Current Combatiblity List:

Mouse:
Penixx PERMICE-201PB (Optical) - works!
Logitech M-859 (mechanic) - works partial, doubleclick slow
Microsoft PS/2 2 Key Mouse (mechanic) - works! Or not, depends on Version?
IBM PS/2 Mouse 3 Key Mouse (Mechanisch) - works!
IBM PS/2 Mouse 2 Key Mouse - non-functional
MIcrosoft Basic Optical Mouse v2.0 (USB/PS/2) - works!
Saitek PM28A - works, sometimes little slow downs
Mini Trackball Keyboard SPR-8695T (Sejin Electronic Inc LTD)- works!
Boeder 67414 - works!

KVM:
Important: The Adapter needs powered "before" the KVM Switch is Enabled.

KeySonic IC712I - non-functional
Rohline CS-14 "MASTER VIEW" - non-functional
INDY CPU-Switch Junior Model: 4-Port. - works!
Athen CS-62A - Mouse found, no movement.

matze79 - it is nice to see someone actually has time to invest in this. Your efforts are most appreciated. I have a few questions for you.

1) How is the tracking quality of the conversion? A handfull of companies have made these PS/2-to-serial protocal converters over the years, namely Raritan, Vetra Systems, and a few KVM companies. The problem I found with their conversion was the loss of tracking quality, which translated into lagged hand-to-pointer movement, particularly with acceleration, and somewhat jumpy pointer motion. This problem was amplified with increasing screen resolution.

2) Could you do an analysis of tracking quality when using the native PS/2 mouse port and compare that when using your protocal converter on the serial port? Is there any human noticable loss of tracking quality? What about at resolutions of 1280x1024 or higher?

3) As another forum member pointed out, we did have a go at this project from another member in this thread, PS/2 to Serial Mouse protocol converter . This was programmed in assembly using a Microchip PIC microcontroller and would presumably have minimal loss of tracking quality. Unfortunately, only the author was able to get it working. I tried for weeks to get his schematic functional on my breadboard, but failed. I suspect there was an issue with importing the code to my PIC devices, but I could not find any evidence of this. Would you be able to provide the programmed microcontroller chips to reduce this uncertainty? Or will there be a place to purchase the whole kit, whereby the user need only solder the components onto the PCB?

4) Do you have a Logitech MouseMan M-S38 to test with your protocal converter? Although it is a trackball PS/2 mouse, it is the mouse I have been addicted to since '98 and is still the only mouse I use.

5) What have you set your sampling rate at which efforts have you made to optimise this?

As a result of my frusturation from all other PS/2-to-serial protocal converters, I decided to modify several of my motherboards to accept native PS/2 mouse control by modifying the onboad keyboard controller circuit. Reference this thread, Native PS/2 mouse implementation for 386/486 boards using the keyboard controller However, I have only been able to successfully perform this modification on AWARD-based motherboards containing a DIP KBC. Modifying the AWARD BIOS to allow for PS/2 mouse KBC support is rather straight-forward, but not with AMI BIOSes. As such, if your converter contains no noticeable tracking quality at 1280x1024 or 1024x768, I would be very interested in reproducing your circuit for use on my AMI-based boards.

1.) & 2.)
You can not expect the same Tracking quality you will get from a PS/2 Mouse.
There is always loose.. maybe in Future everything will run with the logitech driver at 2400 baud.
Currently only Basic MS Mouse is supported with 1200 Bauds.
It currently depends hardly on the used Mouse. Some Work very well, even at high resolutions, others are choppy.

Currently we run on the Software from Simon Augustin which is avaible under Public Domain.
Which works stable.

i started made my own Devices Firmware since beginning of the project and got some improvments but there is still much work todo.
i`m currently limited in time due to my job and need to set prioritys.

3.)
Currently we have a group buy in the german Community but vogons is almost international and it seems there is much more demand.
You can order PCBs on http://dirtypcbs.com/view.php?share=7678&accesskey
In small amounts i can offer programmed MCUs.
But programming them is really easy, you just need a Parallelport, a Socket and few Resistors.
The hexfiles of simons stable version could be found here:
http://www.tempect.de/senil/progra
http://www.tempect.de/senil/progra/ps2to232.zip

The fusebits for the AVR (avrdude) are: (Setting 8Mhz internal Oscilator, Clockdiv/8: off)
-U lfuse:w:0xe4:m -U hfuse:w:0xd1:m -U efuse:w:0xfe:m

Maybe later i'm be able to provide some complete kits..

4.)
No-

5.) see Point 1 and 2.
i need first to get compliant with more PS/2 Devices.

Do your PS/2 Mod work with older Award Based Systems ? (386)

I suggest using at least ceramic resonator or even a crystal as the clock source. Internal RC oscillator tolerance is so large that some devices may fail to communicate through UART without calibration. If you build this and it does not work, this may be your issue.

Also the software could use some bugfixes. A device is never allowed to push data or clock wires high as they use open collector interface. Well, the ps2_gohi subroutine clearly sets pin high before it sets it to input, so there is a few hundred nanoseconds the pin is a high output. This is especially bad if the other side is pulling low and this device is pulling high, large currents may break something.

Thanks for your Suggestions.
i'm currently a bit busy, in few months i have more spare time.
i'm learning much stuff from scratch and this slows down everything a bit 😀

little fix is made.

1void ps2_gohi(int pin) //gohi and golo were tested working fine.
2{
3	pinMode(pin, INPUT);
4	digitalWrite(pin, HIGH);
5}

I think I will wait for the bugs to get worked out and for increased device support before I invest time into this protocal converter.

matze79 wrote:

Do your PS/2 Mod work with older Award Based Systems ? (386)

I have not yet tried the mod on AWARD BIOSes prior to 4.50, however I do have a 386 board which uses BIOS version 4.20. AWARD 4.20 BIOSes still work with MODBIN and the PS/2 option to select/deselect is available. The motherboard in question also has an external KBC which can be replaced with a KBC supporting PS/2 mouse integration. As such, I suspect the modification to the similar 386 motherboards to be fairly seemless. Unfortunately, my free time right now is in short supply.

got my kits today:

arrived really fast from matze79 😁

I attach a partlist export from eagle (mainly for the resistor values) just in case anyone else needs it too:

1Part     Value        
2
3C1       10uf           
4C2       10uf           
5C3       10uf           
6C4       10uf           
7C5       100nf         
8C6       10uf           
9C7       100nf            
10C8       100nf            
11COM      RS232      
12D1       1N4004      
13IC1      MAX232     
14IC2      7805TV       
15IC3      AVR Tiny2313
16JP1      RS232  header       
17JP2      VCC  header          
18JP3      PS/2  header         
19LED_G    LED3MM     
20LED_R    LED3MM     
21R2       1k             
22R3       1K             
23R4       10K           
24X3       Mouse MINI_DIN_6  

Do you get the microcontrollers pre-programmed? I am eagerly awaiting your results. Do you, by chance, have a Logitech 3-button MouseMan M-S38? It was a very popular trackball mouse in from 1998-2000.

^ yep, matze79 sent it programmed and ready to be used
I don't have any trackballs at all , just a bunch of Logitech SBF96 mice , will report when I test them

keropi wrote:

got my kits today

You got a whole kit? Nice!

I was going to order just the PCB once somebody tested it, but I want a kit if they're available!

I've built both kits but I am unable to test until I make a 386/486 test build...
Both my "ready to go" machines are based on the ss7 hot-591P mobo and both just don't want to coop with serial stuff. I find it strange since not even a real serial mouse works.Not a big deal since these mobos do have PS2 connectors but still... Maybe they need special serial-port brackets or something, I can't think of anything else

On the bright side the adapter does pickup the ps2 mouse and the activity led works when I move it or press buttons.

Don't you have a more modern ATX system with the Serial connector on the rear panel? It should be enough just for testing.

https://www.youtube.com/watch?v=ZvfXvMzH19w

A Video of the Version above.

Showing Paint in 800x600 on Windows 3.1.

Sorry for Quality, i try to make it better 😀

Allright finally tested the adapter , let this be a warning to all: there are TWO pinouts for mobo serial port brackets: using a recent bracket with an old mobo does not work.
As expected the adapter is perfect under DOS and games. It just works and response in low-res stuff is excellent, like a real serial mouse.
I did test in win98SE at 1024x768 and 1280x1024 and I had to raise mouse speed to get in par with a real PS2 mouse. There was a little lag as expected but nothing terrible. If I remember correctly this adapter behaves better than the Raritan adapter which is a commercial solution.

I could not get my "IntelliMouse Optical 1.1A" mouse working, buttons work but there is no movement. The mouse also does not light up so I guess that something confuses it. It is a USB/PS2 one with 5 buttons so that might have something to do with it.

My logitech SBF96 mouse is working perfect. Didn't bother to test mechanical PS2 mice, I see no point in it.

The adapter is excellent in games and stuff up to 800x600 as far as I can tell. Bigger resolutions work but needs a little compromise. IMHO one that wants to use high-res stuff needs to upgrade and get a PS2/USB capable mobo , serial and ball mice just don't cut it. This adapter is perfect for 286/386/486 DOS and win3x gamers , it offers a huge improvement if you replace the old ball mouse with an optical one.

Kudos to matze79 for creating this affordable solution! 😊

The high resolution tracking quality was my primary concern. I run my 486's at 1024x768 in Windows. 1024x768 is the native resolution of my 15" LCD.

matze76, can the tracking quality at 1024x768 or 1280x1024 be improved?

keropi, you said you had to raise the mouse speed at higher resolutions. After doing this, would you say that there was any noticable difference between the serial-ps/2 converted mouse and using the same mouse in a PS/2 port? Could you describe the issues in the tracking quality at higher resolutions? Is it simply and initial start of movement lag, or a continuously spotty lag, like a ghost shadow?

You can't have Perfect Tracking on a Serielle Mouse in High Resolution at 1200 Baud.
Even 2400 Baud Logitech isnt that fast, i have a Serial Logitech Mouse and 1280x1024 is Painful.
Who the Hell uses 1280x1024 on a Computer that doesnt support PS/2 ? 😀

The MS-Mouse was designed in 1983. There was simply no need for better Tracking.

With the 8541/A Graphics Card the need for better Mouse came in form of PS/2 with the PS/2 Computer Series in 1987 😉

I have a Logitech Serial MouseMan, which is the serial version of the same PS/2 MouseMan. Tracking at 1280x1024 seems really good to me.
Can the adapter work at 2400 baud?

@feipoa

It's just 1/3 of a second or something before tracking starts, movement is fluid - ie the cursor doesn't jump.

I tested again with the native ps2 port and the same optical mouse, I had to raise again mouse speed so please keep that in mind (seems I got used to modern mice and their super dpi / speed after all...). To achieve the same speed I have the matze79 adapter to full speed and the ps2 mouse one click before full speed , using 1024x768 as well.

The adapter seems really usable to me , can't say how it will seem to you though since I could also live with the raritan adapter after some getting used to 🤣

maybe my Logitech Mouse is somewhat old 😀 could be that something doesnt behave correctly.

Theoretical the Software could be rewritten to use the Logitech Drivers.
This will give nearly twice the Speed.

matze79 wrote:

maybe my Logitech Mouse is somewhat old 😀 could be that something doesnt behave correctly.

Theoretical the Software could be rewritten to use the Logitech Drivers.
This will give nearly twice the Speed.

That would certainly be a beneficial option if you choose to embark on this. Users could flash the firmware on their own depending on which mouse they have.

keropi, sounds somewhat promising, however the Raritan was completely unuseable to me. Either my particular mouse is more sensitive, or you have a lot of patience.