Double Key Press with PS/2 Keyboard

[michaelkamprath]

I got my Minimal 64 built and it works, but the PS/2 keyboard I am using generates two key events with every key press: once on the down stroke, and once on the up stroke of the key. I don't know PS/2 keyboard protocol well enough to diagnose this. Is this a problem with the keyboard, the Minimal 64, MinOS?

[slu4coder]

Hi Michael,
congrats on building the hardware! It is normal behavior for a PS/2 keyboard to send data for key presses as well as for key releases. The Minimal 64 detects and handles both.

I suggest the following: Go to page 6 of the reference manual. You'll find a diagram showing the various scan codes the Minimal 64 expects your keyboard to send, e. g. for 'a' its 0x1c and for 'arrow up' its 0xe0 0x75.

If you release a key, your keyboard will prepend the release code 0xf0 and send the key code again, e. g. 0xf0 0x1c for 'a' released and so on. The Minimal takes care of all that.

Step 1. Find out if your keyboard sends the correct data. If you have a digital scope take a look at the CLK and DAT lines. The format is 1 low start bit, 8 data bits, 1 parity bit, 1 high stop bit. Very easy to decipher right on the scope.

Step 2. Find out if the Minimal receives this information correctly. You can use the following line of code:

test: INK CPI 0xff BEQ test PHS JPS 0xb04e PLS JPA test ; calls API function _PrintHex

:0E8000000522FF5C008057554EB058260080C8
:00000001FF

This will poll the PS2 register for input and print out the result in HEX format. Then we'll see.

Good luck!

[slu4coder]

Thanks for diving into this. I've tested three different PS/2 keyboards I had at hand. They all show timing differences :-( I see your point with simply polling for that datagram, however, I try to avoid a situation where the OS gets stuck in an endless loop just because some expected behavior doesn't occur. Let me think about this a bit before we adapt the OS code. I am happy to hear that for the time being you have found a solution that works for your keyboard.

Btw, if you want to secure "serial number #3" for you in the Minimal 64 manual's build gallery, send me a photo of your build and I'll be happy to include it.

[michaelkamprath]

Note that the PR #15 I submitted does not get stuck in an endless loop. It will bail after some time.

I will send you a pic soon.

[slu4coder]

Sounds great! Unfortunately, right now I am not in a position to test things on one of my two real hardware Minimal 64s. I have cannibalized both boards to build a breadboard UART-only version that - in theory - runs at 13.3MHz using standard 74HCxx ;-) Again, just for the fun of it... why am I doing this?!

[michaelkamprath]

I think you answered you own question "Why am I doing this?" ... "Again, just for the fun of it" ;-)

[slu4coder]

Your code has been pulled into the main branch. Later, I have uploaded a slightly shorter version using CLB instead of LDI 0 STA ... and taking advantage of INB already updating the flags, i. e. no CPI required. I have tested everything on my real hardware (with two different PS/2 keyboards I had available) and it works just as before on my end.