LocalBus wrote on 2022-02-03, 22:39:

Q1 (PNP) is done. I only get 3.7V to 82c206. I think we are finally getting somewhere. I could swear I measured this before!

Is that with the PSU on? I'm pretty sure you had a measure of that before as well.

Will be interesting if that stops the battery voltage appearing at the BIOS ROM. I'm pretty sure the ROM should only be powered by the board 5V, and Q1 is supposed to stop the battery from trying to power the whole board.

snufkin wrote on 2022-02-03, 22:44:

LocalBus wrote on 2022-02-03, 22:39:

Q1 (PNP) is done. I only get 3.7V to 82c206. I think we are finally getting somewhere. I could swear I measured this before!

Is that with the PSU on? I'm pretty sure you had a measure of that before as well.

Will be interesting if that stops the battery voltage appearing at the BIOS ROM. I'm pretty sure the ROM should only be powered by the board 5V, and Q1 is supposed to stop the battery from trying to power the whole board.

That's with the PSU on and I don't have the battery soldered on the board right now, it was just causing confusion. Once the issues are resolved I will put it back.

Yes it was strange that we could observe battery voltage at Vpp pin of the BIOS ROM. You were right all along! The PNP measured kind of OK, but that was only half the truth then 😀

While we are at it, you mentioned about an "external battery connector" - is this common? I have not spotted any obvious location for it.

"Test CMOS RAM Checksum if bad or Insert key depressed; load defaults." most likely means set global flag to display "CMOS corrupter, press Insert key to load defaults" _after_ we already initialized keyboard and graphics.

great about the beeps when holding keys, means IO access is at least working for keyboard controller

I hope Chkcpu comes back. Im a BIOS disassembly virgin and having problems. Loaded it up in IDA and the very first thing I see is

1ror     edx, 10h
2cli
3cld
4in      al, 64h
5test    al, 4
6jnz     short loc_F00DD

part of warm reboot detection, so far so good, but IDA only disassembled about 1% of the file 🙁 While what I got makes sense (reads/writes to CMOS, some floppy initialization, some keyboard code, 486 cache control, PCI) there are zero writes to 80h in there so definitely Im bad at this 😀

It'll (probably) be JP21, next to the CMOS clear JP24. They're pretty common, especially on boards that have barrel batteries. You'll want to double check, but I think the pin closest the board edge will be the +ve pin and the pin furthest away will be the ground pin. It looks like the external battery goes through CR5 then CR4. So a 6V external battery (as shown in the 206 datasheet) should be fine. 4.5V would probably work as well. The diodes will protect the external battery from being charged, so ordinary AA will be ok.

I'm not certain about this... but given you've removed the battery, I think it would be safe to just short out from the collector to emitter of Q1. That will then connect +5 to the 206, which is what should happen anyway when Q1 is working and the board PSU is on (board PSU turns on Q1, allowing +5 on to the battery net). Obviously it won't be able to save any CMOS settings, and the sequence for that battery good pin on the 206 might not be right, but it might give a clue if this is the problem.

snufkin wrote on 2022-02-03, 23:19:

It'll (probably) be JP21, next to the CMOS clear JP24. They're pretty common, especially on boards that have barrel batteries. You'll want to double check, but I think the pin closest the board edge will be the +ve pin and the pin furthest away will be the ground pin. It looks like the external battery goes through CR5 then CR4. So a 6V external battery (as shown in the 206 datasheet) should be fine. 4.5V would probably work as well. The diodes will protect the external battery from being charged, so ordinary AA will be ok.

I'm not certain about this... but given you've removed the battery, I think it would be safe to just short out from the collector to emitter of Q1. That will then connect +5 to the 206, which is what should happen anyway when Q1 is working and the board PSU is on (board PSU turns on Q1, allowing +5 on to the battery net). Obviously it won't be able to save any CMOS settings, and the sequence for that battery good pin on the 206 might not be right, but it might give a clue if this is the problem.

Thanks! Yes I always wondered what was the deal with JP21, as not in jumper pins then but battery connector 😀

Your are absolutely right, now without the battery I could simply short Q1 to get proper Vcc voltage all the way to 82c206!

I wonder how much current the replacement BJT would need to deliver, I found 2N3906 readily available.

Good luck. Hope you get this board working.

I didn't read everything but one weird thing I have run
into before was a bad BIOS chip. Even though I was
able to read the chip with a programmer, the board
would not POST. Replaced the BIOS chip and the
board started working.

I didn't use a POST card on that one so no idea what
kind of error would show up.

LocalBus wrote on 2022-02-03, 23:38:

I wonder how much current the replacement BJT would need to deliver, I found 2N3906 readily available.

Opti 82c206 datasheet doesn't say. A Chips 82c206 datasheet says 30mA. Plus there'll be the current for the RTC oscillator, but that'll be tiny. 2N3906 says it can handle 200mA so should be fine. Although the one that was there has broken and I don't know why that'd be. If there's been some voltage or current spike that's damaged it then there may be other damage to be found.

cyclone3d wrote on 2022-02-03, 23:52:

Good luck. Hope you get this board working. […]
Show full quote

Good luck. Hope you get this board working.

I didn't read everything but one weird thing I have run
into before was a bad BIOS chip. Even though I was
able to read the chip with a programmer, the board
would not POST. Replaced the BIOS chip and the
board started working.

I didn't use a POST card on that one so no idea what
kind of error would show up.

Thanks, I hope so too! I'm hell-bent even to get this working 😀

Yes it doesn't sound too far fetched if the BIOS ROM is starting to go flaky, that it doesn't work on the mainboard but is perfectly fine on the desk.

In this case I think the ROM is just fine; I put all my bets on the faulty PNP transistor feeding the 82C206 with Vcc (+5V).

snufkin wrote on 2022-02-03, 23:55:

LocalBus wrote on 2022-02-03, 23:38:

I wonder how much current the replacement BJT would need to deliver, I found 2N3906 readily available.

Opti 82c206 datasheet doesn't say. A Chips 82c206 datasheet says 30mA. Plus there'll be the current for the RTC oscillator, but that'll be tiny. 2N3906 says it can handle 200mA so should be fine. Although the one that was there has broken and I don't know why that'd be. If there's been some voltage or current spike that's damaged it then there may be other damage to be found.

Just for the fun of it I will connect the DMM in series and try to get a reading of the current draw 😀

It took me some time to analyze this 2A5ULT40 BIOS - POST_B procedure. 😉
A lot of calls to supporting subroutines are made here, so I had to look at many different code fragments throughout the BIOS. I also encountered several unfamiliar variables that I had to look up in earlier POST procedures to understand their meaning.

Most of what POST_B does has to do with the RTC and other CMOS registers. So finding a problem with the 82C206 supply makes sense.

My analysis:
1) POST_B starts by checking the RTC’s 1 second pulse, via CMOS register 0Ah bit 7 (RTC update in progress bit). If present, the RTC’s time and date registers are checked for valid values and if invalid, the clock’s time and date are set to a valid value.
When the pulse is absent, the clock is obviously not running and the validity check is simply skipped.
2) Next, the keyboard controller is directly accessed via I/O ports 64h/60h to check for a depressed Insert key. If so, the “Load BIOS defaults” flag is set.
3) Earlier, just after the initial RAM test, the BIOS has setup a stack-frame to house all its variables. POST_B now copies the 128 CMOS registers to the first 128 bytes of this stack-frame to have a CMOS work-copy in RAM.
However, if earlier checks revealed a bad CMOS checksum or power loss of the RTC, or if the Insert key was pressed, the BIOS goes through a lot of code to fill these 128 bytes of the stack-frame with default values.
Obviously, these CMOS values on the stack will be manipulated by later POST procedures or by the user via the BIOS SETUP screens. At the end of POST, the BIOS will write these stack-frame values back into CMOS for use during the next cold boot.
4) Finally, at the end of POST_B, a routine is called that manipulates several chipset registers. As we don’t have the OPTi 571 datasheet, I don’t know what happens here. But comparing this with other Award v4.50G BIOSes, I think it is an early resource setup for PCI devices.

Although I didn’t find any clues for a possible hang in POST_B, a lot depends on the 82C206 Peripheral Controller because, apart from the Real-time Clock Subsystem, it also contains the DMA Subsystem, the Interrupt Controller, and the Counter/Timer Subsystem.

As the video setup is done in POST_D and POST_E puts the first BIOS messages on the screen, a videocard issue should not be a factor in this POST_B hang.

Jan

Wow. Can you tell if it's doing something like trying to use the 206 timer as a trigger for when to decide the the RTC update pulse is absent? I had been thinking that the 206 might have been ok as I thought DMA would have been tested by this stage, but looking at that generic Award test order I now see that DMA is only initialised early and the actual testing comes after 0B. So it may not have actually needed to use anything on the 206 until this stage. Hopefully getting proper power to it will make a difference.

Brilliant! We have definitely narrowed it down to the 82C206 😀

Jan, you have inspired me to go down the rabbit hole of disassembling BIOS routines. But first things first, let's get this 82C206 powered-up!

I guess the lack of RTC is not a "show stopper" as far as the POST routines are concerned, but maybe step 3) is when trying to access the CMOS registers.

I'll keep you "posted" once I have hot-wired the Vcc to 82c206.

Cheers!

Ok I jumped the Emitter and Collector of Q1 and got some ~35mA current draw. However still no dice.

Took out my Flir One and I can see a small heat spot on the 82C206 around the corner of pin 80 and pin 81.

It just won't fire up properly?

Not sure if there is any space underneath the chip so there is a shorted decoupling capacitor? There is one next to pin 26 which is one of the Vcc pins and the other is pin 78.

So I cannot really say that the 82C206 itself is shorted since it doesn't get warm at all.

Or maybe it is as simple as the 82C206 is gone...

Anyways, I have a good reason to practice my soldering skills to remove the PQFP package. Will use low-melt solder (ChipQuik).

Judging by the FLIR image it looks like a rectangular spot... could I be so lucky that there is an SMD capacitor hiding underneath? 😀

I recon there are UMC replacements?

Boo. Again. The 32kHz clock input to the 206 is pin 75, so maybe that's the cause of the heating in that area. Other than that, it's just IRQ pins around there. Check that the voltage on both sides of R64 (near Q1) is now 5V.

Shorted decoupling would draw a lot more that 35mA, and there won't be any room under that chip anyway (and there's nothing on the back of the board).

Hmm. Other clocks on the 206? Pin 15 should be the System clock. It goes to various places, including pin 74 (datasheet calls this ATCLK) of the 82c822. I think it might also be the ISA clock, pin B20 on the ISA slots. That's with some guessing about where it's going, but I think it ends up on R47 between the PCI and ISA slot, which is about the right place to connect to B20. Might be worth checking if those are all connected together, and if they go to any pin on the Chrontel clock gen (mayyyybe pin 13?).

Yes! you have a problem, Not the IC, the hot spot points to shorted even partial short will do that, you have to check pin by pin on that 206 to find the one that does not look right.
Use diode mode method with red probe on ground, and use black probe to check. Normal is anywhere .200 to .800 diode voltage drop.

ohm measurement will not activate the diode voltage drop or bad IC since the voltage is not high enough.

Diode mode measurements is routinely done on cellphone diagnosis.

Cheers,

snufkin wrote on 2022-02-04, 17:40:

Boo. Again. The 32kHz clock input to the 206 is pin 75, so maybe that's the cause of the heating in that area. Other than that, it's just IRQ pins around there. Check that the voltage on both sides of R64 (near Q1) is now 5V.

Shorted decoupling would draw a lot more that 35mA, and there won't be any room under that chip anyway (and there's nothing on the back of the board).

Hmm. Other clocks on the 206? Pin 15 should be the System clock. It goes to various places, including pin 74 (datasheet calls this ATCLK) of the 82c822. I think it might also be the ISA clock, pin B20 on the ISA slots. That's with some guessing about where it's going, but I think it ends up on R47 between the PCI and ISA slot, which is about the right place to connect to B20. Might be worth checking if those are all connected together, and if they go to any pin on the Chrontel clock gen (mayyyybe pin 13?).

Boo indeed!

Confirmed 5V on either sides of R64.

The SYSCLK I have traced to 82C822 so I think that one is alright.

So maybe time to look for shorts on random I/O pins on the 82C206 then, IRQ pins for instance.

The 82C206 does not heat up so I hope it is not bad.

Oh well, to be continued! Thanks for the support!

pentiumspeed wrote on 2022-02-04, 18:22:

Yes! you have a problem, Not the IC, the hot spot points to shorted even partial short will do that, you have to check pin by p […]
Show full quote

Yes! you have a problem, Not the IC, the hot spot points to shorted even partial short will do that, you have to check pin by pin on that 206 to find the one that does not look right.
Use diode mode method with red probe on ground, and use black probe to check. Normal is anywhere .200 to .800 diode voltage drop.

ohm measurement will not activate the diode voltage drop or bad IC since the voltage is not high enough.

Diode mode measurements is routinely done on cellphone diagnosis.

Cheers,

Glad to hear there is still hope for the IC! Ok diode mode and check on different I/O pins - got it! 👌

Yes believe me, I have been watching a lot of YouTube lately about microsoldering and repairs, which got me thinking about this board to begin with 😀 Still novice though 🤔

If I cannot find any culprit by diode mode measurements, then I will lift the IC to see what's going on underneath.

This is a super cool board..

I’m looking at it, and all like: needs sata card, voodoo 6000, k6-3+, audigy 2 zs, 512mb ram, “laughs maniacally” ….if you get it working that is…

snufkin wrote on 2022-02-03, 18:07:

Do you have any ISA cards you could try with, in case there is a problem with the LocalBus -> PCI bridge?

Btw, something I was confused about, (maybe the manual covers it if you have it) was wether or not that diag card was keyed for 5v pci or 3.3v late model pci

The answer is it is indeed keyed for the oldest pci slots. (If you have the same one I have, which it looks like you do). So it should plug into your pci slots without issue being they are among the first generations of pci undoubtedly.
Idk if you think it may be helpful to read from pci as well. (Just don’t try to plug it into a newer board)

Sphere478 wrote on 2022-02-04, 19:43:

This is a super cool board.. […]
Show full quote

This is a super cool board..

I’m looking at it, and all like: needs sata card, voodoo 6000, k6-3+, audigy 2 zs, 512mb ram, “laughs maniacally” ….if you get it working that is…

snufkin wrote on 2022-02-03, 18:07:

Do you have any ISA cards you could try with, in case there is a problem with the LocalBus -> PCI bridge?

Btw, something I was confused about, (maybe the manual covers it if you have it) was wether or not that diag card was keyed for 5v pci or 3.3v late model pci

The answer is it is indeed keyed for the oldest pci slots. (If you have the same one I have, which it looks like you do). So it should plug into your pci slots without issue being they are among the first generations of pci undoubtedly.
Idk if you think it may be helpful to read from pci as well. (Just don’t try to plug it into a newer board)

Thanks! I always felt that there was something peculiar about this board. Before reading up on this forum and reconnecting with legacy x86 systems, I thought that this was a combined 486 / Pentium board. Provided the chipset it is hosting, I think it would be technically possible.

Haha well I do not intend to go over the top with this build, I would just love to relive the moment I first fired up Doom 2 on my 486 DX2 back in 1994. This would fit the bill, even a working "turbo" switch on this board 😀 (provided I get it working).

Regarding the PC analyzer card, the keying allows it into the PCI slot, but the PCB is very thick so I feel it might crack the brittle plastic.

Hopefully the VLB-TO-PCI bridge is working once we reach to that point; past the CMOS issues we are trying to solve.

All in all I must say I am stoked by the collective experience, support and engagement from all of you, much obliged! 😀

Chkcpu wrote on 2022-02-04, 11:59:

It took me some time to analyze this 2A5ULT40 BIOS - POST_B procedure. 😉

I was hoping for a listing 😀 Can you hint where POST_B starts (offset)? Whole dump contains only two occurrences of 0xe680 (out 0x80, al), and they arent even valid code when looking at bytes around.

LocalBus wrote on 2022-02-04, 15:02:

Brilliant! We have definitely narrowed it down to the 82C206 😀

Im not so sure anymore, at least not the CMOS area. I found few calls to read/write RTC/CMOS at the very start of the bios

F000:E98A read_CMOS
F000:E99B write_CMOS

and they definitely dont block the boot process. It looks to me like the very first thing bios does on warm reset is read CMOS 0xF offset to check for reboot reason https://www.cs.yale.edu/flint/feng/research/B … ly/cmosram.html

1mov     al, 8Fh
2out     70h, al
3out     0E1h, ax
4in      al, 71h  

reading empty space (206 chip not powered up/in low power mode) would result in undefined values (maybe whatever was last on the bus?), but I still dont know how that would make the bios stop proceeding that late.
I would definitely go once more over all 206 pins under magnification, and measure
pin 6 PWRGD high
7 PSRSTB high
9 TEST low
if you have a scope/logic probe 74 AS being strobed after reset (it has same role as Chip select for accessing RTC/CMOS)