VOGONS


FOREX 386 Cache board doesn't post

Topic actions

First post, by MMaximus

User metadata
Rank Oldbie
Rank
Oldbie

I got this board "naked' in a lot a while ago - no BIOS chip, no CPU, no RAM, no oscillator, no Cache.

It looks really good optically, maybe never even used, so I thought it would be a matter of filling in the blanks to make it POST...

LiIPuvol.jpg?2

What I've done so far:

Tried with two different CPUs (i386DX25, AMD386DX40), both working in other boards.
Tried with two different oscillators (50Mhz and 80Mhz), both working in other boards.
Tried with 4x 1Mx9 SIMMS in BANK 0, working in other boards.
Tried with BIOS image from matching board from the Vogons 386 Bios image collection thread, burned onto a W27C512-45Z EEPROM.

No luck so far - I'm wondering if it refuses to post because the cache sockets are not populated? I can't find a jumper to disable cache, only some to change the cache size (from 64k to 256k). I recently "restored" a HOT-307 386 motherboard, and it POSTs without problem even when the cache chips are not present... so I find it weird this would be a requirement.

Another possibility would be the BIOS image being incorrect? Or some physical damage to the motherboard that I can't see.

The diagnostic card doesn't show any error code: only "----".

Hard Disk Sounds

Reply 2 of 44, by Tiido

User metadata
Rank l33t
Rank
l33t

I have a board that looks the same. BIOS is found here : Re: 80386 BIOS image collection
EDIT: seems you already got to this point 🤣.
EDIT2 : I remember my board giving trouble with RAM, I remember needing RAM in bank1 instead for things to boot if not all slots are used. And if mixed size is used the larger sticks need to be in bank1 too rather than bank0 or it won't see all the RAM.

T-04YBSC, a new YMF71x based sound card & Official VOGONS thread about it
Newly made 4MB 60ns 30pin SIMMs ~
mida sa loed ? nagunii aru ei saa 😜

Reply 3 of 44, by Predator99

User metadata
Rank l33t
Rank
l33t

Put a Supersoft ROM on the W27C512 and connect a speaker 😉

Reply 4 of 44, by MMaximus

User metadata
Rank Oldbie
Rank
Oldbie
Tronix wrote:

Several BIOS for FOREX chipset here: http://chukaev.ru54.com/bios_en.htm#386 ....

Thanks for this link. Bookmarking it for the future.

Tiido wrote:

I have a board that looks the same. BIOS is found here : Re: 80386 BIOS image collection
EDIT: seems you already got to this point 🤣.
EDIT2 : I remember my board giving trouble with RAM, I remember needing RAM in bank1 instead for things to boot if not all slots are used. And if mixed size is used the larger sticks need to be in bank1 too rather than bank0 or it won't see all the RAM.

Yep, that's the same BIOS I've used. Interesting observations about the RAM though - I guess I need to try Bank 1 tomorrow...

Predator99 wrote:

Put a Supersoft ROM on the W27C512 and connect a speaker 😉

Thanks. I remember you telling me about the supersoft ROMs a while ago but I somehow forgot about this.

I've just tried this but I think my image isn't right because nothing happens. Also I put the W27C512 containing the supersoft ROM I've just burned in a working 386DX board and nothing happens either. Basically IIRC what I did was:

- Download the supersoft 5170 roms from minuszerodegrees
- use Winhex - Tools - Unify - Bytewise
- select even file first, then odd file
- unify them into a 64kb file (i checked the code at the end of the image for the words to mean something instead of being gibberish)
- burn the image into a W27C512 using a TL866ii

Is this the correct procedure?

EDIT: I've just read that VGA cards don't work with the supersoft ROM. I guess I need to try with MDA... but that doesn't explain why I didn't hear any beeping from the speaker...

Hard Disk Sounds

Reply 5 of 44, by pshipkov

User metadata
Rank l33t
Rank
l33t

I have the same mobo, not working as well.
One of the caps next to the power socket was popped. Replaced it, but still no go.
Going now through lines and components.
Replaced 3 more caps already. Readings were inconsistent, despite the components being of the same type.

I don't know if we can have the time/will to run a cross-check on component readings - but that can probably help both of us identify problems with them.

retro bits and bytes

Reply 6 of 44, by ph4nt0m

User metadata
Rank Member
Rank
Member

Does your POST card show the ISA bus getting clock signal? Cache initialisation happens well after KBC, memory, video.

My Active Sales on CPU-World

Reply 7 of 44, by MMaximus

User metadata
Rank Oldbie
Rank
Oldbie
pshipkov wrote:
I have the same mobo, not working as well. One of the caps next to the power socket was popped. Replaced it, but still no go. Go […]
Show full quote

I have the same mobo, not working as well.
One of the caps next to the power socket was popped. Replaced it, but still no go.
Going now through lines and components.
Replaced 3 more caps already. Readings were inconsistent, despite the components being of the same type.

I don't know if we can have the time/will to run a cross-check on component readings - but that can probably help both of us identify problems with them.

I might if I knew how to... but I'm afraid I know very little about electronics unfortunately. I have a multimeter but I only went as far as checking the voltage of an AA battery with it once 🤣

ph4nt0m wrote:

Does your POST card show the ISA bus getting clock signal? Cache initialisation happens well after KBC, memory, video.

The "CLK" led on the POST card lights up, so I suppose it means ISA bus is getting clock? If I remove the crystal oscillator from its socket, the CLK led doesn't light up on the POST card.

Apart from that I confirm that my supersoft ROM is working - I put it into a working 386 board (OCTEK Jaguar V) with an MDA card, and the diagnostics got displayed on screen. However, it still won't display anything when used with the FOREX board.

Hard Disk Sounds

Reply 8 of 44, by Deunan

User metadata
Rank Oldbie
Rank
Oldbie

You should get a 01 code for the CPU test at least I think. This code runs from BIOS ROM and needs no RAM (therefore no cache either). While it's not impossible that the missing chips mess something up I'd say you have a reset signal issue. I assume you've properly mirrored the BIOS image for the '512 EEPROM?
Do you have RESET LED on your test card? Does it change state when you press reset button?

Reply 9 of 44, by MMaximus

User metadata
Rank Oldbie
Rank
Oldbie
Deunan wrote:

You should get a 01 code for the CPU test at least I think. This code runs from BIOS ROM and needs no RAM (therefore no cache either). While it's not impossible that the missing chips mess something up I'd say you have a reset signal issue. I assume you've properly mirrored the BIOS image for the '512 EEPROM?
Do you have RESET LED on your test card? Does it change state when you press reset button?

I have no way to tell if my BIOS image is correct on the EEPROM, however the programming procedure went ok with the TL866 and I didn't even have to unify it because the file from the vogons thread is already 64kb, so it was quite straightforward. Besides, I used the same Winbond 512 chip to program the supersoft rom, and as I was mentioning above this supersoft rom works in another board - so I guess that rules out a defective EEPROM chip.

The RST led on the test card lights up when I short the reset jumper on the board, and the CLK led stops lighting up. Then the LEDs revert back to their previous state as soon as I stop shorting the jumper.

Hard Disk Sounds

Reply 10 of 44, by Deunan

User metadata
Rank Oldbie
Rank
Oldbie

Well then, perhaps this mobo does require cache to be populated after all.
But there is one more test you can run before that - measure the voltages on all ISA slot pins. This is chipset-dependent but usually you'd see the address lines toggling on BIOS reads as the way it's typically connected is ISA-like (it's a slow 8-bit "memory device"). Let's see if we can find any indication of that happening.
It would be better to use oscilloscope, but I assume you don't have one. And so there isn't much more you can do with this mobo.

Reply 11 of 44, by MMaximus

User metadata
Rank Oldbie
Rank
Oldbie

I have a multimeter so I guess I could try this, however I have no idea how to do it - any chance you could explain the procedure in very basic steps?

Hard Disk Sounds

Reply 12 of 44, by Deunan

User metadata
Rank Oldbie
Rank
Oldbie

It's simple, use this pinout of the ISA slot: https://upload.wikimedia.org/wikipedia/common … SA_Bus_pins.svg
Measure voltages on all XT address pins. And possibly the rest as well since you have no other leads to follow. You can also do the 16-bit slot extension if you wish but that is optional and might be misleading - the address lines on that part work a bit differently.

Reply 13 of 44, by Mut

User metadata
Rank Newbie
Rank
Newbie

Check the temperature of all ICs with your finger, if you find something extremely hot its shorted. In the past I had a similar dead 386 motherboard, the problem was a shorted keyboard controller.

Reply 14 of 44, by MMaximus

User metadata
Rank Oldbie
Rank
Oldbie
Deunan wrote:

It's simple, use this pinout of the ISA slot: https://upload.wikimedia.org/wikipedia/common … SA_Bus_pins.svg
Measure voltages on all XT address pins. And possibly the rest as well since you have no other leads to follow. You can also do the 16-bit slot extension if you wish but that is optional and might be misleading - the address lines on that part work a bit differently.

Thanks for the explanation. When you say "It's simple", I think you vastly overestimate my ability in this field 🤣 I don't really know what I'm doing since this is all new stuff to me - nevertheless I looked at the suntac thread specifically at this post and managed to get a few readings. But first of all can we check I'm doing it right? I hold the black needle of the multimeter over pin B 01 (GND) of an ISA slot, then I put the red needle over A 31 (Addr 00), check the reading, and work my way up to A12 (Addr 19). Is this what I'm supposed to be doing? What voltages should the multimeter read?

Also, what is this "short, fast beep" you're talking about regarding the multimeter? I have a "Fluke 101" but it doesn't make any beep when I read the voltages. There is a yellow button that beeps when I press it... I've looked at the manual but I don't really understand it at this stage.

Anyway here's what I got so far regarding readings: all the pins from the XT address lines give more or less 5v, but 3 pins give 0.12v - A29, A23 and A22, if I'm not mistaken. I measured the first ISA slot, the closest to the power connector.

Mut wrote:

Check the temperature of all ICs with your finger, if you find something extremely hot its shorted. In the past I had a similar dead 386 motherboard, the problem was a shorted keyboard controller.

I have let the board run for a good 30min but nothing seems to get too hot. Even the CPU just gets lukewarm - nothing is too hot to touch.

Hard Disk Sounds

Reply 15 of 44, by Deunan

User metadata
Rank Oldbie
Rank
Oldbie
MMaximus wrote:

I hold the black needle of the multimeter over pin B 01 (GND) of an ISA slot, then I put the red needle over A 31 (Addr 00), check the reading, and work my way up to A12 (Addr 19). Is this what I'm supposed to be doing? What voltages should the multimeter read?

Correct. You don't have to use B01 for ground, there might be other more suitable spots for the black probe - you could even just stick it into one of the Molex connectors (pick one of the 2 black wires). That gives you one hand to hold the red probe, one hand to note down results.
Set the meter to DC volts, either 20V or auto-range (depends on your meter). Measuring ground should give you 0V, and it's enough to get the value with one digit after decimal point accuracy. The rest is noise we don't care about - well not yet anyway.

MMaximus wrote:

Also, what is this "short, fast beep" you're talking about regarding the multimeter? I have a "Fluke 101" but it doesn't make any beep when I read the voltages.

That's continuity test. Most modern meters have a special range in resistance mode that, when the resistance measured is less than certain value (again, depends on the meter but some 100 ohm or less), will emit a beep tone. The problem is some meters, even the more expensive ones, take a moment to actually do the measurement - so the beep is delayed. That's not good for tests where you just sweep the probe tip over some pins until you hear the tone. If you sweep too fast for the meter to detect the low resistance, you miss the spot.

MMaximus wrote:

Anyway here's what I got so far regarding readings: all the pins from the XT address lines give more or less 5v, but 3 pins give 0.12v - A29, A23 and A22, if I'm not mistaken. I measured the first ISA slot, the closest to the power connector.

So that would be address bits 2, 8 and 9. Curious. So the address is 0xffcfb. Due to how x86 power-on reset works, the very first address executed is 0xffff0 (on 32-bit CPU all the higher address lines are 1 as well, mobo chipset usually only tests A31 or A23 for BIOS ROM redirection). The very first instruction is a far jump to somewhere around 0xfe000. Assuming what we see is not just some glitch, I think that jump fails - possibly due to being read wrong from the ROM chip. CPU reads wrong stuff, hangs.

Perhaps a data line is broken near the BIOS socket? But then again, this would also fit nicely with the missing cache causing wrong reads. I think you'll have to find some SRAM chips to populate the cache after all.

Reply 16 of 44, by MMaximus

User metadata
Rank Oldbie
Rank
Oldbie

Thanks for all this useful information - I really appreciate it!

Somehow this board takes 10 SRAM chips so I'm one chip short at the moment - I'm waiting for a spare module I've just bought on Ebay to arrive and I'll have enough chips to fill up all the slots and test again...

Hard Disk Sounds

Reply 18 of 44, by MMaximus

User metadata
Rank Oldbie
Rank
Oldbie

Quick update - I fitted all the cache chips to no avail. The board still doesn't post, no error codes on the POST card, the supersoft ROM doesn't display anything on MDA, no beep codes whatsoever. The board is going back in the pile of non-working boards, but if someone wants to have a go at it, I'll gladly send it away 😀

BvhrBSxl.jpg?1

Hard Disk Sounds

Reply 19 of 44, by PC@LIVE

User metadata
Rank Oldbie
Rank
Oldbie

A question, from the photo I see that you have started the MB without the battery, maybe maybe connecting a part of it.

AMD 286-16 287-10 4MB HD 45MB VGA 256KB
AMD 386DX-40 Intel 387 8MB HD 81MB VGA 256KB
Cyrix 486DLC-40 IIT387-40 8MB VGA 512KB
AMD 5X86-133 16MB VGA VLB CL5428 2MB and many others
AMD K62+ 550 SOYO 5EMA+ and many others
AST Pentium Pro 200 MHz L2 256KB