`setup_amps` optimisation improvement

[tristpinsm]

When testing on the SLAC system, with a different 4K amp as used in SO, the setup_amps optimisation fails unless it starts on the correct value. Need to look into it, but maybe the algorithm could be made more flexible.

[dpdutcher]

Hey @tristpinsm , #473 is also related to this. Can you describe the particular failure mode? What was the different 4K amp you were using, and what biasing parameters does it want?

[tristpinsm]

In this case the issue was just that the new amp needed a negative gate voltage (or something like that), which wasn't possible with the allowed range in the sodetlib optimisation.

[tristpinsm]

right, and my solution was to expose optimisation parameters to the setup_amps call. As in this commit