Sin function bug

[ahkab]

Sin function bug

[ggventurini]

Hello Dan,

Thanks for the nice words. I am very glad you enjoy this library!

Here follows the explanation for that message you see. I agree that it would be arguably good to avoid that condition, so I will patch it away soon.

Verbose answer follows:

The sine function has five parameters:

• vo, offset voltage in Volt
• va, amplitude voltage, in Volt
• freq oscillation frequency in Hertz
• td time delay, in seconds
• theta oscillation damping constant in seconds

For all time functions, you may change the parameters at will in your code like, for example:

mysin.td = 1e-6

The change will be effective immediately.

As of now, you can even define a time function with a subset of the parameters and define the missing parameters subsequently, eg after you compute them.

For example, if you leave theta undefined -- I'd venture to guess this is what is happening in your code -- when you try to evaluate the sine, it will complain some parameters are undefined and hence the time function should be impossible to evaluate. Only when it actually comes to evaluate it, the (pretty trivial) algorithm tries to provide a result anyway -- or otherwise it should stop the simulation -- and theta undefined is interpreted as no damping, ie theta == 0 and the damping code is not executed (as damping with null theta means divide by zero) and all is fine.

Providing a default value hard-coded in ahkab would make this error go away. I believe this to be a good idea, especially for theta, which few people use. I will still think a bit about it before patching to make sure this doesn't bite back in unexpected ways.

I will also post here a drawing as time allows.

All the best,

GV

EDIT: clarifications, rewording.

[ggventurini]

It should be fixed in 19eecc2

Now all commonly used parameters are compulsory and all seldom used parameters have a default value.

Noticed I realized I made a mistake whic got fixed too, the old theta corresponds to, according to the usual conventions, 1/theta. I have added the phase lag phi too.

Here is the docstring from devices.sin:

    f(t) = 

        t < td:  
                 vo + va*sin(pi*phi/180)
        t >= td: 
                 vo + va*exp(-(time - td)*theta)*sin(2*pi*freq*(t - td) + pi*phi/180)

    *Parameters:*

    vo: float
        Offset

    va: float
        amplitude

    freq: float
        frequency in Hz

    td: float, optional
        time delay before beginning the sinusoidal time variation, in seconds. 
        Defaults to 0.

    theta: float optional    
        damping factor in 1/s. Defaults to 0 (no damping).

    phi: float, optional
        Phase delay in degrees. Defaults to 0 (no phase delay).

Please feel free to reopen if you spot any problem.

I will likewise fix the other time functions.