The Bull Micral 500 page is now up: https://www.ardent-tool.com/Bull/Micral_500.html

Together with the WD1007V-MC1 ESDI Adapter: https://www.ardent-tool.com/storage/WD1007V.html

Thanks to Stefan Lemanski for all the source material.

More info to come...

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Now that is detailed, even about the fan. Applause!

I adore the battery pack named "SAFT" (German for "juice").

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Tomas Slavotinek schrieb am Freitag, 7. Januar 2022 um 13:40:24 UTC+1:

The Bull Micral 500 page is now up: https://www.ardent-tool.com/Bull/Micral_500.html

Together with the WD1007V-MC1 ESDI Adapter: https://www.ardent-tool.com/storage/WD1007V.html

Thanks to Stefan Lemanski for all the source material.

More info to come...

Wow, great work, Tom. Thank you!
Better photos and hopefully, the refdisk, are coming.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

And thank you for all the info.

I was wondering about the planar SIMMs... are these industry standard?
And what capacity?

On 09.01.2022 19:55, schimmi wrote:

Tomas Slavotinek schrieb am Freitag, 7. Januar 2022 um 13:40:24 UTC+1:

The Bull Micral 500 page is now up:
https://www.ardent-tool.com/Bull/Micral_500.html

Together with the WD1007V-MC1 ESDI Adapter:
https://www.ardent-tool.com/storage/WD1007V.html

Thanks to Stefan Lemanski for all the source material.

More info to come...

Wow, great work, Tom. Thank you!
Better photos and hopefully, the refdisk, are coming.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

They are intel branded, never seen these before. 1MiB per stick with parity. I'll send you front+back images.

Tomas Slavotinek schrieb am Montag, 10. Januar 2022 um 22:02:33 UTC+1:

And thank you for all the info.

I was wondering about the planar SIMMs... are these industry standard?
And what capacity?
On 09.01.2022 19:55, schimmi wrote:

Tomas Slavotinek schrieb am Freitag, 7. Januar 2022 um 13:40:24 UTC+1:

The Bull Micral 500 page is now up:
https://www.ardent-tool.com/Bull/Micral_500.html

Together with the WD1007V-MC1 ESDI Adapter:
https://www.ardent-tool.com/storage/WD1007V.html

Thanks to Stefan Lemanski for all the source material.

More info to come...

Wow, great work, Tom. Thank you!
Better photos and hopefully, the refdisk, are coming.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Seufz,
i tried what i could finding a refdisk. Asked the Bull Society in BE/FR/DE (btw. very cool people, but had no idea what i'm talking about :) ), searched archive.org, crawled Zenith (was ab susidary of bull back then). I think the ref-disk is lost. Last
resort is system.cfg -> http://forum.system-cfg.com/viewtopic.php?f=3&t=13369 Maybe, maybe. I don't know :) :(
Regards,
Stefan

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Hi Stefan,

I gotta say, you don't give up easily. :)

Unfortunately, the system-cfg people don't know much about the 500
either from the look of things.

I was contacted by a fella that also has some Bull machines, but you
guessed it, no diskettes... The machines are somewhere in a warehouse,
so it may take a while until he gets to them, but I'll let you know if
anything comes out of it. Maybe there is a 500 with a functional hard
drive and a copy of the system programs in there...

That aside, I've checked the Headland GCK181 datasheet, and the chipset
is designed to be register-compatible with Models 50/60 and 80 (T1/T2) -
for the most part anyway. So that's a good thing... Further, by checking
the POST/BIOS code, it seems that the system reuses one of IBM's planar
IDs. I have found the following entries in the routine returning the
number of MCA slots:

0FBFFh - PS/2 Models 50/50Z/55SX (4 slots)
0F7FFh - PS/2 Model 60 (8 slots)
0FEFFh - PS/2 Model 80 386 16 MHz (8 slots)
0FDFFh - PS/2 Model 80 386 20 MHz (8 slots)
0DFFFh - PS/2 Model 70 386 16 MHz (4 slots)
0F9FFh - PS/2 Model 70 386 25 MHz (4 slots)

The Micral 500 has *five* physical slots, so my guess is that it
pretends to be a Model 80 with 8 slots - probably planar ID 0FDFFh (20
MHz) or perhaps 0FEFFh (16 MHz). The remaining 3 slot positions are
simply unused and will always return adapter ID FFFFh - "empty". You
should be able to verify the planar ID by booting to a refstamped disk
and polling the planar POS regs. 100h and 101h (or just run QCONFIG,
SIT, or something similar). The model/submodel info also matches that of
the corresponding IBM machines (F8/00h and F8/01h for the two Models 80).

The part I'm worried about is the CMOS and NVRAM layout. I can some
familiar CMOS/NVRAM offsets in the POST code, but I don't have the
entire thing reversed, and I expect at least some differences compared
to IBM's implementation.

So, I don't expect Model 80 refdisk to work correctly out of the box,
but it may be close... close enough to be patchable. (Unmodified, it
should at least get past the planar/model/submodel check if I'm right.)

On 22.06.2022 23:52, schimmi wrote:

Seufz,
i tried what i could finding a refdisk. Asked the Bull Society in BE/FR/DE (btw. very cool people, but had no idea what i'm talking about :) ), searched archive.org, crawled Zenith (was ab susidary of bull back then). I think the ref-disk is lost. Last

resort is system.cfg -> http://forum.system-cfg.com/viewtopic.php?f=3&t=13369

Maybe, maybe. I don't know :) :(
Regards,
Stefan

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Hey there,

interesting, thought it was more likely a model 70 clone. Thank you, I'll give it try - at least we can figure out the planar ID :D

Regards,
Stefan

Tomas Slavotinek schrieb am Montag, 27. Juni 2022 um 17:49:27 UTC+2:

Hi Stefan,

I gotta say, you don't give up easily. :)

Unfortunately, the system-cfg people don't know much about the 500
either from the look of things.

I was contacted by a fella that also has some Bull machines, but you
guessed it, no diskettes... The machines are somewhere in a warehouse,
so it may take a while until he gets to them, but I'll let you know if anything comes out of it. Maybe there is a 500 with a functional hard
drive and a copy of the system programs in there...

That aside, I've checked the Headland GCK181 datasheet, and the chipset
is designed to be register-compatible with Models 50/60 and 80 (T1/T2) -
for the most part anyway. So that's a good thing... Further, by checking
the POST/BIOS code, it seems that the system reuses one of IBM's planar
IDs. I have found the following entries in the routine returning the
number of MCA slots:

0FBFFh - PS/2 Models 50/50Z/55SX (4 slots)
0F7FFh - PS/2 Model 60 (8 slots)
0FEFFh - PS/2 Model 80 386 16 MHz (8 slots)
0FDFFh - PS/2 Model 80 386 20 MHz (8 slots)
0DFFFh - PS/2 Model 70 386 16 MHz (4 slots)
0F9FFh - PS/2 Model 70 386 25 MHz (4 slots)

The Micral 500 has *five* physical slots, so my guess is that it
pretends to be a Model 80 with 8 slots - probably planar ID 0FDFFh (20
MHz) or perhaps 0FEFFh (16 MHz). The remaining 3 slot positions are
simply unused and will always return adapter ID FFFFh - "empty". You
should be able to verify the planar ID by booting to a refstamped disk
and polling the planar POS regs. 100h and 101h (or just run QCONFIG,
SIT, or something similar). The model/submodel info also matches that of
the corresponding IBM machines (F8/00h and F8/01h for the two Models 80).

The part I'm worried about is the CMOS and NVRAM layout. I can some
familiar CMOS/NVRAM offsets in the POST code, but I don't have the
entire thing reversed, and I expect at least some differences compared
to IBM's implementation.

So, I don't expect Model 80 refdisk to work correctly out of the box,
but it may be close... close enough to be patchable. (Unmodified, it
should at least get past the planar/model/submodel check if I'm right.)
On 22.06.2022 23:52, schimmi wrote:

Seufz,
i tried what i could finding a refdisk. Asked the Bull Society in BE/FR/DE (btw. very cool people, but had no idea what i'm talking about :) ), searched archive.org, crawled Zenith (was ab susidary of bull back then). I think the ref-disk is lost.

Last resort is system.cfg -> http://forum.system-cfg.com/viewtopic.php?f=3&t=13369

Maybe, maybe. I don't know :) :(
Regards,
Stefan

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 30.06.2022 22:37, schimmi wrote:

Hey there,

interesting, thought it was more likely a model 70 clone. Thank you, I'll give it try - at least we can figure out the planar ID :D

Regards,
Stefan

Yeah, when it comes to the form factor, it's very similar to the Model 70.

Models 70 and 80 share some of the core logic (depending on the planar
type) and use the same reference disk. So from a hardware and firmware perspective, it's not a big difference whether it identifies as 70 or 80...

IBM liked to package the same tech differently to target different
market segments. There are multiple "sister" systems in the PS/2 line.
One packaged as a desktop and the other as a tower:

Models 50 and 60
Models 55 SX and 65 SX
Models 70 and 80 (and their little sister - the P70)
Models 90 and 95 (and later the "big bro" Server 500)

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Hey,

you were absolutely right - the micral boots with the 70/80 refdisk and its original BIOS. Seems even the board ID was copy fringed in any way :P However, the first 640k were not detected (0K), the rest of the 4MB are shown. The ESDI Controller is
detected. Let me test a few things, maybe we get it to detect everything :O Thanks!

Regards,
Stefan

Tomas Slavotinek schrieb am Donnerstag, 30. Juni 2022 um 23:58:14 UTC+2:

On 30.06.2022 22:37, schimmi wrote:

Hey there,

interesting, thought it was more likely a model 70 clone. Thank you, I'll give it try - at least we can figure out the planar ID :D

Regards,
Stefan

Yeah, when it comes to the form factor, it's very similar to the Model 70.

Models 70 and 80 share some of the core logic (depending on the planar
type) and use the same reference disk. So from a hardware and firmware perspective, it's not a big difference whether it identifies as 70 or 80...

IBM liked to package the same tech differently to target different
market segments. There are multiple "sister" systems in the PS/2 line.
One packaged as a desktop and the other as a tower:

Models 50 and 60
Models 55 SX and 65 SX
Models 70 and 80 (and their little sister - the P70)
Models 90 and 95 (and later the "big bro" Server 500)

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Add: It boots the hard drive and you can add ADFs to select Adding-Cards. Wohoo!

schimmi schrieb am Dienstag, 12. Juli 2022 um 20:02:40 UTC+2:

Hey,

you were absolutely right - the micral boots with the 70/80 refdisk and its original BIOS. Seems even the board ID was copy fringed in any way :P However, the first 640k were not detected (0K), the rest of the 4MB are shown. The ESDI Controller is

detected. Let me test a few things, maybe we get it to detect everything :O Thanks!

Regards,
Stefan
Tomas Slavotinek schrieb am Donnerstag, 30. Juni 2022 um 23:58:14 UTC+2:

On 30.06.2022 22:37, schimmi wrote:

Hey there,

interesting, thought it was more likely a model 70 clone. Thank you, I'll give it try - at least we can figure out the planar ID :D

Regards,
Stefan

Yeah, when it comes to the form factor, it's very similar to the Model 70.

Models 70 and 80 share some of the core logic (depending on the planar type) and use the same reference disk. So from a hardware and firmware perspective, it's not a big difference whether it identifies as 70 or 80...

IBM liked to package the same tech differently to target different
market segments. There are multiple "sister" systems in the PS/2 line.
One packaged as a desktop and the other as a tower:

Models 50 and 60
Models 55 SX and 65 SX
Models 70 and 80 (and their little sister - the P70)
Models 90 and 95 (and later the "big bro" Server 500)

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

That's great news, Stefan!

Can the 70/80 refdisk actually configure the machine properly or do you
always end up with some POST errors? If there are no errors then the
CMOS/NVRAM map must be really close as well, down to the checksum ranges
and algorithms (assuming the POST code actually bothers checking these).

On 12.07.2022 20:09, schimmi wrote:

Add: It boots the hard drive and you can add ADFs to select Adding-Cards. Wohoo!

schimmi schrieb am Dienstag, 12. Juli 2022 um 20:02:40 UTC+2:

Hey,

you were absolutely right - the micral boots with the 70/80 refdisk and its original BIOS. Seems even the board ID was copy fringed in any way :P However, the first 640k were not detected (0K), the rest of the 4MB are shown. The ESDI Controller is

detected. Let me test a few things, maybe we get it to detect everything :O Thanks!

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Hey Tom, first off, many thanks :)

- the refdisk can configure MCA cards
- it can set time and date and a password
- it does not complain in any ways that it is not an IBM machine
- it has, however, problems with the base memory. it's reported as 0kb in the refdisk and with every boot you get a memory error (you can easily skip)

It's not 100%, but a significant step forward :)

Regards,
Stefan

Tomas Slavotinek schrieb am Dienstag, 12. Juli 2022 um 22:35:41 UTC+2:

That's great news, Stefan!

Can the 70/80 refdisk actually configure the machine properly or do you always end up with some POST errors? If there are no errors then the CMOS/NVRAM map must be really close as well, down to the checksum ranges
and algorithms (assuming the POST code actually bothers checking these).
On 12.07.2022 20:09, schimmi wrote:

Add: It boots the hard drive and you can add ADFs to select Adding-Cards. Wohoo!

schimmi schrieb am Dienstag, 12. Juli 2022 um 20:02:40 UTC+2:

Hey,

you were absolutely right - the micral boots with the 70/80 refdisk and its original BIOS. Seems even the board ID was copy fringed in any way :P However, the first 640k were not detected (0K), the rest of the 4MB are shown. The ESDI Controller is

detected. Let me test a few things, maybe we get it to detect everything :O Thanks!

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 12.07.2022 22:48, schimmi wrote:

Hey Tom, first off, many thanks :)

- the refdisk can configure MCA cards
- it can set time and date and a password
- it does not complain in any ways that it is not an IBM machine
- it has, however, problems with the base memory. it's reported as 0kb in the refdisk and with every boot you get a memory error (you can easily skip)

It's not 100%, but a significant step forward :)

Glad I could help. In return your experiments will help the other Micral
500 owners to get their rare machines up 'n running.

I'll update the page accordingly...

It's interesting that it has trouble only with the base memory - that's
the part that should be more or less consistent across all systems. Does
the skippable error appear during POST or refdisk load? And exactly what
error code/message are you getting? I'll check for a possible fix or
workaround when I have time.

Oh, and when you have a chance, please run the QCONFIG utility and send
me the output. It comes standard with later PC DOS releases or you can
download it here (it works from MS DOS too): https://ardent-tool.com/utils/QCONFIG_2000.zip

Run it with the /D and /O switches - that should give a more detailed
listing and save it to a file (QCONFIG.OUT).

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

The refdisk does not mention any error. it just says: 0kb :) only the bios complains a config error, which is easily skippable.

schimmi schrieb am Mittwoch, 13. Juli 2022 um 00:04:53 UTC+2:

There you go:
---snip---
Operating System: MS-DOS Version 6.20
Date & Time : 2022-12-07 23:59:16
ROM Copyright : Version B1 1a c Copyright Bu
Model ID : F8 Sub-Model ID : 00
BIOS Revision : 05 BIOS Date : 11/09/89
BIOS Type : IBM
Machine Type : IBM PS/2 Model 80
Features : Cascaded IRQ2 Controller
: Real Time Clock
: Extended BIOS Data Area
: MicroChannel
Processor : Intel 386 Processor Speed : 16 MHz
Estimated Speed : 20 MHz
CoProcessor : None
Bus Type : Micro Channel 32-Bit Bus
Keyboard Type : Enhanced
Equipment : 1 Parallel Port(s)
: 1 Serial Port(s)
: 1 Diskette Drive(s)
: 1 Fixed Disk(s)
Serial Port 1 : COM1: 03F8
Parallel Port 1 : LPT1: 03BC
Primary Video : VGA
Diskette Drive A: 3.50" - 1.44M - 80 Track - Type 4
Fixed Disk 1 : 156MB ( 159 981KB) ( 163 821 056 bytes)
: Cylinders 156 Heads/Cyl 64 Sectors/Head 32
Logical Drive C : Size 159 552KB ( 155,8MB) Avail 94 940KB ( 92,7MB)
: Total Units 39 888 Sectors/Unit 8 Bytes/Sector 512
: Avail Units 23 735 Total Sectors 319 104
: Local Drive - File System is FAT
: Volume Label is MS-DOS_6
Total Memory : 0KB (0,0MB)
Conventional : 640KB Free: 602KB
Extended Memory : 64 896KB Free: 0KB
XMS Memory : 3 008KB Free: 3 008KB
XMS Version : 3.0
Adapter ROM 1 : Addr DC000-DFFFF
Planar : ID FEFF - PS/2 Model 80 16 MHz
: POS Data = 9F F5 FF FF FF FF
: Connector 1 = empty Connector 2 = empty
Total Slots : 8 System (DISK): 1 User Slots: 7
Expansion Slot 1: * No Adapter Present
Expansion Slot 2: * No Adapter Present
Expansion Slot 3: * No Adapter Present
Expansion Slot 4: * No Adapter Present
Expansion Slot 5: ID DDFF - IBM ESDI Fixed Disk Controller
: Enabled - POS Data = 55 17 FF FF FF FF
: Int Level 14 IO Address 3518-351F
: ROM Address DC000-DFFFF Arbitration Level 5
Expansion Slot 6: * No Adapter Present
Expansion Slot 7: * No Adapter Present
Expansion Slot 8: * No Adapter Present
---snap--

:D
Tomas Slavotinek schrieb am Dienstag, 12. Juli 2022 um 23:31:45 UTC+2:

On 12.07.2022 22:48, schimmi wrote:

Hey Tom, first off, many thanks :)

- the refdisk can configure MCA cards
- it can set time and date and a password
- it does not complain in any ways that it is not an IBM machine
- it has, however, problems with the base memory. it's reported as 0kb in the refdisk and with every boot you get a memory error (you can easily skip)

It's not 100%, but a significant step forward :)

Glad I could help. In return your experiments will help the other Micral 500 owners to get their rare machines up 'n running.

I'll update the page accordingly...

It's interesting that it has trouble only with the base memory - that's
the part that should be more or less consistent across all systems. Does the skippable error appear during POST or refdisk load? And exactly what error code/message are you getting? I'll check for a possible fix or workaround when I have time.

Oh, and when you have a chance, please run the QCONFIG utility and send
me the output. It comes standard with later PC DOS releases or you can download it here (it works from MS DOS too): https://ardent-tool.com/utils/QCONFIG_2000.zip

Run it with the /D and /O switches - that should give a more detailed listing and save it to a file (QCONFIG.OUT).

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

There you go:
---snip---
Operating System: MS-DOS Version 6.20
Date & Time : 2022-12-07 23:59:16
ROM Copyright : Version B1 1a c Copyright Bu
Model ID : F8 Sub-Model ID : 00
BIOS Revision : 05 BIOS Date : 11/09/89
BIOS Type : IBM
Machine Type : IBM PS/2 Model 80
Features : Cascaded IRQ2 Controller
: Real Time Clock
: Extended BIOS Data Area
: MicroChannel
Processor : Intel 386 Processor Speed : 16 MHz
Estimated Speed : 20 MHz
CoProcessor : None
Bus Type : Micro Channel 32-Bit Bus
Keyboard Type : Enhanced
Equipment : 1 Parallel Port(s)
: 1 Serial Port(s)
: 1 Diskette Drive(s)
: 1 Fixed Disk(s)
Serial Port 1 : COM1: 03F8
Parallel Port 1 : LPT1: 03BC
Primary Video : VGA
Diskette Drive A: 3.50" - 1.44M - 80 Track - Type 4
Fixed Disk 1 : 156MB ( 159 981KB) ( 163 821 056 bytes)
: Cylinders 156 Heads/Cyl 64 Sectors/Head 32
Logical Drive C : Size 159 552KB ( 155,8MB) Avail 94 940KB ( 92,7MB)
: Total Units 39 888 Sectors/Unit 8 Bytes/Sector 512
: Avail Units 23 735 Total Sectors 319 104
: Local Drive - File System is FAT
: Volume Label is MS-DOS_6
Total Memory : 0KB (0,0MB)
Conventional : 640KB Free: 602KB
Extended Memory : 64 896KB Free: 0KB
XMS Memory : 3 008KB Free: 3 008KB
XMS Version : 3.0
Adapter ROM 1 : Addr DC000-DFFFF
Planar : ID FEFF - PS/2 Model 80 16 MHz
: POS Data = 9F F5 FF FF FF FF
: Connector 1 = empty Connector 2 = empty
Total Slots : 8 System (DISK): 1 User Slots: 7
Expansion Slot 1: * No Adapter Present
Expansion Slot 2: * No Adapter Present
Expansion Slot 3: * No Adapter Present
Expansion Slot 4: * No Adapter Present
Expansion Slot 5: ID DDFF - IBM ESDI Fixed Disk Controller
: Enabled - POS Data = 55 17 FF FF FF FF
: Int Level 14 IO Address 3518-351F
: ROM Address DC000-DFFFF Arbitration Level 5
Expansion Slot 6: * No Adapter Present
Expansion Slot 7: * No Adapter Present
Expansion Slot 8: * No Adapter Present
---snap--

:D

Tomas Slavotinek schrieb am Dienstag, 12. Juli 2022 um 23:31:45 UTC+2:

On 12.07.2022 22:48, schimmi wrote:

Hey Tom, first off, many thanks :)

- the refdisk can configure MCA cards
- it can set time and date and a password
- it does not complain in any ways that it is not an IBM machine
- it has, however, problems with the base memory. it's reported as 0kb in the refdisk and with every boot you get a memory error (you can easily skip)

It's not 100%, but a significant step forward :)

Glad I could help. In return your experiments will help the other Micral
500 owners to get their rare machines up 'n running.

I'll update the page accordingly...

It's interesting that it has trouble only with the base memory - that's
the part that should be more or less consistent across all systems. Does
the skippable error appear during POST or refdisk load? And exactly what error code/message are you getting? I'll check for a possible fix or workaround when I have time.

Oh, and when you have a chance, please run the QCONFIG utility and send
me the output. It comes standard with later PC DOS releases or you can download it here (it works from MS DOS too): https://ardent-tool.com/utils/QCONFIG_2000.zip

Run it with the /D and /O switches - that should give a more detailed
listing and save it to a file (QCONFIG.OUT).

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 13.07.2022 0:04, schimmi wrote:

There you go:

Aha, so it reuses planar ID of the 16 MHz Model 80 (FEFFh). Reply did
the same thing with their Model 32...

Thanks!

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 13.07.2022 0:10, schimmi wrote:

The refdisk does not mention any error. it just says: 0kb :) only the bios complains a config error, which is easily skippable.

Is it the "CONFIGURATION......" POST line? Does it give any additional info?

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

So, all in all, with it's G2 Chipset - it's a complete replica ragarding hardware and software (BIOS -> Model 80)?

No wonder it wasn't that common. Maybe copyright issues?

Tomas Slavotinek schrieb am Mittwoch, 13. Juli 2022 um 00:30:17 UTC+2:

On 13.07.2022 0:04, schimmi wrote:

There you go:

Aha, so it reuses planar ID of the 16 MHz Model 80 (FEFFh). Reply did
the same thing with their Model 32...

Thanks!

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Tomas Slavotinek schrieb am Mittwoch, 13. Juli 2022 um 00:44:43 UTC+2:

On 13.07.2022 0:10, schimmi wrote:

The refdisk does not mention any error. it just says: 0kb :) only the bios complains a config error, which is easily skippable.

Is it the "CONFIGURATION......" POST line? Does it give any additional info?

https://photos.app.goo.gl/uFFm3egfJBNqsVbXA

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 13.07.2022 0:47, schimmi wrote:

So, all in all, with it's G2 Chipset - it's a complete replica ragarding hardware and software (BIOS -> Model 80)?

No wonder it wasn't that common. Maybe copyright issues?

The chipset was specifically designed to be almost 100% register
compatible with the Model 50/60/80(T1&2), but there are some minor
differences. Looks like the Micral 500 deviates further by handling
stuff like the memory PD differently (of course it uses standard SIMMs
in place of the proprietary memory cards).

The POST/BIOS implementation looks nothing like IBM's but clearly they
closely copied even aspects that were not documented publicly. So either
they sourced the info from IBM, some 3rd party, or reverse engineered
the original implementation (clean room or otherwise). Can't say
anything about the system programs as we don't have these.

There might be some period articles on how far Bull went with the
licensing...

I'll do more digging later on.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 13.07.2022 0:58, schimmi wrote:

Tomas Slavotinek schrieb am Mittwoch, 13. Juli 2022 um 00:44:43 UTC+2:

On 13.07.2022 0:10, schimmi wrote:

The refdisk does not mention any error. it just says: 0kb :) only the bios complains a config error, which is easily skippable.

Is it the "CONFIGURATION......" POST line? Does it give any additional info?

https://photos.app.goo.gl/uFFm3egfJBNqsVbXA

"164 - Memory Configuration Error"
That's what I was after...

From what I remember the POST error reporting is rather complex
(structured) compared to IBM's straightforward implementation of the
era. Will have to revisit it and figure out what exactly is going on there.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Guess it has something to do with the Bull-BIOS implementation. There's more than the memory counter on regular PS/2s, and that's only the visible side. I don't know. It works, DOS shows all the memory and the refdisk doesn't complain. Now that we know
it's a replica - maybe swapping the roms for a model70/80 might help us, determinig if its a bios or hardware issue?

Tomas Slavotinek schrieb am Mittwoch, 13. Juli 2022 um 01:26:58 UTC+2:

On 13.07.2022 0:58, schimmi wrote:

Tomas Slavotinek schrieb am Mittwoch, 13. Juli 2022 um 00:44:43 UTC+2:

On 13.07.2022 0:10, schimmi wrote:

The refdisk does not mention any error. it just says: 0kb :) only the bios complains a config error, which is easily skippable.

Is it the "CONFIGURATION......" POST line? Does it give any additional info?

https://photos.app.goo.gl/uFFm3egfJBNqsVbXA

"164 - Memory Configuration Error"
That's what I was after...

From what I remember the POST error reporting is rather complex
(structured) compared to IBM's straightforward implementation of the
era. Will have to revisit it and figure out what exactly is going on there.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 13.07.2022 1:44, schimmi wrote:

Guess it has something to do with the Bull-BIOS implementation. There's more than the memory counter on regular PS/2s, and that's only the visible side. I don't know. It works, DOS shows all the memory and the refdisk doesn't complain. Now that we know

it's a replica - maybe swapping the roms for a model70/80 might help us, determinig if its a bios or hardware issue?

Yeah, it's most likely an incompatibility between the POST code and the
refdisk system programs - regarding how the memory configuration is
stored in the CMOS/NVRAM. Or between the refdisk and the memory hardware (though, it's strange that it correctly reports the rest of the memory).

Swapping the ROMs may be an interesting experiment, but the POST will
probably fail on one of IBM's strict check routines. Hard to tell,
though. The hardware/firmware from this era isn't THAT complex and
sensitive compared to the later stuff.

I'll definitely revisit the Bull firmware, I remember it being quite and unusual and interesting.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)