• Please review our updated Terms and Rules here

430FX head scratcher

E

Ellron_Cupboard

Member
Joined
Dec 23, 2021
Messages
14
Location
Colorado
Hello,

I've been messing around with this board for some time on the side. and have yet to get it working, so thought id toss this out there and see if anyone has any ideas.

its a TMC pci54its v2.0 board with an intel 430FX chipset. essentially its unable to initialize the graphics card (one long,two short beeps w/ Award bios or 8 beeps with AMI bios.) i've tried about 7 different video cards in this thing to no avail, 4 16bit isa cards & 3 pci (early) cards.

the Award bios, with only the video card + post card in it will beep (as described above) and halt on floppy drive failure, following one single high and one single low beep (i think this is a system halt noise). With award, it will just continue to cycle on 8 beeps, but also halt at the floppy drive initialization.

now when i hook up a floppy drive and insert a boot disk, with Award it will actually boot off the DOS disk (still have the one long, two short beeps) and with a keyboard plugged in, i can type some basic commands (a: ; dir , etc) and i can hear the head seek -- so i'm fairly sure its actually booted into DOS. ctrl-alt-del also reboots PC at this point too. note that caps lock, num lock , scroll lock doesn't seem to do anything.

with the AMI bios + floppy + boot disk, it cycles between F0 and F1 for post codes beeping indefinitely.

i do know that all video cards work (tested in my other build with a 430HX chipset) along with the Dallas RTC ( cr2032 modded ).

so i guess it seems like this motherboard is booting, but for whatever reason isnt able to identify anything on the PCI or ISA bus w/ exception to the post card displaying post codes.

to date, i've used deoxit on the pci and isa slots, reflowed all pci/isa solder joints. verified 33mhz on the pci bus along with all voltages (save 3.3v). i've sourced about 6 different bios images (586 PnP bios') from the TMC website (via wayback machine). verified the 14mhz crystal going into the clock chip along with some other odds and ends.

also, maybe worth noting. regardless of if a video card is inserted or not. it will always beep as described above. i would expect maybe one long, 3 short beeps with no card installed. i'd also expect the system to halt when it cant identify or initialize a video card -- but alas it doesn't seem to do any of that.

i recently saw a thread on here discussing the 430NX chipset needing a specific type of dallas RTC config to function properly. i wonder if the 430FX is similar?

after writing this out it almost seems like this mobo is a good candidate for the wall of failure.

anyways, thanks!
 
Sounds like the system is configured for no/cga video card. Have you cleared the bios yet?
 
Yeah, I had that thought last night. I cleared the cmos via the dip switch on the board. Same result. I also popped the award rom into modbin and double checked that it was set for vga on the display adapter, which by default it was. While I was at it I set the pci settings all to auto and reflashed the chip. Haf the samw one long two short beep code. Also for grins, I set num lock to 'on' during boot up. Didn't seem to make any difference and the lock lights still aren't changeable via key press.

I was considering a north/south bridge issue but honestly if either of those were bad I would expect no cpu activity.

Southbridge is clearly working as it's communicating with the super io chip along with the bios/cmos. And by proxy the north bridge should be working too.

Removal of the rtc/cmos results in no post as I'd expect.

It's almost like the rom isn't loading the cmos correctly. But attempting other rtcs I've got around yields the same issue.

Maybe it's time to break out the oscope and look at the data lines between cmos and rom. Although if one or more lines was stuck it'd probably not post either... maybe?
 
Can you post a video by chance? And some close up shots around the memory... Keep in mind these boards can be really picky with PCI 2.0 video cards, sometimes even ISA -- Mine wouldnt post whatsoever on Oak VGA cards, but fired up no issue with my Wyse Tseng ET4000ax. What cards have you tested so far?

Also what memory are you using? 72 pin EDO or FPM or are you using PC66/100/133 SDRam?
 
yeah, i'm a little worried this is just a case of a really picky motherboard... but the sheer number of cards ive tried also has me slightly convinced something else is a miss.

so it currently has 32mb of 72 pin EDO each stick in each bank matches, but both banks are not the same brand. i've tried it with only 16mb / one bank filled with no change in result. i guess it is worth noting that with no ram , the system beeps continuously as expected... so i suppose thats good. i've had the though maybe the bios is defaulted to FPM perhaps, need to check. but i'd expect no activity if that was the case?

while i'm thinking about it too, i've tested two different 166mHz and a 133mHz CPU in this thing. same result.

i've tried the following cards in this mobo (in all slots just to be sure):

PCI: -- maybe i need to get a few more pci cards to test i suppose... any suggestions?
STB S3 Trio 64v+
Chips F66548 ( this is a custom card for a dolch of some type )

ISA:
Chips F65520 (custom vga for dolch)
TSENG ET4000AX
OTI 037C
Paradise PVGA1A

i'll toss some pictures up here in a second from my cell phone (do ignore the sloppy work on the RTC battery mod, its just a temporary thing during testing) . will try to direct upload video but may have to look at google drive or something.

i've also been slowly testing the sop logic chips, theres still a few 245's left to test up near the memory. so far everything checked out just fine. cache and tag chips stay cool. i will add no chips get super hot with exception of the main 3v regulator on the heat sink. there is a 5V regulator right near the clock chip that does get pretty hot, i think this is going into the bottom right pin of that IC if memory serves... i had initially suspected this as a culprit and pulled it out of circuit, verified OK. i let that theory go as i recalled that the socket 7 3v regulator is known to get hot. i'd assume the clock chip is also going to get warm to since its generating all the various signals for the mobo.
 
Ok everything looks clean. What kind of ram are you using? EDO? When testing the board keep the battery off if you can. It will help clear the bios and one less thing to diagnose.

The s3 card really should work. I would next get a multimeter and make sure the card slots are getting power. Here is a diagram of the pci and isa slots.

1654825927452.png

PCI pinout - Attie's Wiki

wiki.attie.co.uk wiki.attie.co.uk
 
Last edited:
Ok read that you have EDO, that should be fine. Ya go ahead and read the voltages on the isa/pci bus and report your findings please. Perhaps something isnt right...

Do you have a programmer by chance? I can dig through my pentium boards, think I still have an FX board.
 
Ellron_Cupboard said:
i recently saw a thread on here discussing the 430NX chipset needing a specific type of dallas RTC config to function properly. i wonder if the 430FX is similar?
Click to expand...
I asked a knowledgeable user 0xCats on our discord server (who has done some significant BIOS modding on the 430NX chipset), about this, and he is pretty sure it is a red herring. Despite this I currently have 3 NX boards that are non-functional.

If you are looking for BIOSs we have over 18,000. Here are the ones for the 430FX
www.ultimateretro.net

The Retro Web

The Retro Web hardware database
www.ultimateretro.net www.ultimateretro.net

Here is what we have on your board:

TMC/MyComp/MyNix/Megastar PCI54ITS (VER. 2.0)

TMC/MyComp/MyNix/Megastar PCI54ITS (VER. 2.0) is a motherboard based on the Intel 430FX (PCIset FX Triton I) chipset. Get specs, BIOS, documentation and more!
www.ultimateretro.net www.ultimateretro.net
 
mR_Slug said:
I asked a knowledgeable user 0xCats on our discord server (who has done some significant BIOS modding on the 430NX chipset), about this, and he is pretty sure it is a red herring. Despite this I currently have 3 NX boards that are non-functional.

If you are looking for BIOSs we have over 18,000. Here are the ones for the 430FX
www.ultimateretro.net

The Retro Web

The Retro Web hardware database
www.ultimateretro.net www.ultimateretro.net

Here is what we have on your board:

TMC/MyComp/MyNix/Megastar PCI54ITS (VER. 2.0)

TMC/MyComp/MyNix/Megastar PCI54ITS (VER. 2.0) is a motherboard based on the Intel 430FX (PCIset FX Triton I) chipset. Get specs, BIOS, documentation and more!
www.ultimateretro.net www.ultimateretro.net
Click to expand...
I'm the one who made the thread about the 430NX. I've got 4 boards now that all exhibit the exact same fault, and refuse to work properly with NOS/Glitch Replacement 12887s, but will "work" with their original Dallas chips (just not save any information, and have garbled data for time/date) - If it's not RTC "formatting" related, I'm not sure what it could be at this point.

Sorry to drag off topic for the post lol.
 
I just posted a potential solution to this problem with the NX chipset on your other thread:

In short, the older BIOS images for these boards have some issues, being the first attempt at a PnP BIOS on the Pentium. The later releases fix a lot of weird PnP issues, as well as issues like accepting the newer RTC modules that have more memory. Updating the BIOS on my NX chipset Dell allowed me to use the new Dallas RTC.
 
So pulled out my Dell p166, and it is indeed FX chipset based. I'll hook it up tonight and dump the bios. Sadly the bios is soldered to the mainboard, so can't use my programmer. :\

Mine does not have a dallas chip on it, so maybe it would work.
 
Didn't have a chance to get to it today, but will have a dump for you tomorrow evening. My dell's board is near identical except the dallas chip.
 
hey twolazy, i checked the voltages today on both ISA and PCI. they were a solid +/-5 V & +/-12V no 3.3v on this board verified via the o-scope and dvom. unfortunately the last pin i checked my hand slipped -- guessing i put 12v on a dataline or something. it seems i've killed this board as it no longer posts. i'm going to take a few days off from it and go at it with a fresh set of eyes to get it back to the level of broken it was before today.

i do have a couple of videos on that board beeping. still sorting a way to get them up. thinking google drive i guess could work.

i do have a minipro and a few other flashers around. so pushing some new bios' to test shouldn't be a problem. (thanks for the help on this by the way.)

mR_Slug said:
I asked a knowledgeable user 0xCats on our discord server (who has done some significant BIOS modding on the 430NX chipset), about this, and he is pretty sure it is a red herring. Despite this I currently have 3 NX boards that are non-functional.

If you are looking for BIOSs we have over 18,000. Here are the ones for the 430FX
www.ultimateretro.net

The Retro Web

The Retro Web hardware database
www.ultimateretro.net www.ultimateretro.net

Here is what we have on your board:

TMC/MyComp/MyNix/Megastar PCI54ITS (VER. 2.0)

TMC/MyComp/MyNix/Megastar PCI54ITS (VER. 2.0) is a motherboard based on the Intel 430FX (PCIset FX Triton I) chipset. Get specs, BIOS, documentation and more!
www.ultimateretro.net www.ultimateretro.net
Click to expand...

thanks for that link mR_Slug. i have some additional bios' and documentation i can upload there for the next person to come along.

i'll continue to keep an eye on that NX thread, i'd almost agree with yall that something fishy is with the early PnP bios' and/or the dallas rtc 'factory' programming.
 
Since you can't use a video card, how is it handling a COM port being present?
Try setting up a bootable DOS floppy that points the command prompt to COM 1 and attach a terminal. It's unlikely to work but if the system is booting you can hope it's at least initialized the COM port.
 
Actually, that's not a bad idea at all.. I'll give this a go and report back once I sort out what I broke the other day. Thinking I'll take a stab at fixing it on Sunday when I get a few free minutes.

I guess my question is, if I can get into the system via the com port. Are there any utilities that can be run to aid in any meaningful diagnosis?
 
If the memory of the Dallas module is accessible to DOS, DEBUG might be able to reach into it to read/write values. It's a stretch but that's how we would reprogram the NVRAM in a Sun to clear garbage data.
 
Had a few moments with it today. Looks like I killed the northbridge from my goof the other day. It was quite hot when powering it up. I snagged another off an old parts board, once I desoldered it the pin alignment was quite off so not really worth reusing. I see a few nos 430fx66 ics on ebay for $10-20. I'm going to have to take a break from this and decide if it's worth fixing. For now I just used the 430hx board in the target system with a 200Mhz mmx cpu, which honestly is a better solution anyway. Aside from that, the MakeMhz xbox mod just came in the mail today, so I need to shift focus on that puppy to get some guaranteed good times in for a little while. Maybe with the shift in weather in the fall i can give this another shot!
 
You must log in or register to reply here.
Back
Top