jnemo2004 wrote on 2024-03-29, 12:42:

I received the MS-5119 board that someone wanted to remove the KBC, AMIKEY-2 that was soldered on the board with a screwdriver.

Take a closer look at Q2, Q3 and Q4. It looks like Q2 was moved (by force or when heat was applied to desolder something nearby) and Q3 / Q4 are missing. See if you can spot any sign of part removal on the pads (the photo is a bit blurry to tell). You might need to replace these transistors as well. Q2 might just work as-is but I would attempt to move it back to its proper place, to make sure it doesn't short anything and the pads have good contact with the pins.

Hello,
You are right, the photo was taken before to soldered the transistors in its pads (the three are 1AM). They fell down when I applied heat to remove the KBC chip. I also have soldered C73. I have used one from other board that fit in that place.
Please, do not pay attention to the yellow wire, it was used to follow the track in the rear side of the board.
Thank you very much for your advices.
Regards.
Remark: Sorry for the quality of the photos, I have not a better camera.

Deunan wrote on 2024-03-29, 15:37:

jnemo2004 wrote on 2024-03-29, 12:42:

I received the MS-5119 board that someone wanted to remove the KBC, AMIKEY-2 that was soldered on the board with a screwdriver.

Take a closer look at Q2, Q3 and Q4. It looks like Q2 was moved (by force or when heat was applied to desolder something nearby) and Q3 / Q4 are missing. See if you can spot any sign of part removal on the pads (the photo is a bit blurry to tell). You might need to replace these transistors as well. Q2 might just work as-is but I would attempt to move it back to its proper place, to make sure it doesn't short anything and the pads have good contact with the pins.

jnemo2004 wrote on 2024-03-29, 16:33:

Remark: Sorry for the quality of the photos, I have not a better camera.

Ah, no problem, somehow I didn't see the nice drawing you made. So with that I would say that R83 is 4k7 (472) since it's a pull-down of some sort. That's a safe value that should work I think.
R82 should be much smaller in value, about 33 ohms to maybe 100 ohm. It's some sort of line terminating resistor for the clock signal, to reduce the ringing on transitions. Anything in between is acceptable and without a working mobo it's hard to say what value to use. Start with 33 or 47/51 ohms and see if it works and the system is stable I guess.

In the same way the RN14 is there to reduce noise on the IRQ lines, but the value can be higher than the clock line since the exact nanoseconds don't matter here, just that the signal is eventually recognized. I would say 50 ohms to 330 or 470 ohms. Maybe even 1k would work if you have nothing else to put in there. In fact as long as you are not using any ISA cards that need IRQ (VGA doesn't) then the mobo would work even without RN14. But it would be difficult to solder it after you add a socket for the new KBC, and KBC is needed.

EDIT: BTW the broken F245 must be replaced. It's probably driving the lower 8 data bits of the ISA bus and without that no card will work properly. If you can't find F245 for replacement then ALS or ACT series should work too.

Thank you very much. A great help!!!
The IC 74F25PC has been already replaced by other with the same reference.
I will use a socket for the KBC but I will remove the plastic that joins both sides of the pins and that covers the damaged part of the board to change components if needed.
Now I will try to find the components I need as you have suggested.
Meanwhile I return to the sketch I am making of plate 286.
Thank you and regards.
Remark: I don't have a better camera for motherboard repair since I don't have room, I wouldn't know where to put the microscope.

Deunan wrote on 2024-03-29, 23:32:

Ah, no problem, somehow I didn't see the nice drawing you made. So with that I would say that R83 is 4k7 (472) since it's a pull […]
Show full quote

jnemo2004 wrote on 2024-03-29, 16:33:

Remark: Sorry for the quality of the photos, I have not a better camera.

Ah, no problem, somehow I didn't see the nice drawing you made. So with that I would say that R83 is 4k7 (472) since it's a pull-down of some sort. That's a safe value that should work I think.
R82 should be much smaller in value, about 33 ohms to maybe 100 ohm. It's some sort of line terminating resistor for the clock signal, to reduce the ringing on transitions. Anything in between is acceptable and without a working mobo it's hard to say what value to use. Start with 33 or 47/51 ohms and see if it works and the system is stable I guess.

In the same way the RN14 is there to reduce noise on the IRQ lines, but the value can be higher than the clock line since the exact nanoseconds don't matter here, just that the signal is eventually recognized. I would say 50 ohms to 330 or 470 ohms. Maybe even 1k would work if you have nothing else to put in there. In fact as long as you are not using any ISA cards that need IRQ (VGA doesn't) then the mobo would work even without RN14. But it would be difficult to solder it after you add a socket for the new KBC, and KBC is needed.

EDIT: BTW the broken F245 must be replaced. It's probably driving the lower 8 data bits of the ISA bus and without that no card will work properly. If you can't find F245 for replacement then ALS or ACT series should work too.

jnemo2004 wrote on 2024-03-30, 07:16:

Now I will try to find the components I need as you have suggested.

I looked at some datasheets. Are you sure the chip nearby is Winbond W83787F? It's important because there are some other chips in that family and these have different I/O pin functions.

* If this is indeed W83787F (or 777F) then very likely the RN14 is meant to be unpopulated.
* If it's W83787IF (letter I before F) then the pins connected to RN14 can be used as general purpose I/O (GPIO) but that still doesn't make much sense to connect them to ISA IRQ lines. So I would still leave RN14 unpopulated.
* If the chip is W83877F then it does have some extra functionality, probably to use alternate IRQ assigment for the UART/parallel port. However if you are not going to use IRQ 2, 5, 10 or 11 for serial/parallel ports then RN14 can still be omitted.

So don't worry about RN14 too much, it's only really required for 877F variant, and even then only for alternate IRQ assignments. In all other cases don't install it. For the value I'll revise it a bit, since these are outputs to ISA bus, not inputs as I thought - so also 33 to 100 ohms, just like the clock. Reading the datasheet I'm also pretty confident that 4k7 value for R83 is a safe choice.

Thank you very much for your great support.
Really, it cost me a lot to identify the pin 1 position. It was not 100% clear for me also with the data sheet.
I was looking for the data sheet of the W83787F but google send to the W83787IF data sheet.
For that reason I thought that pinout will be the same.
It was my fault, sorry.
Regards.

Deunan wrote on 2024-03-30, 12:59:

I looked at some datasheets. Are you sure the chip nearby is Winbond W83787F? It's important because there are some other chips […]
Show full quote

jnemo2004 wrote on 2024-03-30, 07:16:

Now I will try to find the components I need as you have suggested.

I looked at some datasheets. Are you sure the chip nearby is Winbond W83787F? It's important because there are some other chips in that family and these have different I/O pin functions.

* If this is indeed W83787F (or 777F) then very likely the RN14 is meant to be unpopulated.
* If it's W83787IF (letter I before F) then the pins connected to RN14 can be used as general purpose I/O (GPIO) but that still doesn't make much sense to connect them to ISA IRQ lines. So I would still leave RN14 unpopulated.
* If the chip is W83877F then it does have some extra functionality, probably to use alternate IRQ assigment for the UART/parallel port. However if you are not going to use IRQ 2, 5, 10 or 11 for serial/parallel ports then RN14 can still be omitted.

So don't worry about RN14 too much, it's only really required for 877F variant, and even then only for alternate IRQ assignments. In all other cases don't install it. For the value I'll revise it a bit, since these are outputs to ISA bus, not inputs as I thought - so also 33 to 100 ohms, just like the clock. Reading the datasheet I'm also pretty confident that 4k7 value for R83 is a safe choice.

Therefore, I understand that I can leave the RN14 empty, w/o the resistors array.
I have to look only for R83 and R82.
Many thanks.

In the first picture it looks like resistors were soldered in R82 and R83 (sorry again for the poor quality).
In the second pictures, the no damaged pads for both looks OK but maybe it was done by me after revised the damages. I am not sure..

I bought this board very cheap because I knew that they had removed the BIOS.
When I received it I realized that not only had they removed the BIOS but also all the jumpers.
I found a BIOS dump and configured the board with jumpers for a processor.
The board didn't work but everything seemed fine until I looked under the KBC and realized that the person who had removed the BIOS thought he could remove all the components that way (with a screwdriver).

Since it's a 787F I would leave RN14 empty. Therefore there is also no need to fix any traces that connect to it, just make sure what is left on the mobo is not shorted somewhere.

R83 (along with R84) configure the 787F chip at power-on, so R83 should be present and pull the pin to GND. This chip can provide IDE interface but if it's unused (and IDE is done through some other chip like Intel south bridge) then there is no need to fix the trace except for this resistor.
R82 is most likely the line terminator for the ISA clock signal, without it ISA will not work (I assume here the Intel chip is the source of that clock). So this trace must be properly repaired. And there is a broken trace between R82 and R83 that looks like some power supply, that too needs to be fixed. But all in all this might be less work than I originaly thought.

BTW if you have a POST card you can try powering the mobo without KBC, the BIOS will probably complain and stop with some code - but since it's a newer mobo with a proper chipset it might even show something on screen with a video card installed. Nothing bad will happen either way (if there are no shorts) so you can try that before installing the KBC. Unless this mobo still uses KBC to process power good signal from the PSU it should not be totally dead without it.

Thank you very much.
There is not shorts and it is a good idea to test the card w/o the KBC
I will look for the removed components R82 and R83
Regards

Deunan wrote on 2024-03-30, 22:26:

Since it's a 787F I would leave RN14 empty. Therefore there is also no need to fix any traces that connect to it, just make sure […]
Show full quote

Since it's a 787F I would leave RN14 empty. Therefore there is also no need to fix any traces that connect to it, just make sure what is left on the mobo is not shorted somewhere.

R83 (along with R84) configure the 787F chip at power-on, so R83 should be present and pull the pin to GND. This chip can provide IDE interface but if it's unused (and IDE is done through some other chip like Intel south bridge) then there is no need to fix the trace except for this resistor.
R82 is most likely the line terminator for the ISA clock signal, without it ISA will not work (I assume here the Intel chip is the source of that clock). So this trace must be properly repaired. And there is a broken trace between R82 and R83 that looks like some power supply, that too needs to be fixed. But all in all this might be less work than I originaly thought.

BTW if you have a POST card you can try powering the mobo without KBC, the BIOS will probably complain and stop with some code - but since it's a newer mobo with a proper chipset it might even show something on screen with a video card installed. Nothing bad will happen either way (if there are no shorts) so you can try that before installing the KBC. Unless this mobo still uses KBC to process power good signal from the PSU it should not be totally dead without it.

Hello, here you have the quick test w/o memory, processor, graphic card, keyboard and KBC.
Regards

I have checked in cpu socket, Vcc2 and Vcc3 = 3.45-3.46v.. Freq. = 49.5 Mhz.

Well, nothing is smoking and the reset LED acts as it should so this mobo doesn't seem to be using KBC for power good. So try adding CPU, RAM and some video card and see what happens. Keep the POST card in the mobo though because, depending on what BIOS code does, it might still hang without any video output with no KBC installed. But you should see some codes on the POST card at the very least.

OK, I will do it.
Regards.

Deunan wrote on 2024-04-01, 10:25:

Well, nothing is smoking and the reset LED acts as it should so this mobo doesn't seem to be using KBC for power good. So try adding CPU, RAM and some video card and see what happens. Keep the POST card in the mobo though because, depending on what BIOS code does, it might still hang without any video output with no KBC installed. But you should see some codes on the POST card at the very least.

Hello, good afternoon. I have done the test and apparently with the same results. I have carried out the test with an SY007 processor (Pentium I 100 Mhz), with 4+4 MB SIMMs, and I have tested with an ISA and also a PCI graphics card.
On the other hand, the motherboard came without a BIOS and I have tried one downloaded from The Retro Web but I don't remember which one I downloaded. I still have to do a more systematic test. This board can carry both AMI and AWARD BIOS. Maybe it's also time to put on the KBC. Greetings

jnemo2004 wrote on 2024-04-02, 11:55:

Hello, good afternoon. I have done the test and apparently with the same results.

So no codes at all? In my experience some BIOS images found on the net are broken so try another one. AMI/AWARD shouldn't really matter as long as it works. The things to check an try:
- make sure the jumpers are in the right place and not missing. For basic operation at least the CPU clock and multiplier must be set (clock might be slower than required but multiplier is best set as required)
- I see this mobo has connector for VRM module, so for non-MMX Pentium make sure it has jumpers in the right place
- cache jumpers and all chips must be present - if there is any doubt try removing all the cache chips and testing without

This board had all the chips that were in a socket removed, including BIOS and cache. Breaking the KBC and the 74F25PC IC trying to remove them as if they were in sockets.
I'm going to check the board in more detail in case there are any more damaged components.
I set all the jumpers according to the manual for the Pentium I 100 Mhz processor.
I'm also going to put the KBC in a socket so that the board is complete.
We'll see what happens.
Thank you very much for your help.

I have installed all the components and repaired everything.
The motherboard still does not give any post code.
There must be some other problem but I don't know where to look for it, apparently everything is fine.
For now I'm going to leave it without doing anything in case I get any ideas in the future.
I'm going back to work on the sketch for plate 286, which I left standing until I finished this one.
Thank you very much and greetings.
Note: What I would like is to be able to verify that the Dallas is working.