VOGONS


First post, by .legaCy

User metadata
Rank Oldbie
Rank
Oldbie

I'm working on restoring a S7 Pentium MMX system, cleaned all of the case, retrobrighted the front, everything perfect.
Now came the time to make it actually work.
First i did a sanity check on the AT PSU, all voltages were correct(of course that i'm loaded some rails to test), then i connected back to the board, and nothing, no beep, nothing at all, CPU get warm.
Then i went through my usual troubleshooting, decided that a recap was needed.
The board has no sign of physical damage, no corrosion, everything seems ok, so i recapped all the THT capacitors(all the smd ones remains the same).
After recapping i quickly inspected and became testing, first without any cpu,ram, and expansion boards, nothing exploded, then i tested with the pentium MMX, 2 known good simms, and a known good vga card, booted up nicely, complained about the CMOS battery but i managed to proceed to the setup and it worked.
Then i installed back into the case, connected everything back, and then it suddenly stopped working, now it doesn't beep, my post card shows just 00, reset is constantly high and IRDY LED is constantly off.
I disconnected everything, then i checked for shorting something into the case, and its all good, but the board don't boot anymore, it stay just as described above.
Anyone got a clue of what might be happening? i also tried with a k6-2 and another pentium mmx and nothing.

EDIT: Look the rest of the thread cause i talk about another board.

Last edited by .legaCy on 2019-04-08, 01:51. Edited 1 time in total.

Reply 1 of 13, by rasz_pl

User metadata
Rank l33t
Rank
l33t

Seen few computers with one metal post too many 😉, might also be a good old "never happened to me ergo doesnt exist" ESD damage.
Take everything out and try again on cardboard box.

Open Source AT&T Globalyst/NCR/FIC 486-GAC-2 proprietary Cache Module reproduction

Reply 2 of 13, by skitters

User metadata
Rank Member
Rank
Member

Recently I watched a fixit video on YouTube where the problem was that the reset switch on the case was bad.
So try disconnecting the power switch and reset switch and see if you can turn it on by shorting pins instead of using the power button on the case.

Reply 4 of 13, by .legaCy

User metadata
Rank Oldbie
Rank
Oldbie
rasz_pl wrote:

Seen few computers with one metal post too many 😉, might also be a good old "never happened to me ergo doesnt exist" ESD damage.
Take everything out and try again on cardboard box.

I highly doubt, cause every moment i used ESD safe equipment including gloves, even when installing it on the case.
ESD safe soldering station, ESD safe desoldering station, everything cleaned with a lot of IPA and scrubbed over with ESD safe brushes.
I never in my life had any equipment suffer ESD damage(maybe it is because i always take the proper precautions).
I'm really leaning towards a bad solder junction on another component because it worked until the board was installed in the case, maybe the movement and slight bends of installing everything could crack a solder joint.
about the reset switch i highly doubt too, cause it wouldn't post even without any switch or led or speaker connected, since it is an AT system the power switch lies on the PSU.
I may try to flux and reflow the solder joints of every THT component, if it doesn't work i will reflow the smd IC's.
It would be nice to have board schematics like we have with apple products, then i would create a channel called legacy rossmann and would livestream old boards repair,and instead of bashing apple, i would bash pcchips.
But back to the topic, i HIGHLY doubt ESD because the weather is pretty dry and i'm getting static shocks every time, so when working with the board i made sure to be discharged.

Reply 6 of 13, by .legaCy

User metadata
Rank Oldbie
Rank
Oldbie
quicknick wrote:

If reset is always high, this is your starting point. Check power_good signal, find out the datasheets of your chipset, trace back the reset circuitry...

will do, well a quick duckduckgo didn't brought me the datasheet of Intel 430 VX, but when i came back from work i will check.
edit: power good signal is 5v so it is being signaled properly.

Last edited by .legaCy on 2019-04-05, 21:19. Edited 2 times in total.

Reply 7 of 13, by Deunan

User metadata
Rank Oldbie
Rank
Oldbie

If the card does display something you are getting into BIOS - so a reset problem is unlikely (unless it's unstable). Different BIOSes do things their own way but in general there would be some sort of (simple) CPU self-test on 286+ machines, that's code 00. CPUs rarely fail, so check voltage settings, reseat it, inspect the pins and socket for any damage. Other than CPU it could also be a problem with the BIOS image - got corrupted or the chip is not making good contact (if socketed).

Reply 8 of 13, by .legaCy

User metadata
Rank Oldbie
Rank
Oldbie
Deunan wrote:

If the card does display something you are getting into BIOS - so a reset problem is unlikely (unless it's unstable). Different BIOSes do things their own way but in general there would be some sort of (simple) CPU self-test on 286+ machines, that's code 00. CPUs rarely fail, so check voltage settings, reseat it, inspect the pins and socket for any damage. Other than CPU it could also be a problem with the BIOS image - got corrupted or the chip is not making good contact (if socketed).

I managed to break the chipset legs with my flux pen
well i feel terrible now
QZ7oLjG.jpg
https://imgur.com/QZ7oLjG

Oh well, i'm a little sad but i have another pile of s7 boards around.

Reply 9 of 13, by rasz_pl

User metadata
Rank l33t
Rank
l33t

not the end of the world, you can dig that pin out with a dremel
https://imgur.com/gallery/Amt1b
or laser https://ca.rstenpresser.de/blag/tag/quadcopter/

Open Source AT&T Globalyst/NCR/FIC 486-GAC-2 proprietary Cache Module reproduction

Reply 10 of 13, by Deunan

User metadata
Rank Oldbie
Rank
Oldbie

I use a small, precision painting brush to apply flux. I always liked the idea of flux pens but never used one - was worried it'd dry out before I used it up. I hate when this happens to my markers.

Reply 11 of 13, by .legaCy

User metadata
Rank Oldbie
Rank
Oldbie
rasz_pl wrote:

not the end of the world, you can dig that pin out with a dremel
https://imgur.com/gallery/Amt1b
or laser https://ca.rstenpresser.de/blag/tag/quadcopter/

I'm not nearly skilled enough to have precision, when i use my dremel bad things happens.
I dont use flux pen often, but my bottle of 951 kester flux is empty.

Reply 12 of 13, by Deunan

User metadata
Rank Oldbie
Rank
Oldbie

Yeah, neither my eyesight nor my "mechanical" skills are that great either so I understand your reluctance to fix this. But it's a great opportunity to improve and learn - it's dead anyway and not much can be salvaged from a (semi)modern PC mobo for other projects. So might as well try. Just take your time with it, don't rush. Maybe try with another cheap first to get the idea how deep the cut must be.

Both of those two pins seem to be a power/ground plane and are connected together - same as around the corner. So it should be easier to solder, just expose some metal and connect the two at the same time with a single piece of wire, you don't need any special materials like the flex cable here.

Reply 13 of 13, by .legaCy

User metadata
Rank Oldbie
Rank
Oldbie
Deunan wrote:

Yeah, neither my eyesight nor my "mechanical" skills are that great either so I understand your reluctance to fix this. But it's a great opportunity to improve and learn - it's dead anyway and not much can be salvaged from a (semi)modern PC mobo for other projects. So might as well try. Just take your time with it, don't rush. Maybe try with another cheap first to get the idea how deep the cut must be.

Both of those two pins seem to be a power/ground plane and are connected together - same as around the corner. So it should be easier to solder, just expose some metal and connect the two at the same time with a single piece of wire, you don't need any special materials like the flex cable here.

Well i have really thin wires to wind inductors and repair traces i can try, my hope was that i broke some redundant vcc or gnd pin.

Edit:
now i'm working on another socket 7 board, i recapped most of it, from post card showing 00 i reached F9 code, but no video display, tried with another video card, ram and cpu, when i change the video card(so i'm not using the onboard) the post code changes, any clue, my post card booklet with the post code descriptions don't have any F9 code.