VOGONS


Reply 341 of 539, by maxtherabbit

User metadata
Rank l33t
Rank
l33t
Rio444 wrote:

Yes, I tested the mod with 286.
It work.

thank you for the response, does the version of PS2SUPPC.COM you posted in this thread also run on the 286?

Reply 343 of 539, by maxtherabbit

User metadata
Rank l33t
Rank
l33t

I drew up an interposer PCB in Kicad to make this a little cleaner. If you have the existing keyboard controller socketed, and NOT under the ISA slots, it should work great as a totally plug and play mod. Here are some renderings - I'm going to do a final review of the design then send it off to fab this week.

If this setup is too tall for your application, you could build it by soldering the keyboard controller IC directly to the PCB, using different pin headers, and trimming everything short after it's soldered. That would get the overall added height to around 3mm or less.

topraw.png
Filename
topraw.png
File size
39.98 KiB
Views
1666 views
File license
Fair use/fair dealing exception
top.png
Filename
top.png
File size
67.53 KiB
Views
1666 views
File license
Fair use/fair dealing exception
bottom.png
Filename
bottom.png
File size
58.13 KiB
Views
1666 views
File license
Fair use/fair dealing exception
stack.png
Filename
stack.png
File size
39.09 KiB
Views
1666 views
File license
Fair use/fair dealing exception

Credit to Rio444 for the schematic upon which this was based. Once I get it tested, it will go up on OSHPark if others are interested.

Reply 344 of 539, by maxtherabbit

User metadata
Rank l33t
Rank
l33t

the PCB works - here's the OSH park link if anyone would like one:

https://oshpark.com/shared_projects/1rUdxa5p

the pinout for the PS/2 header is the one on the right in this photo (because that matched the 3 dongles I had laying around)
I believe this is the Acer/Asus spec

1329676968.or.15360.jpg
Filename
1329676968.or.15360.jpg
File size
53.01 KiB
Views
1637 views
File license
Fair use/fair dealing exception

Thanks again to Rio444 for the schematic and PS2SUPPC software, this mod works perfectly on my 12MHz 286 (the keyb controller that came with the mobo didn't work, but I swapped it for a spare AMI 'F' chip and it works a treat

Reply 345 of 539, by maxtherabbit

User metadata
Rank l33t
Rank
l33t

Something interesting I noticed - not sure if it was swapping keyboard controllers or the mod itself, but my key lock doesn't lock the keyboard anymore on the 286

I know the lock worked before because I tested it, and now when you lock it nothing happens. Not complaining, since the lock is practically useless, but I thought it was interesting to note.

I went ahead and buzzed out the connection from the physical lock switch to pin 34 on the keyboard controller (Keyboard-Inhibit) and verified that the pin is properly being pulled low when the lock is locked, but the keyboard controller just is not acknowledging the inhibit pin as being asserted.

Could someone else who has an AMI MEGA-KB-F-WP being used in PS/2 mode test locking the keyboard please?

Reply 346 of 539, by keropi

User metadata
Rank l33t++
Rank
l33t++

I came to the conclusion that kbc are different or they need different passives/specs to support them.
I have boards that just won't boot with the VIA kbc - no mods installed just kbc replacement.
Try another one if you can and see what happens.

🎵 🎧 PCMIDI MPU , OrpheusII , Action Rewind , Megacard and 🎶GoldLib soundcard website

Reply 347 of 539, by maxtherabbit

User metadata
Rank l33t
Rank
l33t
keropi wrote:

I came to the conclusion that kbc are different or they need different passives/specs to support them.
I have boards that just won't boot with the VIA kbc - no mods installed just kbc replacement.
Try another one if you can and see what happens.

in this case I don't believe it is a passive or routing issue - the donor board that I pulled the MEGA-KB-F-WB from also has the pins for the keylock header wired one directly to ground and the other directly to pin 34 on the KBC socket

my guess is that changing to PS/2 mode makes the controller either disregard the keyboard-inhibit signal or, perhaps even more strangely, look for it to be asserted on a different pin

Reply 348 of 539, by maxtherabbit

User metadata
Rank l33t
Rank
l33t

so both of the AMI 'F' KBC I have (one marked MEGA-KB-F-WB and one marked KB-BIOS-VER-F) will respect the keyboard inhibit line when plugged directly into the motherboard socket, but when run through the interposer PCB they just ignore it

I verified that the interposer is passing pin 34 straight through without shorting it to anything else, so the only other remaining explanation is that when the KBC detects PS/2 keyboard mode, it decides to ignore keyboard-inhibit being asserted on pin 34 (or maybe requires a register bit set to support it)

odd, but hey I got an optical mouse on a 286 so who cares

Reply 349 of 539, by feipoa

User metadata
Rank l33t++
Rank
l33t++

I have noticed that some boards with PS/2 support have things like turbo switch input pins and or polarity set differently. There is also a KBC type setting in the BIOS, but I don't know much about this.

Maxtherabbit, do you have a BOM with part numbers? I might try to assemble one of your units. When I assembled Rio444's unit, I was unable to get it to function for some reason. I used desoldered SMD transistors, perhaps one was bad...

Plan your life wisely, you'll be dead before you know it.

Reply 350 of 539, by maxtherabbit

User metadata
Rank l33t
Rank
l33t
feipoa wrote:

I have noticed that some boards with PS/2 support have things like turbo switch input pins and or polarity set differently. There is also a KBC type setting in the BIOS, but I don't know much about this.

Maxtherabbit, do you have a BOM with part numbers? I might try to assemble one of your units. When I assembled Rio444's unit, I was unable to get it to function for some reason. I used desoldered SMD transistors, perhaps one was bad...

Q1-Q3 MMBT3904LT3G
R1,R3,R5,R6 ERA-6AED472V
R2,R4 ERA-6AED202V
C1,C2 CL21B221KBANNNC
KBC Socket DILB40P-223TLF
Pin Headers to MB 350-80-120-00-001101

All passives are 0805, any 4.7k/2.2k resistors and 220pf caps will work.
For the KBC Socket any DIP-40 socket will do, and all the other pin headers are standard 2.54mm Berg

Reply 351 of 539, by feipoa

User metadata
Rank l33t++
Rank
l33t++

Thanks.

How proficient are you with KiCAD? Would you be interested in drawing up this little project? Custom interposer module for TI486SXL2-66 PGA168 to PGA132 - HELP! Someone else started on it then vanished.

Plan your life wisely, you'll be dead before you know it.

Reply 352 of 539, by maxtherabbit

User metadata
Rank l33t
Rank
l33t
feipoa wrote:

Thanks.

How proficient are you with KiCAD? Would you be interested in drawing up this little project? Custom interposer module for TI486SXL2-66 PGA168 to PGA132 - HELP! Someone else started on it then vanished.

did you ever get the schematic file from him? routing out the board is the easy part

Reply 353 of 539, by feipoa

User metadata
Rank l33t++
Rank
l33t++

You'd probably have to read the thread. Its been so long and my life rather chaotic that I'm starting to forget things. That guy had the false impression that he could make a multi-purpose 486-to-386 interposer module in this manner, and I think he lost interest after realising this type of passive interposer will only work with the SXL series of CPUs. It is a fairly straight forward translation, with much of the questions all worked out in the thread. I sort of recall him having some incorrect routings with a few pins, but (memory?)... I also recall there being some aspect too complex for him - perhaps related to 3+ layers and vias (memory?).

Unfortunately, it is wishful thinking to think that I will ever be able to enjoy the time to learn KiCAD with three screeming and foul tempered children around me constantly.

In vintage computing and collecting, a common theme or goal is to completely max out a given platform. Hot items are the 386's and 486's. Without entirely turning a 386 into a full fledged 486, the ultimate [stable] upgrade is the TI486SXL2-66 run at 80 MHz (2x40). The only device produced in this form factor was by Evergreen and is exceedingly rare. Hence the interest in this custom interposer.

Plan your life wisely, you'll be dead before you know it.

Reply 354 of 539, by maxtherabbit

User metadata
Rank l33t
Rank
l33t

anyone have any experience with a NEC D80C42C250 keyboard controller? can't seem to find a datasheet for it

it came preinstalled in my Everex Tempo 386sx, I tried swapping it with an AMI Megakey and while the keyboard input works the BIOS gives KBC error and the machine is stuck in non-turbo mode
(the only way to toggle turbo on this machine is a keyboard shortcut ctrl+alt+minus or ctrl+alt+plus)

I assume the NEC KBC has some custom code in its ROM to handle the turbo toggle, question is can this KBC support PS/2 mode? (I kinda doubt it)

Reply 355 of 539, by Tiido

User metadata
Rank l33t
Rank
l33t

The turbo function will be controlled by one of the GPIO pins on the controller and you should be able to find it with a multimeter easily. Measure all pins and toggle turbo, eventually you'll find a pin that changes state with the turbo. Isolate it and wire it to traditional pushbutton, swap in a new KBC and that problem has a solution.

T-04YBSC, a new YMF71x based sound card & Official VOGONS thread about it
Newly made 4MB 60ns 30pin SIMMs ~
mida sa loed ? nagunii aru ei saa 😜

Reply 356 of 539, by maxtherabbit

User metadata
Rank l33t
Rank
l33t
Tiido wrote:

The turbo function will be controlled by one of the GPIO pins on the controller and you should be able to find it with a multimeter easily. Measure all pins and toggle turbo, eventually you'll find a pin that changes state with the turbo. Isolate it and wire it to traditional pushbutton, swap in a new KBC and that problem has a solution.

that may well let me force turbo on, but at least something else is an issue since the BIOS was reporting a KBC error

Reply 357 of 539, by feipoa

User metadata
Rank l33t++
Rank
l33t++

I thought I'd provide a brief update. If you recall, for AWARD 4.50 and newer BIOSes, we can use MODBIN to add BIOS support for PS/2 mice. But for AMI BIOSes, we don't have the MODBIN equivalent program AMIBCP, which is very version dependent.

Vogons user jakethompson1 figured out a way to disable the checksum on these AMI motherboards via a BIOS hack and to also enable PS/2 mouse support [in the BIOS]. As such, I no longer need to use Rio444's TSR, PS2SUPPC.COM, to enable PS/2 mouse support in DOS. I also no longer need the a NTDETECT.COM file for PS/2 mosue support in NT4 and NT3.51 on AMI boards which have a hacked on PS/2 mouse via the keyboard controller.

He is also able to unhide AMI BIOS features which includes more than what is available in AMISETUP.

Plan your life wisely, you'll be dead before you know it.

Reply 358 of 539, by maxtherabbit

User metadata
Rank l33t
Rank
l33t
feipoa wrote on 2020-08-31, 10:33:

I thought I'd provide a brief update. If you recall, for AWARD 4.50 and newer BIOSes, we can use MODBIN to add BIOS support for PS/2 mice. But for AMI BIOSes, we don't have the MODBIN equivalent program AMIBCP, which is very version dependent.

Vogons user jakethompson1 figured out a way to disable the checksum on these AMI motherboards via a BIOS hack and to also enable PS/2 mouse support [in the BIOS]. As such, I no longer need to use Rio444's TSR, PS2SUPPC.COM, to enable PS/2 mouse support in DOS. I also no longer need the a NTDETECT.COM file for PS/2 mosue support in NT4 and NT3.51 on AMI boards which have a hacked on PS/2 mouse via the keyboard controller.

He is also able to unhide AMI BIOS features which includes more than what is available in AMISETUP.

that sounds awesome, can we get the details on how to implement it ourselves please?

Reply 359 of 539, by feipoa

User metadata
Rank l33t++
Rank
l33t++

Best for user jakethompson1 to discuss that. I don't know how seamless it is to transfer the findings to other BIOSes. I think he had to review the BIOS in assembly code.

Plan your life wisely, you'll be dead before you know it.