Taken from DS1287A datasheet: "DS1287A is identical to the DS1287 with the addition of the RAM clear pin."

The RAM clear pin is pin 21 (RCLR - RAM clear). If the original RTC module for the board is DS1287 (the one with missing pin 21) and you're using DS1287A(+) instead, the pin 21 of DS1287A(+) module may be shorted to ground and this way stuck in permanent RAM clear (clear BIOS settings) state.

I tried bending pin 21 out of the way and re-installed the RTC module, same problem.

Ok, pin 21 shorted to ground is not the culprit. Your DS1287A+ RTC module is brand new, so it doesn't come with stored BIOS settings for another board.

Do you have POST code diagnostic card?

I do, but the guide I have doesn't identify all of the codes that appear. The last code that shows before it stops and the beeping starts is "0d" which doesn't have a description for AMI bios in the guide it came with. Best I can tell from Googling is it's probably indicating the video adapter was initialized. I've tried with other video cards, doesn't seem to change anything. Even with the video card removed I still get the same 10 beeps and "0d" post code.

I don't know which AMI BIOS version your motherboard uses, but you can check for AMI "0D" POST code here: https://blog.theretroweb.com/2024/01/20/amibi … post-codes-list . I assume your 486-boiard uses either AMI "color" or AMI "winBIOS" and the "0D" POST code points to different problems according to which AMI BIOS is use.

AMI color: oD = 8042 keyboard controller tested, CMOS

AMI winBIOS: 0D = verify batch command

On the other hand, ten short beeps for AMI BIOS = CMOS shutdown register read/write error - the shutdown for the CMOS has failed

I've made some progress. The RTC chip that came with the board is an Odin OEC12C887. If I put that back in the system POSTs.. I get the CMOS battery low error, but it still POSTs. I have 3 Dallas RTC's, 2 from eBay and one from Digikey. The one from Digikey is a DS12887A, the ones from eBay are a DS12887A+ and a DS12C887.

The Odin RTC was already modified to use a CR2023, but one of the exposed pins on the chip broke off so I can't solder a wire to it. I tried the Dallas DS12C887 again, but back to 10 beeps. Not totally sure what is going on, the board doesn't seem to like the Dallas chips and I can't do anything with the Odin chip. I believe the Dallas DS12C887 should be a drop in replacement for the Odin, but it's not working at all.

So back to the CMOS problem.

Mind that ODIN's external battery modification is not exactly the same as the one for the Dallas' RTC: you have to cut internally pin 20 on ODIN RTC module instead of pin 16 (like on Dallas modules).

But why brand new Dallas DS12887A doesn't work as a substitute?

That's the question. This board on retroweb: https://theretroweb.com/motherboards/s/biosta … -1433aea-v#docs

The picture shows a TH6887A RTC chip installed, the jumper manual shows a Dallas DS1287, and the board I have came with an Odin OEC12C887. I can't figure out why it doesn't like the Dallas RTC's.

I once replaced ODIN OEC12C887A RTC with Dallas DS1287A on a s.5 Soyo motherboard and there was absolutely no problem.

Ok, maybe they're not always 100% compatible.

From Dallas DS12887A datasheet:
"Interfaced with software as 128 RAM locations
– 14 bytes of clock and control registers
– 114 bytes of general-purpose RAM"

From Dallas DS12C887A datasheet (I can't find datasheet for ODIN OEC12C887A, but they must be identical):
"Interfaced with Software as 128 RAM Locations
- 14 Bytes of Clock and Control Registers
- 114 Bytes of General-Purpose, Battery-Backed
RAM (113 Bytes in the DS12C887 and DS12C887A)"

DS12887A has 1 byte more RAM.

Оne of the dumbest problems with s.5/7 motherboards - RTC modules.

Just for the record, the shutdown register allows the BIOS to reset the CPU and remember the reason why it reset afterward--if you think back to the story of the 286 CPU not being able to go back to real mode from protected without a reset, it's related to that.

Your BIOS is complaining that the shutdown code it writes there doesn't "stick," so it's in a hopeless situation since it won't be able to remember how far through the POST it is (since the CPU resets multiple times during the POST). So definitely double check that it's the correct RAM clear pin (21) bent out of the way.

I don't know why it doesn't work either. Do you have another motherboard with a discrete RTC that maybe you can drop the new Dallas into, to preload some possibly more sane settings into it, then move it back to this board? And another option is to dump the BIOS, since the CMOS shutdown register test code should be straightforward to find. I think many of us will be curious why there is a compatibility issue.

Just glancing at the datasheet, you might double check that on the socket on the motherboard,
pin 1 and 24 do not have continuity.
pin 18 and 24 have continuity.
pin 12 and 21 do not have continuity.

Unfortunately this is my only board with this type of RTC. I can't get a dump of the CMOS either, I can't get it to boot to anything. The BIOS settings won't save at all. Every time I power it on, I need to set drive A: to a 1.44mb floppy, but as soon as I save the setting and reboot it resets back to a 1.2mb 5.25 drive. Can't get it to boot to anything now.

I got my hands on a nwX287 with the BQ3285 chip which is supposed to be compatible with the Odin OEC12C887 RTC module but it still won't post, still getting the same post beeps. For whatever reason this board will not post without that OEC12C887 chip installed and it doesn't look like it's possible to get a replacement.

Is JP11 open or closed?

It's soldered closed.