Maybe the extra RAM can only be used as EMS or RAMdrive.
Most of the 286s I know of with 8-bit slots are limited to only 1MB of directly addressable memory.

you maybe right, someone else suggest me the same.
i tried to see with qram if something happen, but it's a no go, it just hang when run optimize for the first time.
i don't know if he require something more proprietary driver. if you know something i can try ?
he actually run pc dos 7.

for the 8 bit slots, i think it's mostly because the chassis don't allow 16 bit to fit... they reused the same chassis as the MBC 16/17 series, and it's a short one.
the motherboard itself is very short.

if this can help, he use a phoenix bios version 1.02.10 - 01. with a very minimalistic setup menu, just for set time and date, and confirm if 287 is here or not.
and the chipset seems to be a chips and technologies F82C230.

hmmm i finally found something about the chipset... and in fact, this machine is more a PS/2 model 30 8086 clone than an AT.
http://www.bitsavers.org/components/chipsAndT … orm_Catalog.pdf

so, you think sanyo made the choice to only limit extra ram as ems only ?
sad they don't say anything about this in their manual.

Hi Whaka, I don't suppose you have the manuals and disk images avaialble? I have one of these but no disks so I can't change any BIOS settings.

Good afternoon, dear friends! I apologize in advance for my poor English 😀
I am forced to ask for advice, because I can't think of anything else.
I received such a board (see photo) - Sanyo MBC-17Plus, also known as Sanyo SAT-201B. At first, I wanted to dismantle it and use it as a donor of microcircuits. Because it did not show any (at all) signs of life. The VGA/EGA board did not initialize. As it turned out, the Sanyo BIOS does not write any POST codes to port 80h and does not even give a sound indication of POST errors.
There is very little documentation for this board. It all boils down to configuring a few jumpers and one DIP switch.
However, after checking the oscillograms of the data bus and the address bus, I thought that the patient had a chance to live. And indeed, when I used HGC instead of a VGA or EGA card - Sanyo started! It also started with a CGA card without any problems. I thought: "Okay, there are some problems with the initialization of the EGA/VGA BIOS, or problems with finding extensions at address C000h". Moreover, the Sanyo also refused to see the XT-IDE card, configured for Dx00h.
Having replaced the original BIOS with Supersoft/Landmark Diagnostic ROM, I managed to test the board. No errors were found. At the same time, I remembered the "features" of the Landmark ROM regarding the so-called "self-initializing" VGA cards. And indeed, as it turned out later, with some VGA and EGA cards Sanyo MBC-17 works fine. And with some - it refuses outright. Maybe someday I'll check with my entire collection and publish the statistics, but that'll be later.
I should also add that this board requires a working battery to start. It doesn't give out "Battery state low" notifications, it just freezes.. Strange... But that's not a problem - I unsoldered the connectors and connected a new battery, along with the keyboard and speaker.
But that's where my luck ended... I connected the floppy drive - it won't boot... I tried booting with a 3.5''-1.44Mb floppy drive, a 5.25''-1.2Mb, and a 5.25''-360Kb floppy drive. The "Floppy drive seek at boot" procedure runs normally, after which the floppy drive makes a few head movements - and that's it. It freezes. I tried with different floppy disks (MS-DOS v.3.3, MS-DOS v.5.0, MS-DOS v.6.22, EZ-DOS, CP/M-86) - the same result. I tried with the original floppy disk that came with the Sanyo MBC-17 (actually MS-DOS v.3.20), and contained some utilities for working with hard disks - the same result. The floppy drive makes a few head movements - and that's all. It freezes.
Okay, I thought. Maybe the problem is in the disk drive controller (WD37C65) built into the MBC. I turned off the built-in FDC with jumpers, tried with external FDC cards. With several different FDCs (37C65, as well as with 82077 and some others...) - the result is the same. The disk drive makes a few movements with its heads - and that's it. It freezes.
Having lost hope of booting from a floppy drive, I decided to try hard drives. Since this Sanyo did not want to “see” XT-IDE, I used the most ordinary 16-bit IDE card. It is clear that large hard drives should not even be tried to be connected without XT-IDE, so I took the “correct”, “reference” ST351A/X.
By the way, I have a strong suspicion that the hard disk parameter table in the Sanyo MBC-17 BIOS is somewhat different from the usual ones used in AMI BIOS, Award BIOS, etc. But I have no way to check this - so maybe I'm wrong.
Nevertheless, using a regular 16-bit IDE card and ST351A/X, the process started, but only at first glance. The hard disk moved its heads more actively, blinked the LED indicator, and even issued a notification "Starting MS-DOS". But that's where all the progress ended, the boot did not reach completion, everything froze ...
I repeated these attempts with other hard drives (Miniscribe 8051A, Western Digital 95044-A, WD Caviar280). And the same, with different operating systems. The result is the same - everything freezes on the notification "Starting MS-DOS" or "Loading CPM.SYS".
Almost in despair, I decided to try to boot from MFM hard drives. Given the nuances with the initialization of VGA/EGA cards and problems with XT-IDE, I expected problems with the BIOS initialization of 16-bit MFM cards. But, surprisingly, with two MFM controllers (Cirrus Logic CL-SH260 and Western Digital WD 42C22A) and with two hard drives (Seagate ST-125 and Miniscribe 8425F), Sanyo "found a common language". But only for a short time. It all ended the same - everything hangs on the notification "Starting MS-DOS" or "Loading CPM.SYS".
At this point I gave up trying to get the Sanyo MBC-17 to boot normally. Of course, I had the idea to try SCSI adapter like the AHA-1520/1522 or AHA-1540/1542. But I think the result would be the same...
In addition, I suspected that the problem might be in the Sanyo MBC-17Plus BIOS firmware. I found several BIOS versions for this board - both slightly older than mine and slightly newer. I tried them all - the result was the same...
I tried to get it to work with various early versions of Phoenix BIOS, Award BIOS, AMI BIOS. I couldn't get any improvement with any of them, on the contrary. With almost all the "foreign" BIOSes I received messages about errors in determining the amount of RAM.
I'm giving up ((( But I put so much effort into this board – that I just don't want to back down. Has anyone encountered similar boards? Has anyone observed similar problems? – I would be very grateful for advice !!!

Supersoft/Landmark Diagnostic ROM is really bad when it comes to testing ram, and everything points at bad ram
try flashing this one https://github.com/ki3v/xtramtest

rasz_pl wrote on 2026-06-19, 22:21:

Supersoft/Landmark Diagnostic ROM is really bad when it comes to testing ram, and everything points at bad ram
try flashing this one https://github.com/ki3v/xtramtest

Are you sure this diagnostic ROM will work if used as a BIOS extension on a 286 motherboard? - I tried running Ruud's diagnostic ROM - it didn't work, it won't start.

Vlodek_d wrote on 2026-06-21, 17:03:

rasz_pl wrote on 2026-06-19, 22:21:

Supersoft/Landmark Diagnostic ROM is really bad when it comes to testing ram, and everything points at bad ram
try flashing this one https://github.com/ki3v/xtramtest

Are you sure this diagnostic ROM will work if used as a BIOS extension on a 286 motherboard? - I tried running Ruud's diagnostic ROM - it didn't work, it won't start.

As i know - diagnostic ROM is BIOS replacement not BIOS extension

Babasha wrote on 2026-06-21, 18:47:

As i know - diagnostic ROM is BIOS replacement not BIOS extension

"Method 2: Installing as an "option ROM"
The second way this ROM can be used is by installing it as an "option ROM". The exact method of preparing a ROM chip will vary (and is currently left as an exercise for the reader, as they say), but it is the same as preparing any other option ROM such as XT-IDE.
Once the EPROM is prepared, it can be installed in any option ROM socket on the motherboard. Depending on your BIOS, you may be able to install the option ROM in the BASIC ROM socket (which will generally be empty on clones and filled on actual IBM PC/XT computers).
Then, when booting any BIOS that supports option ROMs, the BIOS will initialize this ROM, which will install itself as a warm start handler, and when the BIOS attempts to boot the operating system, this BIOS will print a message asking you to press the T key to start RAM tests, or any other key (or just wait a few seconds) to continue booting.
If there are other boot-providing option ROMs in the system (such as XT-IDE or network cards with boot ROMs) which are loaded at higher addresses, they will be given an opportunity to boot first, and they may not permit the XTRAMTEST ROM a chance to take over, so you may have to remove other cards in order to get this ROM to run in option ROM mode."

I quoted the Github of the author of the specified project.

but why do you insist on complicating things? just flash it as a bios replacement

rasz_pl wrote on 2026-06-21, 19:54:

but why do you insist on complicating things? just flash it as a bios replacement

This test is written for PC XT and clones. It is flashed into a single EPROM chip.
It is not possible to simply replace the two chips that together make up the BIOS of this motherboard with it.
Landmark has separate versions of its diagnostic ROM for 5162, 5170 and AT clones.

here some tools to split rom into two interleaved halves
A simple BIOS ROM-dump interlace splitter, deinterlace merger and slicer
Re: Suntac 80286 Mainboards
Re: AMI Color BIOS (1993 and earlier) modification in hex editor

dont worry about this being for XT and just flash and run it , instead of figuring out which eprom is odd and which even just try eproms in both combinations
AT is just a fatter XT with different CPU, wont matter
make sure to use CGA video, it uses it for ram
use xtramtest.8k from https://github.com/ki3v/xtramtest/releases first determine size of your rom chips, then copy xtramtest.8k into itself how many times required to obtain combined size of both chips
for example if 32KB do "copy /b xtramtest.8k + xtramtest.8k + xtramtest.8k + xtramtest.8k xtramtest.bin"
if 64KB "copy /b xtramtest.8k + xtramtest.8k + xtramtest.8k + xtramtest.8k + xtramtest.8k + xtramtest.8k + xtramtest.8k + xtramtest.8k xtramtest.bin"
etc
then split, flash, try both combinations

this is the best ram testing program available for 16bit platforms, will definitively tell you if you have ram problems

rasz_pl wrote on 2026-06-22, 06:42:

here some tools to split rom into two interleaved halves
...............
this is the best ram testing program available for 16bit platforms, will definitively tell you if you have ram problems

Thank you very much for the advice. I will try it soon.
I have already decided to unsolder all the DRAM chips and install sockets instead.
And check the chips themselves with a DRAM tester and also in another motherboard or in another video card.
For now, I will refrain from such radical actions)))

Here's a little more about my suffering with the Sanyo board:

Sanyo MBC-17 has a built-in BIOS mechanism for setting system parameters, very similar to the BIOS Setup from AMI or MR that we are used to (see photo). Only when selecting the type of hard drive, the geometry corresponding to this type is not displayed on the screen. Only the inscription "Type23" (for example) - and that's it ....
As I already wrote, I have a strong suspicion that the table of hard drive parameters in Sanyo MBC-17 BIOS is somewhat different from the usual ones used in AMI BIOS, Award BIOS, etc. Because for the last 35 years I have always used ST351A/X hard drives as "Type17", with parameters 980/5/17. And there have always been no problems. With this board, I also specified "Type17", the result was only freezing at the stage of the inscription "Starting MS-DOS". The same situation with other old IDE drives. I also specified them in BIOS Setup with the types that are "native" to them, but the results were the same. From which I concluded (assumption) that the geometry of hard disk types in BIOS Setup Sanyo MBC-17 is different from the geometry of hard disk types in BIOSes of other manufacturers that we are familiar with.
But, in addition to setting the hard disk parameters in BIOS Setup, Sanyo MBC-17 has two specific programs HDTYPE.COM and HDFORMAT.COM, which must be run from the installation diskette. As far as I understood from fragmentary messages on various forums, it is HDTYPE.COM that finally sets the type of hard disk (?), and HDFORMAT.COM formats it according to the set type. I made this conclusion because when I try to run HDTYPE.COM with another motherboard, a menu similar to the BIOS Setup Sanyo MBC-17 menu is displayed on the screen. But, of course, nothing happens then (.
So, I have only one way left: to make this board work with XT-IDE. Because all other controllers (16-bit IDE, 16-bit MFM, 16-bit RLL) refer to BIOS Setup for data about the hard disk geometry. And of course, I still need to try to connect a SCSI controller and some SCSI device. After all, they also do not "use the services" of BIOS Setup to determine the HDD geometry! – If some progress is made – then there will be many options for what to do.
By the way, thank you very much for the advice on XUB addressing and possible conflicts with EMS!! I forgot that with HGC I can use the addresses Сх00h at my discretion!! – I am already so used to the fact that I need to avoid conflicts with VGA, and I automatically configure XT-IDE to Dx00h!
P.S.
I have observed similar problems with loading from FDD more than once. The most striking examples are ATX-Turboxt-v3 by Rodney Knaap or 386SBMC by Alexandru Groza. Usually, they were caused by incorrect timings of the microcircuits that serve DMA. But I have never encountered anything like this on serial boards, especially from such a reputable manufacturer as Sanyo. It is one thing when this happens in home-made projects. Another thing is in serial computers. Strange….

I understood why my XT-IDE card was not initialized before. The fact was that in the BIOS Setup Sanyo MBC-17 there is no option “Floppy Disk Drive – None” AT ALL. To be more precise, it is provided only for Drive B: Some kind of floppy drive, it MUST be connected as Drive A: If the “Floppy drive seek at boot” procedure is not successful, all further actions (including initialization of BIOS extensions) do not occur.
I connected Gotek instead of a floppy drive and CF-Lite card by Kiselev, and XT-IDE immediately started to start, the CF card was detected.
XT-IDE BIOS extension successfully starts in the address range from 0xD000h to 0xDC00h.
However, it did not bring me any results. Same situation: everything freezes on the notification "Starting MS-DOS".

FDPT was historically somewhat troublesome https://www.os2museum.com/wp/the-fixed-disk-parameter-table/
BUT since you have XTIDE bios executing all that doesnt matter as XTIDE takes over.
You could try booting freedos, worked for Re: Support for DIY BIOS for AMD Elan SC300 386SX SoC

rasz_pl wrote on 2026-06-22, 20:03:

FDPT was historically somewhat troublesome https://www.os2museum.com/wp/the-fixed-disk-parameter-table/
BUT since you have XTIDE bios executing all that doesnt matter as XTIDE takes over.
You could try booting freedos, worked for Re: Support for DIY BIOS for AMD Elan SC300 386SX SoC

I tried FreeDOS as well. Same result.

for laughts you can try booting this game https://github.com/io12/bootmine 😀 it fits in single sector, and at least that much loads fine since you are getting "starting msdos" 😀
but I would just flash that ram tester

rasz_pl wrote on 2026-06-23, 09:20:

for laughts you can try booting this game https://github.com/io12/bootmine 😀 it fits in single sector, and at least that much loads fine since you are getting "starting msdos" 😀
but I would just flash that ram tester

Good morning!
I listened to your advice and tested the Sanyo-20IB board with various test ROMs.
Here are the results:
1) SuperSoft/Landmark Diagnostic does not find any critical errors at all (see photo 1)
2) Ruud's Diagnostic ROM stops at errors (?) of the 8253 timer (see photo 2)
3) xtramtest by David Giller and Adrian Black gives the following result - see photo 3
Unfortunately, I did not find a detailed description of xtramtest, so I still do not understand the reasons for the errors it found.
Still, I am not sure that "AT is just a fatter XT with different CPU" - maybe AT (especially with such a strange chipset) should be tested a little differently?