jnemo2004 wrote on 2024-02-15, 13:05:

Under RTC , motherboard looks also OK.

Yes, that was just a shadow on the previous photo.

jnemo2004 wrote on 2024-02-15, 13:05:

My following steps in this next weekend is to remove all ICs in sockets , check them externally and put in all socket contact cleaner .

That's a good idea. This type of socket can develop poor contact sometimes. You can try standing the mobo upright and testing with ohm meter the RTC chip pin on one side and the soldering below the socket on the other side. Each pin should have good connection (< 1 ohm).
I can see the mobo markings saying MC146818A, so in fact they suggest -A version, as expected for chipset integrated into a couple of large chips. I wonder if your HM6818P is a replacement already? Maybe it's broken after all. If you don't have another mobo to swap the chip try looking for a cheap offer on ebay or something. There isn't any functional difference between 6818 and 6818A, it's just that A version has improved power draw and is easier on the battery.

jnemo2004 wrote on 2024-02-15, 13:05:

Also I will do a test with the two BIOS dumps available in this motherboard:
In BIOS Even is equivalent to LO? and Odd to HI?

Yes, that how it usually is. I've seen some cases where LO and HI were reversed for some reason but even if you put the chips in the wrong order nothing will break. The mobo just won't POST with incorrect BIOS data.

Thank you very much. I will do all your suggestions.
I do not know if IC HM6818P is the original or it is a replacement.
Regards.

Deunan wrote on 2024-02-15, 13:57:

Yes, that was just a shadow on the previous photo. […]
Show full quote

jnemo2004 wrote on 2024-02-15, 13:05:

Under RTC , motherboard looks also OK.

Yes, that was just a shadow on the previous photo.

jnemo2004 wrote on 2024-02-15, 13:05:

My following steps in this next weekend is to remove all ICs in sockets , check them externally and put in all socket contact cleaner .

That's a good idea. This type of socket can develop poor contact sometimes. You can try standing the mobo upright and testing with ohm meter the RTC chip pin on one side and the soldering below the socket on the other side. Each pin should have good connection (< 1 ohm).
I can see the mobo markings saying MC146818A, so in fact they suggest -A version, as expected for chipset integrated into a couple of large chips. I wonder if your HM6818P is a replacement already? Maybe it's broken after all. If you don't have another mobo to swap the chip try looking for a cheap offer on ebay or something. There isn't any functional difference between 6818 and 6818A, it's just that A version has improved power draw and is easier on the battery.

jnemo2004 wrote on 2024-02-15, 13:05:

Also I will do a test with the two BIOS dumps available in this motherboard:
In BIOS Even is equivalent to LO? and Odd to HI?

Yes, that how it usually is. I've seen some cases where LO and HI were reversed for some reason but even if you put the chips in the wrong order nothing will break. The mobo just won't POST with incorrect BIOS data.

Good morning, here in Zaragoza (Spain).
I did the following test:
I put the KBC and the RTC in their sockets with contact cleaner in the sockets.
I removed all the memory chips on board
I put an internal battery Ni-Cd 3,6v and a buzzer
I switch on the motherboard.
Motherboard stopped in code 10 and with three short beeps in the buzzer. No image in the screen.
............................
I will do the following:
I have not SIPP memory modules. I will prepare a SIMM socket to be inserted in the SIPP socket.
I will (try) to do a test with SIMM modules instead of memory ICs on board.

Next pictures and video

I did the test.
With two modules of 256Kb , code count stops in 0A
With two modules of 1Mb with parity, the same results with memory on board. code count stops in 20, buzzer continously beeps with short beepers.
Next step, to check with other BIOS.

Memory makes tons more sense than some vague RTC error. 20 should show video on screen. have you tried with another graphic card?

Yes, I got the same message in the screen with code 20: CMOS inoperational. System halted.

rasz_pl wrote on 2024-02-18, 13:35:

Memory makes tons more sense than some vague RTC error. 20 should show video on screen. have you tried with another graphic card?

With the two BIOS images available in The Retro Web, the result is the same. Stop in code 20, message CMOS inoperational. System halted and buzzer beeps conttinously with short beeps. At this point, I do not know what to do. OK. Thanks to all.

Looking at the pics of U7, the 83C201P, it might be a ripped via near pin4, going down into the board. It might be worth checking continuity to where it leads.

Did you try a completely different BIOS? MR-Bios would be a good idea?

Re: How about a MR-BIOS ROM file repository?

Will only be a workaround, but maybe you are lucky and it starts...

Rwolf wrote on 2024-02-18, 17:56:

Looking at the pics of U7, the 83C201P, it might be a ripped via near pin4, going down into the board. It might be worth checking continuity to where it leads.

this via does indeed look weirdly black, charred even

I will do the suggestions but for you what means exactly "CMOS inoperational". Which IC? nearly all ICs are CMOS (I believe) maybe only a few are TTL (I believe also).

In context of BIOS CMOS means battery backed SRAM, that means RTC chip. Wow, thats a lot of abbreviations, at least they arent all TLAs :]
The problem with that statement is BIOS checks SRAM way earlier, and if "SRAM was inoperational" boot would stop way earlier. You could try decompiling bios to see what exactly is happening, another option is loading this bios in emulator(for example 86box) and simulating defects until you find the culprit. Trying MR-Bios is a good idea, might give more descriptive error message.

jnemo2004 wrote on 2024-02-19, 09:13:

I will do the suggestions but for you what means exactly "CMOS inoperational". Which IC? nearly all ICs are CMOS (I believe) maybe only a few are TTL (I believe also).

This error is related to NVRAM in the RTC chip which is used to store configuration data (floppies, HDDs, memory settings, etc.) and is only shown on early mobos (286, some 386), later ones will just beep and hang with black screen. The early mobos would often assume CGA or MDA is connected (depending where some video RAM is found), initialize it and display the error.

I had a look and one possible cause of this problem would be a permanent loss of RTC power, which BIOS detects and gives up. Please make sure pin #22 on the RTC chip (PS - power sense) has battery voltage on it, or at least that it is powered from PSU when the mobo is running. There must be at least 3V there I think.

EDIT: As to why it's calles "CMOS", it's because the early chips (including CPUs) would be older tech, like NMOS or even PMOS, this allowed more dense and faster transistors but at the cost of higher power draw. Obviously RTC cannot have high power draw or it will drain the battery quickly, so these were made in CMOS technology. Often that would be the only true CMOS chip on the early mobo, so it was called that instead of NVRAM or RTC.

Thank you for the info. I will try with Mr. BIOS option. It looks a good idea.
I will start with the suggested by "Predator99"

V000B200 STD_286 Generic (non-Chipset) 286 / 386 /486 Generic 28 6
in a standard 12 MHz 80286 with VLSI Chipset:

Or should I use another one?

rasz_pl wrote on 2024-02-19, 13:00:

In context of BIOS CMOS means battery backed SRAM, that means RTC chip. Wow, thats a lot of abbreviations, at least they arent all TLAs :]
The problem with that statement is BIOS checks SRAM way earlier, and if "SRAM was inoperational" boot would stop way earlier. You could try decompiling bios to see what exactly is happening, another option is loading this bios in emulator(for example 86box) and simulating defects until you find the culprit. Trying MR-Bios is a good idea, might give more descriptive error message.

Thank you very much for your info. I have already checked the voltage in pin 24 and it was OK (4,78 mobo ON and 3,47 mobo OFF). I will check the voltage in pin 22 (PS). Regards.

Deunan wrote on 2024-02-19, 14:29:

This error is related to NVRAM in the RTC chip which is used to store configuration data (floppies, HDDs, memory settings, etc.) […]
Show full quote

jnemo2004 wrote on 2024-02-19, 09:13:

I will do the suggestions but for you what means exactly "CMOS inoperational". Which IC? nearly all ICs are CMOS (I believe) maybe only a few are TTL (I believe also).

This error is related to NVRAM in the RTC chip which is used to store configuration data (floppies, HDDs, memory settings, etc.) and is only shown on early mobos (286, some 386), later ones will just beep and hang with black screen. The early mobos would often assume CGA or MDA is connected (depending where some video RAM is found), initialize it and display the error.

I had a look and one possible cause of this problem would be a permanent loss of RTC power, which BIOS detects and gives up. Please make sure pin #22 on the RTC chip (PS - power sense) has battery voltage on it, or at least that it is powered from PSU when the mobo is running. There must be at least 3V there I think.

EDIT: As to why it's calles "CMOS", it's because the early chips (including CPUs) would be older tech, like NMOS or even PMOS, this allowed more dense and faster transistors but at the cost of higher power draw. Obviously RTC cannot have high power draw or it will drain the battery quickly, so these were made in CMOS technology. Often that would be the only true CMOS chip on the early mobo, so it was called that instead of NVRAM or RTC.

jnemo2004 wrote on 2024-02-19, 14:37:

Thank you very much for your info. I have already checked the voltage in pin 24 and it was OK (4,78 mobo ON and 3,47 mobo OFF). I will check the voltage in pin 22 (PS). Regards.

If you want something to try, I have 2 experiments in mind:

1) Please remove the RTC chip and see if the mobo powers up with the same error. Or is it different now.
2) With the RTC removed please attach 4k7 resistor between pins 11 and 24. You could try sticking the resistor leads into the socket but it might damage it, you can also do a temporary solder to the socket pins on the backside of the mobo. Power the mobo with the resistor (but without RTC chip) and see if anything changes.

If you don't have 4k7 resistor then 10k down to maybe 1k should also work. But do not use resistor smaller than 1k for this.

Voltage in pin 22:
Motherboard OFF : 3.47v.
Motherboard ON: 4.62-4.67v.

Deunan wrote on 2024-02-19, 14:29:

This error is related to NVRAM in the RTC chip which is used to store configuration data (floppies, HDDs, memory settings, etc.) […]
Show full quote

jnemo2004 wrote on 2024-02-19, 09:13:

I will do the suggestions but for you what means exactly "CMOS inoperational". Which IC? nearly all ICs are CMOS (I believe) maybe only a few are TTL (I believe also).

This error is related to NVRAM in the RTC chip which is used to store configuration data (floppies, HDDs, memory settings, etc.) and is only shown on early mobos (286, some 386), later ones will just beep and hang with black screen. The early mobos would often assume CGA or MDA is connected (depending where some video RAM is found), initialize it and display the error.

I had a look and one possible cause of this problem would be a permanent loss of RTC power, which BIOS detects and gives up. Please make sure pin #22 on the RTC chip (PS - power sense) has battery voltage on it, or at least that it is powered from PSU when the mobo is running. There must be at least 3V there I think.

EDIT: As to why it's calles "CMOS", it's because the early chips (including CPUs) would be older tech, like NMOS or even PMOS, this allowed more dense and faster transistors but at the cost of higher power draw. Obviously RTC cannot have high power draw or it will drain the battery quickly, so these were made in CMOS technology. Often that would be the only true CMOS chip on the early mobo, so it was called that instead of NVRAM or RTC.

Same error, motherboard stops in code 20 and in the screen the same message: CMOS inoperational. System halted.
In this situation, should I make the test with the 4k7 resistor?
Regards and thank you

Deunan wrote on 2024-02-19, 14:52:

If you want something to try, I have 2 experiments in mind: […]
Show full quote

jnemo2004 wrote on 2024-02-19, 14:37:

Thank you very much for your info. I have already checked the voltage in pin 24 and it was OK (4,78 mobo ON and 3,47 mobo OFF). I will check the voltage in pin 22 (PS). Regards.

If you want something to try, I have 2 experiments in mind:

1) Please remove the RTC chip and see if the mobo powers up with the same error. Or is it different now.
2) With the RTC removed please attach 4k7 resistor between pins 11 and 24. You could try sticking the resistor leads into the socket but it might damage it, you can also do a temporary solder to the socket pins on the backside of the mobo. Power the mobo with the resistor (but without RTC chip) and see if anything changes.

If you don't have 4k7 resistor then 10k down to maybe 1k should also work. But do not use resistor smaller than 1k for this.

So this bios is incapable of generating lower number error code for bad/missing cmos. I would source another RTC chip, preferably one like in the retroweb photo