I briefly return to the burnt pins (fused plastic), of the Gigabyte GA-7VAXP ATX connecter, noting that there are four, I checked which lines have four pins (out of the total of twenty), and I discovered that it is +5V, basically it should not be a surprise 😱, the S.462 from what I know, requires a lot of current on that line, and in some cases it exceeds 15A, maybe 🤔 the burning is due to a poor ATX power supply, at least this is what I can imagine 💭.
Fortunately, I have old ATX power supplies, with a sufficient amperage on +5V, recent ones are quite enough only if of high power, since from the P4 we began to use the +12V with many amperes, so nothing strange if an old 300W power supply has more amperes on the +5V, than a new power supply of about 500W.

I have a CPU ready for the Gigabyte, a Barton Athlon XP 2800+, it should be from 333 FSB, I don't remember the real frequency, but it seems to me it was slightly higher than 2 GHz, or 2.1 GHz, the multi instead should be around 12X (I don't remember exactly), the integrated L2 cache would be 512 KB, which is the maximum for Athlon S.462, if I remember correctly, a version with 1 MB cache was planned, but it was deleted and never arrived on the market.

But as often happens, I work in parallel with other MB, every now and then I take one to try, and put it on the bench, today it's the turn of an S.423 MB, MSI MS-6529 845-Pro VER:1, Intel P4 423 CPU I don't know what frequency (but I'll find out soon), it has integrated audio, AGP 4X five PCI slots and a CNR, the format is the classic ATX, aesthetically it's excellent, just a little dust at some point.

Unfortunately the result of the boot test, was not a success, there is a problem that I have to understand well, in practice pressing PWR-ON, the LEDs light up (even those of the post card), the fan starts to spin, and immediately turns off, so there is something wrong!

I tried to leave only the CPU, removing everything, but it doesn't work, I removed the four-PIN ATX connecter, the same, so unless there are suggestions on what to check, for now I can't do anything else, and I think the only thing possible, would be to try another CPU, it seems very unlikely that that's the problem, but with the change if it works, it means it was defective.

As for the MSI S.423 motherboard, I saw the CPU installed, it is a Pentium 4 1.5GHz/256/400/1.75V, a fairly normal CPU in those years, and I would have changed it with a 1.7GHz, if the board worked.

The 200 MHz more speed, would not have made a big difference, but maybe with a light overclock, from 100 to 110 (therefore from 400 to 440), it would have raised the frequency to 1.87GHz, which would make a total of 370 MHz, unfortunately I don't have a 1.8GHz, because I could have approached 2.0 GHz, which is a much better frequency in my opinion.

Looking at the card I notice something, which could explain why it doesn't start, in at least one of the Mosfets, I notice a kind of ball, which could be the result of a failure, with relative overheating???

Since the boot test was negative, and having disassembled the heatsink, I decide to put it among the MB to be reviewed in the future, the heatsink with fan, I reuse it on a high MB 423, an Intel D850GB (Garibaldi), with a P4 1.6GHz/256/400/1.75V, to which I added four RIMM modules, 128 MB each, recovered from my first P4 423 2.0 GHz, after upgrading to 1 GB, with four modules of 256 MB each.

For today I have run out of time to dedicate it to the MB, we will see in the next few days, to try the Garibaldi, and we hope to have a better result, with a BIOS screen on the screen.

I'm almost ready, to try the Intel D850GB with P4 1.6GHz, I installed the fan cooler on the CPU, adding a little paste, maybe it's not enough, but I wouldn't want putting too much to come out, so I'd say we're almost there.

I found out that the card is used, on many Compaq brand PCs, I can't say if this MB is from Compaq, but it doesn't matter if it's not, what's important is that it works!

In this regard, I noticed an incorrect setting of two jumpers, honestly I don't know if they can create problems or not, but before trying the board, I'll put them in place, I just have to add a CR2032 battery, I have a recovery one that I measured at + 3V, a sufficient voltage to store the BIOS settings.

But for today probably, maybe it won't be possible to do anything else, if I have enough time, I'll use it to try the startup.

For the Intel D850GB with P4 1.6GHz, I looked for info around, and I found a comparison, between various 423 motherboards of famous brands, there was also this Intel, even with less RAM and a P4 1.4GHz, some cards were able to overclock very interesting, even above 140MHz, others instead did not climb so much, but only a little from 105 to 120 maximum, this makes it clear that the CPU sometimes does not go up, because it is the motherboard that is not suitable.

But reaching high frequencies does not mean having stability, and having both in time, and above all in continuous use (with temperatures within the limits), it is quite difficult if it is not overvolved, in short if the overclock is light normally no problem, if it is high it can be possible by giving up a few MHz, let's say for example if it is more than 140 MHz, going to 133 MHz could be quite safe.

But this Intel D850GB, what can it overclock ???

I don't think anyone, the card is known for its stability and reliability, in short, on the one hand the overclock sufferers are disappointed, but on the other hand the chances that the motherboard is working, I would say that they go up a lot, it would seem like a very robust card, perhaps with a few editable entries in the BIOS, and in this regard I think the BIOS is quite simplified.

With these premises, optimism grows enough, we'll see if it starts, and the theoretical possibilities, I hope, are quite high.

are you working on all those motherboards simultaneously ? Are you selling them after you fix them ?

Ujeen wrote on 2024-12-03, 02:36:

are you working on all those motherboards simultaneously ? Are you selling them after you fix them ?

Yes, I have a large desk, which allows me to work on several cards at the same time, but I have a second workstation nearby, so in conclusion, if the desk is full I move to the other station.
The cards are part of my collection, I repair them but I don't sell them, maybe in the future I could get rid of double or triple cards, for now I'm trying to make some definitive configuration, and with a very precise idea, for example a Windows 98SE PC that I will work on soon, it's an Athlon 64 3700, MB ECS 939, it's still ⚓️ waiting, because I've worked on other boards, and because I want to try some to see which ones work.

got it so you're building a retro pc collection for yourself 😀

For today I have concluded the tests of the motherboards, in fact I have only tried one, but I have not been successful!!!

I finished assembling the Intel D850GB, the one with the P4 423 1.6GHz, I put the essentials, a PCI video card, the CR2032 battery, the keyboard ⌨️, and the usual ISA/PCI post card.

Fortunately, I have suitable power supplies, basically there are three different sockets, the 20 PIN, the 4 PIN (square), and a six PIN (like the AT), this is what I use, an FSP has them supplied, no adapter is necessary.

I haven't connected a floppy reader, and I don't think it's useful, after starting it shows some codes (few), and it stops almost immediately on d4 d3, a sign that something is wrong, exactly I didn't understand what the problem is, I don't know maybe 🤔 the BIOS is an AMI or who knows what, I can't find anything useful written in the MB, I'll find it but for now I have two out of two 423 cards, not working.

On this one maybe 🤔, it is possible to make some further attempts, in the hope that there is something remediable, the other instead could be more complex, and could commit me a lot, without any certainty (or hope) of solving.

Unfortunately years ago I snubbed this 423 socket, otherwise 😅 I would have cards like the ASUS P4T, or equivalents of other brands, I think it was the RIMMs that kept me away from these PCs, they were expensive and configurations like that could even cost twice as much 🫡 a good P3.

Ujeen wrote on 2024-12-03, 18:39:

got it so you're building a retro pc collection for yourself 😀

Yes I have almost all my PCs from the 90s to today, but today it would be about a dozen years ago, I don't have current PCs, the best I have are fourth and third generation i3 and i5, then in addition to those in signature, I have several S.7 various Slot1, and several AMD sockets, plus many cards to fix and try, let's say that if I were forced to stay at home, at least one PC a day I could make it work.

I made another attempt, but I didn't make progress, currently starting, I see d2 d3 and d4 slide, then it stops on d4 d3, I tried other video cards, changed PCI slot, and finally I removed the PCI VGA, and put an ATI Radeon AGP, no progress, but something remains to be tried, for example other RAM, or rather other RDRAM, and at this point I would try to use only two together with two terminators, in practice they would be fake RDRAM, in these MB you have to fill all the banks, and not leave empty.

It would also be to try a CMOS reset, I don't think it's useful, but you can give it a try, unfortunately the BIOS chip is soldered, so I can't remove it and possibly rewrite it, yes because for me that could be the problem, maybe an unsuccessful attempt to update, or a corruption of the content, in short, in special cases like this, you are forced to remove the BIOS chip, but you could avoid it, if there was an adapter with F socket for the chip, and M plug for the programmer.

I tried to understand something more, on the MB D850GB (Garibaldi), basically I haven't tried anything yet ️, but I collected some useful info, first thing, I found some info on the manual (pdf), about RAM, according to what I understand, RDRAMs are supported, both ECC and non-ECC, this is interesting, because I should have a couple of ECC of 256 MB each, plus others of 128 MB (I think), so if I can make it go, I can put more than 512 MB of RAM, and the ones I currently have, I can set them aside for a possible future project (other MB 423), although I think it will be very difficult to find a low-priced card.

Could it be a RAM problem?

Maybe, but trying doesn't cost anything, according to what I read from the manual, d4 is about memory, even if I didn't quite understand which type, so trying other benches is the next thing to do.

What BIOS is it?

It looks very similar to an AMI, but it could be a BIOS made in the Intel house, maybe the base is AMI, however since I don't see anything on the screen, I don't know what version it is or if it's of a Compaq PC, ultimately if I don't get any results, with the RAM and zeroing the CMOS, it would be necessary to try to rewrite a BIOS, the problem of the soldered chip remains, but I'll see if I can build myself an adapter, in this way there's no need to remove the chip, to reprogram it.

I just tried the ECC RAM, I used a pair of 128 MB and two terminators, unfortunately nothing changed at startup!!!

I don't think it's useful to try the 256 MB ones, I'll keep them aside in case I solve it, because I would have the possibility to put a total of 768 MB (512+256), at the limit I could see if I can find two more 256 MB ECC at a human price, they usually ask for figures a little too high for my taste, however in the 423 with P4 2000 (working) I have 1 GB, quantity for a PC like that, more than enough, at least for Windows or Office use, if I didn't read wrong, the maximum installable quantity is 2 GB, so it would be four 512 modules MB each.

The last attempt would be the reset of the CMOS, apart from that, there would be nothing else to do, maybe a splash of IPA in the RAM banks, in short I don't think there are HW problems, at least visually I don't see any suspicious sign, if I can't make any progress, the thing that remains to be done, would be the BIOS Recovery via floppy, this would avoid having to finker on the soldered chip, and as I've already done in the past, it would bring the PC back into operation, unfortunately there will be no video message, due to the lack of ISA slots, you have to go Blindly, not the best but maybe it works???

I think that in this MB Intel Garibaldi, there is no hw failure, it is rather a problem either in the BIOS chip, or in its content.

But thinking about it I should do some more tests, I would like to remove all the RAM and see if starting, the codes displayed are the same, or if it stops first?

The other MB MSI, instead I'm trying to gather ideas, on what or where to look, possible failures, initially I'll look in the main lines, especially the +5V and the +12V, if I'm lucky, one of the two has a problem, but it could also be all ok, and the problem is somewhere else, thinking about it I should try to start with nothing, that is, without RAM and CPU, as I had already tried without RAM, if the problem disappears, it's the CPU most likely ko.

If I can get one back, maybe I'll advance the end-of-year celebrations.

Sometimes moving on to other motherboards, the ones I worked on, they stay a long time waiting for further tests or verifications, usually nothing changes after so many years, but sometimes there are obvious signs of what 😨 change, in particular I refer to electrolytic capacitors, sometimes even if they are visually perfect 🤩, it can happen that they need to be replaced, and this is evident when they are swollen, clearly if we change them all, we may have solved the problem, but the change of some can be problematic, ultimately if they are few you can change them all, if there are many normally, I change the faulty (visually), more those in the area or nearby.

Regarding the motherboards, I happen to put them away, and there are a lot of time waiting to be reviewed 📰, but looking for a card, it happens that I find others that need some work, but it was not possible to do it, always on the subject of electrolytic capacitors, it happens that I run out 😩 the stocks, and after maybe 🤔 working on something else, I skip those cards and move on to something else.

Then there are boards, which have some missing or damaged SMD, and it can be a problem, if they are in areas that are not very accessible, professional equipment helps in these cases.

Other cards, on the other hand, do not have any failure, just reprogram the BIOS, to have the card working, usually the failure is due to an incorrect update.

Examining after a few years one of the motherboards, the (very likely) reason for the non-functioning appeared, in particular I am referring to the ASROCK K8Upgrade-NF3, the card was visually perfect, but it did not give any sign of life, even using the particular Future-Upgrade board, which I have set aside, both in version 939 and in the AM2, this means, that with the card inserted, the part of the S.754 and its RAM, is excluded and the rest of the board is connected to the Future-Upgrade board.
What appeared?
Some electrolytic bulges, at least five are visible, but others may have gone, even if for the moment the bulge is not obvious, the best thing to do is to replace them, possibly all, and try the board again, some time ago I had updated the BIOS, with the latest version available, if that was the problem I will not know, because I have not tried it, but if after replacing the capacitors and the updated BIOS, it will not work (am I wrong?) I'll know that the BIOS was ok, and there's some other problem, if it works instead, I'll have a reason to celebrate 🎉 🎈 🥳

Among the 478 boards there is an ASUS P4S800-MX motherboard, a fairly compact card with integrated video, this limits the possibility of expansion a little, but having almost everything integrated, you can already use it like this, without any additional card, in short there is already everything you need.

I don't remember if the card was tested, some time ago, and if it gave any sign of operation (or none), but I think I'll try it as soon as possible, it's not a card for which I have a certain interest, but I would like to know if it works or not.

All in all, you can use all P4 478 CPUs, and this means that at most you could put, a 3.4GHz Prescott P4, although perhaps an approximately 3 GHz NW would already be enough.

To try the MB, however even a Celeron is enough, this type of 478 CPU, I never use it, on all PCs I use P4s, but in a card like that, there could be a Celeron 2600 or a Celeron D 2.66GHz, maybe the last one would be ideal by raising the FSB to 667, with a frequency that rises to 3.33GHz, all in all the result might not be far from a P4 NW, at least I think.

Apart from the Dell PC of a family member, I don't have any PC with BTX motherboards, but some time ago in a batch of cards, I took some together with others mostly ATX, among all the BTX I have a couple of Dell PCs, and at least three of HP PCs, the latter have a particular ATX connector with few PINs, and it is not possible to use ATX power supplies without an adapter (which I don't have), while on the Dell the power connectors are ATX 24+4 PIN, in the past for a short time I had to remove the Dell power supply (failure waiting for the Spare part), and use one taken from an assembled PC.

Returning to the Dell motherboard, it is a BTX LGA775, with DDR2 RAM and CPU installed Core2 Duo E6400, 2.13GHz/2M FSB 1066, there is no integrated video, but there are three PCI slots and as many PCI-E, the latter are of three different types 16X 4X and 1X, there is integrated audio (although limited to only two jacks), and I think LAN is 100 .

I have an original Dell ATX power supply, which I have to check, maybe it doesn't work? But in any case I think I can make it work, just to try this MB, and do some tests and bench, obviously if it works.

wow I've never seen this form factor before 😀 with a cut in this Dell PCB

Ujeen wrote on 2024-12-10, 17:00:

wow I've never seen this form factor before 😀 with a cut in this Dell PCB

Yes it is not very common, in this case it is a BTX motherboard, the dimensions are I think normal for this type of card, the part with the indentation instead, I don't think it is usually found in these BTX, it probably serves to leave room for the drives (CD-DVD), but I'm not sure about this, for a while this new format, tried to replace the ATX, initially spread among the big manufacturers, but later we returned to the ATX cards, which we still find today in all PCs (Barebone and allin1PC excluded).