Those old 720k floppy drives have a few electrolytic caps in them and should be replaced.
Dead floppy drive? Need a 2nd opinion
As far as the old HD's have never had any good luck fixing the controller boards or if they did work after fixing it was not for long. Motor bearing or platter issues soon followed.

I thought about this (this is a 1.44MB drive though), but if you pay attention, the caps are all located on the second board, not the one where the stepper motor is connected to (and I've looked on the other side and there are none). Now maybe that's the problem, but that doesn't seem to make much sense to me.

I may end up swapping that drive afterall, finding an identical one shouldn't be too hard.

OMG. This is the first PC I ever owned and as knowing nothing at the time (I as 11 at the time) we didn't realise how poor a specification this PC had. (An 8088 cpu in 1991!)
My father bought it and it came as such:
8088 cpu
640k RAM
Built in VGA graphics
NO Sound card (only PC speaker-I added a Disney Sound Source)
1.44mb Floppy disk drive
NO HDD - had to boot into DOS with Floppy
NO Mouse (added a serial one later)
NO windows (Only Deskmate - 7 Floppy disk 'OS' with limited functionality)
A Good Colour Monitor
I added a joystick card afterwards.
Epson Dot-Matrix

Nice ! Do you remember how the bios setup was accessed ? Or do you still have the setup disks around ?

Deksor, if the bios is available online (hint hint) someone could dive into it and figure out what data is expected to be found inside of the dallas chip, which port it is located on and a simple program could be written to get the required data back into the dallas chip.

Did you ever get the hard drive working again ?

Sadly no, I haven't found any identical HDD from which I could steal the pcb

However I made a strange discovery when looking inside the bios ROM : there are signs of a BIOS setup.

However the computer doesn't let me access it in any way, it asks for a setup disk.

Now I'm quite confused.
Did Tandy take an ami bios (that's what it is) and disabled the integrated setup ?
Was this by (bad) design and the setinit program could be resumed as a debug command ?

This needs more investigation but I'm not sure I can do that.
I tried to run debug to run code at the offset the setup seems to start but that didn't work 😒

Can you post a dump of the BIOS? Would like to take a look at it.

Whoops, sorry, you can download it here https://www.ultimateretro.net/en/motherboards … 10766#downloads

odd things, the bios seems to have the option for english franch 😁 and german, yet I find only english and french text. no german.

Just noticed, the mainboard uses the acer m1101 chipset. I think that's the same one as my old xt board has, which is a rebrand of ... i forgot... probably some UMC chip, yep umc 82c088
it can switch between turbo & xt frequencies by writing to a specific port if memory serves me right. and without loading the keyboard driver, if the bios supports it, should be able to switch using ctrl-alt-numpadminus or plus .

Here's to hoping predator99 can figure out which port the dallas chip is on and what the checksum and other entries ought to be.

I took a look at the dump. The 1st half is a Video ROM. The 2nd half is the system BIOS.
It runs in PCem in the "Atari PC3" machine.

It loads XT-IDE but doesen detect a HD (this should be fixable on a real PC..). It boots DOS from floppy.

So whats the issue you actually have? Doesnt your PC boot at all? I think you can also try a generic XT BIOS like
https://www.phatcode.net/downloads.php?id=101

The problem I have is that I can install things on the HDD, but I can't boot from it, and I'd like to access the CMOS setup because I'm curious what settings I can mess with in there (That'd also let me get rid of the message "Run SETINIT.EXE" at POST)

So you can boot DOS from floppy and access the hard disk as C:?

Adding a XT-IDE Bios on a network card should fix this?

I am not sure if the CMOS setup is really implemented in the BIOS. Yes, the strings are visible. I will try to check this. But you can also see
"720 KB 3?", 1.44 MB 3?", Not Installed Options:"
"German : Select German"
I dont think a 1.44 MB floppy is supported. There also no further german strings inside...

It seems CMOS is configured exclusively with SETINIT.EXE. Therefore I dont think it can be done within the BIOS, too.

It *does* support 1.44MB floppy

Using XT ide would probably work, but that doesn't feel right, because the main bios should be capable of this too 🙁 (and other things such as the RTC don't seem to work properly)
This bios is so weird.

Also I don't think I'd even need a NIC for XT-IDE if I wanted : there's a second ROM socket, and the socket the ROM is in can fit bigger ROMs, so I wonder if I couldn't just put a larger ROM and it'd work as well ?

Deksor, from my experience with my xt board with an m1101 chipset, the second rom socket, on my board, was mapped to where the BASIC rom is supposed to be

getting rid of the SETINT message would imply figuring out where it reads the RTC ram and maybe does some checksum calculation, then figuring out how to write a good checksum back into the RTC ram.
ergo which port the rtc is located at on the XT, you could also do this with a logic analyzer hooked up to the nwx287 AD & AS pins
or using something like unpc to read all the io ports and look where the data resembles RTC registers.

Goodmorning !

I took a look at the BIOS...:

At the beginning it does this:

1F000:E015 BB00A0        MOV	BX,A000                            
2F000:E018 8EDB          MOV	DS,BX                              
3F000:E01A 32FF          XOR	BH,BH                              
4F000:E01C 813F55AA      CMP	WORD PTR [BX],AA55                 
5F000:E020 7406          JZ	E028                               
6F000:E022 813F4D46      CMP	WORD PTR [BX],464D                 
7F000:E026 7505          JNZ	E02D                               
8F000:E028 EA080000A0    JMP	A000:0008                          
9F000:E02D FC            CLD	                                   
10F000:E02E 8CC8          MOV	AX,CS                              
11F000:E030 8ED0          MOV	SS,AX        

Doesnt make much sense to me: It looks if there is a ROM at A000:0. If yes, there is a jump to it? Thats where the Video RAM should be?

All these error messages are printed here:

1F000:8487 B402          MOV	AH,02                              
2F000:8489 32E7          XOR	AH,BH                              
3F000:848B 7404          JZ	8491                               
4F000:848D 81CD8000      OR	BP,0080                            
5F000:8491 B028          MOV	AL,28                              
6F000:8493 E680          OUT	80,AL    [b]+++POST Code 28     [/b]                     
7F000:8495 F7C50020      TEST	BP,2000                            
8F000:8499 7408          JZ	84A3                               
9F000:849B BEF185        MOV	SI,85F1     [b](CMOS INOPERATIONAL)          [/b]             
10F000:849E E8FD40        CALL	C59E                               
11F000:84A1 EB48          JMP	84EB                               
12F000:84A3 F7C51000      TEST	BP,0010                            
13F000:84A7 7406          JZ	84AF                               
14F000:84A9 BE9B86        MOV	SI,869B 	[b](CMOS checksum failure)        [/b]               
15F000:84AC E8EF40        CALL	C59E                               
16F000:84AF F7C50800      TEST	BP,0008                            
17F000:84B3 7406          JZ	84BB                               
18F000:84B5 BE1F86        MOV	SI,861F   [b] (CMOS battery state low)  [/b]                      
19F000:84B8 E8E340        CALL	C59E                               
20F000:84BB F7C50400      TEST	BP,0004                            
21F000:84BF 7406          JZ	84C7                               
22F000:84C1 BED286        MOV	SI,86D2      [b] (CMOS time & date not set)                     [/b]
23F000:84C4 E8D740        CALL	C59E                               
24F000:84C7 F7C58000      TEST	BP,0080                            
25F000:84CB 7406          JZ	84D3                               
26F000:84CD BE5287        MOV	SI,8752        [b](CMOS floppy type mismatch)     [/b]               
27F000:84D0 E8CB40        CALL	C59E                               
28F000:84D3 F7C54000      TEST	BP,0040                            
29F000:84D7 7406          JZ	84DF                               
30F000:84D9 BE1187        MOV	SI,8711           [b](CMOS dispyal type mismatch)    [/b]             
31F000:84DC E8BF40        CALL	C59E                               
32F000:84DF F7C52000      TEST	BP,0020                            
33F000:84E3 7406          JZ	84EB                               
34F000:84E5 BE5786        MOV	SI,8657           [b](CMOS system options not set)     [/b]            
35F000:84E8 E8B340        CALL	C59E                               
36F000:84EB B029          MOV	AL,29                              
37F000:84ED E680          OUT	80,AL  	[b]++++Post Code 29                [/b]            
38F000:84EF C606720000    MOV	BYTE PTR [0072],00                 
39F000:84F4 E85F5B        CALL	E056                               
40F000:84F7 81E5FFBF      AND	BP,BFFF                            
41F000:84FB 55            PUSH	BP                                 
42F000:84FC 81E5FF1F      AND	BP,1FFF                            
43F000:8500 5D            POP	BP                                 
44F000:8501 742F          JZ	8532                               
45F000:8503 E8505B        CALL	E056                               
46F000:8506 F7C5FC00      TEST	BP,00FC                            
47F000:850A 740A          JZ	8516                               
48F000:850C BEA587        MOV	SI,87A5    [b]Run SETINIT program            [/b]            
49F000:850F E88C40        CALL	C59E                               
50F000:8512 81CD0040      OR	BP,4000                            
51F000:8516 BE9585        MOV	SI,8595    [b]Press a key to boot    [/b]                    
52F000:8519 E88240        CALL	C59E                               
53F000:851C F7C50020      TEST	BP,2000                            
54F000:8520 7509          JNZ	852B                               
55F000:8522 B03F          MOV	AL,3F                              
56F000:8524 E81140        CALL	C538                               
57F000:8527 3CA5          CMP	AL,A5                              
58F000:8529 7407          JZ	8532                               
59F000:852B B400          MOV	AH,00                              
60F000:852D CD16          INT	16                                 

61F000:852F E8245B        CALL	E056    

However, the actual check is not done here. It depends on the content of BP which error message is printed. And this is loaded somewhere else...
You can also see not all text thats inside the ROM is used.

This is weird because I definitely have french messages on my screen. Your screenshot has one too ! Unless you mean that some french messages appear and some don't ?
Maybe the ROM is shadowed and modified in memory ?

Considering this is a 1991 XT bios and what we can see in this ROM, to me it feels like it was rushed (They probably didn't put too much money into making such a machine, because who would buy a new XT machine in 1991 except someone that didn't have a very large budget ?)

Indeed, I noticed my mistake as I noticed my screenshot above: In the dump there is always the english message and next to it the identical one in french. Both are printed...

Shadow in a Xt..? Dont think so. Maybe A000 is the location of the 2nd ROM socket. But thats where the VGA-RAM should be...

But also interesting is that the Video-ROM is in the 1st half of your ROM dump. Therefore it should be mapped to F000:0. Cant check in PCem if its really there because I had to delete this part to get it running. BIOS-ROM starts at F000:8000.

I guess I could check this with some debug-fu on my machine ?

Shadow in a Xt..? Dont think so. Maybe A000 is the location of the 2nd ROM socket. But thats where the VGA-RAM should be...

Well I know it's uncommon, but everything in that machine is uncommon so that may not be entirely out of question ^^

Maybe I should try some VGA program to make sure VGA works

Here I am looking at the code, wondering why that OUT 80,AL looks so familiar.
then it hits me, duh, post codes 😋

would it be possible, with some glue logic in hardware to map the first part of the rom to A0000 and the second part at F8000 ?
I guess it would only be one bit that needs to flip at the right time.

but

I'm more interested in what gets done before those error messages get printed.
Which IO ports gets read that might be read from the RTC and see what gets done with the data. reverse engineer what the SETINIT should do.