StateSpace to SisoZpk conversion removes states

[albheim]

Got asked why this was behaving weirdly, and I'm not sure. Seems like the struct_ctrb_obsv removes some states when converting from state space to zpk, but was not immediately clear to me what was going on in there.

using ControlSystems

A = [-0.016303260598174024   0.008606869939999928   0.002945859872611465   0.00038747346072186837
        0.00211965267541959   -0.00211965267541959    0.0                    0.0
        6.776556776556777e-5   0.0                   -6.776556776556777e-5   0.0
        0.0                    0.0                    0.0                   -0.0024333333333333334]
B = [0; 0; 0; 1.0]
C = [1.0 0 0 0]

sys = ss(A, B, C, 0)

@show tzero(sys) pole(sys)
@show tzero(tf(sys)), pole(tf(sys))
@show tzero(zpk(tf(sys))), pole(zpk(tf(sys)))
@show tzero(zpk(sys)), pole(zpk(sys)) # This is different than all the others, still seems similar but some states removed

[baggepinnen]

I've had problems with struct_ctrb_obsv before as well, #409
maybe it needs a rewrite. I'm not even sure what it does.

[olof3]

Most recently discussed some days ago in #475

[albheim]

Okay, closing this then and we can continue discussing in the other ones.