PC@LIVE wrote on 2024-10-03, 15:09:

This morning I repaired the audio cable of the POST CARD, not having one in reserve, I managed and fixed it, I hope it lasts a f […]
Show full quote

This morning I repaired the audio cable of the POST CARD, not having one in reserve, I managed and fixed it, I hope it lasts a few years, even if it is not indispensable, sometimes it is useful for the beeps of the MBs on trial.

Staying at the M550, so I didn't remember well, the card starts, but it doesn't display anything, no beep, you can see some POST codes, the last ones displayed are:

D4 d3

What does it mean??? Or what does it mean???

Surely there is a problem, we will see later if it can be solved or not!

Let's start by saying that the BIOS is AMI, so the meaning changes, from BIOS to BIOS, to understand it they are not universal, if for AWARD means one thing, for AMI it can mean another, if we are practical of BIOS or specifically of BIOS AMI, we will also be able to know what problem there may be, unfortunately it is not my case, and I just have to interpret what is written in the POST CARD manual, and translate it, I believe that the first code displayed in all MBs is d0, the following ones I do not know if they follow a numerical order, or a visually causal order (i.e. You move on to other codes by skipping several).

Here the sequence of the new codes, is quite short, one second two, maybe three, however after d4 nothing happens anymore, but I haven't tried to leave it on for a long time ⏱, so I don't think it will go on at some point (waiting).

According to what I read, and understand by translating, the d3 disables the cache before detecting the memory. Performs memory sizing. And if the sizing is not performed, activate a boot lock code. I think this is the important point, but maybe this step is performed, and therefore no problem

For d4, the info is quite little, it starts with, tests the 512 KB base memory, runs something that stores in the first 8MB (cache???), finally sets up the stack. Well, I didn't understand much, apart from the 512 KB memory, I think they are valid concepts, for X86 PCs from XT up, but that most ignore because they run at startup.

The thing that could be interesting, however, is the d5, in which we talk about BIOS that is copied from ROM to RAM, here this could be the problem, if the ROM is not readable, it will not copy anything, for this problem, we have a possible solution, download a BIOS and rewrite the chip, with a programmer, maybe if we have the possibility, we can write to a new BIOS chip, and we can load into the corrupted BIOS chip, a different version, if it can be useful to us.

If it was an Nforce board I'd say its dead, but this isn't the case. Maybe try different ram/cpu?

kotel wrote on 2024-10-03, 17:35:

PC@LIVE wrote on 2024-10-03, 15:09:

This morning I repaired the audio cable of the POST CARD, not having one in reserve, I managed and fixed it, I hope it lasts a f […]
Show full quote

This morning I repaired the audio cable of the POST CARD, not having one in reserve, I managed and fixed it, I hope it lasts a few years, even if it is not indispensable, sometimes it is useful for the beeps of the MBs on trial.

Staying at the M550, so I didn't remember well, the card starts, but it doesn't display anything, no beep, you can see some POST codes, the last ones displayed are:

D4 d3

What does it mean??? Or what does it mean???

Surely there is a problem, we will see later if it can be solved or not!

Let's start by saying that the BIOS is AMI, so the meaning changes, from BIOS to BIOS, to understand it they are not universal, if for AWARD means one thing, for AMI it can mean another, if we are practical of BIOS or specifically of BIOS AMI, we will also be able to know what problem there may be, unfortunately it is not my case, and I just have to interpret what is written in the POST CARD manual, and translate it, I believe that the first code displayed in all MBs is d0, the following ones I do not know if they follow a numerical order, or a visually causal order (i.e. You move on to other codes by skipping several).

Here the sequence of the new codes, is quite short, one second two, maybe three, however after d4 nothing happens anymore, but I haven't tried to leave it on for a long time ⏱, so I don't think it will go on at some point (waiting).

According to what I read, and understand by translating, the d3 disables the cache before detecting the memory. Performs memory sizing. And if the sizing is not performed, activate a boot lock code. I think this is the important point, but maybe this step is performed, and therefore no problem

For d4, the info is quite little, it starts with, tests the 512 KB base memory, runs something that stores in the first 8MB (cache???), finally sets up the stack. Well, I didn't understand much, apart from the 512 KB memory, I think they are valid concepts, for X86 PCs from XT up, but that most ignore because they run at startup.

The thing that could be interesting, however, is the d5, in which we talk about BIOS that is copied from ROM to RAM, here this could be the problem, if the ROM is not readable, it will not copy anything, for this problem, we have a possible solution, download a BIOS and rewrite the chip, with a programmer, maybe if we have the possibility, we can write to a new BIOS chip, and we can load into the corrupted BIOS chip, a different version, if it can be useful to us.

If it was an Nforce board I'd say its dead, but this isn't the case. Maybe try different ram/cpu?

Thank you very much friend, I will try what you suggest, but I will also try to rewrite the BIOS, at least I will have tried everything possible, I will start with a K6-200 CPU or a 200MMX, for the RAM actually it may not be suitable, because it is a 128 RAM MB, I'll try a 32 MB one, I'm pretty sure it shouldn't have any problems, or I'll try some 72 PIN RAM, I'm sure those work, then we'll see what happens.

I made some progress with the M550, and I changed the CPU and RAM, the CPU just to be sure that the K6-233 didn't have any defects, and despite the fact it seemed to work, maybe at some point it caused problems(?)

Instead of the K6 I put a P166MMX (like Celeron), I changed the jumpers for multi 2.5X and VCORE 2.8V, in reality nothing has changed!

But at least I'm sure that the K6 CPU should be 100% working (at least hopefully), so I change the RAM, because in fact the 128 MB modules are not supported (sup to the manual), I try a 64 MB EDO SDRAM, it seems strange but from CPU-Z it was identified as EDO, however it doesn't work it's practically the same, so a search between 32 MB modules started, I find one, but I don't know if it's ok, because I've never tried to put it in a PC, instead I find a 64 MB PC100 module, double-sided total 16 chips, and I decide To try to use that, I start and here it seems to start, it reaches the point where it reads the floppy reader, and starts beeping with a continuous beep.

The floppy light stays on, this is something I've already seen in the past, and it's possible that with a VGA ISA you can see a message on the screen, I currently have a 2 MB PCI SiS, but the PCI slot is not yet ready to display, that ISA could be (I'll try again).

So in fact I didn't remember badly, it seems that you try to make a Recovery BIOS, at this point I would say, it doesn't cost me anything anyway, I download a BIOS from the web, I load it on a floppy, and try to see if it reads it and restores the (probably) damaged one (?).

Alternatively, there is always the possibility of removing the chip, and reprogramming via USB programmer, an operation that I have never done with this type of chip, but which could be quite simple and fast.

As soon as I have news, I will let you know, I hope to be able to show a nice BIOS screen, and that we can recover this MB M550, I have some K6s waiting for installation.

For the BIOS restore of the M550, I downloaded a file from the web that should be corrected, the file is 128 KB, but to pass it to a floppy, I had to take out (from the Black Beast's den) my Compaq P4 3,067 NW, transferred the file renaming it AMIBOOT.ROM, I put it in the M550 player and turned on the PC, unfortunately I never saw anything on the video of what it was doing.

However, after a while, that the green light of the drive went out, and having heard four beeps just before, I turned off and restarted, well everything seems normal, except for the lack of video signal, then at a certain point, it comes to codes, just before there are beeps that I can't interpret, in short it remains waiting for something, but it's not in bulk, because you can perform the keyboard reset, taking up, having tried various ISA and PCI video cards, I would say that the corrupt BIOS problem could be solved, but it's strange that it doesn't You can't see anything on the screen, certainly if I had another BIOS chip, I could try to put it and see what happens, if it works, it means that the BIOS chip is damaged, but this is just a hypothesis, there could be something else wrong, maybe who knows trying other RAM(???).

PC@LIVE wrote on 2024-10-04, 15:47:

For the BIOS restore of the M550, I downloaded a file from the web that should be corrected, the file is 128 KB, but to pass it to a floppy, I had to take out (from the Black Beast's den) my Compaq P4 3,067 NW, transferred the file renaming it AMIBOOT.ROM, I put it in the M550 player and turned on the PC, unfortunately I never saw anything on the video of what it was doing.

However, after a while, that the green light of the drive went out, and having heard four beeps just before, I turned off and restarted, well everything seems normal, except for the lack of video signal, then at a certain point, it comes to codes, just before there are beeps that I can't interpret, in short it remains waiting for something, but it's not in bulk, because you can perform the keyboard reset, taking up, having tried various ISA and PCI video cards, I would say that the corrupt BIOS problem could be solved, but it's strange that it doesn't You can't see anything on the screen, certainly if I had another BIOS chip, I could try to put it and see what happens, if it works, it means that the BIOS chip is damaged, but this is just a hypothesis, there could be something else wrong, maybe who knows trying other RAM(???).

The winbond chip you've used is an 256k and the image is 128k. From what I was told on TRW you'd need to merge 2 copies of that bios into one then flash it, although I don't know how the floppy BIOS images work and it might just extract everything the BIOS needs (which the size of that file would the the full 256k).
I don't think its gonna be ram related, but it won't hurt to give it a go. It might be complaining about cmos checksum failure/no bootable disk.
Dumb question, but did you use an ISA gpu? Maybe try gpu in different slots?

kotel wrote on 2024-10-04, 17:02:

The winbond chip you've used is an 256k and the image is 128k. From what I was told on TRW you'd need to merge 2 copies of that […]
Show full quote

PC@LIVE wrote on 2024-10-04, 15:47:

For the BIOS restore of the M550, I downloaded a file from the web that should be corrected, the file is 128 KB, but to pass it to a floppy, I had to take out (from the Black Beast's den) my Compaq P4 3,067 NW, transferred the file renaming it AMIBOOT.ROM, I put it in the M550 player and turned on the PC, unfortunately I never saw anything on the video of what it was doing.

However, after a while, that the green light of the drive went out, and having heard four beeps just before, I turned off and restarted, well everything seems normal, except for the lack of video signal, then at a certain point, it comes to codes, just before there are beeps that I can't interpret, in short it remains waiting for something, but it's not in bulk, because you can perform the keyboard reset, taking up, having tried various ISA and PCI video cards, I would say that the corrupt BIOS problem could be solved, but it's strange that it doesn't You can't see anything on the screen, certainly if I had another BIOS chip, I could try to put it and see what happens, if it works, it means that the BIOS chip is damaged, but this is just a hypothesis, there could be something else wrong, maybe who knows trying other RAM(???).

The winbond chip you've used is an 256k and the image is 128k. From what I was told on TRW you'd need to merge 2 copies of that bios into one then flash it, although I don't know how the floppy BIOS images work and it might just extract everything the BIOS needs (which the size of that file would the the full 256k).
I don't think its gonna be ram related, but it won't hurt to give it a go. It might be complaining about cmos checksum failure/no bootable disk.
Dumb question, but did you use an ISA gpu? Maybe try gpu in different slots?

No I didn't change the BIOS chip, I simply did the restore via floppy, and it seems to have succeeded, but since I never saw the image on the screen, perhaps the original chip, despite the rewriting, is malfunctioning(?)
I used both VGA ISA and PCI, in both no image, that's why the idea came to me to try other RAM, because the current one, a 64 MB with 16 chips (8 per side), could create problems, I'll know if with one with 32 MB and 16 chips, I'll see something, but in that case, I'll have finished with the repair, I'll have to move on to phase two, that is, test the operation, and change some things, add a sound card, in short, complete the configuration.

Well for today I stop, I made some attempts but I can't see anything on the screen, the only thing I'll do is connect another monitor, and see if something changes, theoretically nothing should change, but it's not said, the goal is to see a screen on the monitor, honestly I don't understand the reason or why you can't see anything, the PC seems to work, and answer the commands, you can as I wrote reset the keyboard, this means it's waiting for something, but not seeing the screen, I don't know what it asks.

The keyboard is working, because I see the leds flashing, and when I press F1 in the post card it displays a different code (87), looking for the meaning, it seems to have the password, even if it seems strange to me, the BIOS should be as factory, and the old settings should have been lost

In my opinion I am quite close to the solution, I should look in my notes if I find cards with similar problems, and how I eventually solved them, but without video, for me it is almost impossible to understand what happens, there are just before the post codes stop, the two short beeps followed after by a single short beep, I looked for what they mean, but I can't find anything with 2+1 short beeps.

In the end ✔️ I solved it!!!!!

The M550 works!!!!!🌟🌟🌟🌟🌟

In practice the restore via floppy 💾, it worked (!!!!!)

The only thing you couldn't understand was why you couldn't see anything on the video?????

Well it was very simple, normally I use an old 👴 Philips LCD, but on some occasions I use a large (and quite recent) monitor, well with this large monitor, you can't see anything, with the old Philips you can see it!!!! !

Maybe it's likely that even with the large one, you'll see, but I think it's necessary to act from the menu and blah blah blah various things, set something, like the resolution, in short, for the use I make of it (PC Office), it's fine 😌 like this.

From today I can start with the usual IDE 40PIN CF adapter, bench and see all the details that interest me from QTPRO.

A brief consideration, on the AMI BIOS, what is present is of the Win-BIOS type, let's say that it doesn't change much, because the entries are the same, but there is something strange, from the main page, where you only see icons, by clicking on the secondary ones, you see some lines, like in normal BIOSes, so in short I would say that the final result could be decidedly better, and perhaps 🤔 to distinguish themselves from other producers, they have only changed the main pages, the rest has not been changed much , and for this reason, the final result is just 😥 below the mark, in my opinion, it's a job more or less half done.

At a short start with the tests and some bench DOS, but first I saw on YouTube the latest video of "Necroware", in which some MB S.7 are the protagonists, it is a coincidence that we worked on S.7 cards, for my part I am quite satisfied with the result, I do not miss the S.7, already having some, the only one I do not have is an ASUS TX-97 or similar, finding one today is both expensive and difficult

First of all, I mistakenly put the multi 2X, instead of 2.5X, in fact you read a frequency at startup of 133 MHz, it doesn't matter because in this way I will make benches at 133 and others at 166, I don't think you can do them at 200, because the CPU should have the multi locked up.

The next phase is to put the K6 233 back, maybe starting with 200 MHz, and then go up to 233, but this will also be an intermediate step, the CPU that I will use will be a K6-2 at about 400 MHz, the minimum VCore will be slightly higher 2.5V, it will heat up slightly but it is a voltage of about 10% more than normal (2.2V), which should be well tolerated.

Later the upgrades will be for the RAM, from the current 32 MB I would like to go up to at least 128 MB, it depends on whether the RAM I have is read in full, and so if I use two 64 MB, I get the 128 MB (which was the initial idea), alternatively I should settle for 64 MB (2X 32), but we'll see...

In the YouTube video I saw, there is even a MB identical to my ASUS P55T2P4, and it has the same problem with the Dallas chip (although in my MB it is ODIN brand), and it shows how to replace the internal battery, adding a battery holder, a job that for those who have a minimum of skill, can do without particular difficulties.

Little curiosity, the M550 card was bought at the end of March 2023, but it remained for more than a year, in a box waiting to be fixed, fortunately I found it, and I dedicated myself to the main problem, which immediately seemed to me related to the BIOS, it was actually like that, and fixing it was both simple and cheap, the card cost less than €20.00 including shipping, to fix it I used a new floppy (1.44 MB), it was not necessary to remove the chip to reprogram it, and above all I did not have to look for phantom component failures, no welding is It was done, but the CR2033 battery was empty, I put it on a charge.

Out of curiosity I removed the jumper from the VCORE, the voltage drops to about 2.5V, I tried to turn it on and it works, the PMMX at 133 MHz at least for a simple boot, it also goes at low voltage.

I can expect a few days, of tests with various HW and SW, this is perhaps the part that gives the least satisfaction, but if you have clear ideas, you avoid trying things that will not fit in the final configuration, so I think a VGA PCI S3 of 2MB or 4MB, ISA sound card, probably a PCI LAN, then you need a PS/2 port, plus a PCI card with USB ports, I think this is the minimum.

Quick question. Do you have the "test" bios'es that cycle through POST codes for AT/ATX mainboards? I had found them on vogons but I forgot what thread it was.

kotel wrote on 2024-10-05, 15:18:

Quick question. Do you have the "test" bios'es that cycle through POST codes for AT/ATX mainboards? I had found them on vogons but I forgot what thread it was.

No and I didn't know that they existed, and that they were here on Vogons, however if they are useful to you, I can try to look for them, I don't think they are useful, because in one way or another, if you manage to identify a MB, you will at least find a copy of the BIOS, sure maybe it's not the same (identical), but you can still create a second BIOS chip, and see which of the two is better.

I also saw Tony359's YouTube videos, in his channel there is the latest video with various MB, including an S.7 that had some problems, in the end if I understood correctly he solved it, among the other MBs, there is a Batman S.4, I think it is the same or very similar to mine, I think he will have to work on that soon, if he needs a hand 🖐, I'm at your disposal.

My suggestions are probably more practical than technical, but having a working one, I could take some measurements, at least you can make a comparison.

Returning to the M550, I did the first diagnostics and ran some bench DOS, for now at only 133 MHz and 32 MB of RAM, for the moment I put back the jumper for VCORE 2.8V, I will remove it when I change CPU with a K6-2, it will be a little above the factory required voltage, but it should be well tolerated.

I hope I can do a lot of bench, I downloaded from Phil's site, his bench Ver.1.6, which I imagine is the last release, but I also downloaded those for the 386, which I think are a simplified version and suitable for PCs of the late 80s early 90s.

I'm pretty sure that from this PCChips M550 motherboard, you can make a nice PC, it certainly has its limits, but in my opinion it promises well.

The final configuration should be like this (I guess):

AMD K6-2 about 400 or 450 MHz

128 MB PC100 CL2 or 3 (2X64)

512 KB cache L2 (on board)

VGA PCI S3 Virge from 2 or 4 MB

Sound card ISA type SB16

PCI USB card

As a hard drive I don't know whether to use the IDE CF or SD adapter, or take a large HD and limit it to only 8 GB.

Maybe if you want to suggest alternative solutions, go ahead, maybe I'll miss something.

I made another bench, always at 133 MHz, I put the photo.

I have not yet ️ passed the Phil's files, in a CF card, but I connected a second IDE 40PIN CF adapter, in which there are other DOS bench, or different versions are stored, basically there is not too much difference, but anyway these are only a part, the others unfortunately do not fit inside a floppy, I have to put the card in a USB reader and copy the package of about 12 MB.

I noticed that it is not possible (at least in this motherboard), to insert two IDE-CF adapters of the cable-free type, that is, that is inserted directly into the IDE port of the MB, unfortunately they are so close that the side with jumper comes into contact with the back of the pins (where the jumper is inserted), this could also cause big problems, because it selects the voltage, if you want you can make a solution mod, shorten the back pins by milling them.

For the moment I haven't made this mod, but later I'll think about doing it, because I might need to transfer files, and if I don't have other adapters available, I'll have to use two of the same.

As soon as I finished making the bench of Phil's, there are absolutely not many, some are the same at various resolutions, unfortunately using a fairly "bare" configuration, in the Quake benches or similar, there is no audio, for this reason I think I will have to decide which sound card to use, and install it, currently I only work in DOS, mostly Ver.5, everything works quite well, in some other CF there is Windows, but it is version 3.0, and I don't think there are benches, that work with that version.

I'll put some images, I actually have several, but I'll see to write the results, later I'll compare them with those performed at a higher frequency, for the moment I'm at 133 MHz with a P.MMX.

Considering that the M550 is a socket 7, the use that can be made of it, would be at most Windows 98, maybe even 2000 or XP could go, if you have a CPU of about 400 MHz (or more), and 128 MB of RAM, 64 would be a bit few (I imagine), this I could do, with a K6-2 and two RAM of 64 MB, however there is a little problem, the capacity of the disk should be about 8 GB, which is the maximum that the BIOS can recognize, and allows you to have enough free space.

But in my opinion it would be better, Windows 98 and DOS (together), a dual boot that can be done with a modified file, in this case even 4 GB would be enough, if you don't need plenty of free space.

Now I would switch to the sound cards, I have few ISA sound cards, some nothing special, I would immediately discard those with IDE connector (or similar) for CD-ROM, those I reserve for PC 486 or lower, here it is not necessary because I have two integrated 40PIN IDEs, and then usually the IDEs of the sound cards are slower (in my opinion).

But I have several PCI sound cards, some are too modern and I think they are unsuitable, maybe considering that there are 4 PCI, you could use one for the sound card, but I would limit the possibility in the future of adding some cards, summarizing I would definitely have on the PCI:

VGA S3 2-4MB and a 10-100 LAN (for connection to other PCs)

Regarding the PCI USB card, I saw that it is not necessary, you can use the integrated ones (USB) in the MB, I think an ASUS MIR card can work, I don't have the original cables, so I don't know how they were made.

Now for a while ⏱, almost everything free, will be dedicated to the S.7 M550, I will do all the Phil's benches that I can, because some I have already seen that are not good for a 166, for a 133 instead yes, I noticed it because I moved the jumper, and I selected the 2.5X multi, reaching the factory frequency (of that CPU), I don't think you can go to 200 or more, I think it's blocked to go up, but trying it doesn't cost me anything, I'll try too, but not now.

Having neither Windows nor Programs, I don't notice any difference in DOS, surely there is a little more speed, you can see it from the Speedsys score, which went up from about 100 to about 125, but you can compare the scores I had obtained at 133, to see the difference, at this moment I don't take note of the individual scores, then I will make a summary with the scores at the various frequencies and with different CPUs.

Initially I will use the Pentium MMX, then a K6-233 and later a K6-2 of about 400 MHz, wanting for the sole purpose of curiosity I could use other non-Intel CPUs, a CYRIX MII-300 or other 686MX.

Since my Pentium MMX 166 can only do 133 and 166 MHz, I speak with equal FSB (66), because the multi does not have the 1.5/3.5X and the 3X active, in practice you can't make it go to 200 or 233 MHz, I decide to change CPU, and switch to another 233 MHz P.MMX, with this I can go to 200 with the multi 3X.

Currently, even if under DOS, I see a greater speed, compared to the frequencies of 133-166, not much at eye 👁, but it is noticeable.

Obviously, I will use Phil's benches, to see the scores of how much they go up (or down), for now I made a couple, and concluding these, and those I will do at 200 MHz, I would say that with the Pentium MMX I finished

I will move later to AMD K6 before and K6-2 after, but maybe before the K6, I try some Cyrix, just to see the scores, the first Cyrix MII-300 I had on an MB VX, it was fine with Office programs, unfortunately the MB had linear regulators, and after about three months it went ko!

As for MMX CPUs, I have to correct, what I wrote earlier, where I wrote "1.5/3.5X", to be precise no MMX has a 1.5X multi (that I know of), so for those CPUs only 3.5X, the 1.5X multi is instead used on normal Pentium.

With this I am saying, that you can't have Pentium with multi 3.5X and MMX with multi 1.5X, even if they were versions with the unlocked multi, maybe they were planned in the design phase, but they never got on the market.

However, various overclocks are possible, to obtain frequencies higher than those of the factory, usually the easiest thing is to switch to FSB 75, unfortunately even if I can I will not use that setting, except that the frequency generator, keep the PCI at standard values (33 MHz), in that case it would be interesting.

For the sound card, I decided which one to use, I chose an ISA SB CT4170, with Vibra16XV chip, it should be from 1997, or maybe later, in my opinion it should be suitable for that PC, which is currently an MMX 233, with 32 MB of RAM, in the meantime I put it among the cards intended for the M550, together with an S3 Virge PCI.