Trying many things I figured out that the records were identical rotating or not the shaft of the rotary encoder...
So definitely the code working on the paper does not work. I suspect that HAL does not like to have an io pin read and written in a same cycle.
So I decided to find a workaround and found one by reading the hal manual about encoder.
I looked for a consequence of the index being triggered to rearm index-enable a bit later.
I found that counts were set to 0 when index triggered. (Makes sense because that is a purpose of an index in case of homing) and used it as follow :
loadrt hostmot2
loadrt hm2_eth board_ip="10.10.10.10" config=" num_encoders=1 num_pwmgens=0 num_stepgens=0 sserial_port_0=0000xx"
loadrt threads name1=servo-thread period1=500000
loadrt wcomp
loadrt not
loadrt conv_s32_float
loadrt tristate_bit
loadrt updown
addf hm2_7i92.0.read servo-thread
addf conv-s32-float.0 servo-thread
addf wcomp.0 servo-thread
addf not.0 servo-thread
addf tristate-bit.0 servo-thread
addf updown.0 servo-thread
addf hm2_7i92.0.write servo-thread
setp tristate-bit.0.in 1
setp wcomp.0.min -100
setp wcomp.0.max 100
net s32_count hm2_7i92.0.encoder.00.count conv-s32-float.0.in
net count conv-s32-float.0.out wcomp.0.in
net in_min_max wcomp.0.out not.0.in
net reset not.0.out tristate-bit.0.enable
net enable tristate-bit.0.out hm2_7i92.0.encoder.00.index-enable updown.0.countup
start
loadusr halscope
loadusr halmeter
And the plot in attachment.
Thanks to PCW that put me on track.
