The "Medicat-System" label is likely has nothing to do with the medical stuff. It looks like a typical sticker put by a reseller company to track the warranty period.

Do you have one of those POST-cards to see the bios current activity? Might help in identifying problem.

I think you are going to want Parity SIPPs or to find a jumper that turns parity off. Those will have 3 chips or 9 chips on them. Another thing to try would be to "borrow"4x256 and 1x256 from any other machine you have that has that type and seeing if it will recognise those in the DIP sockets with no SIPP installed.

Your comb-shaped RAM modules are called "SIPP modules" (single in-line pin package). The board should run with two modules, likely BANK0 + BANK1. Your modules are 1MB, 70ns, no parity. The missing parity bits might cause the issue you are seeing, although that doesn't explain why you get a different symptom with or without the video card installed. The base 64K test should be a lot earlier during POST than the video card initialization.

You should be aware that some video cards beep during POST if no monitor is connected, so possibly the three beeps are not from the AMI BIOS, but from the cirrus logic card. The beep codes by the AMI BIOS should consist of beeps of exactly the same length and the same frequency, possibly preceeded by an introduction beep that might have a different length and frequency. You must not count the introduction beep towards the beep code. If you have it, you might try whether a different video card causes other beeps, or no beeps at all.

Another test is powering on the mainboard without any RAM at all. You should get that base 64K error beep in that case, unless the HT12 memory auto-config code locks the system before the base 64K test starts.

Your memory modules use CMOS chips, so it is expected that they don't get warm unless you run memory cycles at maximum speed all the time. It's normal for the 286-16 (which is likely produced in Intels HMOS technology, which is a refined NMOS process. AMD just manufactured the Intel design as second source) to get quite warm. So the temperatures you mention are what you expect, even on a working board.

Kindly thank you for the informative responses, and for helping me out. Now, thanks to you I understand the RAM concept on these motherboards better. Parity sure rings a bell, I even have some 30 pin SIPP sticks, but they are not in a comb form. Unfortunately I have no DIP memory, as this is my only 286 board. I planned to order some one day, seems like I will need to. Also dont have the diagnostic card, but I will purchase one for sure, I spent too much time scratching my head about whats wrong with these boards 😁
I discovered what likely caused the strange beeping behavious whether the GPU was connected or not. I managed to literally trace the problem, and it was a broken trace from a small impact that just broke a single trace (PIN17) near the SIM2. That trace was connected to SIM1 aswell (also PIN17), and the docs states it's one of the address lines, so this might explain the strange silent behaviour.
As I fixed that trace, there are now always 3 beeps when RAM is in SIM1, SIMM1+SIM2, SIM1+SIM2+SIM3+SIM4 (also tested SIM1+SIM3, although not sure if it makes any sense), no matter GPU. It there is no RAM, the board goes silent.
I've ordered some another parity RAM sticks, because I dont want to mess my already owned ones. I plan to make combs out of them like in the pic. I pressume it should work if I solder that header to the RAM stick?

If I'm not mistaken that's exactly what you are showing but with thinner pins.

You can get standard 30 pin SIMM holders which will fit straight in to the SIPP locating holes. This will open up your choice of ram modules a bit. Just make sure they a correctly orientated before using them though.

Yes, that is what I have done with SIPP sockets, plugged SIMM sockets into them. For temporary use, you can sometimes get away with getting a dual pin header, squishing the two sets of pins together one side, plugging one side of the bottom into the sockets then putting the SIMM in between the squished together pins on top. However, it is not very secure and knocks and vibrations may dislodge it.

Hi again, thank you very much for helping me out!I took your advice, but still no success 🙁
I've purchased a parity rams, put them (sim1, sim1-2, sim1-4) - still 3 beeps.
I've also purchased a POST diagnostic card, and what's interesting I have a "100C" code, which according to manual means (in my own translation):
10 - The byte zised command to the keyboard controller has been stored. Next, generate the command to lock and unlock pins 23 and 24.
0C - The input buffer of keyboard contoller has been erased. Next, send the BAT command to the keyboard controller.

So this suggest, there is something wrong in the communication with the keyboard controller during its test (thus the BAT command). I've tried to change the controller, I put the Fast Keuboard BIOS v5 but this didn't change anything. Do you think it is a good direction?

Rafal wrote on 2021-08-17, 19:49:

Hi again, thank you very much for helping me out!I took your advice, but still no success :( I've purchased a parity rams, put […]
Show full quote

Hi again, thank you very much for helping me out!I took your advice, but still no success 🙁
I've purchased a parity rams, put them (sim1, sim1-2, sim1-4) - still 3 beeps.

I've also purchased a POST diagnostic card, and what's interesting I have a "100C" code, which according to manual means (in my own translation):
10 - The byte zised command to the keyboard controller has been stored. Next, generate the command to lock and unlock pins 23 and 24.
0C - The input buffer of keyboard contoller has been erased. Next, send the BAT command to the keyboard controller.

Discard the manual provided with the card. The card doesn't understand anything about the POST process, it just displays whatever code numbers are sent by the BIOS. Every BIOS manufacturer uses different codes. Sometimes, the codes get reorganized by major version updates even at the same manufacturer. Also some vendor-provided hooks in certain boards output extra POST codes, which you can't find in any list.

Your best choice is a comprehensive POST code list, e.g. the one at www.bioscentral.com. Checking the AMI BIOS subpage shows for old AMI BIOSes:

• 0C: System timer OK
• 0D: Referesh timer OK
• 10: Start 64K base memory test

Your card shows the last value on the left and the previous value on the right. So the computer is currently at the 64K base test, and the previous checkpoint was that the system timer is OK. The POST card model you use (which is very common, and I use it, too) can miss some POST codes if they are sent in quick succession, so possibly 0D was also sent but not caught by the card. The three beeps you are hearing "base 64K memory test failed" match up perfectly: The BIOS started the base memory test (POST code 10), found the memory to be bad, issues three beeps and halts the system.

As you now have (probably) good memory modules, and the BIOS still is unable to access that memory correctly, I guess your board has some hardware damage. Possibly it is a broken trace to memory on the board, most likely concerning the RAM address bits. Your board obviously can read the BIOS EPROM successfully (BIOS checksum check is before base 64K RAM check), and as the BIOS is 16 bit (a "high" ROM and a "low" ROM) on your board, this means the 16-bit data bus from the processor to the ROM works perfectly fine, as well as the addressing of the ROM. A stuck "chip select" line on the ROMs, having them interfere with memory data might also be the problem. If you are interested in hardware troubleshooting, please tell us about your experience and tools (e.g. a cheap multimeter or a digital scope).

I have seen many SIPP modules not working in such 286 boards. Try to get some DIPs and fill one bank.

Aw, I should assume the manual is for more modern boards 😀 So there is a 100% problem with access to the RAM, so basically two options I could try:
1. Analyze the traces on the board - I do a lot of board repairs, via replacement, trace repair, recaps, small IC replacement, transistors, and so on. Even small SMD components.
I have a DMM, but Im not sure where the memory address / data / control lines should go. I guess I should mess around with it - so data lines should go to the HT12 chipset right? And address lines to the CPU or also HT? (Im have no experience in PC, I do a lot of Amiga 600 / 1200 repairs so different thing). Could you please help me with troubleshooting that traces?
2. Try DIP RAM - Meanwhile I will try to order some DIP packaged RAM (seen one offer in one far away country so the shipment would take some time).

Predator99 wrote on 2021-08-18, 11:48:

I have seen many SIPP modules not working in such 286 boards. Try to get some DIPs and fill one bank.

The SIMM/SIPP specifications were kind of up in the air, incomplete/not-firm, even as the first boards were being made quite likely.

Rafal wrote on 2021-08-18, 13:35:

2. Try DIP RAM - Meanwhile I will try to order some DIP packaged RAM (seen one offer in one far away country so the shipment would take some time).

Did you search for 514256 ...think they should work? You can also look on old VGA cards for some matching ones...

Please don't be disappointed by the slow progress.
I'm confident that it can be repaired, it judt takes some time.
Maybe one of the 74xxx chips is defective, at worst, not sure.
The chipset itself is unlikely to to be damaged, I believe.

From my experience, 286 systems were though and hard to kill.
Other systems like the C64 were much more prone to failure.

Predator99 wrote on 2021-08-18, 11:48:

I have seen many SIPP modules not working in such 286 boards. Try to get some DIPs and fill one bank.

Afaik, 256k SIMMs/SIPPs should be as slow/limited as real DIL chips. 😉

Edit: 286 systems require pairs of two, due to the 16-Bit interface.

Jo22 wrote on 2021-08-18, 15:31:

Predator99 wrote on 2021-08-18, 11:48:

I have seen many SIPP modules not working in such 286 boards. Try to get some DIPs and fill one bank.

Afaik, 256k SIMMs/SIPPs should be as slow/limited as real DIL chips. 😉

Edit: 286 systems require pairs of two, due to the 16-Bit interface.

I am not talking about speed but compatability. I have several similiar boards I have tried to get running with SIPPs without success. After inserting DIPs they worked. I would try this 1st...

Take a look here
80286 BIOS image collection
How many boards with SIPP sockets populated do you see here? 😉

Predator99 wrote on 2021-08-18, 17:28:

I am not talking about speed but compatability. I have several similiar boards I have tried to get running with SIPPs without su […]
Show full quote

Jo22 wrote on 2021-08-18, 15:31:

Predator99 wrote on 2021-08-18, 11:48:

I have seen many SIPP modules not working in such 286 boards. Try to get some DIPs and fill one bank.

Afaik, 256k SIMMs/SIPPs should be as slow/limited as real DIL chips. 😉

Edit: 286 systems require pairs of two, due to the 16-Bit interface.

I am not talking about speed but compatability. I have several similiar boards I have tried to get running with SIPPs without success. After inserting DIPs they worked. I would try this 1st...

Take a look here
80286 BIOS image collection
How many boards with SIPP sockets populated do you see here? 😉

Well I have one... tested with 2 256K sipp modules and 4 1M sipp (9 chips) modules. Re: 80286 BIOS image collection
Though !!! because I had more simms than sipps I mangled 4 simm sockets, tinned the pins a bit, and plugged them into the sipp sockets 😋
sadly for me, the board only supports a maximum of 4MB of memory, which is 4 sipps or simms of 1M 9 chips. I was so hoping to have 8 or 16MB... but no. 🙁

I see you posted this board Re: 80286 BIOS image collection with 2 different simm modules in one bank, and two extra 1 bit dips ? must be some crazy good chipset on that one.

this board of yours has two sipp modules Re: 80286 BIOS image collection

and many more boards with sipp sockets, though no modules,
and even some board with no ram whatsoever.

Rafal wrote on 2021-08-18, 13:35:

1. Analyze the traces on the board - I do a lot of board repairs, via replacement, trace repair, recaps, small IC replacement, transistors, and so on. Even small SMD components.
I have a DMM, but Im not sure where the memory address / data / control lines should go. I guess I should mess around with it - so data lines should go to the HT12 chipset right? And address lines to the CPU or also HT? (Im have no experience in PC, I do a lot of Amiga 600 / 1200 repairs so different thing). Could you please help me with troubleshooting that traces?

I hoped the HT12 datasheet is freely available, but it seems it is not. So generally, you expect the row/column multiplexer to be inside the chipset, this means RAM addresses should go from the HT12 chip to the SIPP sockets, possibly via a unidirectional buffer (e.g. 74'244). Also, RAS/CAS originates in the chipset. The 16 data lines probably are directly connected to the 286 local bus, possibly via a birectional buffer like the 74'245.

Hehe, yes I made the photos of most of my boards before I tried to get them running...many of them are still not running at all.

Predator99 wrote on 2021-08-18, 14:27:

Did you search for 514256 ...think they should work? You can also look on old VGA cards for some matching ones...

I was looking for HY534256S-70, saw another thread on forum with a mobo with these chips. Also it has a smaller DIP16 chips, i guess these are for parity? Do i need the DIP16 or Its just enough to fill one bank with DIP20?

Jo22 wrote on 2021-08-18, 15:31:

Please don't be disappointed by the slow progress. I'm confident that it can be repaired, it judt takes some time. Maybe one of […]
Show full quote

Please don't be disappointed by the slow progress.
I'm confident that it can be repaired, it judt takes some time.
Maybe one of the 74xxx chips is defective, at worst, not sure.
The chipset itself is unlikely to to be damaged, I believe.

Thanks! I'm stubborn, and I never give up on my repairs 😁

mkarcher wrote on 2021-08-18, 19:19:

I hoped the HT12 datasheet is freely available, but it seems it is not. So generally, you expect the row/column multiplexer to be inside the chipset, this means RAM addresses should go from the HT12 chip to the SIPP sockets, possibly via a unidirectional buffer (e.g. 74'244). Also, RAS/CAS originates in the chipset. The 16 data lines probably are directly connected to the 286 local bus, possibly via a birectional buffer like the 74'245.

Yes, I was looking for that aswell, but the only thing that i found is this document from a different board but with HT12. On the other hand, the 286 package has a good pin out documentation already.

I did just a quick check with a DMM and it turns out that more address traces are not connected to the chipset. So I made a closer look but I see no other damage (I already fixed one tiny address line break before).
For example the A10 is not connected to HT12, and does not beep on any of the DIP sockets aswell. I need to take closer look and maby make some bridges, but still don't know where exactly...
But in general, it should just like you wrote:
- address from CPU to HT12 (green)
- memory address (MA?) from HT12 to RAM (blue)
- data CPU, HT12, RAM (red)