#876 New capacitor model breaks `m=NNN` multipliers

[Munaut Sylvain]

The {m} gets replaced by 1 if m is not specified but if one if specified, it gets replaced with a float value like 4.0000000000e+00 with a bunhc of spaces before and after and that breaks the G source created by the new model. So instead of trying to create a scaling factor combining the m and 1e9, I just added m=%s and use the m parameter of the VCVS.

[Holger Vogt]

Fix is pushed to git branch pre-master-48