ALWAYS test your code on real hardware!!

[Leffmann]

Originally Posted by Toni Wilen

Explained in previous reply. (No handshake quickly enough = next key code will be received) I think I'll change this, this behavior is from very early version of UAE, this really should be emulated better today

Ok that explains what I'm seeing here, thanks.

[Asman]

Originally Posted by Toni Wilen

[...]
Check that CIA-A serial port (SP) interrupt is set (or use ports interrupt), if set, read SDR, then set bit 6 in CIA-A CRA, wait at least ~70us, clear bit 6. (this is one method to handle handshake, there is also another one)

I have question. I found in Hardware Manual that wait should be at least 85us (http://amigadev.elowar.com/read/ADCD.../node0198.html). Is it true or I'm missing something.

[phx]

Yes, it is true that some keyboards may need a slower handshake.

85us is from the HRM and I try to follow this recommendation. Although in the NetBSD kbd driver source they even use 200us with a comment about "bloody Cherry keyboards".

[NorthWay]

Woa! I have one of those Cherrys and they are lovely.
Too many l33t haXXors coded key scans that plain didn't work with it.

I think the handshake wait needed is something like 4 scanlines.

[phx]

Four scanlines on a PAL display are a little bit more than 200us. So it is confirmed: the recommended 85us are not sufficient for all keyboards.

[Jope]

.. or we just forget all about that Philips controller in the A2000A keyboard, pretend it never happened. :-)

[Asman]

Thanks for answers. I will stay with 85us.

[NorthWay]

I wont say for certain it was 4 scanlines - it has been decades since I coded it and I just followed the blue HRM spec. Bitrot and brainrot and all that...

[Lonewolf10]

Originally Posted by NorthWay

I wont say for certain it was 4 scanlines - it has been decades since I coded it and I just followed the blue HRM spec. Bitrot and brainrot and all that...

Well I have always used ~320us which (in my calculations, below) is 4 scanlines...

;  PAL
;  ===
;
;  50 VBLs / second = 0.02 seconds per VBL
;  250 scanlines per VBL = 0.00008 seconds per scanline
;  0.00008 seconds = 0.08 milliseconds (mS)= 80 micro seconds (uS)

... if my maths is off, please feel free to point it out

[Toni Wilen]

250 scanlines? Even NTSC screen has more scanlines.

PAL scanline length is always 64us. Number of scanlines is normally 313 (Long field mode, 49.92Hz) or alternating 312/313 (interlace, 50.00Hz). 312 is also possible (without VPOSW tricks) but rarely used (Short field mode, 50.08Hz)

[Lonewolf10]

ok, thanks for the correction Toni

[UberFreak]

Originally Posted by phx

Yes, it is true that some keyboards may need a slower handshake.

85us is from the HRM and I try to follow this recommendation. Although in the NetBSD kbd driver source they even use 200us with a comment about "bloody Cherry keyboards".

Ha, this might explain why I could never get my Elbox PC-Key 1200 to work with the old German A2000 keyboard!
It just goes unresponsive after a few key presses.

Works fine with later A2000, A3000 & A4000 keyboards.

[NorthWay]

I fired up WinUAE and found this that I used once:

	lea	$bfe001,a0
	rept	61		; PAL timing!
	tst.b	(a0)		; 8 cycles 68000, expanded to 10
	endr

Could be that I had subtracted some delays from register accesses I had already done.