VOGONS


First post, by philmac

User metadata
Rank Newbie
Rank
Newbie
IMG_3495.JPG
Filename
IMG_3495.JPG
File size
1.82 MiB
Views
357 views
File license
CC-BY-4.0

I'm after some help with attempting to repair the above board. Normally I'd just replace boards like this, but trying to track down a replacement on eBay was difficult (I only found one viable replacement and I'm now using it). This board isn't a standard ITX so if the replacement fails, it's curtains for the PC using it so I'd like to try and get this fixed.

Symptoms are:

1) On powerup, the power LED, HD LED stay lit (full) and the fan spins up to full power.
2) There's no graphics output, either when using the built-in GPU or a seperate AGP card.
3) No beeps from the PC speaker.
4) The machine only powers off if you hold the power button in for 4+ secs.

Removing all other components (storage devices, PCI card etc) has no effect.

The memory, CPU and PSU are all known working components (I've tried spares without success, I've replaced the board with the spare and the machine happily powers up). I've inspected the board and to the best of my knowledge, I can't see any failed components:

IMG_3497.JPG
Filename
IMG_3497.JPG
File size
1.26 MiB
Views
357 views
File license
CC-BY-4.0
IMG_3496.JPG
Filename
IMG_3496.JPG
File size
1.69 MiB
Views
357 views
File license
CC-BY-4.0
IMG_3500.JPG
Filename
IMG_3500.JPG
File size
1.53 MiB
Views
357 views
File license
CC-BY-4.0
IMG_3498.JPG
Filename
IMG_3498.JPG
File size
1.53 MiB
Views
357 views
File license
CC-BY-4.0

I've got some basic electronic skills and repair tools (I own a multimeter and a soldering iron, I've repaired/replaced basic audio connectors before, but nothing as complex as this). Any recommendations on tools I should get, or things I should check with the board itself?

Are there any repair services in the UK who could do this kind of work, or am I best tackling it myself?

Any guidance or advice appreciated.

Reply 1 of 22, by zyga64

User metadata
Rank Member
Rank
Member

Check CR2032 battery. I had similiar symptoms on socket 478 board once, because battery was dead and caused short circutit.

1) VLSI SCAMP /286@20 /4MB /TVGA9000C /CMI8330
2) i420EX /486DX33 /16MB /Trio64V+ /AZT2316
3) i430HX /P233MMX /64MB /VirgeDX+3Dfx /Vibra16s
4) i440BX /P II 400 /256MB /FX5500/AWE64
5) i865G /E5800 /2GB /Ti4200 /YMF724

Reply 2 of 22, by philmac

User metadata
Rank Newbie
Rank
Newbie
zyga64 wrote on 2020-05-28, 06:23:

Check CR2032 battery. I had similiar symptoms on socket 478 board once, because battery was dead and caused short circutit.

Great minds think alike - forgot to add that I'd replaced this 😀

Reply 3 of 22, by Doornkaat

User metadata
Rank Oldbie
Rank
Oldbie

Get the EEPROM out of its socket and try to reflash it, then put it back.
Check for shorts.
Check VRMs for correct voltages.
Check for correct voltages (and ripple) on the CPU socket.
Check signal lines.

Having the board professionally repaired is probably too expensive to be worthwile.

Reply 5 of 22, by philmac

User metadata
Rank Newbie
Rank
Newbie
Doornkaat wrote on 2020-05-28, 07:27:
Get the EEPROM out of its socket and try to reflash it, then put it back. Check for shorts. Check VRMs for correct voltages. Che […]
Show full quote

Get the EEPROM out of its socket and try to reflash it, then put it back.
Check for shorts.
Check VRMs for correct voltages.
Check for correct voltages (and ripple) on the CPU socket.
Check signal lines.

Having the board professionally repaired is probably too expensive to be worthwile.

Any particular EEPROM tool/reader you'd recommend? I've never tried removing ROM chips before, usually I leave well alone if it's in working order.

PC Hoarder Patrol wrote on 2020-05-28, 07:52:

I'd be thinking that every one of those OST caps is a potential source of (open) failure without visible sign - are you up for the time & cost of a recap?

The time and cost doesn't bother me at this point, given the relative scarcity of viable replacement hardware. I don't mind whether I fix it or someone else does, as long as the board doesn't get trashed in the process.

Are the OST's known for failing without bulging? I've only ever seen caps that are visibly bulging/bursting on failed boards.

Reply 7 of 22, by philmac

User metadata
Rank Newbie
Rank
Newbie
PC Hoarder Patrol wrote on 2020-05-28, 08:07:

Sadly yes - I'll quote @PCBONEZ from another thread here "OST caps are famous for failing without bloating or leaking."

Thanks, I'll bear that in mind. Any recommendations for replacements (if they're faulty)? Reliability rather than cost is the consideration, so I'm happy to get the best I can if it extends the life of the board by another 15+ years.

I've never recapped a board before, but I've got an MSI 6119 I'm going to practice on with known bad caps (and a bag of Panasonic replacements to go on it), so hopefully I'll get the hang of it...

Reply 8 of 22, by PC Hoarder Patrol

User metadata
Rank Oldbie
Rank
Oldbie

Brand-wise you'd be looking at Rubycon, Nichicon, Panasonic or UCC (United Chemi-Con). As to specific brand series, maybe someone more experienced with actual recapping might chime in on that.

Reply 9 of 22, by Doornkaat

User metadata
Rank Oldbie
Rank
Oldbie
philmac wrote on 2020-05-28, 07:59:

Any particular EEPROM tool/reader you'd recommend? I've never tried removing ROM chips before, usually I leave well alone if it's in working order.

I have never used a dedicated tool. I always perform a hot flash. Here's my personal approach step by step:
-Prepare a floppy with Uniflash 1.40 and the BIOS image I want to flash.
-Find a board with the same EEPROM type+socket the EEPROM to be flashed uses.
-Set that board to boot from floppy and to shadow its BIOS into RAM.
-Take the EEPROM to be flashed out of its socket using two screwdrivers.
-Loosen the EEPROM in the board I'm using to flash so it still makes contact but I can easily take it out.
-Set up the board I'm using to flash with only PSU, FDD and keyboard.
-Start the board, and boot into DOS using the Win98 boot Floppy.
-Extract the board's EEPROM (using a non conductive pin/toothpick if I can't get a good grip).
-Insert the EEPROM I'm going to flash.
-Eject the Win98 boot floppy and replace with UniFlash+BIOS floppy.
-Start Uniflash 1.40 from floppy and use it to flash the new BIOS image.
-Power down the board and swap EEPROM back to original.

Concerning caps: Old capacitors are never not suspect and inside small cases like the XPCs they tend to age faster because they get warmer.
Still simply replacing capacitors without trying to diagnose the problem first does not appear very expedient to me. If you have any way of diagnosing the circuitry I recommend doing that first. I don't want to rule out capacitors as the cause of your problem but it is at least probable you put in all that work and get nothing in return because the culprit is actually a failed transistor.
Consider replacing the aged caps on the replacement board though - just for peace of mind. 😀

Reply 10 of 22, by Doornkaat

User metadata
Rank Oldbie
Rank
Oldbie

And considering replacements: Always try to match the characteristics of the capacitor you're replacing. Not just capacitance and voltage rating.
I see at least some capacitors on that board are OST RLX series. Look for any markings on the other caps that indicate a series and google for datasheets.
You usually want to match capacitance and dissipation factor+ESR+impedance as closely as possible.
You usually want to match or exceed voltage, temperature and ripple current rating as well as all tolerances.
That being said VRM curcuitry on motherboards are usually rather tolerant towards changes in capacitors and your board will probably run flawlessly if you just go for ultra low ESR caps from a reputable brand with same capacitance and voltage rating as the originals. But if you want to do it 100% try to match as many values as possible.
Hope this helps! 😀

Reply 11 of 22, by evasive

User metadata
Rank Member
Rank
Member
PC Hoarder Patrol wrote on 2020-05-28, 08:07:

Sadly yes - I'll quote @PCBONEZ from another thread here "OST caps are famous for failing without bloating or leaking."

Rite.... Like those on my FM10 board then.
Oh wait, I first need to pry out the Chhhhsssssiiishtz caps from that little power supply as they DO bloat already 🙁 Maybe replace the TAEPO primary cap as good measure? Even though those seem to hold quite nicely over time.

Reply 12 of 22, by philmac

User metadata
Rank Newbie
Rank
Newbie

Thanks for the advice everyone, I've ordered some extra tools (including a PCI POST tester) and will give diagnosis a go.

Hopefully it's just the caps, at least swapping them out would be straightforward.

Reply 13 of 22, by Xplo

User metadata
Rank Newbie
Rank
Newbie

Is this the FN45 V1.0 or V2.0?

And did you change the DDR voltage to 2.7v at any point? It is well known that changing this option can kill FN45’s within a short space of time. I think this was exclusive to V2.0 though. And it was the vdimm regulator that fried out. Shuttle in later bios revisions removed the option to adjust DDR voltage because of this. Working off memory on all of this

Reply 14 of 22, by luckybob

User metadata
Rank l33t
Rank
l33t

Before you recap the board, do yourself a favor and REALLY wash that board.

Literally put it in the dishwasher. (take the fan off & battery out first, obviously) Rince the board in 90% ipa and leave it out in the sun for an hour. and try again. Test the board with a PCI video card. Test it without a cpu, without ram. See if you can get the board to do ANYTHING differently.

it will help you when you DO get to the recap stage, but you should check the board for something more incidious. Screwdriver damage. Socket 462 (A) was notorius for having heatsinks that took a 600lb gorilla and a sharp blade screwdriver to install. Even a minor nick on one trace can render the board dead. I use my phone for this. I set it to macro mode and I take max-res images as close as possible to the motherboard. Looking directly at the white area above and below the socket. If a screwdriver did take a chunk out of the board, that is where it will likely happen.

Also, when/if you do recap the board, Don't neglect the green caps around the AGP slot. There is a large and important power regulator there and they can easily be destroyed if someone had put in a 3.3v card into a 1.8v slot.

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

Reply 15 of 22, by pentiumspeed

User metadata
Rank Oldbie
Rank
Oldbie

OST kill them! Bad capacitor brand! Replace them with quality low-ESR capacitors. I had same thing happen to P4 motherboard I had. Bad capacitors kept PC dead and did not beep. Replacing them revived the motherboard and used it for years.

Cheers,

Great Northern aka Canada.

Reply 16 of 22, by philmac

User metadata
Rank Newbie
Rank
Newbie
luckybob wrote on 2020-06-02, 21:30:
Before you recap the board, do yourself a favor and REALLY wash that board. […]
Show full quote

Before you recap the board, do yourself a favor and REALLY wash that board.

Literally put it in the dishwasher. (take the fan off & battery out first, obviously) Rince the board in 90% ipa and leave it out in the sun for an hour. and try again. Test the board with a PCI video card. Test it without a cpu, without ram. See if you can get the board to do ANYTHING differently.

it will help you when you DO get to the recap stage, but you should check the board for something more incidious. Screwdriver damage. Socket 462 (A) was notorius for having heatsinks that took a 600lb gorilla and a sharp blade screwdriver to install. Even a minor nick on one trace can render the board dead. I use my phone for this. I set it to macro mode and I take max-res images as close as possible to the motherboard. Looking directly at the white area above and below the socket. If a screwdriver did take a chunk out of the board, that is where it will likely happen.

Also, when/if you do recap the board, Don't neglect the green caps around the AGP slot. There is a large and important power regulator there and they can easily be destroyed if someone had put in a 3.3v card into a 1.8v slot.

Cheers for the advice, an IPA rinse is definitely on the cards at least. I've tried removing absolutely everything, unfortunately it doesn't change the observed behaviour.

Screwdriver damage is less of a concern with this board, the heatsink was case specific and screwed through the board into the case behind it (no spring-loaded clips, no bladed screwdrivers):

DSCF0015.JPG
Filename
DSCF0015.JPG
File size
224.34 KiB
Views
150 views
File license
CC-BY-4.0

The CPU on this board had been fitted from new and hadn't been changed until I replaced it to rule it out.

The board has never had a 3.3V AGP card fitted either, it's only ever been used with a Radeon 9600, an x1600 and then finally an HD3650 (all 1.5V cards), or using the built-in GeForce MX400. According to the shuttle manual, it has some form of 3.3V protection built in and refuses to power up if you fit one to it (I can see a protection LED on the board marked out next to the AGP slot, but I've never seen it light up).

Reply 17 of 22, by computerguy08

User metadata
Rank Member
Rank
Member

Get a PC POST card. It really helps you in these kinds of situations.

I'm surprised nobody suggested this so far.

pentiumspeed wrote on 2020-06-02, 23:55:

OST kill them! Bad capacitor brand! Replace them with quality low-ESR capacitors. I had same thing happen to P4 motherboard I had. Bad capacitors kept PC dead and did not beep. Replacing them revived the motherboard and used it for years.

They are not that bad. I think you haven't seen the power of leaky CapXons or GSCs yet. All caps will die on a P4 no matter the brand (at least the ones near the CPU heatsink). Age and heat are not their best friends.

Reply 18 of 22, by shamino

User metadata
Rank Oldbie
Rank
Oldbie

Check all the voltages you can. Make sure you have good 5V, 12V, 3.3V, 5Vsb, Vcore, etc. If you find a pinout for the DDR slots you can also check the ~2.5V that should be there.
Some voltages are easy to grab from an unused drive connector, but others need to be backprobed on the ATX and "P4" connectors.
Even if the board was booting I'd still suggest doing this with a multimeter because the readings from onboard sensors aren't reliable. It's common to have huge discrepancies from what a meter will tell you.

Typically Vcore is switched by a group of 2 or 4 or more MOSFETs near the CPU socket, and half of them will have Vcore at the large tab. The other half will have the input voltage there (5V probably).
You can also find Vcore at the inductor between those MOSFETs and the CPU, or on the positive legs of the Vcore caps. You can't assume that all the caps near the CPU are for Vcore, but it will be fairly obvious which of them are when you find a bunch of them in electrical parallel. To really confirm that you can also check for continuity to the Vcore pins at the CPU socket (have to find a pinout to do that).

Reply 19 of 22, by philmac

User metadata
Rank Newbie
Rank
Newbie

My PCI tester card dropped through the post today, so I've tried powering up with it in the slot.

The display rapidly cycles through a series of codes too quickly to count, then stops on F6, which is PhoenixBIOS shorthand for "Boot to mini DOS". It never gets any further. Behaviour is always the same despite the CPU in the socket and whether I've put a DIMM in one of the slots.

I'll pull out my multimeter and start testing the voltage/current readings.