Question about point source image simulation

[Di He]

Hello,

I am a student trying to generate some mock image of microlening from lenstronomy. I build a simple model, with a point mass lens and a point source light. But I find out something strange about the ImageModel. Sometimes the imageModel gives me a normal image, but sometimes it turns out to a image of Nothing at all.

I tried to figure something out by running `PointSource.image_position` with the same parameter, and it turns out that I didn't get the same results every time. Sometimes the function returns an empty array but sometimes it doesn't.

The parameter I use is 

`kwargs_lens = [{'theta_E': 0.06382052310385468, 'center_x': -0.001320424615941812, 'center_y': 0.003961273847825464}]` and `kwargs_ps = [{'ra_source': 0, 'dec_source': 0, 'source_amp': 100}]`

I guess it might have something to do with the LensEquationSolver. Does anyone know the reason for it, and how can I fix it?

Thanks!

Di