thanks to the bump I finally ordered some MAX232CPE to build this adapter 😁
once they arrive I'll report back with either success or burned components 🤣 🤣 🤣

ufff, very very old post, but necessary to update it, to continue improve vintage computers!

I see this post and the principal message has lost many links/images... for example i don't know how to use the 4 jumpers (4dip switches)
ok good, i have a new idea for this...
I reborn a AMSTRAD PC1512, all hardware works correctly, but in the original mouse the ball is deteriorated. no works correctly in movement
i see this post and I think, oh i see more schematic convertors for Amiga like this (port for mouse), for example but i mistake to work with this PC
well, i have a grand idea to improve this to the perfection
1) autopowered by the COM port, yes we can (old mouse do this), try to reduce the consume by use a LM2931 instead a LM7805 or better: a pair of transistors to do a regulated and effective 5v power
2) yes you can replace MAX232 by transistors too for level conversions, result is low consumption too!
this 2 mods can do the board a little larger, but the result is interesting, do the unique IC necessary is the PIC
I attach in this message 2 pdfs this the picture schematic and references oh this 2 improvements...

And if anyone has a idea of a converter for my Amstrad mouse port, (to no use the COM) , is coming grateful

Modma. Welcome to the forum. Did you try the circuit presented by sklawz? I was unable to get it to function. I wonder if it was related to the specific mouse he and I were using.

MAX232 is cheap. If I recall correctly, current output from RS-232 is less than 10 mA, so a power source will be needed to drive the PIC, mouse, etc. There was another chap on the forum who designed another circuit + firmware and was going to sell the prototype boards. I wonder what ever happened to that? Anyone remember the link to bump?

i also build this circuit, but had too issues with kvm switches 😒

the adapter i build before works mostly fine, except for kvm stuff too.

PS/2 to Microsoft Serialmouse Adapter Converter / Updated First Post / Firmware Update added

You know, I had no idea this particular project existed until I was trolling around Chipkin.ru looking for cool IBM PC related PCBs and found a reference to this schematic.

I've been playing around with this for a couple of weeks now, and apart from the comparatively low update rate of 20 packets / second vs the regular serial mouse's update of 40 packets a second, this adapter works really nicely. I put it together on the breadboard and it works well, nice that it doesn't need the oscillator parts. Programming a PIC16F628A with the TL866 is fairly trivial:

The purple usb thing is just the cheapest micro-usb power adapter I could find, they work great for small projects since the power output is small enough that they can't damage much.

I put it together on a protoboard and made a nice 3d printed case for it with some fancy dip switches to set the mouse options:

The latency of it compared to the arduino adapters I've tried is significantly better, on a vga res screen, the updates feel instantaneous. On a 2560x1440 screen, the mouse movement feels more like its moving over a network, like Teamviewer-ish. In Windows 3.1, with algorithm 2, the DPI halved and movement rate doubled, the mouse movement with my intellimouse optical is excellent, there is no perceivable latency, in my opinion.

I think I've also determined why Feipoa and others could not get the adapter working. It's very picky about the mice that are connected because it starts initialising the PS/2 mouse too soon, only about 1/2 of my mice actually work with it. I found I only got the MS Intellimouse optical (my favourite old mouse) to work accidentally once. All other times, the adapter would appear to initialise, blinking its bi-colour LED between orange & green then the cursor would do nothing at all. It could work with the Genius Netscroll+ Eye and some other old mice but not the ones I wanted to use.

The solution I've found is to power up the microcontroller shortly after giving power to the adapter with a separate switch for its power, so the mouse can power/start up fully first. This isn't great and would be fixed better with code, but I don't yet know how to put a delay in before the main loop / recompile the code.
What set me off in this direction was this post from the other person on here who knows how to make PS/2 mice work well: Re: PS/2 to Microsoft Serialmouse Adapter Converter / Updated First Post / Firmware Update added
If anyone else would like to try it, here's my schematic in tinycad, which I used for pointers when putting the PCB together:

The LEDs on pins 1&2 are very handy for telling what's going on. I thought I didn't have any bi-colour LEDs, but it turned out I had one from a router I'd dismantled for parts years ago.

The jumpers are:
J1: Mouse algorithm 1/2 (newer mice should use algorithm 2
J2: Mouse movement rate, on = 2x movement rate
J3: Mouse DPI 1, should 1/2 movement rate
J4: Mouse DPI 2, should 1/4 movement rate

I'm fond of this adapter, because it works with devices that don't have ISA ports spare (laptops & SBCs) and there's no driver setup required, this adapter detects and works just fine in the Windows 3.11 setup program with no extra mouse driver loaded 😀

I tested some more mice with this adapter, the Microsoft Basic Optical mouse works really well with it, it detects without any hassle and the motion is really good.

I'll have to see how it behaves with a KVM at some point, although my belkin omniview SE already does PS/2 to serial conversion, it'd be a useful test regardless 😀
During my tests of different mice, I found that with the mice I had available, only a few were PS/2 compatible. Maybe I had a bad sample set, but perhaps PS/2 isn't the best protocol to convert to in 2020. I think USB microcontrollers that can do USB host have come down to the pricepoint & availability that it's probably best to do that, and I've started to look into it. But there's a lot of work involved with all of that, that I have yet to put time into.