that the board will be recycled or destroyed, nullifying any collectable value anyway.

Some people may also argue that actually running the system will make the collectable value go down.

Reply 17 of 90, by Errius

User metadata
Rank l33t
Rank
l33t

I tried two ISA floppy cards, a no-name generic card and an Adaptec AHA-1542CP. I still could not get the floppy to work. This also wasted an ISA slot (there were only 3) so even if it had worked it wasn't an ideal solution.

Is this too much voodoo?

Reply 18 of 90, by h-a-l-9000

User metadata
Rank DOSBox Author
Rank
DOSBox Author
feipoa wrote:
When in DOS 7.1, I type DEBUG, then carriage return. Each new line is now proceeded with a dash - -o3ff55 returns ^Er […]
Show full quote

When in DOS 7.1, I type DEBUG, then carriage return. Each new line is now proceeded with a dash -
-o3ff55 returns ^Error
-i3ff returns FF
-o3ff aa returns nothing (but the cursor goes to the next line)
-i3ff returns FF

Does this mean the serial is not working at all? I ran these commands with the serial mouse plugged into COM1.

The first command should have a space between o3ff and 55, then there will be no error. But as it looks the serial port is not present in the I/O address space. FF is the value you get when reading an unused port. Normally the serial port scratchpad would be read/written.

Unfortunately there is not much information gained - it appears the BIOS is unable to initialize the chip.

1+1=10

Reply 19 of 90, by feipoa

User metadata
Rank l33t++
Rank
l33t++

Its a long shot, but I will try a UV Flash which has the board's original BIOS on it. Other than that, I will try to replace the Southbridge when I'm feeling at least half confident.

Errius, you need to use an ISA I/O card which has the same UM8663 super I/O chip. Then ensure that you still have the BIOS set to floppy A:\ 3.25" and you must leave the motherboard's onboard FDC enabled. Also, you will want to first disable all other features on the ISA I/O card to start with. This is what worked for me anyway.

It would be really nice to resolve this issue, or at least determine what is causing the fault. I'm sure most collectors here have at least one motherboard with a dead floppy or serial controller.

Plan your life wisely, you'll be dead before you know it.