Strange behaviour inserting particles

Hi,
I'm having problems inserting particles.
I'm using pour/dev/packing command.
After few iterations they jump to +Z very quickly like rockets and all crash.
My file follows:

************************************************
atom_style granular
atom_modify map array
boundary m m m
newton off
communicate single vel yes
units si

region reg block -0.015 0.015 -0.015 0.015 -0.32 10 units box
create_box 1 reg
neighbor 0.0200 bin
neigh_modify delay 0

# Material properties required for new pair styles
fix m1 all property/global youngsModulus peratomtype 20.e+9
fix m2 all property/global poissonsRatio peratomtype 0.20
fix m3 all property/global coefficientRestitution peratomtypepair 1 0.001
fix m4 all property/global coefficientFriction peratomtypepair 1 0.4
#fix m5 all property/global characteristicVelocity scalar 2.
fix m6 all property/global cohesionEnergyDensity peratomtypepair 1 2.0e+8
#fix m7 all property/global k_finnie peratomtypepair 1 1.0

# New pair style
pair_style gran/hertz/history 1 1 #Hertz with cohesion
pair_coeff * *

# TIME
timestep 1e-6
#fix ts all check/timestep/gran 1 0.1 0.1

# gravity
fix gravi all gravity 9.81 vector 0.0 0.0 -1.0

# Box walls
fix boxwalls_x all wall/gran/hertz/history 1 0 xplane -0.015 0.015 1
fix boxwalls_y all wall/gran/hertz/history 1 0 yplane -0.015 0.015 1
fix boxwalls_z all wall/gran/hertz/history 1 0 zplane -0.32 10.00 1

# Distribution for insertion
variable radius1 equal 0.0020
variable radius2 equal 0.0015
variable radius3 equal 0.0010
variable density1 equal 1500
variable alpha equal 0.2
variable Nattemps equal 1000

fix pts1 all particletemplate/sphere 1 atom_type 1 density constant ${density1} radius constant ${radius1}
fix pts2 all particletemplate/sphere 1 atom_type 1 density constant ${density1} radius constant ${radius2}
fix pts3 all particletemplate/sphere 1 atom_type 1 density constant ${density1} radius constant ${radius3}
fix pdd all particledistribution/discrete 1. 3 pts1 0.57 pts2 0.26 pts3 0.17

group nve_group region reg

region bc1 block -0.015 0.015 -0.015 0.015 -0.32 0.00 units box
region bc2 block -0.015 0.015 -0.015 0.015 0.10 0.42 units box
region bc3 block -0.015 0.015 -0.015 0.015 0.52 0.84 units box

fix ins1 nve_group pour/dev/packing 1 distributiontemplate pdd vol ${alpha} ${Nattemps} vel uniform 0. 0. 0. 0. 0.0 region bc1
fix ins2 nve_group pour/dev/packing 1 distributiontemplate pdd vol ${alpha} ${Nattemps} vel uniform 0. 0. 0. 0. 0.0 region bc2
fix ins3 nve_group pour/dev/packing 1 distributiontemplate pdd vol ${alpha} ${Nattemps} vel uniform 0. 0. 0. 0. 0.0 region bc3
fix integr nve_group nve/sphere

# Output settings, include total thermal energy
compute 1 all erotate/sphere
thermo_style custom step atoms ke c_1 vol
thermo 1000
thermo_modify lost ignore norm no
compute_modify thermo_temp dynamic yes

# RUN
restart ${restarttime} restart
run 1
dump dmp all custom ${dumptime} post/dump.model_1_ins id type type x y z ix iy iz vx vy vz fx fy fz omegax omegay omegaz radius

# Set particles
unfix ins1
unfix ins2
unfix ins3
run ${runtime}
************************************************

If I increase the radius I have no problem (but I need this size of particles).
If I increase the skin I got overflow and it dies. If I decrease the skin I got error and it dies.
Where am I wronging?
Am I forcing something?

Thanks for any help

Andrea