lepold fc500r usb/ps2 kebord haked so im can tap into signal from scrol lok key
ps2 adapter
windows 7 computer
benq 2411p 144hz tn panel monitor
del u2412m 60hz ips panel monitor
program that make scren go white when pres key down running at ~16000fps (exept vsink on test wher its at 144hz of course)
photo diode taped to monitor
100mhz scope with cursors that alow mesure time intervals acurately
channel 1 = signal from scrol lok key, voltage go up when key is down, scope triger on it
channel 2 = signal from photo diode, voltage go up when scren is more white
100 samples per test
tests im did
usb+144hz vsync off
ps2+144hz vsync off
ps2+60hz vsync off
ps2+144hz vsync on
![Image](http://zzz.st/up/Ls0ZYtnN/IMG_20201217_192618269.jpg)
http://zzz.st/up/NtFfUjkB/IMG_20201217_174857729.jpg
http://zzz.st/up/sAlDy6x0/IMG_20201217_181430464.jpg
http://zzz.st/up/VKQR4PKE/Untitle24d.jpg
resluts (numbers in mili seconds)
![Image](http://zzz.st/up/BwAipTTu/stuf.png)
![Image](http://zzz.st/up/qKWVRE3G/gnnublot.png)
numbers http://zzz.st/up/MdDR4B9C/data.zip
how much of a difrence dose it make
duno but
u can compensate for a constant value added to the latency such ass with ps2 144hz vsync on vs off (to som extent, it becums harder the biger the constant is of cours)
a wider distribution results in a biger random error aded to evry keystroke
if u for example lok at (max-min)/2 u get worst case erors (for this data) asuming u aim for midle of distribution and ur aim is perfect
500fps merge with first alo at difrent frames in 2ms intervals, from definitly 2early to definitly 2late (frame# 6511,6512,6513,6514,6515,6516)
4 survive 2 dident (died in "oposide" directions) if its hard to se in vidio
https://www.youtube.com/watch?v=8HZ5tF033Do
1000fps merge with first volt at difrent frames in 5ms intervals, from definitly 2early to definitly 2late (frame# 590,595,600,605,610,615,620)
2nd volt is automatic
https://www.youtube.com/watch?v=DZ5i_2OF5Ks