VOGONS


Am386DX/DXL not posting

Topic actions

Reply 60 of 78, by Eep386

User metadata
Rank Member
Rank
Member

To be honest, Trident 8900D-R's are a bit of an odd duck. They were apparently producedtoward the very end of the 8900 series' market life as a way to eke one last little bundle of money from an aging technology. As with most later 8900D's they seem to be decent enough performers though, about on par with early-revision ET4000AX's.

Life isn't long enough to re-enable every hidden option in every BIOS on every board... 🙁

Reply 61 of 78, by mkarcher

User metadata
Rank Member
Rank
Member
Predator99 wrote on 2021-01-05, 17:02:
Eep386 wrote on 2021-01-05, 16:33:

The fact that you get a colored screen or blinking cursor indicates that the board sees the VGA BIOS and is able to initialize the card to some degree before something else brings everything to a halt.

Thats indeed a good point. Totally new to me that there cards with integrated BIOS and I have seen many...

Another card with integrated BIOS (stub): The Tekram DC6x0 controllers with ST100AII, ST200AEI or ST300ALI chip. It's only used with Firmware 2.x, whereas Firmware 1.x disables the integrated BIOS, and needs a separate BIOS chips.

Reply 62 of 78, by Neolyum

User metadata
Rank Newbie
Rank
Newbie

This was another question I had - how could I see a cursor?
But I really had this once - now I just get the black screen everytime, the board hanging on 2C.
If this card has a BIOS, what could be the issue?

And I can't find any manual for this card online, none matches my layout, does someone has one or knows where to look? 🙁

Reply 65 of 78, by GigAHerZ

User metadata
Rank Oldbie
Rank
Oldbie

You mean your motherboard's bios?
It may very well be 27C512 type instead. 😉

"640K ought to be enough for anybody." - And i intend to get every last bit out of it even after loading every damn driver!

Reply 66 of 78, by Neolyum

User metadata
Rank Newbie
Rank
Newbie

So I flashed a Trident TVGA8900D (no R!) onto a 27C256 - no change. Black Screen, nothing else. I tried to change some jumpers on the card, also no change. Could the VGA the Problem as the board seems to boot fine sometimes? The card was sold as "working" to me, but who knows. I have a second EPROM which I could try, while waiting for the other one to be erased (380nm light for a longer time should be enough, although not perfect, right?)

Yesterday the card gave me a blue screen one time, don't know if it's relevant.
What can I try furtheron? I have been thinking, with the keyboard BIOS presumed dead and the board itself not seeming to be stable, I thought it would be less stressful if I just buy another board. I own a AMD 486 DX4-100, one 486 66 and a Pentium 75. Should I buy a board for the pentium or the DX4?
Before that I would try to desolder the Keyboard BIOS socket and fix the traces, but I don't know if I'm capable of that, my hands are not quite calm 😜

As attachment my BIOS, if someone's interested ^-^

Attachments

Reply 67 of 78, by Predator99

User metadata
Rank Oldbie
Rank
Oldbie

Try to program a system ROM with Supersoft and listen to the beep codes. No need to install VGA, its not supported anyway.

Attachments

  • Filename
    Supersoft-XT-64k.zip
    File size
    6.06 KiB
    Downloads
    6 downloads
    File license
    Fair use/fair dealing exception

Reply 68 of 78, by Neolyum

User metadata
Rank Newbie
Rank
Newbie

it first beeps 5 hi/lo 9 lo for cannot initialise video (expected) and then 1 hi/lo 7 lo for DMA controller 1 and then halting.

Edit: now it beeped 1 hi/lo and 8 hi/lo for DMA 2 additionally and 3 hi /lo 4 lo.
I used this PDF as reference http://www.minuszerodegrees.net/supersoft_lan … rs%20Manual.pdf

Reply 69 of 78, by Neolyum

User metadata
Rank Newbie
Rank
Newbie

Hi again 😀 I‘ve made some progress with the board 😀 after some months just laying around, I found some time today to further test it. And indeed it showed some other signs today! It seemed to boot with the keyboard controller inserted, which it never did before.
And one time, it did show me some text on the screen 😁 very briefly though, it Looked like the memory test being performed, I could read „Wait...“ before it rebooted and ended in a boot loop to Debug Code 13, like it did before with the keyboard controller inserted.

So I remembered the tip to solder a wire to fix the broken trace, found the pin and soldered a jump wire. Tomorrow morning I will find out exactly which pins on ISA and the socket these were.

On the first boot after it did initialize my VGA so I could see the Memory test screen again, sadly I didn’t have my camera ready, I could read 128KB OK, then the screen got blank and only
„8042 GATE-A20 ERROR
SYSTEM HALTED“
was shown, and it beeped endlessly.
The soldered wire did not seem to make any difference, I will continue to check the connections tomorrow.

I can see if the vga card is being initialized correctly on the „IREADY“ LED on my debug card, if it is on the whole time on code „2C“ (should be the code for handover to vga bios), it will work, if it goes dark, it won‘t.

If I reset the board it goes into the bootloop, only a cold start works. Why is that?

A question: I wanted to desolder the keyboard bios socket; how do I do that? If I heaten up one pin to remove it, I can’t move it of course, because all other are cold, if I heaten up the secound one, the first one is cold in seconds xD

Are there any other things I need to do?
Does this mean, the Keyboard Bios is indeed dead and I need a new one? Or could it still be a trace issue? Where could I buy the correct one for the board?

Reply 70 of 78, by Deunan

User metadata
Rank Oldbie
Rank
Oldbie

On earlier mobos (up to PCI 486 or so) the 8042 keyboard controller does several things. It converts the I/O from and to keyboard to serial protocol, that's it main function. But it can also control A20 line gate/mask, and being an MCU it can detect if the reset is a power-on event, or hard reset, or warm reboot via software. And the problems you get all seem related to these functions.

I've never had a keyboard controller go bad on me, though I'm not saying it can't happen. First and foremost though I would suspect the mobo, especially if there was a battery spill. Inspect the reset input to the KBC, make sure it does follow the signal from mainboard (but it can be inverted). All 8 data lines should be connected to the '206 chip (on your mobo it's called MORSE 92A206S) either directly or through that 74F245 nearby KBC. That '245 is the main bi-dir driver for the lower 8 bits of the data word for entire ISA bus. So if ISA is flaky make sure that chip (and it's signals) is not corroded from the spill.

Reply 71 of 78, by Neolyum

User metadata
Rank Newbie
Rank
Newbie

I retried cleaning the area of the battery with vinegar and iso alc and began to check the lanes. On the side facing to the rear I could find a connection for all but the 7th, 8th and 10th pin, on the "upper" side there seemed to be many unconnected parts.. When I powered it up, I saw a signal on most of the pins. With the keyboard controller plugged out, I had a signal on the pin, which had the most corrosion (and on which I soldered a jump wire onto), with the keyboard bios plugged in, there was no signal.
On the pins where the battery leaked the traces are ok, all seemed to be connected.

After the cleaning, the board doesn't like the keyboard bios anymore, sometimes it didn't even show a single debug code 🙁 without it everything seems normal.

How can I track the reset signal? Is there a pinout available for the keyboard bios, so that I can see, which pins are what and can better test the signals?
I don't have an oscilloscope, would this be a good opportunity to buy one? xD

Edit: for the undocumented S1-Jumper next to the chipset- does anyone know what it might be for? Would it be safe to shorten it and look what it does?

Reply 72 of 78, by Deunan

User metadata
Rank Oldbie
Rank
Oldbie

Reset input on the 8042 is usually on pin 4 - an easy test is to connect a meter to that pin, power up, and then press the mobo reset switch (the one that's connected via pin header on the front). You should see the level change and typically the reset is negated so an active signal is L (lower than 0.8V) and inactive is H (higher than 2.0V).

Scope is always good to have if you plan on doing more of these repairs, or just could use one for other projects as well. But just for that mobo? No. A lot can be tested with a cheap digital multimeter.

As for data bus, it's just 8-bit to the KBC, usually pins 12-19. You should be able to find connections between these and chipset (or the F245 nearby).

Reply 73 of 78, by Neolyum

User metadata
Rank Newbie
Rank
Newbie

Reset works. High is 4.9V, Low 0.05V.
When the Board is hanging on 2C, the data pins 1-7 are high, D0 is low. When the board is in boot loop, the voltage on D0 is fluctuating as well, but considerably lower than the other data pins (between 0.2 and 1.5V, 1-7 are on 1.7-3.5V). Is this a bad connection, even though I soldered a jump wire? D0 seems to be connected to pin 11 on the 27C256 BIOS, D1 to 12, D2 to 13, all the others seem to be connected as well.(Edit: D3-D7 are connected to the 206S) I just found out, that the Pin for D0/Pin12 is bent away which seems to prevent a good signal to the BIOS.
And that was the problem! The board booted up, no error, nothing else ^-^ At least one time. I didn't have the keyboard connected, so I was stuck at "Press F1" 🙁
The next boot - hanging at 2C, without Graphics Card going to 31; 1 long beep, 8 short beeps, then going to 40 and hanging. This means missing video adapter (obviously).

Edit2: just some observations: If I apply a little pressure on the upper right part (where the caches are), the board stays on "--" or "06", and does not work at all (I don't like that to be honest). On normal operation (atm hanging in 2C), the first cache chip gets hot, the second warm, 3 and 4 and the MT5C6408-20 below stay cold. Is this expected? Will the board start without the cache? Will it work, if I set the jumpers on 32KB, so not all cache is used? Or will this not change anything?

What is preventing the system from booting now, as it prooved to be working (at least one in 50 times)?

Attachments

Reply 74 of 78, by Deunan

User metadata
Rank Oldbie
Rank
Oldbie

Did you try removing and cleaning/brushing the cache chips' pins? Could be oxidation. The mobo might not boot properly without cache but should not stay at -- but rather post an error code about it. So possibly there is either bad connection there or mobo damage - look for cut traces close to edges and corners, could've been dragged over (or stabbed with) something sharp.

Do try removing the cache chips one by one, it might prevent full boot but note the behaviour, is one chip causing different codes for example? Other options are to swap them places (except the TAG if it's different size, pay attention to chip markings) to see if one is always getting hot, could be faulty. Unless it's burning hot it's not an issue but I'd expect all the chips to have similar temperature.

Reply 75 of 78, by Eep386

User metadata
Rank Member
Rank
Member

Don't forget to turn off the cache in BIOS setup. If you try to boot a board without cache, but forget to turn off the cache in BIOS setup, it'll profoundly not-work.

Life isn't long enough to re-enable every hidden option in every BIOS on every board... 🙁

Reply 76 of 78, by Neolyum

User metadata
Rank Newbie
Rank
Newbie

I removed all cache chips and the jumpers for the settings (thinking, maybe there are sane defaults xD) and it booted 😀 keyboard working, all how expected 😁 on call to the bootloader it said, "cache memory bad, do not enable cache", but it was working anyway.
Does this board work with only 3 of the 4 cache chips?

Edit: I tried to boot from a 2GB hard drive. On the HDD it says 4095 Cylinders,16 Heads and 63 Sectors. The BIOS wants to know wpcom and lzone, are 0 and 4094 correct values?
On accessing the hdd the bootloader dies with an "invalid system. Press any Key..." message in german. Is this data from the hdd or the bios? why does it not work? It *is* possible, that the hdd is bad, in another computer it is not detected (and even prevents booting), but when I press any key, I hear something working in it for a short time.

Another thing; I tried the second set RAM I got, it counts to 16MB (and I am quite sure, that is it only 4), can I presume that the ram is dead?

Edit2: Yet another Problem: the board won't boot (stays at --)if I connect a floppy drive. I am quite sure, it is working though. could the IDE controller be the problem?

Reply 77 of 78, by Woody72

User metadata
Rank Newbie
Rank
Newbie

I'm no expert in motherboard repairs but it was red flag to me when you said the cache chips were wildly different temperatures. I think you're on the right track now.

Modern PC: i7-9700KF, 16GB memory, RTX 3060. Proper PC: Pentium 200 MMX, 128MB EDO memory, GeForce2 MX(200).

Reply 78 of 78, by Deunan

User metadata
Rank Oldbie
Rank
Oldbie
Neolyum wrote on Today, 13:22:

Does this board work with only 3 of the 4 cache chips?

There should be always 4+1 or 8+2 - and that means data+tag chips. Though some mobos have 8+1 config as well, and there is at least one chipset (can't remember which one) that had the tag memory embedded so only data chips were required.

Neolyum wrote on Today, 13:22:

Edit: I tried to boot from a 2GB hard drive. On the HDD it says 4095 Cylinders,16 Heads and 63 Sectors. The BIOS wants to know wpcom and lzone, are 0 and 4094 correct values?

WPCOM stands for write pre-compensation, this is only for MFM HDDs. A value of 65535 will disable it, though on non-MFM drives it's ignored anyway. LZONE is landing zone, usually you'd set it to last cylinder (or whatever the HDD sticker says) but it has no meaning on CF cards and modern drives that auto-park heads on power-off. The values you gave are OK.

Neolyum wrote on Today, 13:22:

On accessing the hdd the bootloader dies with an "invalid system. Press any Key..." message in german. Is this data from the hdd or the bios? why does it not work? It *is* possible, that the hdd is bad, in another computer it is not detected (and even prevents booting), but when I press any key, I hear something working in it for a short time.

Boot from a system floppy and and do FDISK /MBR - for some reason DOS installer will not do this and there is no proper boot code in the first sector (or there is something left over from whatever was on that CF/HDD before).

Neolyum wrote on Today, 13:22:

Another thing; I tried the second set RAM I got, it counts to 16MB (and I am quite sure, that is it only 4), can I presume that the ram is dead?

Could be, but make a photo of all the sticks and write down the chip markings - it's better to be sure before you throw away what might be a 16MB set.

Neolyum wrote on Today, 13:22:

Edit2: Yet another Problem: the board won't boot (stays at --)if I connect a floppy drive. I am quite sure, it is working though. could the IDE controller be the problem?

Just the drive or do you mean the controller card? Did you make sure the cables are attached the right way around? Some floppy drives (Sony for example) have the connector keyed wrong. That might upset the controller card and in turn prevent the ISA bus from working.