Resolution for optical encoder

[baileysw]

Rather than using a potentiometer for feedback and face the problems of the pot wearing out, inaccuracy, etc, I am trying to use an optical encoder for feedback.

I am attempting to use the two photo-choppers from a old PS2 mouse. There are references for Arduino code for just using the photo-choppers directly into interrupt pins on the Arduino and I may have to resort to that, but initially I am communicating directly with the PS2 mouse chip (EM84510EP) via a 38400 baud serial link. This supposedly has the advantage of reducing switch bounce and providing the counts from both choppers for you. Only time will tell if I miss any counts using this method.

Initial tests show that for 180 degrees movement of the chopper wheel, I receive around 44 counts from the chip. Assuming I drive the chopper wheel directly from a windscreen wiper gearbox output, the rest position will be at horizontal with fully up at 90 degrees and fully down at -90 degrees - i.e around 22 counts from the mouse chip in each direction.

Does anyone have ideas on whether this will be adequate resolution? If not, what number of counts do you think would be adequate.

If inadequate, I could use a larger gear attached to the wiper shaft driving a smaller gear driving the mouse chopper wheel, giving more counts.

Looking forward to some advice and/or other information from anyone who has done something similar.

Regards,
baileysw

[danielDK]

Hi - i did exactly the same with my setup - and it worked quite well! i had 160ticks per revolution.. and at 1000 rpm - it didnt miss any.. The cool thing when using ps2 is the counts are done in hardware within the mouse, that leaves your arduino or whatever free to do other stuff

Different mice has different resolution - Logitech gaming mouse had 400 ticks per revolution ..
I found it difficult to Mount the optical encoders and wheels in a secure manner so i went with the omron e6c2-cwz5b instead.. Found them used for 30$ a piece... they allmost have too high an resolution at 360x4 tics per revolution.

Cheers
Daniel