I'm back at work on some motherboards, currently repairing and trying to repair various boards, most are Intel P.2-3-4 generation and AMD S.462-A and 754 equivalents.
Unfortunately, to repair some of them you need spare capacitors, not having them available I am forced to keep them aside, obviously they will be repaired when I have the spare parts.
Among the cards repaired waiting to be tested, as I wrote I have a Lucky Star K7MKLE, a first generation A socket, PC133 RAM and integrated video (without AGP), a card I imagine very economical times, but that apart from the lack of 'AGP, should be ideal for Home-Office use, I have installed a Duron 1300 and adding one or two RAM (max.), I should do a boot test, I hope everything goes well and works without problems.
I repaired another socket A-462, it is a MB for barebones PC, it is a BIOSTAR M7NBA, this card is of the latest generation for Athlon XP, it has DDR RAM and there is also integrated video here, even there are two VGA ports, here if you want you can install an AGP video card, being intended for a barebones PC it has connectors both on the front side of the card and on the back side. I have to see the supported AthlonXP type, I guess it works without problems with a 1700 + / 2000 +, unfortunately I don't have the heatsink originally used in that PC, it seems to use that of the P4 (478), and since socket 462 has the hooks for a traditional heatsink, I removed the basket where the skt478 heatsink attaches, I will use it on an Intel CPU motherboard, then in this board I will put a traditional AMD 462 CPU heatsink (almost all are compatible with the SKT7 and 370). As RAM I think I use a DDR333, it remains to be seen if it will boot and work, if successful, I could see if I change CPU to a faster Athlon XP, or if I get something more suitable for the type of PC, such as a low consumption processor, I am thinking of one for laptops or a Geode (?).
Having recovered the basket (or what is it called ???) to hook the heatsink (S.478), I put it on an ASUS P4S800-MX that I have to try, visually it is quite good, they are noticeable only in some places in the back of the areas darker (burnished), they are at the chipset and the mosfet near the AGP slot, at the moment I have not done further investigations to see if I find anything else strange, so I start the card without RAM and without CPU. The video card is integrated but if needed you can change it with a traditional VGA AGP by inserting it into the available slot, the card starts up but obviously does not show any error code in the post PCI card. Then I put RAM and CPU, a P4 2.4GHz and a 256MB single-sided DDR400, the PC seems to want to start but immediately stops on the error code C0 -, removing the RAM and starting with the CPU alone, the PC starts and it stops on code C1 C0 in addition to emitting a long continuous beep (repeated continuously), this makes it clear that the RAM is absent, but putting it back or trying others it returns to the previous error code C0 - (without any beep). Here I would be of the idea that you have to take a look at the RAM area, to see if the voltages are present, or if some capacitor is knocked out (even if visually ok), if instead everything is ok the fault could be somewhere else, but for the moment I don't focus on anything else, I don't think it can be a RAM problem, unfortunately I don't have double-sided RAM, but I guess single-sided ones are fine.

Among the S.478 motherboards (I have several), since the ASUS P4S800-MX has not started and needs further investigation, I switch to an HP motherboard, among the writings present the most significant are:
HP 845GE and NR138 VER: A06
later it will be possible to have further useful info (Product Name HP d230 MT), roughly it is in excellent condition without any visible damage, there is a dark area in the back in correspondence of two mosfets, but apart from this the rest is visually a given that most of the electrolytic capacitors are Rubycons, this could suggest that they have been replaced, since at that time there were many cheap (and poor) brands, if instead they are factory the motherboard turns out to be of quality .
The first test performed is the one with nothing on board, no CPU and RAM cards, obviously since the video is integrated there is still a video output, about this for those who are not satisfied with the i845GE video chip you can put an AGP card, also there is the integration of the Audio and LAN card, even here if necessary you can use two of the three PCI slots to put the appropriate cards. I also insert the Post Card and do not connect the speaker, since it is already present in the motherboard, I connect the ATX power supply, and I try to start by shorting the PWR-ON PINs, the card starts up, a sign that it has no problems and does not there is nothing in short, this obviously prevents possible damage to the CPU and RAM (in case of bad luck).
I connect both the CPU (with heatsink and fan) and the RAM, I try the start again, on the post card various codes start to appear, but unfortunately it always stops at some point, the most commonly displayed is 27 25, I do a test without RAM and confirms that it works (audible beeps), I try various RAM and in the end I put a couple of 256MB DDR333 (total 512MB), after several attempts I get a screen on the screen, honestly I notice it when the PC freezes, maybe it was preceded from others like the HP logo, however after several attempts it seems there is a bad contact somewhere, or maybe it is an imperfect weld, when it manages to start it works but I don't know how long, it could be fine for minutes or hours, but not knowing exactly where the defect is it is impossible for me to fix it (re-soldering), in the past I have read about defective 478 sockets, this is of the type with surface ball soldering (BGA), if a ball is missing you should use an equipment ra specific to fix it, even on the ASUS P4S800-MX there could be the same problem, since the S.478 uses the same type of ball welding, I know that some use a domestic oven to solve the problem, or alternatively you could use a hot air dryer, but for both methods, if you have no idea how long and what temperature to use, the result may not be what you imagined, and ultimately if the procedure fails, the card could be permanently damaged.

So while waiting for the Asus A7V600 card to arrive, I went in search of some cards that I keep aside, they are difficult to repair cards, some have missing smd, so the only way to fix them is to find the missing smd, the ideal would be to have a second board for comparison, after that there are other works to be done, such as replacing capacitors and repairing some damaged (because scratched) tracks on the back.
Among the cards there is an Asus P5V-VM SE HD, at the moment it is not repairable for the reasons mentioned above, but having an ASUS S.AM2 card in repair (M3A), I make a comparison to see if I can recover some capacitors , luckily on the AM2 there are only 3 (out of 4) capacitors to replace, while on the P5V-VM I find five in excellent condition from the Panasonic (1000uF 16V), so I remove four and change them all on the M3A (even if one is good).
The M3A therefore only needs to be tested, in theory the repair should put things back in place, and the card should be able to release all the power of the installed cpu, in this regard I have an Athlon64 LE I think from 2GHz, I think I use that for see if the repair is successful, for the P5V-VM instead I put it waiting to find a genelle, or possibly if I find some detailed image of the back, I will be able to understand what is missing and in that case I will have to get some other capacitors to put them back.
Among the other S.462 cards, there are other ASUS cards, one is quite difficult to fix, another maybe it is not worth repairing it, finally there is an A7N266-VM whose repair requires some effort, some tracks on the back they are scratched, but it does not seem complicated to fix them, while most of the large capacitors must be replaced, as I do not have new ones, I could use some recovered from other MBs, they are from OST, a brand known for its poor quality and durability, but perhaps considered the attempt to repair it and the chances of fixing it (not very high), since from checks they appear to be in excellent condition, I would say that you could put them on and see if the card comes back to life.

Today I start a repair job on a perhaps recoverable motherboard, as I wrote earlier is an ASUS A7N266-VM REV. 1.05, the problems to be solved are the following:
All 3300uF 6.3V capacitors must be replaced, because almost all swollen, those visibly in place, after removal and verification were in poor condition, some even exceeded 10000uF, the latter value well beyond the tolerance. I also change the two from 1500uF 6.3V, because one is ruined from above (it has suffered several bumps), both tests turn out to be a little above the tolerance, more than 1700uF are detected, so I also change these with others for recovery.
For those of 3300uF use of the capacitors of the OST, recovered from an ASUS for P4 478, I check some of them to see what state they are in, the measurements show ESR values ​​of 0.10 OHm and variable capacitance from just under 3300 to about 3700, therefore I guess they are fine, eventually if I can fix the card (it works again), I can change them with others of the same quality (at the moment I don't have any available).
Those from 1500uF 6.3V are replaced by Teapo 1500uF 16V, despite the higher voltage they show quite similar values, both around 1500uF, so ok, the ESR is always about 0.10 OHm, so I install them, but the same speech is valid. first, if all goes well I can change them with others of the same quality and voltage as the originals (6.3V).
I had previously fixed a couple of semi-detached SMDs, re-soldering them, but I also notice a capacitor with bad soldering, soldering the balance recovery ones too.
Normally the work would be finished, but here unfortunately in the back there are some places where the tracks are scratched (and some are broken), the breakpoints are not very visible, through an enlarger you can see that the copper trace is ruined , sometimes it is blown away, sometimes it is quite ruined but it is not clear if it is broken or not.
I start with the repair of a broken track near socket 462, the copper draws an S and just in a curve it is interrupted, I use an abrasive paper to scratch the paint, unfortunately I forget to protect the neighboring tracks, in the end it is not a problem but I would have gladly avoided stripping them, I try to tin the copper to try to join the track where it is broken, unfortunately at the moment I cannot because the dimensions are really tiny, it is not easy to solder to the eye, I will try again with another soldering iron with a slightly larger tip , maybe it is possible that a ball of tin can join the interrupted section, unfortunately, not having professional equipment, I can only try until I can connect the points where it is broken, while maybe this could be done without too much difficulty.
If I can fix this S, I try to clean some other area to see if there are breaks, where there are more nearby traces I hope the scratch is superficial, but we will see, in case for those broken I think to make a bridge, it would be easier solder distancing a little, and if aesthetically it will not be great, it does not matter, it will be fine for me if you restore continuity.
Finished the work on the tracks, you have to try booting the board to see if it works, if the no start problem was due to the capacitors, everything should be fine, unfortunately the track problem in the back could make the repair a little more difficult, but resolving and resoldering the interruptions, it may go back to beep and display the BIOS screen on the monitor.
As a CPU I think I use a Duron of about 1GHz, 266MHZ DDR RAM, I have aside the heatsink for the chipset, I think I disassembled it long ago, to use it by adapting it to some old CPU, you don't need the AGP card since the video it is integrated, I have already downloaded the manual to set the jumpers correctly, there is some confusion for BSEL0 and BSEL1, it is not clear if the PIN 1 is on the right or on the left (that is, if this is the case: 123 or 321), but at the limit the RAM would go to 133 MHz, while the CPU if it tolerates overclocking it would run with FSB 133 (instead of 100), I think it is possible with a Duron 1000.?

I continue with the repair work started for the ASUS A7N266-VM REV. 1.05, now there is only some soldering to be done on the back of the board, some tracks are scratched, it is necessary to reconnect the points where the copper is no longer there.
The first one I try to fix is ​​the one near the CPU, which draws an S, yesterday I could not connect it, today I try again and after a few attempts, I can reconnect the track with plenty of tin, a curve of the S has been lost because it is full of tin, I do not know if it is important, but I think that the priority was to restore the point of interruption, and this in the end is what I did, if I wanted I could try to re-weld, but with the equipment I have hardly available I would be able to do better, if having eliminated a piece of S would cause operating problems, then I could try to re-weld, hoping to do better, to connect the ends of the interrupted track.
I move to another part nearby, under a hole near the CPU, next to a bigger trace, I find a thinner lower one, obviously broken, this time I protect those nearby, and I start to sand the broken one, the point becomes visible of breaking, I weld the copper of the track with tin, after several attempts, also here with abundant tin, I manage to join where it was broken.
I take pictures to see the result, but surprise! The image highlights other traces to be placed a little above, three thin very close and two more close above. It seems that they are going to the CPU PINs, surely you have to fix these too, but there are other nearby areas with possible breaking points, in total there could be a dozen interrupted points, I have to understand if it is better to use a different method, that is instead of doing the tin bridges, perhaps it is better to make them with an insulated copper wire, soldering somewhere else on the track, aesthetically it will be unpleasant, but as I wrote before you just need to reconnect the breakpoints.
In the meantime I found a chipset heatsink, it was if I'm not mistaken a P4, it has the same dimensions and connection points (same distance from the holes), just clean it up and do the same on the NB chip, there is a kind of pink rubber, not it comes off with alcohol, I try to scratch it but there is still a thin layer that I can't remove, maybe it will still be fine, but I don't know how to remove it completely without damaging the surface of the NB.
Yesterday after a few days of waiting, the ASUS A7V600-X REV arrived. 1.00, it is in excellent condition, a couple of capacitors are a little bent but it doesn't look like anything repairable, I think it is enough to straighten them, it has an unspecified Athlon CPU on board (frequency? FSB?), The pcb visible under the Ii is brown, I hope it is one of those with FSB 400, because I only have FSB 200 CPUs 266 and 333, I don't want to disassemble the I said to see what it is, I hope it can be a Barton 3000+ or ​​3200+, in any case we will see what will read on the first restart in the BIOS screen.
There are both the CPU (fanless) and the memories, two 400MHz DDR modules of 256MB each (512MB Tot.), Have on an Olidata label, this makes me think that the MB comes from an Olidata PC, they mounted in that period especially ASUS motherboards, the fan for the CPU I could have it, but in case I would adapt one of other sizes.
I thought it had the integrated video card, but it isn't there, maybe it's better this way, I can see among the few available if I find a suitable one, I think I have an ATi 9200 and a Radeon 7000, maybe the 7000 is already good for that PC , in the next few days I gather everything and try the start, maybe I try to try some other cards, having already some of them ready for some time.

Unfortunately, after a few attempts on the Asus A7N266-VM, I can't make progress, I re-soldered some broken tracks, it was not easy, but abounding with tin I can somehow reconnect them. Those that I cannot connect are three side by side in a space of about 0.75mm, the single track measures about 0.1mm (or a little more), the point where it is broken and the proximity to other nearby tracks, makes it very difficult to repair, I wanted to solder two of the three, the upper and the lower one, using the same system, scratching the paint of the tracks so that the copper is exposed, and then tinning with the soldering iron (normal type), I can only get the tin to the tracks, but I can't make a tin bridge to join the interruption points, probably for this job it would be ideal to change the system, that is to use an air welder, but not having it available I would like to be able to weld with a traditional welder. I think I'll try again in the near future, maybe I could facilitate the union, with a single very thin copper wire soldered on top, but it's not easy to find such tiny ones, maybe those of a transformer winding would be fine, among those I have available I don't find so thin, I will try to use one that somehow can be attached near (or perhaps above), in order to restore the connection of the upper and lower tracks, for the central one instead, I will make a bridge with the wire to connect it a little distant with tinned traces. Finished here I might have luck and finish the repair, but there are other points to check, if I don't find other breakpoints, I would be fine, and I could try booting the motherboard.
On another similar motherboard (ASUS A7V400-MX), I need to find a recovery component that is fine, I need to replace an SMD (D5), it would be a diode with 3 pins, I don't know what was written on it, but I could recover some someone from scrap cards, unfortunately they are not Asus cards, because in that case I could have found an identical one (having luck), I think I can put one that can be adapted, but it is not easy to understand from acronyms (like KL 4 or L42) that diode either, they could be the same or completely different. Here too there may be some broken tracks in the back, from what I see maybe two or three may need a weld, I'll take care of that after putting the D5 back in place.
As soon as I finish the repairs of these two ASUS S.462 cards (hoping that I can get them to work), I will take care of two or three more ASUS S.462 cards, for the moment I just took a look, besides those I shouldn't have any other S.462 for repair, so I'll try to fix some S.370 or Slot1 motherboards, most have simpler problems to fix, and I hope no traces in the back to be soldered.

PC@LIVE wrote on 2022-09-29, 19:52:

...
I repaired another socket A-462, it is a MB for barebones PC, it is a BIOSTAR M7NBA, this card is of the latest generation for Athlon XP, it has DDR RAM and there is also integrated video here, even there are two VGA ports, here if you want you can install an AGP video card, being intended for a barebones PC it has connectors both on the front side of the card and on the back side. I have to see the supported AthlonXP type, I guess it works without problems with a 1700 + / 2000 +, unfortunately I don't have the heatsink originally used in that PC, it seems to use that of the P4 (478), and since socket 462 has the hooks for a traditional heatsink, I removed the basket where the skt478 heatsink attaches, I will use it on an Intel CPU motherboard, then in this board I will put a traditional AMD 462 CPU heatsink (almost all are compatible with the SKT7 and 370). As RAM I think I use a DDR333, it remains to be seen if it will boot and work, if successful, I could see if I change CPU to a faster Athlon XP, or if I get something more suitable for the type of PC, such as a low consumption processor, I am thinking of one for laptops or a Geode (?)...

Among the cards already repaired but never tried, I have some that I will try shortly, among these I have a Barebone PC card, it is a BIOSTAR M7NBA SKT 462-A, since I have a free space on the bench I decide to try it, I put on a DDR 333 RAM (256MB), an Athlon XP 1800+ CPU, and nothing else, as video is integrated, as well as LAN audio, etc.
Starting, the post card displays various codes, stopping at code 41 10, a couple of screens are displayed on the screen, the first concerns the info on the video card, the second is a screen similar to other views not so long ago. The message in which the PC freezes is the following:
BIOS ROM checksum error
Detecting floppy drive A media...
Drive A error. System halt

Then it starts to beep endlessly, and this reminds me that among the cards fixed not so long ago, there was the PINE that had the same problem, solved by rewriting the BIOS.
I will start looking for the correct BIOS of this BIOSTAR, I hope I will not have difficulty in finding it, probably someone has updated it with an incorrect version, or it could be that there was some problem and that it is ruined, creating this error in the ROM .

After a bit of research it seems that there is no BIOS available for this motherboard, I persist in the search but I do not find any file online that could be useful to me, I start the search again and finally I arrive at a site full of files for old motherboards, unfortunately even here the BIOS files do not lead anywhere, but in addition to those there are exe files, fortunately these can be downloaded, and the content is just that of the BIOS, so at this point I download the files and the final step will be the rewrite of the chip.

Regarding the BIOS of the BIOSTAR M7NBA motherboard, I solved it by reprogramming the chip, using a file found on the web (elhvb.com), I randomly put one of those available, I later discover that it is the most recent one, I had understood that the problem was due to the BIOS (previous experience), the rewriting was not problematic, and also quite fast.

Now it displays three screens at startup, the first with information on the VGA (integrated), in this case nForce chip and 32MB RAM, the quantity is variable, the value can be increased or decreased depending on the RAM used in the PC, lower values they are ideal when RAM is low. The second screen is the full-screen BIOSTAR logo with the Nvidia logo on the bottom right (small). The third screen is the usual main screen of the BIOS, with information on the BIOS and errors detected, from there you can then continue by pressing F1 or enter the BIOS (CANC).

The positive thing is that the continuous beep disappears, now there is only one at startup, entering the BIOS via DEL, the main screen shows a menu that is not at all poor, on the left you can intervene on almost everything, on the right instead there is only the possibility to vary Frequencies and Voltages, the rest are screens for psw - exit - L O D, there is also a line for Upgrade BIOS, and I think that someone went there to do something but the result was not what was hoped for.

As I said I used an Athlon XP 1800+, at the first start it is recognized as 1150MHz, obviously the frequency is low, you have to go to the BIOS and bring the FSB from 100 to 133, among the settable values ​​there are also 166 and 200, this however it does not guarantee that at most (ie 200) the other PCI AGP and other type values ​​are correct.

HERE we talk about it in enough detail: https://bjorn3d.com/2003/10/biostar-ideq-200n-sff-system/

Today made some progress with the MotherBoards under repair, having several waiting to be tested, I choose one that I had recently repaired, but had never been tried.
It is a LGA 775, ASUS P5PL2, it already had a CPU installed, a Pentium 4 2.93GHz/1M/533, in this board I replaced a capacitor in the PCI-E area (470uF 16V), because it had a high ESR, in this card does not have integrated video as in many other similar ones.
Then I add a PCI-E video card with ATI chip, a 512MB 533MHz DDR2 RAM, and the intel cpu heat sink, I also add the usual POST-CARD, connect the keyboard and power supply, and try to start by shorting the two PWR pins -ON.
The fans start but strangely the codes of the POST-CARD are stopped on -- --, however on the video I have the startup screen and a beep is heard, a sign that the PC has started up, I reset from the keyboard and this time the codes appear on the POST-CARD, I enter the BIOS to check the values ​​detected in the HW monitor, everything seems fine, so I decide to switch off to connect an IDE-CF adapter.
Upon restarting the PC detects the drive and the PC starts the DOS contained in the 256MB CF, from there I start some diagnostic and bench programs, obviously I get results, which I cannot evaluate if they are good or not.
I continue with the RAM update by putting a pair of 512MB RAMs (still DDR2 533), in total I get 1GB of RAM, but when I start the PC it tells me that it is not Dual channel, so I switch off and change places to a RAM, restart and this time there is the Dual channel.
The next step is to add two more 512MB RAMs, and bring the total to 2GB, which should be the maximum that can be installed on that motherboard, then I could change the CPU with a Pentium D, I only have one of these CPUs, a 3.4GHz /4MB/800, I think it could be a little better than the 2.93GHz P4, both for the higher operating frequencies and for the two cores (versus the single one).

Today in my free time, I went back to dealing with the s.478 card, I found two more rams among my stocks, one identical to the pair already installed, the other of a different brand, it is however always 512MB at the same speed (533 MHz) and same CL, I start by adding the same module as those already installed previously, when I start the PC the total memory rises to 1.5GB, and the Dual channel becomes asymmetric, I start DOS and actually the memory has increased, but for the At the moment I don't have an HD I can't notice any visible differences.
I continue and add (obviously with the PC off) the fourth module, even if of a different brand it is still detected, the amount of memory rises to 2GB, and should be the maximum for that MB, the dual channel switches to Interleaved, the PC is therefore with all the memory banks full, the same argument made before is valid, I don't detect any visible differences.
The next step will be to connect a HD, and install a Windows, I have to decide which version, but this depends on the capacity of the HD, after that I will have to install some programs to start the benches, currently the CPU is a P4 2.93GHz/1M/ 533, I'm pretty sure overclocking is possible, if you can handle it you could go up to 3.67GHz going from 533 to 667 FSB.
The next step will be to change the CPU to a PentiumD 3.4GHz, for the rest the only thing that could be done would be to change the VGA, I currently have an ATI but I don't remember the model, it's still a fairly cheap card, nothing special , in W7 it achieved scores of 2.0 and 3.0, or slightly higher.

I'm also working on socket 7 MBs, unfortunately I only have a couple to fix, but considering that about 25 years have passed, finding similar ones today is quite difficult, the cheap ones have long since disappeared, you can find them at repair but they are not cheap, while the working ones have quite high prices, even those with AGP and FSB 100 or more, cost even more, especially if they are famous (and quality) brands.
I wanted to get an idea of ​​what to use in those cards, i.e. HD CPU RAM and related frequencies and capacities, as well as of course the type of VGA to use (undoubtedly PCI for intel chipset), but also for everything else, sound card or other cards add-ons, power supplies, etc.
By chance I have an old computer magazine, I had recovered it last summer together with various HW, I choose this one in particular because it says "25 motherboards" on the cover, I get to interesting articles on CPU BIOS CMOS and more, finally towards the last pages I find the motherboards under test, among these I choose a couple, because I have the same faulty ones, they are in AT format, or rather Baby AT to be precise, obviously I chose those with intel 430 chipset.
The next step will be to put one on the bench, and check the voltages of the VRM, I start from the card with VX chipset signed P5F76, I put an image, the card was working, but after a change of RAM it no longer started, I'll see if I can find something out of the ordinary in the voltages (RAM included).

Returning to the two socket 7 motherboards, I begin the attempts to resolve the fault, on the Freetech (Falcon) P5F76, a 430VX with linear voltage regulator, the board is aesthetically in good condition, I can't find it visually damaged, just a PIN of a chip is a bit crooked, but it was already like this when the card was working, moreover it doesn't seem detached and it doesn't touch the nearby PINs, unfortunately the last attempts to restart it were also useless, the post card doesn't stop displaying - - --, I don't know why starting without RAM or CPU (or both) it doesn't make any beeps, usually the lack of RAM makes you emit a few beeps, here instead nothing, while years ago it used to beep.
I still haven't figured out exactly what the problem is, maybe but I could be wrong, you should check the PowerGood, maybe it's the one that blocks startup, with other tabs I have similar behavior, post tab displays -- --, sometimes you just need to change RAM and CPU and the PC starts, it's useless here, I tried a P133 (working) and an ISA video card (ok), on another PC they work, on the VX they don't, the only thing left is the RAM, I have tried both 72PIN modules and DIMMs, nothing, both with chips on one side and with chips on both sides, nothing.
The only thing I would like to do is check some PINs of the ISA PCI and RAM slots, in my opinion the CPU is powered correctly, I have tried various combinations of different voltages, and I detect the set voltages that are present, I don't know exactly in which PINs look up, but I've done similar checks in the past, so I'd have to go look up what PINs had voltage.
Of course, any suggestion could be very useful.

Moving on to the second socket 7 card, it's a Soyo SY-5BT (5BT5), with TX chipset and switching power supply, it looks like it could be the latest REV. of that card, since it has 12 PINs (instead of 10) to select the VCore, the two added PINs are used to select the 2.2V, while before the minimum selectable was 2.8V, except that there was a combination of jumpers not listed in the manual.
I start the repair by re-soldering both the battery holder and the keyboard socket, other than that there would be a missing chip cache PIN, but for the moment I don't fix it, first I want to see if the board works or what problem it has, so I connect all the needed for booting, I use a 64MB DIMM, a P133 and the ISA video card, as well as the necessary, i.e. power supply - video - keyboard, set the jumpers and switches, I try to boot, but nothing! The post card scores -- --, I turn it off and I realize that there is no battery, I put one in and try again, it starts!!!
The first screen appears, and it reports some errors, nothing strange apart from the keyboard one, it doesn't seem to work, I try starting again by putting another one, but nothing changes, so after various attempts and checks, I manage to enter the BIOS and then to solve, I memorize and try to start, unfortunately it freezes on the first screen, I already know what the problem is, the cache, I enter the BIOS and disable the external one (L2), save and restart, everything seems ok.
I connect the usual CF-IDE containing a DOS, and change VGA, put a PCI, the PC starts up and I can try some programs for info on the system.

Image DOS Program

I can't tell why, but today Soyo 5BT5 didn't boot anymore, except once, this makes me think, because this no boot problem is same in other motherboards.
If it were a failure of some component, I don't think it could start, except that it was a bad contact, but thanks to the POST card I have some useful info on what's wrong, taking for granted what can be seen from the LEDs, it seems that the reset has a problem, because the LED stays on (and doesn't flash), to see the difference I put a photo with the PC started up (ok) and with the PC not starting up (KO).
Everything I use seems to work, because I use them in other PCs and they go, so I would rule out a CPU RAM problem or something else, I could think of a missing voltage problem, but in the five mosfets (or similar) I have normal voltages, as far as it concerns the electrolytic capacitors, they are visually in order, but it is not said that by checking them they are in order, regarding the quartzes, I don't know if in the post card the CLK led indicates if they work (or not).
As for the BIOS, since I don't have a programmer for that chip, I don't have the possibility to make a second chip, and understand if what's there is a defect.
I have noticed that the motherboard is not beeping, no RAM, no CPU, and no RAM and no CPU.
I've tried booting with 72 pin RAM, but nothing changes, I'm guessing it's not a RAM issue, only thing I can do is switch RAM and CPU and see if it boots, but I don't think it will.
I would appreciate some suggestions, if you have ideas on what to do, I reset the BIOS via jumper, but there was no change, from what I see it is as if at startup, the reset LED remains on instead of flashing, other LEDs are on, but normally I don't know if they blink or stay off?.

Go understand why, without having done anything, the Soyo 5BT5 motherboard restarted, like nothing it came back working, I just waited more than 24 hours since the last failed boot, now it always starts, I did some benches (DOS) and Tried diagnostic programs, board has nothing wrong with it, other than a cache chip with pin 8 missing (which it already was).
To allow booting I disabled the external cache (L2) from the BIOS, this affects the speed of the PC (it slows it down a bit), but it allows me to use it, even if now, the configuration is quite bare bones, the final one will have more cards, more ports and more drives, I think I'll add a sound card (to decide whether ISA or PCI), LAN card, USB ports, 3.5" 1.44MB floppy drive, maybe I could add a Voodoo type accelerator card, but that depends, first you need to fix the chip cache.
Having this PC working, I can start a bench marathon, with Socket 7 it is possible to install CPUs of different brands (something impossible today with today's PCs), among those available I chose: CPU Intel P54 i200, Intel P55 P233, AMD K6 300, AMD K6-2 400, AMD K5 166, Cyrix 6x86 166, Cyrix 6x86MX 233, Cyrix MII 333, IDT C6 200.
I will use these CPUs to do a series of benches, but to simulate the lower frequencies, or I will lower the FSB or the Multi, if you want it is possible to go up the FSB up to 75MHz or more, but it depends on the CPUs, some have no problem going up, others maybe not, in any case I don't want to exceed the CPU frequency, for example a P233 (3.5X 66) I could set it to 225 (3X 75), I know it can also go to about 262, but the initial purpose is to test the frequencies lower than the factory ones, then later you could try the higher ones.
Unfortunately I don't have a K6-III with integrated L2 and 2.2V VCore, I have a couple of 2.4V 400s, and this board has jumpers to set only 2.2V (or 2.8V), there is no intermediate setting, in the manual there is no way to get it, so if by changing the position of the jumpers it is possible (?) to have VCores not listed in the manual, for example to get 2.4V which jumper to move or add (or remove?), maybe if there is someone who has suggestions, for me they are very useful, they save me from experimenting and trying at random, with an unsafe outcome and relative waste of time.

Some images, using the same CPU at various frequencies, 133 120 100 MHz, on SOYO 5BT5 with L2 cache disabled

With the SOYO 5BT5 I haven't made any progress, I think it's time to fix the cache, you have to reconnect a pin (n.8 😎) from the mb track to the chip, the problem is that it's broken flush with the chip, so you have to do a milling work (from above) of the plastic above, let's say something similar to what is done on the Dallas chips, only that the PINs there are bigger, and therefore it is quite easy both to solder them and to discover them, here they are both tiny and very close , but it costs nothing to try, we'll see if I can solder a new pin, I'm pretty sure the cache can work again.

Today in my spare time I worked on Gigabyte, it's a socket 462 (A), to be precise it's a GA-7ZMMHC REV 1.0, I wrote about it some time ago, it was waiting for a change of capacitors, 7 (out of 13) of the 1200uF they were swollen, all mostly near the RAM banks, honestly I didn't try the card before disassembling them, so I don't know if it worked or if it had any problems, however I decide to change even those that visually look fine, and in fact some are not they were bad, but they had the capacity out of tolerance, changing them may have no effect, but will avoid future problems (soon?).
The removal was a bit complicated, some had very hard solder, adding more new solder did not improve the situation (since that too hardened), but despite various attempts, three holes remained blocked, this is a problem that it happens every now and then, I thought it depended on the soldering iron, that's why I bought a new one not long ago, but even with this new one I have the same problem as before, i.e. the tin doesn't melt completely, on the surface just a little but underneath it stays hard, perhaps it depends on the vacuum pump, it doesn't have a suction cup, it has a hard plastic spout, and it doesn't adhere perfectly, there are still some micro-cracks that could considerably reduce the suction capacity, however despite various attempts it was not possible to free the holes .
Years ago I read that a micro-point for a mini-drill could be used, with a diameter of 0.8-0.5mm, I looked for a similar one, but here in the area, the minimum size I found is 1mm, i.e. it's too big , and that's not good, I think there are other techniques to free these holes in a short time, and without damaging the seats, but unfortunately I have no idea how to do it, so I proceed with a long and slow wear of the pond (with a similar needle) , the end result is not like the free ones, but with a lot of patience you can create a hole large enough for the capacitor pin.
After freeing the holes, I can proceed to soldering the new capacitors, I found some Nichicons of 1200uF 6.3V, I think they are even better than the ones used in the factory (G-LUXON), but this operation takes time, and today I have some available, I think tomorrow I can find the necessary time, but I have two other boards with capacitors to change, one is Gigabyte socket AM2, the other is ABIT socket 370, the simplest one should be the Gigabyte, while the ABIT it has 18 capacitors to replace, plus perhaps a few more (which must be checked), but I'll talk about these two boards later, when I start the interventions.

Today I finished the work on the Gigabyte skt 462, I soldered all the new capacitors, a decidedly simpler job, but not so easy in some points, especially where there was hard tin, definitely not the best aesthetically, but it doesn't matter because the contact is there (at least in the back).

After that I wanted to start the replacement work on the ABIT SE6, but I didn't have enough time, so I had to postpone it until tomorrow (or who knows when?), the board is interesting, I've already repaired a similar one (SA6), this one has the integrated video and one RAM slot less (3 instead of 4), the maximum installable RAM is however limited by the chipset (815) to only 512MB, at least according to what is reported in the manual.

I keep finding interesting info on the Soyo 5BT5, I noticed that to add the 2.2V there is an smd resistor signed 393 before the jumper, the exact value if I am not mistaken is 39K, well on the PCChips M571 the 2.2V are obtained with an additional resistor of 39K , therefore equal, and it is in parallel to R79, this last thing is not very clear to me, but I have to see if I can find a diagram of M571, in order to see what value R79 has, and if there is one equal in the 5BT5, I guess so, because they look quite similar to me (the two motherboards) in this section.

I've been repairing several motherboards, as long as they have swollen capacitors, it's quite easy for me to fix them, sometimes they need other work, trace repairs, lost smds, bios to rewrite, sometimes you just need to find the correct jumper settings, or the right RAM.
Currently I did some work on various motherboards, in a Gigabyte GA-7ZMMHC, I replaced all the 1200uF 6.3V electrolytic capacitors, I had a bit of difficulty with some holes, which I could not free, because the tin has hardened and I could not melt it.
It's a socket 462 (A), it currently runs on a Duron 1000MHz, for RAM I can only use PC100 or PC133 DIMMs, I believe it's an early generation 462 CPU card, it has integrated video, I'm pretty sure you use part of the RAM as video memory, there is also onboard audio, but no LAN, to add it you need a PCI card or find a CNR.
Previously I restarted a socket 7, SOYO SY-5BT, unfortunately it has a cache problem, one chip PIN (No. 8 ) is missing, not having a spare chip, I have to try to do a job, like that which is done on the DALLAS chips, in practice you have to remove a little plastic (above), to weld a wire, which restores contact with the motherboard (of the missing PIN).
It's not an easy job, if you have any ideas or suggestions, I would greatly appreciate them.
Today I worked on an ABIT SE6, I removed all the capacitors (from 1500uF 6.3V), in total there are 18, although visually OK they were all mostly in loss, one with ESR off scale (>40 Ohm), and a 'other open, I haven't soldered the new ones because I want to check the smaller ones (from 100uF 16V), there may be some to change, but I'll know after I test them, I hope they are ok, even if the brand (JACKCON) does not bode well, being the same as the ones I changed.