aaronkatrini wrote on 2021-05-30, 06:18:

One time I found this mATX PC-Chips with a Socket A and universal AGP for really cheap, it was my first time using a motherboard with SiS chipset. I don't know the exact model, but it is very common, but if you happen to find one, just run as far away from it as possible!
The design is so stupid, it has these two "big" electrolytic capacitors on the very left of the AGP slot, making it impossible to plug in some AGP cards.
Although the speed wasn't great, the major problem was compatibility with some AGP cards, especially with 3dfx and Matrox cards, where after installing the Video drivers, after a reboot you'd get just a white screen. This problem was documented on the internet, and apparently there was a BIOS Update that would fix this issue, but even after updating the Bios the problem would persist, at least with the Voodoo cards.

M810 series with SiS730 chipset most likely. I have the not-so-pleasant experience of doing coursework on one of these (slow, unstable). When I complained they said: "we will write these off 4 years early, we're sorry for your experience, replacement systems have been ordered already"

https://www.elhvb.com/mboards/pcchips/pics/m810lmr.jpg

Anything with a proprietary PSU. I hate when I need to try a different PSU but I can't because it has a weird proprietary connector.
Also boards that only fit a certain case so they are practically useless if you don't have that case.

• PCChips M747: major instabilities all the time, until it silently died one day. An incredible annoying experience, I made a thread about it.
• ASUS A7V8X-X: the VIA chipset drivers would make Windows crash randomly, I don't remember if I managed to resolve it by updating. But the worst thing about the board is that on every cold boot it would decide that a previous "overclock" failed and force me to set the CPU speed again in the BIOS... bullshit, I wasn't overclocking anything, no amount of BIOS updates solved the issue. Trash.

I've yet to come by something dreadful yet, but some boards have given me a little headache.

Recently I've acquired an Intel S7 board that's still not working the way I'd like. I'm not sure yet where the problem lies... But it looks like it's not storing BIOS settings.

My IBM Aptiva runs a K6 233 without trouble on the Ali IV chipset. Although the board should support 75 and 83MHz fsb I've never managed to set it up stable with a bus speed above 66. The K6-2 400 @66x6 does work fine, but any amd or Cyrix 75x something it higher won't run stable. I keep this one out of nostalgic reason.

My IBM 300pl only runs with a MMX CPU and 168p EDO, so getting an upgrade is a real contest. I think of getting rid of this one, but on the other hand it still is IBM.

My PC chips M577 (super7) is stable and very forgiving about CPU/ram choice. But overall performance is 20-30% slower than expected. But I still like it as a testing board.

And, speaking of lacking performance: ever had an IBM model 30 286?

evasive wrote on 2021-05-31, 07:44:

aaronkatrini wrote on 2021-05-30, 06:18:

One time I found this mATX PC-Chips with a Socket A and universal AGP for really cheap, it was my first time using a motherboard with SiS chipset. I don't know the exact model, but it is very common, but if you happen to find one, just run as far away from it as possible!
The design is so stupid, it has these two "big" electrolytic capacitors on the very left of the AGP slot, making it impossible to plug in some AGP cards.
Although the speed wasn't great, the major problem was compatibility with some AGP cards, especially with 3dfx and Matrox cards, where after installing the Video drivers, after a reboot you'd get just a white screen. This problem was documented on the internet, and apparently there was a BIOS Update that would fix this issue, but even after updating the Bios the problem would persist, at least with the Voodoo cards.

M810 series with SiS730 chipset most likely. I have the not-so-pleasant experience of doing coursework on one of these (slow, unstable). When I complained they said: "we will write these off 4 years early, we're sorry for your experience, replacement systems have been ordered already"

https://www.elhvb.com/mboards/pcchips/pics/m810lmr.jpg

Yes, it was one of those boards... don't ... just don't get one! 😀

Anything with a via chipset is a nightmare.

I also agree that ASUS quality slipped after the p2b.

Anonymous Coward wrote on 2021-06-01, 03:14:

Anything with a via chipset is a nightmare.

I haven't had problems with MVP3.

Since I see you are using an SATA SSD, I suppose you are using a separate controller? That might explain some things. A lot of grief was related to their crappy 4 in 1 driver.

My issues with various boards:
ECS K7VZA: Not bad, but the BIOS for it is ABSOLUTELY HORRIFIC. It takes seconds to turn of because the ATX soft-off button doesn't work as intended. Flashing it with a Chaintech 7AJA2 BIOS fixed it, and made it much more stable (yes, it's been recapped)
Jetway J-7BXAN Rev 2.1 - ACPI support only up to 98SE. 2000 and XP literally act like the board used is an AT mobo with no ATX function whatsoever.
EpoX EP-7KXA - no 133FSB support DESPITE there being an actual jumper, because the clock generator is unable to run a 133FSB, at least on a Thunderbird core Athlon.
ASRock G41M-VS3 - would randomly crash with a HD7870. Same GPU would work fine on a GB H55M-S2V and a ECS G41T-R3
Gigabyte 7VAX - random crashes in Win2k and hangs at random times. Caps are Rubycons all around.
ACorp 6VIA81P - seems to have died with a reset loop issue, and whenever I get it to POST, having a keyboard plugged in will make the BIOS press random keys for no apparent reason. Shame as despite having the VIA693 chipset, it ran 133FSB quite nicely, about on par with my 440BX based Soyo 6BA+IV
Luckystar P5MVP3-99I - Primary channel is dead (stalls after drive detection, does recognize it but then complains that it's empty when it's not), secondary works but only in compatibility mode. 2k will not install because of those issues, only 98SE or ME. If ITE 8212 RAID cards work on old MVP3 mobos, I might snag one and run two 160GB drives on it and leave the mobo's secondary channel for the CD-ROM.
Zida LX98-AT - annoyingly picky about SDRAM sticks. That's the only thing I can complain about it, otherwise it's a very solid babyAT form factor 440LX mobo.

evasive wrote on 2021-05-31, 07:44:

M810 series with SiS730 chipset most likely. I have the not-so-pleasant experience of doing coursework on one of these (slow, unstable). When I complained they said: "we will write these off 4 years early, we're sorry for your experience, replacement systems have been ordered already"

https://www.elhvb.com/mboards/pcchips/pics/m810lmr.jpg

Hey, I have that one too! Mine seems to have been branded Totem TM-S730LMV and does seem to be built a bit better. Wonder if I should give it a try with any OS. It uses a Matsonic BIOS, and from what I recall, Matsonic was on the higher line of ECS/PCChips rebrands, at least from my experience.

Pegatron 2A94h A3 (HP IPIEL-LA3). It only accepted Hyundai HMT125U6TFR8C-H9 and refused to work with anything else (tried 9 other RAM brands and 15 models). I hate it when I have tons of memory lying around and I have to order more because the board is picky.
It doesn't have a PCI slot either. I still use it because boards with 4xDDR3 SIMMs up to 8 GB and PCI are expensive for a 13 years old system.

PcBytes wrote on 2021-06-01, 08:34:

ACorp 6VIA81P - seems to have died with a reset loop issue, and whenever I get it to POST, having a keyboard plugged in will make the BIOS press random keys for no apparent reason. Shame as despite having the VIA693 chipset, it ran 133FSB quite nicely, about on par with my 440BX based Soyo 6BA+IV

Check the PS/2 keyboard data and clock pins without a keyboard plugged in. They should both be at 5V, as they are default high.

I just fixed a motherboard where the keyboard data pin was being driven to 0V via a dead ceramic capacitor. It caused the keyboard controller to go insane during bootup, with continuous beeping as the keyboard controller in the chipset was picking up system activity as noise and trying to interpret it as keyboard data.

Of course, it may not be worth saving a VIA693 motherboard. The one I have here is very slow.

canthearu wrote on 2021-06-01, 10:48:

Check the PS/2 keyboard data and clock pins without a keyboard plugged in. They should both be at 5V, as they are default high. […]
Show full quote

PcBytes wrote on 2021-06-01, 08:34:

ACorp 6VIA81P - seems to have died with a reset loop issue, and whenever I get it to POST, having a keyboard plugged in will make the BIOS press random keys for no apparent reason. Shame as despite having the VIA693 chipset, it ran 133FSB quite nicely, about on par with my 440BX based Soyo 6BA+IV

Check the PS/2 keyboard data and clock pins without a keyboard plugged in. They should both be at 5V, as they are default high.

I just fixed a motherboard where the keyboard data pin was being driven to 0V via a dead ceramic capacitor. It caused the keyboard controller to go insane during bootup, with continuous beeping as the keyboard controller in the chipset was picking up system activity as noise and trying to interpret it as keyboard data.

Of course, it may not be worth saving a VIA693 motherboard. The one I have here is very slow.

Tested your method and I noticed that in the reset loop I just described, I get 0V on clock, and data constantly fluctuates between 4.73 and 5.00V. Would like to get it going as it is very stable and relatively fast compared to my other 693 based board, a DTK PRM-27IV.

Three letters. E-C-S. Had a Soyo dragon motherboard that was a complete nightmare too.

fgenesis wrote on 2021-05-29, 16:21:

V4P895P3/SMT V1.0 is one sucker. Got it somewhat workable after trying for a long time, process detailed in a wall of edits. Even a semi-broken 286 board was easier to get working compared to that thing.
Plus it simply doesn't POST when EDO memory is inserted, as it requires FPM. That alone took forever to notice since i had never had to deal with a board that didn't like EDO. Gotta learn something new every day!

So weird! I have that exact revision and it's my favourite Socket 3 board — very stable and capable of running any CPU (up to 5x86 which is not a given for VLB boards) at 0ws, 40MHz FSB and tightest timings. The lack of EDO support is rather a norm for VLB boards, it's later PCI chipsets like SIS496/497 that usually support it.

I'm pretty sure there's something physically wrong with your board or the ROM is corrupted.

I am skeptical of the "doesn't work with EDO" I think it's like "Doesn't work with PC-100" ... it's just that higher density DRAM more common on the newer standard and it's those the board doesn't support.

jheronimus wrote on 2021-06-01, 15:51:

So weird! I have that exact revision and it's my favourite Socket 3 board

please please please dump the BIOS ROM and post it in the linked thread, i've been searching for that one forever and it hasn't been dumped yet!
maybe that bios will make a difference. will see once i get my hands on it.

BitWrangler wrote on 2021-06-01, 15:59:

I am skeptical of the "doesn't work with EDO" I think it's like "Doesn't work with PC-100" ... it's just that higher density DRAM more common on the newer standard and it's those the board doesn't support.

It doesn't. 8 MB FPM? All good. 8 MB EDO? Nope. The BIOS init doesn't even get past the 2 or 3 init steps, most of the time it displays one number on the debug card and the 2nd half stays "--", so it locks up really damn early.

How many chips on the 8MB EDO though?

EDO support is limited to a few late 486 chipsets, and it's not at all like using PC100 SDRAM on a board that supports PC66 for example. The only difference in that case is the speed, while using EDO on a board that doesn't support it will probably cause it not to POST or to have strange behavior. EDO RAM is accessed in a slightly different way (although boards that support EDO can usually also use FPM).

PcBytes wrote on 2021-06-01, 14:52:

Tested your method and I noticed that in the reset loop I just described, I get 0V on clock, and data constantly fluctuates between 4.73 and 5.00V. Would like to get it going as it is very stable and relatively fast compared to my other 693 based board, a DTK PRM-27IV.

OK, then you should check the clock and data resistance to ground and 5V. (using Ohms mode on your multimeter, make sure power is off)

Both should have a lowish resistance to 5V (1k or 10k). This is the pull up resistor. If resistance is really high, trace to and check the pullup resistor, it might be broken or bad connection.

Both should also have a substantially higher resistance to ground than to 5V. If resistance on the clock pin to ground is a lot lower than the resistance to 5V, this can also explain why you get 0V on the keyboard clock. This can be caused by a shorted filtering capacitor like what happened on the board I was fixing.

You may find that the pull up resistor or filtering capacitor are in SMD resistor/capacitor packs, or that there is a specific keyboard SMD protection IC that does both. You will have to trace it out and see what connects where.

Also check the 5V pin, make sure it is 5V when PC is powered on. If it is 0V, the keyboard fuse is probably blown, and keyboards will act really weird (and not generally work) because they will still pull a bit of power off the data/clock pins.

fgenesis wrote on 2021-06-01, 16:17:

jheronimus wrote on 2021-06-01, 15:51:

So weird! I have that exact revision and it's my favourite Socket 3 board

please please please dump the BIOS ROM and post it in the linked thread, i've been searching for that one forever and it hasn't been dumped yet!
maybe that bios will make a difference. will see once i get my hands on it.

I'm not sure I'll get around to it until weekend (right now the board is stashed away), feel free to PM me if I forget to do this. Just to make sure: you did try the BIOS from here?