I've always really enjoyed your repair videos. This one is great to show you where to start. I think what a cool think to do might be to make a repair playlist with this as the first video, and then. Either, with future videos, or maybe past videos slowly organize them in a way where someone can start with this video with the very basics, and then graduate to more advanced topics. Once there is enough content perhaps you can even market it as a training course!
Thank you! Glad you like it. The idea is great, but I'm still learning how to make quality content and many older videos are not quite as good. May be I just should start a new playlist for this. I'll think about it. Thank you for the idea.
my step before power on 2 - check everything and anything. all lines, all components, all legs of the ICs, look for any cut, bruise, touching legs (caused by any pressure o a hit), foreign bodies/particles (i´ve found very small metal things,) etc. not only can provoke failure to boot, but killing shorts I do this BEFORE put any power into the system
Yeah, that's a good hint as well. As I said in the video, looking for mechanical damages can be made in the first step. In the most cases touching pins are not critical and power pins would produce a short, which you should measure on the power connector. But if you want to be really sure, it's a good idea to make it first.
BTW, a handy hack for erasing old EPROMs: if you have a SonicCare electric toothbrush, they often come with a stand that has a UV sanitizer lamp. This produces UV-C light and is perfect for erasing EPROMs. I just pop one in (remove the label first) facing towards the bulb, turn it on, and wait about 15 minutes. After that they come out 100% erased (all bits set to one).
mm. as I dont have an UV eraser or any uv source, I did a "little" experiment. I put 12 1Mbit eeproms (27C010) to be erased with sunlight. I checked them every day. After 4 days (~8 hours of sun everyday, sometime some clouds) the first one had already been erased (initially it was not fully written, so it was partly "empty"). The erased ones were removed from the sun. the last one was completely erased after 12 sunny days. then all, together, received an additional day of sunlight . in some time, if everything is OK, I will erase more than 20 eeproms the same way, for further use.
I also experimented with it and at least in German summer sun, after a month I only got couple of bit flips. May be somewhere more in the direction of equator it works better, but in Germany, no way :)
@@necro_ware I'm right below the Tropic of CancerI will use those eeproms in a couple of disk emulator cards. meanwhile I got in sight 4 LEDs in a zaplamp . will experiment with those too
@@necro_ware Exactly - the amount of UVC in sunlight is far too low to erase EPROMs in a reasonable timeframe. Good thing though - or every one one of us would be dying from skin cancer!
That is valid in my opinion when You cant read/write bios chip in an external programmer. My mandatory repair step is to pur bios chip in thebprogrammer and write latest bios and rechceck
There is never a guide possible 😀 each board and issue is different. I have for example 20 462 socket boards to repair and each one of them has a different issue. A basic troubleshoot video for such board generation is highly doable
@@SzymekCRXno post. the cpu and nb get hot, cpu fan spins. no post beeps or video signal on agp or pci. cause of the cpu and nb getting hot i think the board can be repaired
I wonder if the master himself (or someone else) could recommend what next to try in my situation with a broken Gigabyte GA-486US board. It doesn't work at all and just produces dashes on the post card. There is no damage that I can see and no battery leakage. The clock signals all look fine and the CPU has 5v. Reset line goes high then low, but there is absolutely no activity anywhere - Measuring at an ISA slot, all data lines are stuck permanently high. A18, A17 and A29 are all at 0v and don't move. All of the other address lines stay at around 3.8 - 4v.
The fact that you can clearly see the fiberglass weave on that spot means something hit this thing pretty hard. Surprising that there wasn't much more damage than a couple of messed up caps and one small trace break.
You don't see fibreglass that much these days. If you watch Apple and modern video card repairs the PCB is closer to compressed paper. I've seen iPhone PCB turn to dust, ie, burn, with a high current short on STS Telecom channel. Modern GPUs are the worse, because pads fall off or go lose under BGAs all the time , and that's not mentioning the cracks at the holding down lug that northwestrepair is repairing all the time.
@@SonicBoone56 I'm sure on the close-up view through the microscope that you can see the very edge of another trace inside the board to the right-hand side of the damaged area. Maybe I'm seeing a reflection or something but it's around the 20:05 time in the video. I'm guessing that it's not going to be highly populated with traces that close to a screw hole, so this board is quite lucky.
What an excellent video. Imagine if we had videos and resources like this 10+ years ago. So many boards could have been saved. Always love your perfectionist trace repairs.
10+ years ago, it wasn't viewed as worth the effort. There were so many boards, parts, and complete systems available for basically the cost of hauling them away.
Thanks. Your Videos reactivate my „love“ for old PCs again. Successfully fix a 486 MoBo yesterday. Normally want to assemble the PC but get stuck playing Dune II on my test Bench :) Enemy.Harvester.Destroyed!!
VLSI was actually a custom gate array / ASIC manufacturer. Intel just uses VLSI to realize their own chipsets. Maybe in later revisions, Intel let them print their logo onto these chips, whereas in the earlier revisions they didn't.
This comes at a good time - AT boards are getting harder to find (working ones anyhow) and I have a small pile of 286/386/486 boards (even a socket 7) that I couldn't bring myself to throw out. As soon as its cheaper for me to buy a multimeter/scope like that rather than a board I need to complete a system, I'm getting one and trying my hand at this. Very motivational Necroware ;)
Lol 😂 thanks. There are and everyone gets a chance to have an opinion. Especially if some not just pushes "thumbs down", but also tells me, what I could try to make better next time. Unfortunately in many cases I just have to guess.
Super klasse! Ich mag jedes Reparaturvideo. Jetzt brauche ich dringend das Prototypen Messboard zum testen der Busse und auch einen Chipreader für die BIOS Bausteine. Im Keller liegt noch das arme Shuttle und Board aus dem PCD-4H, neben einem 386SX Brett und einem 486er Doner, was im Reset stecken bleibt.
Very useful and a good starting point, indeed! I got six dead AT motherboards, hope to get at least some of them working one day. Will study more from your videos. I guess I need to buy an oscilloscope.
It seems like this channel is the only one on RU-vid that repairs retro computers. There are others, like @RetroSpector78; I subscribe to these two channels, haha. I started college in the year 2000, still encountering EDO RAM and ISA slots. At that time, all I had was a Pentium 133Mhz with 16MB of RAM, and a 1.2G HDD that I partitioned into two. Watching this video really brings back beautiful memories, how much I miss those days, hunting from shop to shop looking for second-hand items, because that's all I could afford back then. It was still the era of 1.4MB floppy disks and CRT tube monitors. Now everything is online, if something breaks, you just replace it, no effort to repair unless the item is expensive.
Ah, there are many great channels. Adrian has been mentioned already, then there are tony359, bits and bolts, vswitchzero and many more great creators. Check those out if you are into this kind of hobby.
@adriansdigitalbasement, @bitsundbolts, @CRG, @Epictronics1, @miketech1024, @mrsolderfix3996, @necro_ware, @NoelsRetroLab, @RetroSpector78, @ScanLinesAU, @scrap_computing, @TechTangents, @The8BitGuy, @TheRasteri, @vswitchzero, and many more that focus on builds, testing, etc.
Probing right at the crystal can change the frequency - your scope probe will add some capacitance to ground. Better to measure at an output of the chip driving the crystal
Right, but sometimes it's not possible, because the crystals are connected directly to the chipset and used internally. I forgot to mension, that the crystals should be measured with high impedance probe.
Lucky star 5v-a1 motherboard sees only 1/4 of installed ram. 8mb of 32mb sdram pc100 stick. Ram is working fine on other machine. Any suggestions what can be wrong?
Today I just completed the repair of a 386 motherboard. (vogons: "some help to repair/restore a leaking battery ravaged motherboard" post) It is the first one I have done with such a level of damage and it took me a while. but when I was measuring everything for the last time before doing the first test, I realized that I had put the power connector on the motherboard backwards!!!!! ouch!!! It was thanks to verifying the 5V. When I went to connect the PSU I saw that they were on the opposite side, hahaha. Thank goodness I took that last measurement and my short-term memory jumped right away. Tomorrow I will unsolder them again. It is something that is a little difficult for me without a desoldering station. the the test will come, power and signals with a scope. I dont have a POST debug card. then with the CPU (will use a TI486DLC/40) and video card. fingers crossed
I have old 386SX board that has 10 beep error code, which means: "CMOS shutdown register read/write error". I've checked all traces around minor battery leak. They have continuity, but I've fixed one most damaged one with soldered wire. Still no luck. Thermal camera shows everything OK. Thanks to this video I've checked oscillators and ISA data lines using multimeter with Hz option. Not as good as oscilloscope, but it shows some activity at least.
This video is a great stating point for people getting into repair! I love how you put everything together so I can send this to people wanting to learn.
Do you recommend a video where you washed a sticky and/or very dirty motherboard? I have a few in that state and would like to clean them up "the right way" before trying any kind of repair. I will eventually watch all your videos because although I'm just a newbie on computer repair and some videos don't directly relate to my VERY basic repairs, they're still very interesting to watch 🙂
That's cool. And even more cool that despite being quite hard to find the old motherboards are much easier to repair. It was so until probably early Socket 7 motherboards with the small exception of the intel mobos of S5/S7 family like Zappa with plenty of SMD components.
Excellent video. Thanks for sharing! It would be interesting to see what pin on the CPU the broken trace connected to, to have an idea of why the board was completely dead. Given that none of the address lines had any activity, it seems like the CPU was completely halted. Maybe a halt, or interrupt, or ready pin or something?
broken trace and tantalum by cpu socket look like the result of someone trying to remove the cpu with the wrong tools. I'm guessing trying to lever it up with a screwdriver and they applied a lot of force and the screwdriver slipped.
Intel has bought the VLSI Chipset and used it. First it was a license but later owned by Intel. Its Intels first "own" chipset with a high integrated silikon and the start of all Chipsets made by the blue guys. Keep the Board if you like such Details :) I like the music douring the trace repair ! And im missing some insane Overdrive content ❤
Nice work as always! I have run into a few boards that act dead without a cmos battery, so that might be something to look out for. It would have been interesting to see what the post card would say in that situation, but didn't have one at the time. Thanks for the video. I always enjoy seeing what you have been working on.
Great walk through Scorp! Excellent idea to empower fellow retro aficianados who just started to collect old crap and want to learn to repair it themself ;-)
@necro_ware , do you have any experience, when graphic card is removed from ISA slot while 486 PC has power on and has fully booted... (waiting boot disk)? I accidentally did that, I did not noticed because system do not have any fans, I saw green led on MB later. How big damage this error could do? Now 486 wont do anything, I guess EEPROM is dead, not sure....
I guess, you were unlucky. It wouldn't necessarily damage anything, but you can get voltage peaks, which can end up bad. Probably, this is what happened to you. This howto should be a good start if you want to bring it back. Analyze the ISA bus and see what's happening
Always an awesome video by Necrware ! 🙂 What is that super video card you have there 5000+ on speedsys? My Cirrus Logic GD5422 is considered a fast card and only goes to 2300
Hi, thank you. It's a matter of the ISA bus speed and wait states. If everything is set for a faster throughput you should come to those values with the most real 16-bit cards. For the video I used WDC90C30 and with a CL-GD5420 I get similar values.
I just noticed on your logo/intro, the board looks like it's supposed to be a 16-bit ISA card, but aren't the components then on the wrong side of the board? They changed the orientation of the slot bracket relative to the card starting with PCI, so that the components face down in a traditional tower, but on ISA they would've been facing upwards.
Many datasheets for common ICs found on motherboards specifically mention to use tantalums. They often have a note saying if you use electrolytics it needs to be a much larger value cap. (This is most likely because larger capacity electrolytics usually have lower ESR).
In high precision high frequency circuits you usually combine many large electrolytic caps with some small ceramic caps. This is because large electrolytic caps have higher ESR, smaller ones lower. This topic is complicated, but on power rails of mainboards of that age, you can easily replace tantalums caps through electrolytic ones. The reason why they used tantalums back then, was because electrolytic were not as good as today. They leaked and dried out, but today tantalums are barely used, because they are too expensive.
Well, sometimes it's really luck, when the damage is small, as in this case, but in many cases it is actually quite a lot of work. Watch my Packard Bell PB450 repair series, that was dozens of hours of work.
Excellent video. Another thing to check for early in the troubleshooting process is jumper configuration, it has happened relatively often that I got boards that appeared dead or were otherwise malfunctioning that just needed to be configured properly. Sometimes even a single missing jumper can completely bork a board. A tip to more easily follow complex jumper configurations is to highlight the lines of the desired settings either on the pdf itself or on paper after printing the jumpers cheat sheet
hello NW. can I have a little help from you? can you take a look at this post in Vogons and give me any hint? "some help to repair/restore a leaking battery ravaged motherboard"
Thank you for this video. One question please. Is it not needed to solder the cristal body to that large pad? (or the body of the cristal is wired to one of the legs?)
No name from the next supermarket. You can use 10%, but it is quite aggressive. For beginners I suggest 3-4%, then you have 15-20 minutes to scrub and look around.
around $70 in Amazon, I see. some older models (by only 2 years, lol) of the same manufacture or similar ones, around 40-50. some have upgradable firmware
One thing about measuring the clocks with the scope: it should be square waves. On 2-pin crystals you should observe a sine wave, because it is a simple passive device. But the the places where you measure are most likely somewhere down the path through logic chips -> it should be a square wave. I suspect that your scope doesn't have the bandwidth to show squares waves at these frequencies (2 times more is a bare minimum, but it only enables us checking if the signal is changing). So if you want to check a 50MHz digital signal 100MHz is a bare minimum. 4 pin crystals: these are actually a crystal + a generator inside, the signal should be square wave at the pin already.
It's 32kHz, quite slow, should be fine. What you say is correct, but this oscilloscope makes it up to 33MHz reliably. I think, the crystal was faulty and I got some back leaking from the connected IC.
Sorry, I jumped into the conclusion that you are saying about all the clock signals. On the RTC signal it should be a sine wave, since it is a 2 pin device, right!
The ones that check for cache and stuff? Check Phil's Computer Lab for benchmarks and utils. He's done a good job of categorizing some of those and putting them on disk images for convenience. But they are widely available other places too. You just have to know what to Google for. Win World is another good resource.
@@nickwallette6201is there any mirror? dont know why I cant access the Phil's Computer Lab web site. I´ve got some tools and benchmarks from around but the Phil's pack may be better
Ремонт конечно простенький. Но само описание подхода к ремонту самое то для новичков. Жаль, конечно, что мало видео по ремонту старого железа попадается с подробностями. Да и в текстовом виде с фото их маловато будет
Спасибо! Да, так и есть, в данном видео ремонт сам по себе не хитрый, но я этого конечно заранее не знал. Меня просто часто спрашивают, какие шаги я предпринимаю, когда чиню что-то и я хотел показать, что я обычно делаю. Если вас интересуют варианты посложнее, у меня есть видео с разными уровнями сложности. Вот эта починка например мне лично особенно понравилась: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-i9DTA81xUN0.htmlsi=Kd9L-kmZAulpdJiG
I have fixed dozens of boards and other cards based on your methods! 👍 This is a very good tutorial! The TRW is also a cool project. I am contributing quite often. Thank you for your videos! Just one thing to add for the viewers - if you went through all the usual suspects and the board is still dead or acting weirdly - check ALL the pins of the big QFP chips. They go loose pretty often on those 30+ yrs old boards. Sometimes it is like the chip almost desoldered itself. Based on my experience this is a super common failure!
Thanks for doing the repair videos, You have a talent for teaching in a 'NON GEEK" terminology, using very relatable analogies of whats going on. I have really enjoyed your videos and learned a lot. I have some older boards I need to re-visit now. Great Content. 😂
Excellent video as always. Thanks for a end-to-end tutorial, but also all the other repair videos you've made so far. It got me into repairing my old vintage hardware with great success so far. Regarding the ISA bus measurement: I typically put the tip of my probe towards the top and outside of the slot, and gently press against the top from the pin from the outside inwards. This makes sure I never short pins, since my probe is not going that far into the slot.
god, i love it! for me, its mezmerizing, I could watch that kind of repairs for hours. its so good to safe these old boards. hail to the old days. cheers buddy and thank you for your time and patience!
nice. Last year I watched all your repair videos, and from other channels too, to soak me up and learn skills and tricks before starting my repairs and restorations. These past month I have been repairing/restoring several motherboards, cards and other parts, all at the same time, crazy, hahaha The one that has taken me the most work has been a 386 motherboard ravaged by leaky battery, maybe eating it away for years. I have had to reconstruct the veins on both sides. Of some only the footprint remained. Last night I finished with the lines, and today I will do it with a resistor array that received a hit and half of it was destroyed. Since I don't have the same 10x10k array, I only have an 8x10k array that I will adapt to replace the 5 destroyed resistors. I did something similar with a floppy drive with a destroyed array, I reconnected the common line and replaced the only destroyed resistor with a discrete one. At the same time, I have been repairing and restoring a PS/2 model 50, one part of which I literally rescued from the mud and trash. More than 20 years in dirty/dusty storage room left it absolutely disgusting. The owner, a neophyte, took it apart and slip the screwdriver three times, destroying several lines and 5 resistors. I already rebuilt the lines (difficult because they are very thin) and all that remains is to put the missing resistors. Thanks to having the schematics I know what its values are. It's a task for later. I always take a while to complete a repair or restoration, what scares me the most is the moment to test it, hahaha. and so I have others in the restoration queue, and all this for now without a debug card, or oscilloscope (I need them for some , I am waiting to receive them, the scope similar to this one). right now I have a day off, and are disassembling a Olivetti M4 (Pentium) that an acquaintance gave me. it need a deep clean and some fixes. I´ll keep it but will replace its guts for some older ones.
