I was curious to see the Chaintech 5TDM2 in operation, so recently, I recovered an i133 CPU (Pentium P54), which was on a P5F76 that I have among the MB to fix, I put on the 5TDM2 a couple of RAM 72 PIN of 8MB each, which I had many years ago on my PCs, for the post card, I used this time my Riser ISA, it allows horizontal mounting of the cards, in this way you can see better, both the codes and the LEDs, after connecting everything else, that is, the video card power supply etc..., I try to start.

Initially, it seems to want to start, you see a code appear on the post card, the code displayed is C0 —, so it crashes, while every subsequent restart attempt, it doesn't show any code, for now apart from a few attempts without CPU and RAM, which give the same result, I'm rearranging the ideas, what to try.

While the card was in operation, I checked the voltages on the Mosfets, there were three Mosfets, and all had correct voltages, two about 5V and one about 3.3V, so I think that the voltages are correct, and that the problem may not be hardware, because the recognition of the CPU is automatic, and if something is wrong, there is no possibility of switching to the manual (via jumper), for now I just do some tests with different hardware, another CPU other RAM, maybe you could unlock, I hope to get some results, because I would like to see this card in operation.

I continue with the attempts to get the Chaintech 5TDM2 to work, for now nothing has changed, the most I can get, is the post code C0 —, leaving the board on, it seems that at a certain point it does an auto reset, maybe it is the sign that something in the BIOS startup went wrong, or in the automatic recognition of the CPU (?), anyway it is something that repeats I think endlessly, I forgot no beep from the speaker, so ultimately I would say that I can do some other tests, but I don't think it will serve anything.

For now I have tried, a 16 MB SDRAM, in both banks, the CMOS reset, and the Post Card in the PCI slot, in all cases no progress, indeed normally you don't see any code (— —), starting the power supply the keyboard LEDs turn on and off, so at least this is positive.

I tried to press various chips, before starting, but even this gave no results, even if at this point, I have the strong suspicion, that there is a software problem, it could be a BIOS problem, corrupted or damaged, or simply updated by mistake, with that of another motherboard, not knowing what version was present, I think that exhausting the tests of alternative hw, the next step to do is to remove the BIOS chip, clean the chip pins well, try again, and if nothing changes, I should reprogram it, hoping that this solves the problem, and That I finally get a video BIOS screen.

I saw this exact same issue on my socket A board with nf2. In my case the board wasn't playing nice with an 100mHz FSB CPU. An athlon xp fixed the issue.
Maybe it's the CPU in your case too? Although you might have tried it already..

kotel wrote on 2025-03-03, 10:19:

I saw this exact same issue on my socket A board with nf2. In my case the board wasn't playing nice with an 100mHz FSB CPU. An athlon xp fixed the issue.
Maybe it's the CPU in your case too? Although you might have tried it already..

Very interesting 🧐 Thank you very much ☺️
In fact I'm going to try other CPUs, a P.MMX for sure. I'll try other CPUs, you never know that the one I currently have on, could have reached the end of the life cycle. Soon we will have updates, I hope in the evening, to be able to make some progress.

When you are going to remove and reprogram the 5TDM2 BIOS chip, please make a backup of the present flashchip contents first and post it here.
I very much like to take a look at it and I will check if it is corrupted.

Jan

Chkcpu wrote on 2025-03-03, 16:44:

When you are going to remove and reprogram the 5TDM2 BIOS chip, please make a backup of the present flashchip contents first and post it here.
I very much like to take a look at it and I will check if it is corrupted.

Jan

Yes I think it won't be long, for now I'm trying different CPUs, to rule out that the problem is a CPU at the end of life, later I'll do a cleaning of the pins of the BIOS chip, if even so it doesn't work, I think it's defective, or the chip or the content, if it's the content I should be able to save it 🛟 easily, even if first I have to check if the chip is among those I can program, I think so, but we'll know more soon.

Currently all attempts to run the Chaintech 5TDM2, have been useless, the most I can get, is the post code C0 —, leaving the card on, it seems that at some point it does an auto reset, just to remove the doubt that the CPU was the problem, I replaced the P133 with another equal one, the post codes, those remained.

To eliminate the doubt that maybe the PC could give some different code, I also replaced the other P133, with a P166MMX, just to understand which parts work and which don't, with the P166MMX the voltage on the Mosfet is about 2.9V, I would say correct, this is the sign that the automatic recognition part of the VCORE works, maybe the PC startup program doesn't work, because I think the CPU recognition also works, since the voltage of the VCORE has dropped from 3.3V to 2.9V, so I would still try something, if nothing changes, I start to Search among the available BIOS, the most updated, from what I understand, it seems that the K6-2 or similar, are supported, and the minimum VCORE seems to be 2.1V, a value that can be used with other CPUs of the K6 series.

The only thing I didn't understand what FAN78 is, maybe it's a kind of additional card to reduce the VCORE (???).

Since for the moment I can't make progress with the Chaintech, I decide to try the PCChips M560 chipset TXPRO, the card is in condition like new, apart from a light powder, which I brushed, the only visible problem, is the DALLAS-like chip, I think they tried unsuccessfully to make the mod., to use the traditional CR2032 batteries, so every modification to the BIOS should be lost, I made this modification in the past on a PCChips M549, I think it has the same chip, in the M560 there is an VIA VT82885V, in the M549 there is an VIA but I don't remember The model, however I know that in case it is broken, there are alternative chips that can be used as a replacement.

For the first boot test, I removed everything from Chaintech, and I put it on the M560, I used my ISA riser, as you can see in the photo, after connecting everything, I try to start, it's working!!!

After showing the post codes in succession, there is the BIOS screen to video, and I can enter the BIOS by pressing F1, the BIOS is AMI like Windows, apart from the aesthetics the rest is practically almost the same as the traditional ones, and for today I stop here.

From tomorrow I will start doing some tests, and some bench DOS, currently the CPU goes to only 133 MHz, because a jumper is missing to select 2.5X, I will fix it soon, even if initially I will use it at reduced speed, then after finishing the benches, I will bring it to 166 MHz, and maybe even higher, but I will have to use a heatsink with fan, for now I used a chipset dissi without a fan, just for a quick boot test.

To fix the RTC chip, I would have to recover the PCChips M549, fortunately at some point, I stopped putting the cards in the box, for some time ⏱ I use a different system, with overlapping boxes, I normally use those of the puff pastry, which contain almost all the formats of motherboards, to these boxes you can attach a little paper tape, with the brand and model of the motherboard written, for the M549 instead I needed a higher box, I have a normal one of almost cubic size, and it fits, in Conclusion it's easy to locate and you don't have to rummage, if I can in the afternoon I'll take it out, and check the RTC chip, it should be similar if not the same, to that of the M560, we'll see...

As you may have noticed, the RAM, in the first screen of the BIOS, is only 16 MB, it is not a recognized module with reduced capacity, it is just a 16 MB, I have others, not many that I will try, I want to have two or three available, to use them in the first power tests, unfortunately I imagine they are PC66 or at most PC100, so they would only be good in the first PCs with SDRAM, for the others I will see if I can find 32 MB PC133 modules.

For this M560, I have not decided how much RAM to use, a lot will depend on the CPU, and especially on the possible frequencies of the FSB, there are both 75 and 83 available, I have to see if by chance there are the same frequencies with PCI at about 33 MHz, if there are it becomes very interesting, I could even put a K6-500, using the multi 6X (2X) and FSB 83 with the PCI at 33, unfortunately on motherboards with Intel chipset these settings do not work, but here they should work, at least with Cyrix type CPU, unfortunately with these CPUs it is not possible to go At 500 MHz, at most you can reach almost 300 MHz, excluding any overclock.

I recovered the PCChips M549, with great surprise, the chip was ... identical!!!

I remembered an VIA chip, even if in reality it was initially covered by a plastic cover, with written on it:

Houston Tech 12888A

Real Time 9639A1 16869V

Inside there was a small battery, CR1225 3V, soldered to the VIA VT82885N chip, obviously it was completely discharged, so every time the PC started, any BIOS setting was lost, and the latter was returned to the original factory settings.

At this point, I have to try to weld two wires, so that I can connect a 3V battery, not necessarily a CR2032, if you don't lose every modification to the BIOS, it means that the problem was only due to the power failure, if it doesn't work anyway, the chip could be damaged, because initially it was mounted backwards, fortunately before starting, I put it back in the right position.

Currently on the M549 I have a 1 MB CL video card, a SB ISA Vibra 16 sound card, and an AMD K6-2 CXT 360 MHz CPU, because Windows 95 doesn't like 400 MHz, it takes a file to fix the problem, but that card has the same problem as the Soyo 5BT5, the L2 cache doesn't work, you have to disable it to make the PC work, so for now I'm satisfied with 360 MHz, in the future I'll try to solve the problem, maybe some contact of the cache chips is bad.

I continue the tests with the PCChips M560, currently there is a P.166MMX at 133 MHz and 16 MB of SDRAM, enough to run DOS programs, the first step is to continue with the BIOS until it starts from the O.S. drive, well I memorize the new settings, and I continue confident that the PC starts normally, unfortunately it is not like that, it freezes practically immediately, showing on the post card an error code 71, which I don't think is very useful, after some useless attempts, I decide to try to disable the L2 cache, this works and the PC starts!!!

Strangely enough it's not the first S.7 PC that has L2 cache problems, I don't think it depends on the RTC chip, but I can't rule out anything for now, so the PC is slower, I don't know exactly how much, but at least I can do the benches of Phil's, this is what I got:

2) 95.2 | 3) 64.7 | 4) 34.0

5) 30.3 | 6) 13.9

A) 198.64 | b) 58.44 | c) 30.5

D) 13.7 | e) 12.1

L) 1068.55 and 2236.32 | n) 99.23

These values have not been compared with those of similar PCs, I do not expect high values, indeed if anything the opposite.

For now I'll continue with other benches, later I'll see if it's possible to solve the L2 cache problem, honestly I don't consider it working if it doesn't work, I hope it's not a big headache, and that it's just a bad contact, I'll see if I can find something anomalous, all in all if it works (even the cache!) I could do some interesting tests, however it seems that the L2 cache chips are a bit delicate, I have at least three cards where the cache does not go, in short two PCChips with defective cache, I can think of the famous fake cache, I hope that in these cards (M549 and M560) there is no (?).

Frequency change on the PCChips M560, now with both jumpers, I have 166 MHz, which would be the correct frequency of the Pentium MMX, the next step should be, the change from FSB from 66 to 75, I haven't seen if the frequency generator has 75 and 83 with 2.5X divider (for PCI), this allows you to keep the cards at the right frequency, the memory I don't know, probably not!

With the frequency of 166 MHz, I ran Phil's benches, there was to be expected a gain in scores, and in fact it is like that, it would be interesting to see with the active L2 cache how much it affects the results, unfortunately at the moment it does not work, in the future I will try to understand why it does not work, the most likely thing is that there is one of the failed chips, if the solders are perfect

Let's see the scores of the benches:

2) 192.9 | 3) 70.9 | 4) 39.7

5) 33.7 | 6) 15.2

A) 211.59 | b) 62.19 | c) 33.6

D) 14.9 | e) 13.2

L) 1736.42 and 2795.42 | n) 124.02

Regarding the Cache L2 chips, they are UT6164C32Q-6, I have to see which pins are of the power supply, to see if there is the correct voltage, measuring the voltage will not be easy, but I will try to understand which circuit is that powers them.

At least for now, I postpone the mod of the RTC chip, for the moment it is not necessary to store the settings, when I have finished with the cache, I will fix the RTC chip.

Finally I noticed that on the cache size jumper, on the 2-3 there is a condenser solder (?) SMD, maybe he could be the problem???

So regarding the PCChips M560, the problem of the L2 cache not working, is limiting in global performance, exactly I don't know how much, but I think at least 10% or even more.

How to solve it? It depends on what is missing, currently I have found something abnormal, on the capacity selection, on the pins for 512 KB there is a resistance 0 OHm 🕉, and this is right, but on the selection for 256 KB, instead of being empty, there is an SMD (condenser?), here I think that somehow it can interfere, and therefore I would remove it, if the cache works it is clear that it created problems, ultimately I am of the idea that it must be removed, also because I have not seen it in other boards, we will see if I can solve it by removing that SMD.

Going back to the M560, apart from the very thin PCB, which flexes a lot by inserting the cards into the ISA and PCI slots, it seems that it is not bad at all in terms of performance, maybe the chipset will not be like the Intel one, but as a global performance it could not be very far away, and here the talk about the cache could be important.

As a PCChips card, it seems that it is not among the terrible ones, I am referring to those cards with fake cache chips, here the cache chips I am sure that they work, because on a recently fixed board (M550 i430TX), there are identical cache chips and they work, so I think first remove the SMD solder next to the 0 OHm resistance 🕉, then I should check if current reaches the chips, I noticed a possible second problem (factory?), there is a JP8 jumper, which is used for the voltage of the cache chips, there is nothing connected, so or not It comes current, or it's on a fixed value, which could be 5V or 3.3V, if this doesn't work out either, I'd have to look if by chance there's something else that escapes me.

The M560 I have is a 4.1A version, I think it's the last of these cards, and it could have some small improvements over the previous version 3.1.

Regarding the minimum voltage of the VCORE, it is possible to set 2.5V without any jumper, with this voltage you can put a K6-2 or III, which work equally with a slight overvolt, maybe they heat up a little more, but if you do not exceed 2.6V, which is the maximum VCORE for these K6 CPUs, regardless of whether they are 2.2V 2.3V or 2.4V, indeed the 2.4V K6-2 450 seems to be the first version of these CPUs (maybe not CXT?), and therefore they could be more problematic than the other 2.2V, but I don't have one available, and so don't say Whether it's the same or not.

I will definitely increase the RAM, the 16 MB module, I will use it as a module to try other motherboards, because otherwise I would have to look for a different module each time, with the doubt that it works or not, in short I thought for this card that 64 MB or even 128 MB are ideal, and maybe I could use only one memory module, so if necessary I can double it with a second module, I would reach 128 MB or even 256 MB, which is more than enough for this PC.

I just finished upgrading the M560's RAM, removed the 16 MB dock, which I will reuse in other motherboards during boot tests, and put on a module that I had discarded in other PCs.

It is a 128 MB single-sided SDRAM, PC133 CL3, in this PC it will work at most as PC83, for this reason it is very likely that it can be used as CL2, if it does not already, for this you should check from BIOS, and eventually you can change the CL, using the lower one, that I know is 2, less than that I do not know if it is possible.

I currently have 128 MB of RAM, later I could add if necessary, a second 64 or 128 MB module, in this way I would get 192 MB or 256 MB, these quantities are already sufficient for Windows XP, but for a PC S.7 they are much more than what is needed, usually 64 MB are to be considered enough, for DOS even 32 MB is quite a lot, but if I had a quantity of only 32 MB, I would use at most Windows 95, for Windows 98 64 MB or more, they are enough to make it run quite well.

Going back to JP8, I did some checks, and in fact on pin 1 and 3 (although in reality there are not), I found about 3V and 5V, while on pin 2 0V, this means that via jumper on 1-2 or 2-3, you get voltage to pin 2, which should power the L2 cache chips, which probably doesn't happen, this would explain why it doesn't work, at this point I don't know whether to first put the three pins of the JP8 with a jumper, or try to remove the SMD soldered next to the 0 OHm resistor 🕉.

Checking the cache chips better, from the pdf it appears that the VCCs are at about 3V, so this is the first check to do, but first you have to see if there is continuity with the 2 pin of the JP8, in that case there are two things, or the current comes anyway and therefore that jumper was useless, which is why it was removed, or the current does not arrive, and therefore it is necessary to solder the three pins (two are enough for the 3V selection), so that the VCC pins receive the current at the right voltage.

Before welding and adding the jumper on JP8, I would still like to make sure that in the VCC pins there is something like 3V, if they are less it is clear that there is a problem, but above all it explains why it does not work, if instead I found about 5V, the discussion would be more complex, I would have to understand if there is any defective component, and why the voltage is higher than that of normal operation.

The VCC pins to be checked are the following:

15 41 65 91

You have to see if they are all connected to each other, I guess so, and if there is with the motherboard on, a voltage higher than 3V, I don't know if by disabling the L2 cache from BIOS, there is voltage or not (?), I would say that it must be measured with the active L2 cache, if there is at least 3V, it remains to check the GND pins, if they are in place, it becomes more complicated to understand what problem there may be.

From a check on the pins of the cache chip of the M560, I noticed that they are connected directly to the PIN 1 of the JP8, and that therefore there should be the same voltage, they are about 3.6V a somewhat high value but still within the 3.1V-3.6V found in the pdf for the VCC.

From this I understand that it is probably useless to connect the PIN 2, and it is very risky to insert the three pins in the JP8, because you could accidentally connect the 2-3, and send about 5V to the VCC PIN, so I would say that the doubt that the chips are not powered has vanished.

The only thing I will do, in the next few days, will be to remove that SMD on 2-3 of R119, because through the test diodes of the tester measures something between 400 and 800, in reality those PINs should be isolated, instead so they are connected, and this could be a problem, visually it looks like that factory, but all the photos I saw of the M560, they only have the resistor 0 OHm 🕉 on 1-2, maybe it will have been connected incorrectly, and I don't think it makes much difference if it's a capacitor or something else, basically it doesn't have to fit, and if after the removal the cache L2 works (we hope well), I will be able to continue with the updates, both CPU and cards, I will change the video one, and I will add the audio one, then maybe also some other USB-PCI type card, as a disk drive, I should try a fairly large memory card for Windows 98SE.

Moving to Chaintech 5TDM2, I had to lift the label of the BIOS chip, the chip is an ATMEL AT29C010A, I think it is a fairly common type of chip, and that therefore you can rewrite, without any particular problems, I downloaded a couple of BIOS, I will try as soon as possible to rewrite the chip, theoretically I should be able to save the content, but this depends, if it is too damaged I do not know what remains of the file, maybe it would just be truncated at some point .

I just finished the work on the M560, well the result was not what I wanted, that is, the L2 cache still does not work, despite having removed that SMD on PIN 2-3, in short, the problem exists, but it was not that one there, so I will have to do some other tests, not that I have much left to do, but I will do a check to see if the PINs of the chips move, in that case it could be some welding to create the problem.

Yesterday I tried a different VGA, a PCI Miro with 1 MB S3 chip, the exact model is Crystal 12SD and the S3 chip is a Trio32 (86C732-P), clearly it will not be the VGA for this PC, this would go on a 486 PCI, when I have fixed one.

Since with the M560, I couldn't make the L2 cache work, I did some tests, I tried to disable the L1 cache, thinking that the L2 remained active, instead that L2 is disabled together with that L1, I ran only one bench of Phil's, Speedsys getting a score of just over 20, let's say like a 486 DX2, but honestly I don't care that the PC slows down, indeed if anything the opposite, that it accelerates a little, and from this point of view, I didn't make any progress, unfortunately the L2 cache doesn't work, but I got an improvement by changing Video card.

I tried a 4 MB Matrox Millennium PCI, but this card won't go in this PC, it will probably go on the Pentium Pro, and from that I would recover a Matrox Mystique, which if I wanted I could use in another PC, recovering another VGA PCI, in short I have no idea which video card I will put?

Before trying it I fixed some PINs, because they were a little crushed, I straightened them and now they are quite straight.

Performed Phil's benches with the PCChips M560 and Matrox Millennium, there is a certain difference in the scores, some are slightly better, others earn more, for a more correct comparison I report the values obtained with the same frequency (166 MHz), but with the VGA SiS 6215 PCI that I had at the beginning of the tests on the bench.

I didn't do any bench with the Miro Crystal 12SD PCI 1 MB, because it's a card that I'll use in more dated PCs, if you want it could expand to 2 MB, adding two chips in the empty sockets, but at least for the moment, it's not necessary.

The results obtained are these, the first value is obtained with the VGA SiS, the second with the VGA Matrox.

2) 102.9 | 149.5 3) 70.9 | 119.8

4) 39.7 | 39.8 5) 33.7 | 40.4

6) 15.2 | 17.8

A) 211.59 | 235.62 b) 62.19 | 68.09

C) 33.6 | 39.7 d) 14.9 | 17.8

E) 13.2 | 14.1

I don't think it's necessary to make a difference in the individual scores, but it's obvious that some bench improves the scores a lot, and this depends on the better video card, which by the way the Millennium, you can see better in my opinion, not that the others look bad, but the colors seem more vivid, at least in this monitor that I use on the bench, a Philips LCD 170C, maybe using another monitor could change?

Going back to the 5TDM2 BIOS chip, it should have this pin out: