Re: [python-control-discuss] really weird error for SISO step response

[Ryan K. <kra...@gv...>]

OK, my problem seems to be fixed.  Documenting my fix here in case this
helps someone in the future:

- upgraded to Mac OSX 10.14, then again to 10.14.6
    - I was doing this to install xcode, which I still have not been able
to do because of disk space
- manually downloaded the commandline tools for xcode and installed it
- updated gcc via homebrew
    - this seems to have fixed an error from pip regarding gfortran
- updated slycot via pip:
Successfully installed slycot-0.3.5

Everything is now happy and I can evaluate the step response of an RC
filter....

--
Dr. Ryan Krauss
Associate Professor
Product Design and Manufacturing Engineering
Grand Valley State University


On Fri, Aug 30, 2019 at 6:21 PM Ryan Krauss <kra...@gv...> wrote:

> Good times.  I have updated to 10.14.6.  I have 39GB of free hard drive
> space.  But apparently this isn't enough to install xcode..............
>
> --
> Dr. Ryan Krauss
> Associate Professor
> Product Design and Manufacturing Engineering
> Grand Valley State University
>
>
> On Fri, Aug 30, 2019 at 11:27 AM Ryan Krauss <kra...@gv...> wrote:
>
>> I guess I have to upgrade to OSX 10.14 to get xcode back......
>>
>> This could take a minute.....
>>
>> --
>> Dr. Ryan Krauss
>> Associate Professor
>> Product Design and Manufacturing Engineering
>> Grand Valley State University
>>
>>
>> On Fri, Aug 30, 2019 at 11:26 AM Ryan Krauss <kra...@gv...> wrote:
>>
>>> I just ran a pip upgrade for the control module.  I am running 0.8.2.
>>>
>>> I just tried upgrading slycot via pip.  pip says I am currently running
>>> 0.3.0.  I tried to upgrade to 0.3.5 via pip, but I receive error messages
>>> about gfortran and xcode.  I am kind of surprised that I don't have them
>>> both working.  I will work on that.
>>>
>>> I am getting False for this variable: slycot.__SLYCOT_SETUP__
>>>
>>> --
>>> Dr. Ryan Krauss
>>> Associate Professor
>>> Product Design and Manufacturing Engineering
>>> Grand Valley State University
>>>
>>>
>>> On Fri, Aug 30, 2019 at 2:16 AM Murray, Richard M. <
>>> mu...@cd...> wrote:
>>>
>>>> Which version of python-control and slycot?  I just ran against the
>>>> latest master and slycot 0.3.4 and it didn’t generate an error.
>>>>
>>>> -richard
>>>>
>>>> > On 29 Aug 2019, at 11:04 , Ryan Krauss <kra...@gv...> wrote:
>>>> >
>>>> > I am trying to do a super basic step response for a first order TF
>>>> and am getting an error I have never seen before.  The notebook is
>>>> attached, but here is the code and the error message:
>>>> >
>>>> > %matplotlib inline
>>>> > import matplotlib.pyplot as plt
>>>> > import numpy as np
>>>> > import control
>>>> > from numpy import sin, cos, tan, pi
>>>> > from control import TransferFunction as TF
>>>> >
>>>> > p = 30
>>>> >
>>>> > G = control.TransferFunction(p,[1,p])
>>>> > G
>>>> >
>>>> > control.step_response(G)
>>>> >
>>>> >
>>>> ---------------------------------------------------------------------------
>>>> > ValueError
>>>> >                                 Traceback (most recent call last)
>>>> >
>>>> > <ipython-input-7-7235fbf92fdd> in <module>()
>>>> > ----> 1 control.step_response(G)
>>>> >
>>>> >
>>>> >
>>>> > /usr/local/lib/python3.6/site-packages/control/timeresp.py in
>>>> step_response(sys, T, X0, input, output, transpose, return_x)
>>>> >     450     >>> T, yout = step_response(sys, T, X0)
>>>> >     451
>>>> >      """
>>>> >
>>>> > --> 452     sys = _get_ss_simo(sys, input, output)
>>>> >     453     if T is None:
>>>> >     454         if isctime(sys):
>>>> >
>>>> >
>>>> >
>>>> > /usr/local/lib/python3.6/site-packages/control/timeresp.py in
>>>> _get_ss_simo(sys, input, output)
>>>> >     378     If input is not specified, select first input and
>>>> >  issue warning
>>>> >
>>>> >     379
>>>> >      """
>>>> >
>>>> > --> 380     sys_ss = _convertToStateSpace(sys)
>>>> >     381     if sys_ss.issiso():
>>>> >     382         return
>>>> >  sys_ss
>>>> >
>>>> >
>>>> > /usr/local/lib/python3.6/site-packages/control/statesp.py in
>>>> _convertToStateSpace(sys, **kw)
>>>> >     895             # transfer function to state space conversion now
>>>> should work!
>>>> >     896
>>>> >              ssout = td04ad('C', sys.inputs, sys.outputs,
>>>> >
>>>> > --> 897
>>>> >                             denorder, den, num, tol=0)
>>>> >
>>>> >     898
>>>> >
>>>> >
>>>> >     899             states = ssout[0]
>>>> >
>>>> >
>>>> >
>>>> > /usr/local/lib/python3.6/site-packages/slycot/transform.py in
>>>> td04ad(rowcol, m, p, index, dcoeff, ucoeff, tol, ldwork)
>>>> >     446         e = ValueError(error_text)
>>>> >     447         e.info = out[-1]
>>>> > --> 448         raise
>>>> >  e
>>>> >
>>>> >     449     if out[-1] > 0:
>>>> >     450         error_text = "The leading coefficient of a
>>>> denominator polynomial is nearly zero; calculations would overflow; no
>>>> state-space representation was calculated.
>>>> ABS(DCOEFF("+str(out[-1])+",1))="+str(abs(dcoeff(out[-1],1)))+" is too
>>>> small."
>>>> >
>>>> >
>>>> >
>>>> > ValueError: The following argument had an illegal value: dcoeff
>>>> >
>>>> > --
>>>> > Dr. Ryan Krauss
>>>> > Associate Professor
>>>> > Product Design and Manufacturing Engineering
>>>> > Grand Valley State University
>>>> > <illegeal_dcoeff.ipynb>_______________________________________________
>>>> > python-control-discuss mailing list
>>>> > pyt...@li...
>>>> > https://lists.sourceforge.net/lists/listinfo/python-control-discuss
>>>>
>>>>
>>>> _______________________________________________
>>>> python-control-discuss mailing list
>>>> pyt...@li...
>>>> https://lists.sourceforge.net/lists/listinfo/python-control-discuss
>>>>
>>>