For today it is scheduled, a test with the Chicony iDX4-100 PC, I found a 4 cm fan 🌬, which I will connect directly to the molex of the drives, unfortunately in this type of motherboards, there is no socket for the 3 PIN FAN connector, or 2 PIN, these appeared if I'm not mistaken on the S.7, in short, a few years after the 93-94, of course anyway at a certain point, the higher CPU speed, made the PCs louder because the fan was essential for operation without any problem, but initially the systems for hooking the Heat sink, they were not the best, some could easily disengage, because in fact there were four clips, and they had to be hooked directly to the CPU, many then did not use the thermal paste.

For the test I will use a game like Duke3D, currently I have loaded the drivers of the SB Vibra16XV, in DOS everything is fine maybe I should just turn up the volume a little, but I can no longer access Windows 3.1, I will have to try to understand if there is a problem, sometimes I remember that you had to do several tests to be able to make the audio work on Windows 3.1, from this point of view the subsequent Windows work better, and indeed for this kind of sound cards, the drivers are already included, and you don't have to load any floppy.

Among the DOS driver files that I installed, there should be a configuration and test program, which could be very useful, first I should see the current configuration, and this could be useful, in case there is a conflict with something else, in that case it would explain why Windows no longer works.

I'll try the same audio drivers on the PCCHIPS M912, although I don't remember if the audio already works, and therefore in that case it wouldn't be necessary to load them, because initially I didn't connect the speaker jack, but once I'm done with the file transfers, I'm trying to make all the installed HW work better, unfortunately for some things, since it's a dated HW, in case of operating problems, you have to try another one that is definitely working, and if the result is a problem with the card (audio, video, etc...), getting the spare parts today, it would be very expensive, some cards like the video ones, can Be very difficult to find for sale.

Here we are again, this morning I connected the fan on the DX4-100 (MB Chicony), I put a thin layer of silver thermal paste, and I started the PC, so far nothing strange, then I started a short Game session with Duke3D, initially I let it run in demo mode, then I started playing a couple of levels, in short I wanted the CPU to heat up a little, and for the moment I haven't seen any locking, everything has always been smooth as oil (!?).

Yes, because in case of problems, you need to understand what causes them, it's not that maybe, solved the thermal problem of heat disposal (CPU), other problems occur due to something else, however usually if we solve the problem that causes the PC to crash, this will work well again.

So now I can say that in the mid-90s, some problems that manifested themselves were not due to a HW defect, but to poor ventilation, or to the assembly of the heat sink without thermal paste, there were even those who put labels 🏷 on the CPUs, I've seen this done even on P4s, however today it is quite easy to remedy it, if we don't find heat sinks suitable for 486 CPUs, you can adapt those of the Chipsets, which are slightly smaller, even if much higher and more efficient, and maybe it may not be necessary to connect a fan.

Currently, the Chicony PC has two SIMM 72 PIN of 8 MB each, in total there are 16 MB, at this point I am considering whether to add another 16 MB, to have a total of 32 MB (8 MB X 4), in this way I can say that the PC would be expanded enough, not at most but not so far, the maximum would be CPU S.3 133 MHz and 64-128 MB RAM, but I don't know I prefer to use the S.3 133 MHz on a 486 PCI, which together with a VGA PCI, and a lot of RAM, could be more suitable for Windows 98SE.

But that will be the idea, which is possible to realize, only after running the only 486 PCI motherboard I have, initially I will start with an AMD 486DX4-100 CPU, it would be slightly slower than the iDX4-100, because it has a cache of 16 KB (instead of the usual 8 KB).

Later I have the possibility to upgrade, but a lot depends on how the PCI works, some go at the same frequency as the FSB, others instead go to half frequency, so it can happen that at FSB 40 MHz, the PCI goes to 40 or only 20 MHz, the first a little high for some video cards, the second too low, and could slow down the PC a lot.

Still on the subject of 486, I performed some tests on the Chicony with iDX4-100 (it's no longer called 486), I tried Windows 3.1, well it works but without audio, so I think it's necessary to load the drivers directly from Windows, it's a job that was done via floppy, but in Windows there are already some audio drivers, like Sound Blaster or other similar tabs, so I have the doubt that it's enough to select one of those, rather than using a floppy with the drivers.

The ISA Audio card is SB 16 Vibra, I think it is the last version produced for ISA, if I'm not mistaken later we switched to SB 16 PCI, here clearly you are obliged to use ISA cards, because the MB is ISA-VLB, I point out that in DOS the audio works, so unless there are strange problems, it should also work in Windows 3.1.

I haven't closed the PC yet, because I upgraded to the RAM, I had two RAMs of 8 MB each (16 MB total), I added two more of 8 MB, now in total I have 32 MB, I already know that someone might think, that now you could also update Windows, moving from 3.1 to 95, but for the moment I would like to keep Windows 3.1, the disk is just over 500 MB, I wouldn't want to fill it, I prefer to keep a lot of free space, of course if I wanted I could use the program that doubles the disk capacity, in the past I used it, But in reality it depends a lot on the files, some can be compressed a lot, but if most of the files are compressed only a little, the doubling of capacity is not actually achieved.

So for the moment I would say that, this PC solved the problem of audio drivers, can be considered as a definitive PC, that is, there is no need to add anything else, but maybe at the time it had an ISA LAN (10M), here is this could add, but overall it has an abundant amount of RAM, which a part can be used as a RAM drive, and could speed up the execution of DOS programs, since the hard drive is quite dated, and even if maybe at that time it could be relatively fast, so using the RAM DISK, the slowdown due to reading and writing is eliminated .

I closed the Chicony PC with iDX4-100, for now I don't have to do anything else inside, I just have to activate the sound from Windows 3.1, and this is done without having to open the PC.

I go back for a moment to the PCCHIPS M912, I saw on the MB that the version is 1.7, here if you wanted we would have to remove the fake cache, the fake chips should be removed one by one, once all the holes are free you can proceed to solder the sockets, while waiting to get working L2 cache chips, but I also imagine the TAG 🏷 chip, which I imagine is also the fake one.

In this PC I have 20 MB of RAM and a 5V IT'S 486DX2-66, I have to check if there is the possibility of using CPUs with lower voltages (from 3V to 4V), in that case I could switch to FSB 40, to change CPU with DX2-80 or a DX4-120, currently it is a little slowed down, the L2 cache of 256 KB that you read on the second boot screen, is a bluff of the BIOS, it makes you believe that there is but it is not so.

Another upgrade I will do is to change a 72 PIN RAM, in total it is 20 MB, and it should be a 4 MB and the other 16 MB, I was thinking of changing the 4 MB one with a 16 MB, to get a total of 32 MB, in this motherboard there are four sockets for SIMM 30 PIN, but for me it is useless to put them when there are more 72 PIN sockets, maybe they could try, but given the rarity of the 4 MB modules (or more), I really don't think I'll use them here, assuming I have them.

I change the motherboard, I try the MB PINE PT-2068.1 with i486SX-25@33, I change the CPU with the 486DX2-66 (blue heat sink), unfortunately despite some attempts it does not work, returning to the SX25 instead it works.

I try a second DX2-66 (this time AMD), sometimes it gives some starting signs, shows a couple of codes, but nothing on the video, so I think the problem is the socket, I have to make a few more attempts with other 5V CPUs, all 486DX2-66, I hope at least one can work, I wouldn't be thrilled if the only one that works is the 486SX.

There is another 486DX2-66 waiting for a check, it is a ZIDA 4DVP with 128 KB L2 cache not working, years ago it worked and I had expanded it to 256 KB, then it didn't work anymore, I don't know if those L2 cache chips work, I don't have a tool for verification, but I think I've tried others that are definitely working, and I had the same problem, so we'll have to understand if there's a fault nearby.

To recover a DX2-66, I would have to replace the one on the PCCHIPS M912, of course I would have to switch to something faster, that is a DX2-80 or possibly a DX4 100-120, being a VLB motherboard, they can be used with FSB 40, with a maximum of two VLB cards, in this case having a VGA and a Controller, I would be fine, it is also true that there are combo video cards (with controller), but I think they are intended for those PCs like the 486DX50, after all the common use of VLB cards was video and controller, other types of VLB cards are quite rare and personally not I've seen some, although it's likely that there's someone who has one or more in their collection.

Here I noticed that in some situations, that is, in some resolutions, you see black vertical lines, sometimes even quite big, I think it may be a video card problem, or maybe the contacts need to be cleaned, I don't think it's a driver problem, since I use DOS, rather it could be a problem with some video RAM chip, but you have to make a try with a similar video card, if the other one is good, then you could start doing some tests on the problematic card.

On the PINE PT-2068.1 with SX25@33, I transferred other bench DOS programs, initially I had connected the Chicony iDX4-100 PC, but shortly after starting the transfer, an error occurred in the file transfer, I did some tests, but there must be something wrong, maybe one or both of the RAMs I added, it could have something wrong, I will try with the QTPRO diagnostic, first with the four RAMs and then removing the two to get to 32 MB, in case everything goes back ok, I should try two more RAMs, hoping they are fine.

I'll see if tomorrow I can do some tests with the ZIDA 4DVP, in that PC years ago the L2 cache failed, and if I'm not mistaken there is a problem with one or two VLB slots, one definitely works, because there is the 1 MB VLB CL GD5424 video card, the doubt is whether I used an ISA controller, or a VLB (?), years ago I didn't have other VLB controllers, so I used the ISA, but later I bought batches of cards, with a couple of VLB controllers, I even have a couple that are malfunctioning, in the future I could be Forced to repair them, given the current prices.

I hope I don't get bored with all these 486, I know lately I left out the PCs I had worked on, but soon I could change Socket, returning to something later, I'm slowly reorganizing the old PCs, to try to differentiate them a little, I don't want PCs all of the same type, so from the four 486DX2-66 VLB, at least a couple will be updated, in fact only one is missing, the other upgrade was made days ago, with the change of Intel CPU, from 486DX2-66 to DX4-100.

Regarding the PCCHIPS M912 V.1.7 motherboard, I checked if beyond the fake L2 cache chips, there was the TAG chip, this turned out to be fake (him too!) So an additional chip will be needed, for the cache instead at least four are needed, but there is room for another four, if we find eight (8+1) chips of the capacity and type (same), we could proceed with the removal of the fake ones, and the welding of the sockets, for the latter in the motherboard the use of DIP32 or DIP28 chips is provided, so the sockets could be welded to 32 PIN, in which eventually we can install the DIP28 chips.

For the BIOS, I'm sure it needs to be updated, because it shows an L2 cache that isn't there, and if I'm not mistaken, it's not possible to make any changes from the BIOS, such as access times or anything else, the BIOS chip is not soldered, this is already very positive, I think that if it was of the rewritable type, you could only make an update from floppy, because it's a DIP28, they change the size compared to the L2 cache chips, I don't think you can reprogram with a programmer for DIP32, on this thing clearly I wouldn't know if it's possible or not, however it wouldn't be from Exclude.

An advantage of the update would be the support for the fastest CPUs of AMD and Cyrix, we are talking about 5X86 with speeds of over 100 MHz, but for these it is necessary that the motherboard is able to provide the CPU with less than 4V, and looking at the motherboard, you see a small black heatsink, which covers the voltage regulator, if this is not fake too, you can easily install any 486 you have, just paying attention to the jumpers, because an incorrect setting can cause the CPU to not start, so be very careful, and above all check before Install the CPU, that the voltage is correct, here it is obvious if the CPU is always 5V, the arrangement of the voltage jumpers should not be changed, and you can do without checking.

The advantage of the real L2 cache, will it be obvious or not? I think it depends a lot on the type of use we will make of the PC, in general, however, I prefer to have a real and working L2 cache, because the difference is greater between 0 and 128 KB, rather than between 128 and 256 KB or 256 and 512 KB, that is, not having one is definitely worse than having a few but good, while having a little or twice as much would only make a marginal increase, of course no one prevents you from installing the amount you want, according to the manual you can even reach 1! MB of L2 cache, but on a 486 having 128 or 256 KB was normal.

I start with the PINE PT-2068.1, I tried some 486DX CPUs, but none of them worked, the PC never started the post code sequence, with a couple of CPUs after several attempts, it had displayed some code, but then it stopped almost immediately, I don't know why it happens, maybe it's the socket that has some not perfect pin contact, but I don't think it's possible to upgrade at this point, in fact I should put it back to 25 MHz.

Among the benches I ran Phil's, for the verification of the cache, he got scores that I can't say if they are adequate or not, but since there is no L2 cache, I don't think there is much to evaluate.

Just because I needed a 486DX33, I took out a second PINE PT-429 motherboard, this one is in baby AT format, and has the VLB, unfortunately there is no Voltage regulator, but there is a 128 KB L2 cache, the chips are real and also the TAG 🏷 one, I will try to put a CPU of at least 66 MHz, among those I have available.

I noticed that it has a reserved space, the Cyrix Coprocessor, which must be welded under the CPU, should be destined for the Cyrix 486S, which had compared to the i486SX, 2 KB L1 cache and the possibility of adding the Coprocessor, although I have seen in the past, intermediate sockets in which it had to be welded.

If I can I would like to do some bench, I also found an HD connected, there should be DOS and Windows 3. X, it's been several years since I last tried it, and I think there's a problem with the keyboard, which had operating problems, that is, no key worked, and it was not possible to access the BIOS, and store the changes.

Well for the moment I continue with the 486, I took them out to review some things, one for example is the CPU, in some PCs despite having a DX2-66 5V installed, it is possible to change and use CPU from 4V or less (up to 3V), in others instead it is not possible to use these CPUs, except using an interposer (which I don't have).

As I wrote, I have a ZIDA 4DVP with a Cyrix 486DX2-66, in the past I had an AMD 486DX2-66, but I removed it, because it had failed and the CPU had ended up on another PC (the Chicony!), later I found this Cyrix with a heatsink included, this allows you to do without a specific heatsink for 486, there is no fan but I think you don't need it.

Looking at the frequency generator, Chronotel brand and the silkscreens, I was sure that there were not only three FSBs available, and that beyond 25 33 and 50 MHz, there was at least 40 MHz, the MB manual was found to be more detailed, as you can see, there are both 40 MHz and 20 MHz, maybe there are also other frequencies (secret), like 16 MHz or something above 50 MHz (?).

Also from the manual, you can see that the motherboard is one of the first VLBs for 486, on a page at the end, there is an update on the new CPUs available, including the DX2-66.

But the important thing would be to fix the L2 cache, then at that point you can start thinking about some upgrades, the most interesting one would be to change RAM, both to increase the capacity, but also to decrease the access time, from the current 70 ns to at least 60 ns.

Maybe a lower access time is not necessary, if we stayed with the 33 MHz of FSB, but if I wanted to bring it to 40 MHz, installing a DX2-80 here maybe, faster RAM could be useful, if we select 0WS, which allows a slight increase in bench scores, but at this point it is likely that even the L2 cache chips, will have to be re-evaluated, and if necessary replaced with others faster than those currently present.

The last batch of motherboards of 2025, is this one with three MB, two ASUS and one Intel, the Sockets are 754 - AM2 - 478.

I had no interest in getting other motherboards to repair, but this batch was cheap, and the boards are also equipped with CPUs, on the 754 you can read Sempron 3000, the other two have the heatsink on and you can't read anything, in that AM2 the shield is included, let's see them one by one.

I start from the S.754, ASUS K8U-X with some electrolytic capacitors swollen in the VRM area, apart from this it seems in normal condition, after the replacement of the electrolytic capacitors, it is very likely that it can work again, the chipset is ULI, if I'm not mistaken it is ALi chipset, which later I don't know why, they started calling ULI, for the rest it seems to me well integrated, with audio and LAN on board.

The other ASUS AM2 is an M2V - TVM/V, it seems in even better condition, the chipset is VIA maybe it is not the best, but being an AMD card, it is quite common to find such a chipset. This card, unlike the other two, has PCI-E slots, for the video card there is the 16X, but for other cards there is a 1X, complete the slot equipment, two normal PCI slots, very useful for using old cards, even if the uATX format limits the number of slots, here we have enough, considering that there are integrated Audio and LAN.

Finally the Intel motherboard, a D845GEBV2 S.478, a classic Intel board that does not offer much choice in the BIOS, are usually quite limited in options, but on the other hand it should be very reliable and with fairly normal performance. Given the use of DDR slots for RAM, it should be suitable for P4-NW CPUs, probably 400 and 533 MHz FSB, I'll look at the manual but I think it doesn't support 800 MHz ones, although I wouldn't mind at all, but for the presence of only two RAM banks, I don't think the FSB 800 is supported.

While waiting to put my paws on the batch of three motherboards (754 AM2 and 478), I looked at the manuals of these boards:

Https://theretroweb.com/motherboards/s/asus-k8u-x

Https://theretroweb.com/motherboards/s/ ... tvm-v2-ae2

Https://theretroweb.com/motherboards/s/ ... ownville-2

As anticipated, the one I already know which CPU has, is the ASUS K8U-X, on which there is a 1600 or 1800 MHz Sempron 3000+, the real frequency of this CPU is much lower than that of the PR, however these AMD 754 CPUs were the cheapest of the Hammer series, I didn't quite understand the reason, of the output on two sockets 754 and 940 (later unified to 939), but later the 754 disappeared along with the 940 and 939, when it was replaced by the AM2.

A flaw in my opinion of this S.754, is the use of single core CPUs, it is true that dual cores at that time may have been expensive, but maybe MB could have been made with two S.754, so that the second CPU could be added later.

Of course then you have to see if the chipset was made to manage one or more CPUs, however today for DOS and retrogame use, it's also fine with the single core, some of the last MB had PCI-E instead of AGP, while for RAM there is only DDR.

For the other two motherboards, I hope it's not a disappointment, regarding the installed CPUs, but I already know that I'll hardly find top-of-the-range models, indeed for the 478 I wouldn't be surprised if there was a Celeron, while for the AMD AM2 it would be nice to find a Dual Core, rather than a Sempron.

But while I wait for these boards, I can dedicate myself to other motherboards, like the 486 or similar, by similar I mean those boards that are made, to install both the 386 and the 486, of this one I have one with the Cyrix 486DLC40, but if I wanted I could install a real 486DX, it wouldn't make much sense to put a 486SX, because you can't add the coprocessor, but I never updated it, because I have several 486, and only that DLC that is an enhanced 386, so I'll keep it like this.

I'm observing the sequence of post codes, from starting the PC to the DOS prompt, PINE PT-2068.1 motherboard and BIOS Award, and it's much longer than what you see, many codes are so fast, that we can't see them, basically we see about a dozen of them, some of them display for a while, like when there is a dead battery, and you have to enter the BIOS to store the settings, but the list is quite long, each code should be checked what it is, this would give a sequence of what the BIOS performs at startup.

Some things are basically identical to those of today's PCs, but perhaps these and even those after 486, include other codes that are not used here, and the list could be even ️ longer.

There are codes, which instead should be the same for all PCs, for example bF 41 is visible for the floppy drive, others instead may be visible, when there is a checksum error, or in other cases that require entry into the BIOS.

Let's say that all these codes executed in sequence, we don't even notice them, but when a problem arises, and it stops at a certain point, they can help us, but first we have to understand if the problem is HW or not, that is if it reports a RAM problem, and we try various working modules, it is possible that there may be a broken track somewhere, the chances are greater, if we have a working bench and another not.

I remember that in the presence of BIOS AMI, the codes could be different from those of Award, and the same could apply, for Phoenix or other types of BIOS.

Instead, regarding the sequence of post codes, I believe that even with the same type of BIOS, there may be different sequences between two motherboards, due for example to the chipset or different RAM banks, this is what I imagine, but between two identical motherboards, for example two ASUS P2B, with the same CPU and RAM, I think the sequences should be, code for code the same

Waiting for the batch, which will arrive in a few days, I continue with the 486, today it is the turn of the ZIDA Z4DVP with Cyrix 486DX2-66, the card has an Opti 495 chipset and three VLB, but the L2 cache present (128 KB), is not working, you have to deactivate the external cache from the BIOS, with the cache active it gets stuck on the written WAIT on the first BIOS screen, the corresponding codes are 96 95, preceded by a long series of codes, which are still fine, since it starts without problems until the point where it crashes.

This card does not have 72 PIN RAM, it is a big limit, if you want to expand the memory, because the 4 SIMM 30 PINs of 4 MB each, have never been cheap, and towards the end of the 90s it was sold to me because it was cheaper to buy a motherboard with SIMM 72 PIN.

Unfortunately, since the L2 cache failed, the PC seems to me to be decidedly slower, in some benches it goes almost half as much as before, I don't think it depends on anything in particular, I tried to change some settings to speed it up, but if it's effective you get a very slight improvement, nothing special, however with the equipment I have available, I can't do much, we would have to check the individual cache chips and TAG 🏷, to understand if one or more don't go (?), and it would be interesting to check if the frequency reaches the hooes, but for this you would need an oscilloscope (which doesn't I have), there are very cheap models but I don't think they can measure higher frequencies, I think only a few MHz, in this case there should be 66 MHz, then of course everything could be ok, that is, correct frequency and working chips, in this case we need to understand if the fault is around, resistors and capacitors should be checked, it can be a long and boring job, but it can be worth it if you can solve it.

To run the benches of Phil's and others, I used the PINE PT-2068.1 disk, but I connected it to a VLB controller, as a VLB video card I have a 1MB CL GD5424, it is not expandable to 2 MB, except that the 8 chips are replaced, with others of double capacity (if you find them).

The RAMs are identical from 1 MB each, and all from 70ns, in total they are 8 MB, enough for Windows 3. X, but few for Windows 95, although it could be installed, I wouldn't, I would rather try other OS, which require little RAM to work.

But if the cache works, I would change the CPU, unfortunately it is limited to 5V CPUs, those with lower voltages (4-3V), you can't put them except with an interposer that reduces the voltage.

Surely since you got used to the various Windows 9X, and gradually abandoned DOS, the RAM endowment has grown a lot, and also the speed in MHz, a PC 486DX2-66 with 8 MB was quite fast in DOS, but to be fast in Windows, you need something more powerful, personally an S.7 with K6-2 (or higher) and 64 MB of RAM, I consider them quite fast, but a P3 of about 700 MHz (or more) with 128 MB of RAM (or more), would be enough to be able to use end-of-of-the-century software, without particular problems, then of course it depends on What we will use, but for most programs (DOS included), it should run quite smoothly.

Every now and then I insert links to YouTube channels, on some particular uses of the PC, or on how to revive old configurations, or on hw updates, well I follow some channels with interest, sporadically if I can I say mine on how to solve, but in general, I follow and look with curiosity, but often I already know that what I see done, I can't do it because I don't have the tools 🛠 that I need, the repairs I do are mostly, quite simple but still I need a certain skill, which I have acquired over the years, and above all I have a certain ability to solve, Even if I'm not a technician or a professional repairman, almost everything I know about MB, comes from my favorite hobby, I've long passed the 100 Motherboards (or PCs) on which I put my paws, initially only SW problems, then HW resolution with change of components or boards, and then finally, with HW resolution with MB component replacement work.

I don't know if in the future I will be able to get useful tools 🛠, for more complex repairs, but maybe I could make some attempts, on scrap motherboards, to use alternative techniques, for example in case of various tiny damaged traces, who has a microscope, can repair them by making bridges, what would the alternative be?

It depends on how many there are, but if there are more than three, and they are very close, you have to observe the path, and find points where you can weld a bridge, so that it distances a little from the bridge next to it, the important thing is to restore the connection, then if you can do it in a short space, so much the better.

And I have some motherboards, with similar problems, and I often spent time on it, but at some point I stopped, because the level of difficulty was a little too high for me.

I have an ASUS S.462/A, which has the AGP slot to replace, I should try to remove it, and then solder an AGP slot recovered from another MB, but the replacement of the AGP slot does not ensure that the motherboard works, maybe it is better to eliminate the part of the burned pins, and try the MB with a PCI VGA, if it works then we could continue, and replace the damaged AGP slot.

I wrote something in the past, on this motherboard S.462/A ASUS A7V400-MX, it was part of a batch of motherboards that were not well packed, indeed it seems to me that there was no packaging between the various motherboards, the result is that almost all of them were damaged, and in some the scratches are such as not to allow a possible repair.

All in all, this could be repaired, I don't know how long it will take me, to finish the repairs, I finished the ones above, there were two SMDs to re-seal, then for the rest I don't see anything in particular, except the electrolytic capacitors of poor brands, as in other motherboards of the period.

The issue changes by looking at the back of the MB, I have identified three points where there is at least one interrupted track, the continuity must be restored, the main obstacle is the tiny dimensions of these tracks, and consequently also of the wire that I will use, the damage is very difficult to see, only by enlarging a lot you can see exactly where the copper (of the track) is missing, of course then it is not said that these tracks are part of an indispensable circuit for the operation of the board, but repairing them in my opinion is better, because otherwise if you notice any problem, we know that the possible cause Be one of those tracks, and we will still have to repair them.

Some of these I could fix them, with a generous welding of tin, but only in case there is only one track broken, while the really difficult ones, can be repaired with tiny bridges.

But doing this without a microscope, is quite difficult and the result will never be aesthetically the best, that's why this type of work, I don't do it often, in fact I give up doing it, because it takes too much effort, and it is very complicated if executed without the help of a magnifier, which however helps a little, of course it is also true that in the past such a card was at low prices, but the general conditions in my opinion, are quite good (apart from the scratches), so I would avoid using it as a donor card.

The repair of the ASUS A7V400-MX, was postponed because I spent some time on another motherboard, which I found looking for other motherboards to repair, it is not as old as a board, but it is already almost 20 years old, it is a T socket (LGA775), with CPU supplied with Intel E2160 Pentium Dual-Core, with frequency of 1.80GHz cache L2 from 1MB and FSB 800 MHz.

The brand of the uATX format MB is MSI, but as I discovered later it comes from an HP Compaq PC, the model is MS-7336 VER:1.0, the slots present are few but still enough, considering that there are integrated audio video and LAN, to use more powerful VGA there is a PCI-E 16X slot, plus two other PCI-E 1X and one PCI, perhaps a second PCI slot would have been more convenient, but they preferred the PCI-E 1X, perhaps because now it is used more (with the new cards).

On careful inspection I don't notice anything strange, the card is in excellent condition, maybe it was recovered from a broken or disused PC, it is clear that if it was broken, the problem must have been the power supply, even details such as the welds, are very brilliant, so I decide to try this card, but I take everything to another room, because I would waste a lot of time if I freed up space on the bench, so instead I can assemble in a few minutes what I need to start up.

Initially, I only connected the post card, the power supply, a RAM and a heat sink (without a fan), turned on the power supply, it seems to turn on and off immediately, it's a bit strange but I try to see what happens if you join the PWSW pins, and in fact it seems to start, because I see several codes appear on the post card, but then I turn it off, because I want to connect the keyboard and video, also to see if there is something on the screen.

In fact you see a screen with a message, which warns that the CPU fan is not present, shortly after the PC turns off by itself, adding the fan I try again, this time the SYSTEM fan is not present, shortly after it turns off! I connect the second fan, and restart, finally I see the first BIOS screen with the main information, and from there I can access the BIOS to change some settings.

The PC is as anticipated, an HP Compaq dx2300 Microtower, the BIOS is V1.08 05/24/07, the line at the bottom is like this:

05/25/2007-Broadwater-6A79LM4FC-00

This info will be useful in the future if it is necessary to update the BIOS, I don't think it will be necessary, but I'm sure there are later versions, they usually come out to correct any problems.

Unfortunately for today I end up here, in the next few days I should look for a SATA disk, to connect it and install the OS, but first I would have to add a second DDR2 RAM, to activate the dual channel, and to have 2 GB total.

On the MB format uATX MSI, MS-7336 VER:1.0, I performed some tests, to see if adding a second RAM everything is fine

I start with various modules all DDR2 667, the substance is the same, the PC does not start, maybe it may depend on the slot, maybe the contacts are not the best, strangely, however, none of the modules get stuck on the same codes, but they show different numbers on the post card PCI (+ISA), this could be the result of having mixed RAM from different manufacturers, I usually prefer to use identical modules, I did not do the test to start removing the working module, because for the moment I did not have to try the individual RAMs, so there could be among the RAMs that I have saved (for future PCs), some RAM Not working.

Anyway I continue with the tests, and it seems that the subsequent RAMs manage to get further (with the codes), but not enough to get to the BIOS screen, for a few moments I thought that the problem of this PC is a non-working RAM bank, which would limit the more heavy use of the PC, however I get to evaluate the other DDR2 RAMs to try, I choose one that in my opinion is quite likely to go, since I use similar ones in many of my PCs, I put it on and I hope it works!

At startup, everything goes well, and after a certain number of post codes I see the BIOS screen on video, now unlike before, there are 2 GB of RAM, and above all there is the dual channel active!

How much the dual channel improves the performance I don't know, definitely not by 50%!, also because the difference will be visible with large memory occupation, and for this I think Windows should be installed as soon as possible, for this kind of PC, I would go to Windows 7, XP seems a bit dated to me, while Vista would perhaps be more correct, but most do not recommend its installation, and so at least for now I go directly with Windows 7.

Maybe I didn't write about the Intel chipset, which has an integrated VGA, the model in my opinion is quite cheap, those who have needs of a certain level, can install the preferred PCI-E video card, and if you are not satisfied with the results obtained, you could evaluate a change of CPU, the current one (E2160) I don't know if it's the original, I saw that some PCs originally had 3.6 GHz P4HT, if you wanting to go up you can switch to a C2D Conroe, but I don't think it's necessary if you have other PCs (faster), an alternative instead would be the overclocking of the E2160 forcing The FSB at 1066, it wouldn't cost anything and you can always go back to FSB 800, in case of any problems.

Speaking of S.462/A, I just recovered the ASUS A7V333-X REV:2.01 with burnt AGP slot, they are in particular the first pins to have melted the plastic (brown), maybe it is the result of using a too powerful VGA, or the contacts were not optimal, and this caused an increase in temperature on those PINs.

Apart from that, the board is in excellent condition, there is a light layer of surface dust, and none of the electrolytic capacitors shows swelling principles, usually the ones that need to be replaced are those of 3300uF 6.3V, or some of the 1000uF 6.3V, here fortunately they all seem to be fine.

Recently I discovered on TRW, an evaluation index, which for this board is very high, 9.8/10 the score, maybe if the equipment was more complete, because in the PCB there are empty areas prepared for RAID and ports 1394, or if there was support for CPU FSB 400, it could even reach 10/10(?).

But I don't know exactly how the score works, that is, if it is obtained from the on-board equipment, or if it takes into account other things, such as performance or reliability or stability.

Let's say it won't be the best for AMD 462/A systems, but it was definitely among the best cards of the period (I imagine), the support for FSB 333 CPUs, means that you can install CPUs with PR from about 3000, the real frequency is about 2.0GHz (more or less), unfortunately there are no Tualatin with these frequencies, and in my opinion I wouldn't compare them to P4, although it is possible to do it, certainly it was very close to the last evolution of the 462/A, the subsequent S.754 had a relatively short life, and the 462/A was one of the largest AMD Sockets Success.

The difficulty of removing the damaged AGP slot would be less, if in the back of the chipset, there were no SMDs (which there are here!), the use of hot air is the best solution, it remains to be understood the necessary temperature, to melt the tin and not burn the PCB, I don't think there are many alternatives to unsold the slot, one would be to cut from above the plastic, and leave the individual pins, which would be unsolded one by one.

Of this S.462/A ASUS A7V333-X REV:2.01 with burnt AGP slot (which I have between my legs), before replacing the AGP slot, I should check that there is nothing else that can cause operating problems, because otherwise we will not know if the problem was already present, or was caused during the replacement of the AGP slot.

Let's be clear, if you are looking for a problem, in a motherboard in excellent condition, it is very likely that in the end you will find it, but it could be something that does not prevent the operation, and in my opinion this is more important, so in fact I would have found a possible problem, right next to the writing "MADE", it is almost invisible, but one or two tiny tracks, could be damaged, or one even interrupted, for this you will need the microscope, but looking at the tracks where they lead, for the moment I would not be willing to do a possible restoration, because they connect the FLOPPY1 port to the ITE IT8703F-A chip, if Even if there was a problem with reading or writing floppy disks, it wouldn't be a problem at all.

In the afternoon, I will try to recover a replacement AGP slot, I have a scrap motherboard, quite similar to this one, but first I would like to do some tests, removing other slots, like the PCI, because I would not want to irreparably damage the replacement AGP slot, the biggest difficulty could be the temperature of the "Phon", I have two possibilities 350 degrees or 550 degrees, the 350 degrees is able to soften the plastic, and should be enough to melt the tin.

Here maybe in this regard, I often see in the videos that they use the "Flow", or renew the welds with new tin, I think both solutions are valid, but I don't know if both should be used at the same time, we will see if it will be difficult to remove a slot, in case it cannot be removed easily, you could make a further attempt, trying to renew the pond.

Of course on the scrap card, I will recover everything I can, assuming it is simple, things that I do not need at the moment, but that might come in handy, for example DDR RAM slots, or the ATX power connector, or even some connector for FAN.

Of course then, heating from the back of the PCB, I could have side effects, such as the detachment of the SMD present on the upper side, in this case (assuming it is possible), would I find myself removing the AGP, with Mosfet and other nearby removed? I don't think so, but it wouldn't be a problem at all.

So, I recovered the scrapped motherboard, it was not possible to repair it, so I occasionally removed what I could need, but there is still ️ so much to be able to withdraw, some things however, both for the size and for the PIN number, are not easily recoverable.

I am not very convinced that the use of the "Phon" can be the ideal solution, for the removal of slots, normally in YouTube videos you can see, that hot air is used on chips, small and medium-sized, while for RAM slots (and ISA PCI etc...) they use a welder with built-in vacuum cleaner, and act PIN for PIN, both in removal and replacement.

If they can't clear the holes, they make some attempts, sometimes adding new tin, this may not work using the "Flux", or it should be used anyway (?).

The high number of PINs to be unsolded (one by one) is not a problem, after all it is a matter of time, the somewhat complex thing is to remove the pond enough, so that it does not remain connected to the MB, which is also complicated by the tiny dimensions, I would avoid forcing the partially unsolded PINs, because they could come off or damage the tracks, it would not be a problem if you damage some track on the scrap board, it would instead be a big problem if we damage the board to be repaired.

But performing some tests, in an attempt to remove the ATX connector, I saw that even if it is relatively easier to remove, it is not so easy to remove, probably the equipment I have available, is not suitable for such a job, after trying with tin vacuum cleaner and braid, none of the holes are free, and therefore I would say that, to work on the PINs of a slot, you need something more suitable or effective, and I would not waste time, to do a job with inadequate 🛠 tools.

However, there could be an alternative, which would also be more comfortable, first of all, instead of recovering the AGP slot from a motherboard, you could buy it new, this would certainly halve the time of the welding work, it would still remain to remove the burnt AGP slot, so maybe it would not be a bad idea, totally destroy the brown plastic part, and conveniently remove the individual PINs, one by one, then of course you have to be careful, but in the end maybe the important thing is to remove it, it doesn't matter if I extract it whole or in pieces, I believe that the path I will follow, It will be just this one!!!

Of course, however, I can always try the "Phon", on the scrap card, in case of success I can be more than satisfied, and I should not look for and wait for a new AGP slot, to replace the burnt one.

Before doing anything, on the ASUS A7V333-X motherboard, I tried to understand the PINs at the fusion (of the plastic), what function they have, first of all, I identified them as A1 and A2, they are both in the row of the upper part of the AGP slot, those of the lower row are called Bx instead, I was quite sure that one of the two was connected to a supply voltage, and the pinout confirmed it.

A1 is surprisingly a +12V, I didn't believe that such a high voltage would arrive, I expected a +3.3V, or at most a +5V, but I imagine that the +12V serves the video card in some way, maybe it is transformed into other voltages, through regulators on board the VGA.

A2, instead it is a "TYPEDET#", perhaps it has no power function, and considering that the A1 PIN, instead is a PIN that feeds the video card, I would say that it is very likely that this is not the PIN that caused the problem, in short, it seems that the probable culprit is only the A1 PIN.

Since A1 is a power PIN, I think it is quite likely that using a particularly powerful video card, the A1 PIN was overheated until the plastic around it melted, certainly no certainty that this is so, but it is possible that it was not the only cause, maybe the PIN did not have an optimal contact, and the final result was the damage to the AGP slot.

I have no idea, if the burnt plastic can somehow become conductive, I hope not, however I think the best thing is to replace the AGP slot, rather than try to repair it, you could also try, to put the pins back in place, but it could be a long job and probably does not solve it definitively, this problem could reappear because the surface of the PIN, could have lost part of the gilding.

Destroying the burnt AGP slot, it could facilitate its removal, it is not at all a simple job, right next to the PINs in the fusion zone, there is a 1000uF 6.3V electrolytic capacitor, here I would have expected to find a 16V, if this is crossed by the +12V that goes to the A1 PIN, but you don't see much with the AGP slot present, however, there is a Q20 next to it signed B1202, and connects the predisposition of the PWR_FAN, in which there are actually the +12V, so it would not be unlikely that the +12V to the A1 PIN comes from Q20.

I looked for a new AGP slot for sale, but strangely I didn't find any, here maybe it's not so easy to find new ones, and so it's easier and cheaper to recover one from a scrap motherboard, in short, I wouldn't say it's simple, but a lot depends on how tiny the holes for the individual PINs are, at first glance 👁 they seem to me the same as those of 1000uF electrolytic capacitors, or slightly smaller, I'll find out the real size by removing the AGP slot that I should transplant, I hope I don't have any surprises.