S2 Photon timing downsampled for large time in instruction

[FaroutYLq]

When taking a instruction with huge timestamp (also in int64), you will have weird S2 waveforms, where the photons timing are downsampled. More discussion here. It is true to all S2s simulated.

Example for simulation using big time instruction

import cutax
st = cutax.xenonnt_sim_base(output_folder='/scratch/midway3/yuanlq/wfsim')
st.set_config({"fax_file" : "/project/lgrandi/yuanlq/wimp/s2sim/021952-9-se-all-20.csv"}) # feel free to check
se_peaklets = st.get_array('021952_se', 'peaklets')

for p in se_peaklets[:100]:
    if p['dt']==10:
        plt.plot(p['data'])
plt.xlabel('Sample [10ns]')

Example for simulation using small time instruction (what people usually do)

st = cutax.xenonnt_sim_base(output_folder='/scratch/midway3/yuanlq/wfsim')
st.set_config({"fax_file" : "/project/lgrandi/yuanlq/wimp/s2sim/jy_se_f6191.csv"}) # the first 100 rows of /project/lgrandi/jyangqi/se_sim_stuff/se_f6191.csv

se_peaklets = st.get_array('se_f6191', 'peaklets')
for p in se_peaklets[:100]:
    if p['dt']==10:
        plt.plot(p['data'])
plt.xlabel('Sample [10ns]')

The issue is suspected to happen when int64 were casted into float.

[ramirezdiego]

Solution in https://github.com/XENONnT/fuse/pull/171.