Let's fix a dead VARTA ravaged 486 motherboard
HTML-код
- Опубликовано: 13 янв 2025
- #repair
If you are into retro computer from the late 80s and into the 90s, you are probably fully aware of how destructive the NiCD "VARTA" batteries are to these systems when they leak ... and this 486 motherboard had one such battery leak and cause some severe damage to the motherboard.
So, let's try to get this motherboard working again to keep it out of the recycle bin!
-- Links
SMMC 0065 where this motherboard was first featured:
• 0065 Amazing replaceme...
Second channel video where I first took a look at this monitor:
• This trashpicked monit...
Leaking VARTA photo by John R. Southern:
commons.wikime...
Creative Commons Attribution-Share Alike 2.0 Generic license.
Adrian's Digital Basement Merch store:
my-store-c82bd...
Adrian's Digital Basement ][ (Second Channel)
/ @adriansdigitalbasement2
Support the channel on Patreon:
/ adriansdigitalbasement
-- Tools
Deoxit D5:
amzn.to/2VvOKy1
store.caig.com/...
O-Ring Pick Set: (I use these to lift chips off boards)
amzn.to/3a9x54J
Elenco Electronics LP-560 Logic Probe:
amzn.to/2VrT5lW
Hakko FR301 Desoldering Iron:
amzn.to/2ye6xC0
Rigol DS1054Z Four Channel Oscilloscope:
www.rigolna.co...
Head Worn Magnifying Goggles / Dual Lens Flip-In Head Magnifier:
amzn.to/3adRbuy
TL866II Plus Chip Tester and EPROM programmer: (The MiniPro)
amzn.to/2wG4tlP
www.aliexpress...
TS100 Soldering Iron:
amzn.to/2K36dJ5
www.ebay.com/i...
EEVBlog 121GW Multimeter:
www.eevblog.co...
DSLogic Basic Logic Analyzer:
amzn.to/2RDSDQw
www.ebay.com/i...
Magnetic Screw Holder:
amzn.to/3b8LOhG
www.harborfrei...
Universal ZIP sockets: (clones, used on my ZIF-64 test machine)
www.ebay.com/i...
RetroTink 2X Upconverter: (to hook up something like a C64 to HDMI)
www.retrotink.com/
Plato (Clone) Side Cutters: (order five)
www.ebay.com/i...
Heat Sinks:
www.aliexpress...
Little squeezy bottles: (available elsewhere too)
amzn.to/3b8LOOI
--- Links
My GitHub repository:
github.com/mis...
Commodore Computer Club / Vancouver, WA - Portland, OR - PDX Commodore Users Group
www.commodorec...
--- Instructional videos
My video on damage-free chip removal:
• How to remove chips wi...
--- Music
Intro music and other tracks by:
Nathan Divino
@itsnathandivino
This board is a Freetech/Flexus 486F24. It is on TRW, but under older revision with 82C460 chipset. If you can't find your mainboard on TRW, don't use the chipset, just enter the amount of slots and the type of the CPU. You will have to scroll a bit more through the results, but you will more likely find what you are searching for. Many entries were back ported from the old databases, where many entries were wrong, set to unknown etc.
Even though it's Genoa branded?
@@2dfx Genoa had a version of the Freetech/Flexus 486F24 - it's the same board. I have one (with 82C460 chipset) complete with the Genoa sticker (and I submitted the photo to The Retro Web)!
Mind blowing to me that a simple bump in revision lead them to use an entirely different chipset. Unless there is a lot more similarities between these two SIS than I realize...
also known as GENOA TURBOEXPRESS 486VL-3 SIS 85C461 BABY AT VLB AMD DX-2 66MHZ 8MB MOTHERBOARD but i do think there was a few versions
Nice to see you around here :D
and thanks for the comment
I started out thinking, "Why bother with his generic motherboard?" and by the end I was almost emotional that it worked. How crazy is this hobby? Thanks, Adrian. Great video!
Hah! Thanks. And yeah, I know it seems like 486 boards are very plentiful, but the reality is that not everyone has easy access to them, so I figured I could show that repairing these damaged boards is possible.
The only conclusion I can come to with the Memory issue, was that only of the 30pin ram sticks was not seated properly when you had them first installed, or there was corrosion.dirt, whatever on one or more of the terminals in the ram sockets. And then when you removed the ram and put it back in, effectively reseating the ram sticks allowed for a better connection and thus posted. It can happen even only those old ram sockets.
either that or there was just a bit of oxidation on the pins and with the memory insertion then removal then reinserted did the trick
I totally agree. This happened to me way back in the 90's while working with these things. Just pop the chips in and out a few times out of desparation and it suddenly worked. @@nathanwedgeworth6192
or the corrosion he scraped just south of the slots might have helped.
So in other words, you should always deoxit those sockets? ;)
@@LeftoverBeefcakeWords to live by right there.
Congratulations on all the new patrons! The screen is filled with all those new names! I'm truly heartened to see that your leap of faith is paying off. :)
It really blew my mind adding those slides into the start of the video. The names just kept coming! ❤
I found my name. It's like a Where's Waldo only text based.
that was a LOT. Very awesome 🙂
@@KayakTN That's one of the reasons why I moved up to the 32-bit tier a few months back; so I could find my name. ;)
Haha yeah the 32/64 bit ones are super easy to always see your name :-) And you can change your Patreon display name so you can put messages like "Hi Mom!"
Got to love those mysterious and unexplainable self fixes that come out of nowhere.
I had no idea that memory had such a sense of humour and self-deprecation that it required a built-in parody line.
I know this is probably a joy, but he meant "parity" as in error checking
Adrian's the ultimate homie for this unloved technology. If I was a 30-year-old damaged and discarded motherboard, his house is the place I'd want to end up too. Good vibes all around.
I built numerous old 386/486 systems when they were new. There were lots of 'quality imports' coming out of China. I had a box full of 2 pin jumpers. It was a royal pain. I don't miss them.
I have found that distilled water will dissolve the green crust better than IPA and it will come off with a stiff brush rather than harsh scraping. Once clean you can flood the area with IPA to remove the residue of water.
Could you use the brush and deoxit or baking soda?
Which makes sense considering copper salts tend to be water soluble! Some people prefer the 70% IPA over the 99% for stuff like this to get a best of both worlds for effect.
I thought white vinegar was the best way to dissovle that stuff.
@@robmoye7373 white vinegar is for neutralising the battery electrolyte, but those salts from the electrolyte will still be present and need to be cleaned and any copper from the traces will also be present. This was post cleaning which would’ve also involved removing the battery electrolyte, so this was all just extremely corroded traces. From more minor battery spills the traces are safe once you’ve neutralised and removed the electrolyte.
If you are using distilled water their shouldn't be any "residue" for you to clean up. Distilled means it has nothing in it. No minerals. If you have anything to clean up it was already there.
I always enjoy how giddy you get when you get these things going. I know you mentioned dumping the ROM in DOS is easy, but I wish you would've quickly shown how to do so. I know the output was on the screen, but would've been nice to see the actual command. Fun to watch.
What I get from Adrian's video when it comes to fixing such complicated hardware is that there will always be a moment when it will magically start to work. And that's usually more scary than anything else.
Adrian, you are a national treasure for the retro community. What a great video! I love the engraving tool idea too.
I'd say international treasure.
That engraver just ruined the remaining traces. It was stupid. Should have used chemicals to neutralize the corrosion first.
Board worked great after, so proof is in the pudding.
This is one of those cases where the board is just missing from the internet, not documented anywhere. The most similar one we have on TRW is a 486F24, with the Genoa branding again and a similar layout, but it uses an 85C460 chipset.
I think they have matching jumper configuration.
Pretty sure, that is the same one, just updated revision.
@@Petr01d - Jumpers don't seem to be in the same position. They may have the same labels, though.
That was so fun to watch, thanks for the ride.
Now I wish I held on to TWO 486 boards that were heavily corroded. I didn't know back then that repair could be possible. Thanks for being a good teacher.
I just watched the video of you saying you were going full time, and I was honestly shocked. the quality of your videos and frequency of uploads made me think you were already full time. great work and very happy for you!
I do not post much on these types of videos, however i believe from time to time I need to express my Happiness when I see someone with your level of passion , and dedication to keeping these little pieces of computing history alive for the future collectors and creators. If I could give you a thousand thumbs up I would , however You Tube will not allow it . . No surprise there.
Keep up the great work , and the wonderfully entertaining videos Adrian . You have a special talent . I am glad you share it with us .As I am with all of your peers. The 8 Bit Guy , Crazy Ken . ETC...
Despite the unknown memory problem, this was satisfying.
Tip- Sharpie ink interferes with soldering! Glad you were marking completed ones, not ones to be done. XD
I think you did scratch off some of the corrosion with the plastic stick or by touching it. Maybe you removed a short just a little bit before you tested it completely. My experience with the infamous dallas chips has been the same: after you fix the battery it doesn't work, and after you start messing with it it suddenly keeps time again.
Excellent work on all the bodge wires! I would recommend grinding/cutting the trace near the solder points on both ends of the wires to disconnect the crusty leftover trace. Additionally, a nice viable alternative when repairing corroded traces is to use stripped wire (or plain, non-enameled wire), drill out the vias, and route the fresh wire exactly on top of the corroded trace through the drilled vias, using uv setting glue (or epoxy) to keep it in place, and applying lacquer afterwards - this method reduces risks associated with having the loose bodge wires (movement, breakage, getting caught accidentaly, etc.). Regardless, excellent video, thoroughly enjoyed it.
37:55 Could it simply be that the RAM did not have good contact before?
37:37 Slot 4 and 6 (from top to bottom) each seem to have one darker pin than all the others (at least it looks like it)
I had tried 3 sets of RAM (off camera) and none helped, so it seemed to be after testing that things worked, so maybe just poling the pins made it have better contact? I assume those dirty spots were actually some mild corrosion too... not 100% sure. We'll really never know :-)
It’s such a joy when we finally get these back in working order! These videos really entertain me.
Thank you for another great video Adrian!
Possibly a bad connect on the ram chips causing post to stop at 13. You did pull them to test pinout. I hate these types issues. Great job once again!
Adrian. Before performing this type of service, you MUST neutralize the alkaline electrolyte leaking from the battery using vinegar. Without this step, corrosion spots will appear in the future, unfortunately.
@@Totttty55Well... Trash for some, treasure for others. You don't measure the value of time spent repairing an outdated piece of technology as a waste of time: It's fun! It's knowledge acquisition! And for us, viewers of this channel, it is entertainment of the best quality: without politics, without ideology, without aggression, without insults and without caring whether it is recent or past technology. Thank you, Adrian, for providing it to all of us who love and admire your work and videos. I hope it continues like this (specially in this new phase of your life). Count on us. Count on me. Roger, from Brazil.
Usually Adrian neutralizes and knows the chemistry associated with this issues... seems to skip all the cleaning in the more recent videos.
@@cls9474IIRC he said a while ago these things get cleaned as needed after their packages are opened, then put on the shelf to fix later. I’m sure recording all that cleaning would be annoying, as would be finding the right spots to splice in the footage for the later video (a smaller version of which is why the mail opening format changed).
@@Totttty55 It wasn't trash, it worked perfectly after repairs. And all the corrosion was removed, so I don't see how any further corrosion could be possible.
@Monty22001 I think Adrian has shown to be super methodical previously, so I can accept your comment, but it would have been better if you had added your preferred methodology of repairing the mish mash of trace issues? Then viewers can consider your point better.
I think your repair was good the first time and the problem was an open between a ram pin and the mobo socket. I've had that issue a few times now. Socket some good known ram into the board and the mobo won't post. Start thinking something happened and it isn't until I pull the ram and re-insert it that the problem "fixes itself". The surfaces of the contact pins can get a micro layer of corrosion that isn't visible but can, under JUST the right (wrong) circumstances cause a lack of continuity until the connection is reset, with the friction of removing and re-inserting re-establishing the connection between pins.
Love the videos, new sub here.
This was one hell of a journey, congrats on all your hard work restoring this motherboard!
Great job of restoring that 486. I appreciate the time and dedication you exerted while repairing an old board. I also appreciate the excitement you express on your successes. I always learn from what you do, and intend to use what I have learned on my Amiga.
Honestly, the metal bit is a bit unnerving. There’s another bit that Necroware uses that's originally meant for polishing the rails of model railroads which is green and has a rubberized texture. It seems to be a bit more gentle to the traces.
He mentions during the video that he'd run out of those Dremel polishing wheels and was waiting for more to arrive hence the need to use the engraver as a substitute.
Adrian, have you ever tried enameled solderable magnet wire aka Poor-man's Multiwire? Hot solder and flux will reflow the end. No stripping necessary. Wire lays on top of the trace. A few drops of superglue in strategic places will keep it in place. Usually AWG 32 to 34 is sufficient. . Great for high-speed over-the-groundplane dead-bug breadboarding as well.
It's quite a feeling seeing a board post after you were certain it was a lost cause. Great job!
I really needed that demonstration in dealing with corrosion as I'm currently working on a keyboard with a bunch of it! Super appreciated and can't wait to see more from your basement!
I think I have an answer to your what the heck happened question. When you were testing the mem slots it's possible there was a small piece of corrosion or foreign material in the slot which was dislodged when you had the probs in. After when you installed the memory the issue was gone. I've run into this a few times in my professional career. Hope it helps. Great video by the way. Cheers!
When a retro-computer repair channel is suddenly transformed into a beautiful and relaxing space of microscopic modern art. (15:09) 😁🎨🥰👌
The 90's were so ... _Wild_ ... when it came to standards
I had ordered a 120Mhz 486DX AMD - at the height of that Pentium1 period
in the mid-90's I was building Laser Barcode-scanners (spread spectrum handheld radio lan devices) for Symbol technologies. So I did quite a lot of SMD soldering & automated PCB testing. Worked with nearly all aspects of manufacturing their PCB's in the Bohemia manufacturing plant, on Lakeville, in Long Island
That's quite a bit of patching you've done
Good Work - strange how it's "Fun" to sorta fix this old junk sometimes---
I did used to own a 386DX 40Mhz AMD as well - it was my FIRST PC.
I'd thought it might be nice to rebuild that old box , the 386. The Board could hold 16Megs of ram which was a lot back then, and I COULD actually run windows 95 on it
Hi Adrian 21:14 you can't get the solder blob to bridge. Next time try without flux. Think of how when you're soldering chips with multiple pins close together you need to use flux, why? To help prevent bridging 😊
My earlier comment suggesting GadgetUK164 seems to have disappeared. Great channel if you enjoy watching trace repairs.
In high school, I had an AMD 486 DX4-120MHz. I kept the boxes all through college and when I gave it away, the box said something like Tomato board. We all had a laugh that my computer ran on produce other than apple. Lots of 486 brands indeed.
I love Tomato boards. They tend to be the most adorable tiny little motherboards. I think they may have been a low-cost option that used each generation's platform, toward the end of its service life. They often have features that weren't very common at first, and do it in the least possible PCB area.
15:13 Adrian plays Bob Ross. 😁🎨🖌🖼
You give me the Inspiration and the confidence to try to fix my own 486 whit battery damage.
,keep up the good work 👍😀
Again superb repair. I have almost the same exact board, the 486dx tho at 33mhz and the 3 pci version as well. Thanks for sharing. Mine has the same varta issue as well but I don't think as bad. Its been removed and vinegar applied. Still need to get it going. Thanks bud for sharing 😊
Kudos to you Adrian. I really enjoyed this repair.
I noticed on some of the waveform traces that they were briefly staying in "illegal logic zones" , like a paused at around 1.5 volts (They were using a 5 V TTL standard I believe)
The only thing that I can think of, is that after the trace repairs, lines electrically connected, yes but no longer terminated in the lines characteristic impedance. Even back then, computer signals were fast enough that required proper termination. I remember my first PC, which was a 80286 based system running at 12 MHZ, and it was made in a six layer board to address those issues.
My guess is that it will continue to work intermittently, depending on what precise path you route the wires, and that by just moving the wires you might lower or increase stray capacitance
and it might stop working again.
Weird voltages like that are probably normal for a data bus. When the CPU isn't reading or writing data the bus is left floating.
Love a good repair video, you just never know where it will take ya!
It's like I'm watching an episode of Epictronics/Bits N Bolts/Necroware! I suggest in the future you give the board a bath in a neutralizing agent before grinding, that should help a lot. Also fiberglass pens are great to use first on traces to clean them. Coleco Adam!!
Ah a video with microscope. Always like this type of video. One of my fav. Well done mister!
It pains me to think about how many 486 systems I just simply threw out in the late 90s. A lot were given to me by family friends when I was a teenager to see if I could make something of them - probably just an easy way for them to get rid of them since most places wouldn't accept them as trash. Most of the time I just tossed them out, but if they had some good parts I kept them. Eventually I had a pretty mean AMD 5x86 OCed to 160 and maxed out with everything the board would support - 64MB RAM, Stealth64 VLB, Adaptec SCSI and drives, etc - kept it alongside my main K6-2 to play older DOS games that sometimes wouldn't run well on the K6-2 - the turbo switch still worked mostly as intended with the 5x86. Tossed that one out in the early 2000s and I regret it all the time.
Great job rescuing another from the dumpster... Wish I would've done the same back then.
I also have one of those tiny engravers and they're fabulous tools for trace rework. Much faster and easier to use to cut traces or expose bare copper than anything else I've tried, although it is a fairly aggressive tool to use for cleaning corrosion as you do here so you do need a light touch and to be careful when using one for that purpose!
More repairs like this please.
The more scope work and difficult to repair, the better.
It would mean more hair loss for you, but I can live with that.
I don't get to do repairs like I used too back in those days. Now, nothing is worth repairing.
A fiberglass pen would have been better for clearing the corrosion rather than the glass engraving bit you used, It would have done alot less damage to the traces overall considering the engraving bit has diamond on it for glass, It may take longer but with IPA and effort you would have seen what could be kept and what needed repair.
It doesn't matter. It's fixed.
He really didn't go deep with it. Those traces were gone
I think the same amount of wires would have been used at the end, since large chunks were missing on almost every trace. I personally would not have trusted whatever thin layer of copper was left and also bypassed the whole section, no matter what.
Fiberglass pens also leave a lot of small glass shards, which isn't great :/
That''s why i suggested using IPA as well to clean it afterward but heaven forbid people would have to take time to do it right instead of doing it right now.@@dan3a
I love the right angle routing on the motherboard, so retro!
In cases like this, I would inspect the silk screen markings. I got a 486 motherboard attacked by the evil barrel battery and the model was between the ISA slots. Just the model.
Other way would be ... dumpin the BIOS and checking the strings with some tool or with a Hex editor.
Keep it up.
Impressive list of patrons! :)
Outstanding work, Adrian. This is therapy that money just cannot buy!
Never heard someone so ecstatic to hear a no memory beep, lol
Great video! Would love to see rhe footage of repair in real time. Understand the time-laps didn't work. But real time video of repair is why I enjoy your channel. And I do realize time restraints so maybe a mix? But great job!
Can't help but notice you sound a little under the weather. Just wanted to say how much I appreciate your hard work you do. Also I have a dead 486 board same sort of death.
What a great video Adrian, loved the detailed macro shots where you repaired the traces, congrats on restoring this 486 Board (with a little bit of surprise-repair)!
When testing, pick up the board and bend & flex it to see if it locks up or crashes, you may have a marginal via, hairline crack or bad solder joint.
When I used to fix CRT's we had a small rubber ball on a stick that we would use to tap around on things.
Ha I didn’t expect you succeeding. That’s incredible. Nice job!
Wow, great tip on the engraver. I'll have to add one to my kit, thanks!
Like a few others here, my first guess for the RAM problem is that there was a speck of dust caught in one of the RAM sockets and it was enough to break contact with one of the pins which caused the failure of the POST memory test. It's probable that once you removed the RAM and handled the board for a while, the speck of dust fell out or at least shifted position which allowed full contact once you put the RAM back in.
Great video again! As II am restoring some 486 based measurement devices myself, I am left with two questions:
1) Is the small rotary tool is missing in your tools list? Would like to buy it with a small benefit for you.
2) Is there an image of that test-disk with all the DOS based tools to check PCs and identify IO and peripherals? Or at least a list of tools you use regularly?
The last question is especially as HP, R&S and other have often made add-on BIOS ROMs or added in special hardware.
I do recall when learning back in the day the sop first step for a dead board was to put a logic probe on the keyboard controller to see if it had clock and was active because the board would not start if it wasn't working.
Thanks for sharing your ideas Sir. Watching from Philippines
Quite an accomplishment as these era boards on not the easiest to work on I think. Congrats and thanks for sharing another great video.
I love The Retro Web, but that motherboard search fails to find my board many times. So I only filter for chipset and rather browse the results manually (a little more scrolling, but greater success ratio). Something may be wrong with the slot filter, or the data is incorrect for many motherboards. I always try combinations (are VLB slots also counted into the 16 bit ISA slots or not?), but does not help. Edit: dumping the BIOS and searching for model numbers or manufacturers also worked for some old motherboards, and then I found it on TRW with a different approach.
Awesome work! Super cool to see. I have 3 of those techmedia 486 boards that are dead but show signs of being able to work, if someone who is more capable of repair than me does the work.
ROFL! :D The evil laugh while doing the bodges makes this instantly my top1 favourite Adrian's video :D
I loved this video. Why things fix itself is sometimes a riddle that we never solve. And it's really annoying coz u don't know if it comes back to bite you later. Might I suggest tho that simply using the sockets a few times, since it's been sitting for a while, maybe fix conductivity issues? Like u know... blowing on carts doesn't fix them, but probably the fact that u inserted it a couple of times extra.
This gives me hope... I have a 486 motherboard that also had a leaky battery and it is also not posting. I'd have to ask my dad for help because I don't have proper tools for such repair :P Thank you so much :) cross your fingers that we'll fix our motherboard :)
I hadn't heard of the Retreo Web before! That's an amazing resource!
kind of disappointing when those things happen, isn't it? :) But rewarding when you see that working again! Thanks for the video! I love the rubber tool because it's so gentle and polishes the traces perfectly. But it also removes most of the solder mask. The engraving tool allows you to be more precise and just work on traces. I don't have one but I am tempted!
I use 1200 grit wet & dry sandpaper (wet with iso) for the corrosion. Cut a strip and fold over the back end of a chopstick.
You can also use a pcb drill bit in a pin vice to clear corroded through holes. Then you can feed a bodge wire thorough the board via.
Be careful of via internal connections (hold the pcb up to the light to check). Don't attempt this on pcb's with more than 4 layers.
Keep bodge wires as short as possible, preferably just bridge the break as you may screw up the signal integrity.
I guess you poke the SIMM socket's pin and some of them just resume "touching" with the RAM module, thank you for the video, remind me when I work in the 486 motherboard aftersales support.
The "poking" was really the only thing I changed.... sucks I had not knowing. Could also be some corrosion underneath too, so maybe it'll die again in the future.
No fibreglass pencil?
The go-to tool for this sort of job! Inexpensive and effective. Much MUCH kinder to delicate PCB tracks than a dremel or other grinder. Ideally neutralise acid with vinegar first (and wash, dry etc) but they are fine to use on a PCB "dry" and won't grind away the PCB substrate. Ideal for (ahem) cleaning SIMM and DIMM connectors too. Also IC pins.
IT LIVES, shades of Young Frankenstine. another great video Adrian (fellow Portlandite)
Correct settings for that motherboard is for 486dx cpu type JP9 pins 1&2 closed, pins 3&4 closed, JP10 pins 1&2 closed, JP22 open, JP23 closed. For setting the clock to 33MhzJP6 closed, JP7 closed and JP8 open.
Do you have any documentation for this motherboard?
@@gregp93 Not for this exact version but I have similar motherboard with same jumper layout and jumper numbering. Main difference with jumpers is that on his JP8 does not have a jumper which limits the bus clockspeed options on his motheboard.
38:18 - You poked the pins around, re-seated the RAM, moved things, dislodged corrosion and dust. I've had endless problems with 30-pin SIMMs and bad connections. Simply re-seating them has fixed so many issues. Perhaps you need some deoxit? 😛 Love these videos. Keep them coming. 🙂
Hi Adrian - did a little digging, that Winbond W85C168 is a pin compatible replacement for the Motorola MC146818 (not the MC146818A). As a result, there are other chips that are identical in functionality and pinout: Harris CDP6818, Hualon HM6818A - all of them are identical CMOS RTC's with 114-bytes of RAM.
Adrian and Necroware both posting 486 board repair on the same day, we are being utterly spoilt.
Nice!
Sometimes I wonder why I don't see repair videos from you, but then I think about how that requires obtaining broken/damaged components, the fact that creating a video can make a repair take four times longer to do. But maybe we'll one day see you bodging traces in order to get some benchmarks and reviews on something that isn't too common.
I'd have to learn a lot first! I've done basic recapping and some soldering of projects (Serdaco kits), but nothing as difficult as SMD or repairing traces. But yes, never say never :D@@E-Box
Don't short change yourself! It's actually not that hard. It mainly requires patience and then the stuff I used in this video. A lot of times, the damage is less severe so you won't need so many repaired traces. And besides, think about it this way: the board is dead anyway. So even if the repair is a failure, it's not worse off than it was before you started. :)
Great video! Thanks for posting ..How did you do the ROM dumps @ 48:43 TIA
This video may hold the record for the number of "freakin'"s being used. 🤣
@30:57 I don't think it's a memory issue. Last post code is 13. That's the last good post code. So either postcode 13 fails... or more than likely it's the next test - 14 or Keyboard controller.
Keyboard IC is near the battery and is a likely corrosively damage or still has a bad trace.
Another tool which might come in handy with these corrosion jobs are fiberglass pens. Looks like a pen, but has fiberglass for "lead".
Note on the memory sockets, the intermittent error on the memory may be due to dust or oxidation on the memory socket pins.
It was very satisfying to watch the cleaning under the microscope. That gleaming copper, mmm.
Maybe the memory sockets were dirty and while you were using the o-scope probes you dislodged a particle of dust that was blocking contact. When you reinserted the RAM, it then made contact. Maybe you accidentally blew on it - LOL. Who knows? I do feel a certain amount of joy here. I upgraded to AMD 486 boards when they came out and were known as solid at the time. They were so much fun. I wish I had kept the boards, but I had to sell the machines for Pentium boards to catch up for work.
I was thinking about the RAM issues that seemed to mysteriously be fixed. Quite possibly all the probing scratched off some corrosion that was causing a bad contact? You did see some corrosion near the sockets and there was suspicion that some other board may have leaked on it. Not impossible there was some minor (not easily visible) corrosion causing a pin not to make good contact.
[Adrian's Head Empty] Adrian: Uh? Brain: Say catchphrase. Adrian: It's freakin' working? Computer: [cough] Adrian: It's freakin' working!
24:18 For a sec I thought the board required a blood sacrifice.
I have a trick I use when I make a solder bridge over a gap. I blow on the solder while pulling the iron away, and it draws a line of solid solder.
One of the first things you should do is to clean up de ram contacts because they usually make bad connections and thats what i think you have there and when you pull them out and put them in again they did the contact properly... this kind of things solved a lot of posting problems up to today's motherboards. You do great work on solving this kind of things but sometimes the problem is more simple than it looks... great work though..
Where possible I like to thread wire through via or component holes, if a via is too far away scrape the mask off a bit further back on some undamaged sections of the trace and “surface mount” the copper repair wire. I repair a lot of 1kZ engine ECUs with cap damage like this
That's especially important with multilayer boards, since bypassing the via may leave internal traces unconnected. Many experienced people have gotten away with bypassing the vias, which works if one side of the board has the trace intact or when there aren't any connections to internal layers through the via.
Maybe when you reinserted the SIMMs they actually made contact??
Between the cut, I actually tried 3 sets of RAM and even a full 8mb of RAM, but it never made a difference..... maybe 4th time was a charm or the testing was. We'll never know sadly.
I had the same problem with 386DX2/4 or 486 VLB Systems in their days that some ISA cards had problems. Some of them just had a jumper for 1 waitstate and then they worked.
I sure don't miss plugging in boards that made my system act funny... but worked *perfectly* with a friend's motherboard.
40:07 when you are so surprised that you become 8-bit noise generator
The only problem I can see with all that corrosion on multilayer boards is, how many traces are messed up on the inner layers you can not see? I used to design these circuit boards back in the day. and most were four to six layer boards. external corrosion does not always mean other layers could be affected as well.
That's always a question. Assuming this is a 4-layer board the inner layers are probably just ground and power so a little corrosion won't hurt too much.
I think that you moved RAM sticks in and out removed some dirt or crust from the RAM pins and that's why it started to work after you checked all the pins.
This video was posted precisely when I started trying to revive two of my own 486 laptops.
One was a Compaq Contura Aero 4/25 (FCC ID CNT75MBZE) that was my beach laptop until the day it got hit by a fast moving ball some 20 years ago. The screen, bezel, part of the body, the flex cable that connects the mouse to the main board and one of the hinges got destroyed. And the HDD seems to be stuck (spins but head does not move). Could never get replacement parts so now I am trying to get a screen connected to its dock connector and... success :) . Now it only needs a CF, some wires to replace the flex cable and a body rebuild. I am considering using wood for that :|
If someone reads this and also needs this info, the pins on the connector go from 1 on the top right to 60 on the bottom left. My manual has the pins correctly numbered but the image is wrong. Pins are 17=vsync; 18=hsync; 31=blue ground; 32=green ground; 33=red ground; 46=blue; 47=green; 48=red.
The second laptop was a Microstar Computer NoteStar NP743 (FCC ID KEWNP700) for which I could never find any useful information anywhere. It was given to me without a screen and the CMOS battery leaked ages ago in a strange way. Instead of just letting the fluid out, it sprayed all over as if it was under pressure and someone made a tiny hole. The result was that over half the main board is mildly corroded. This one will take some effort. Something funny about this laptop is that the CPU is socketed and the socket is socketed :P