Model and document sensor and electronics-chain readout non-linearities

[cwwalter]

Although quite small (based on test-stand measurements) we should eventually add the ability to put in electronics non-linearity.

As of 09/19/2018 information from @jchiang87 is that:

"The most recent raft tested at IR2 has linearity < 0.2%. The acceptance spec < 3%"

This issue also serves to document our current understanding. This information should be propagated to the imSim feature matrix on the Wiki.

[cwwalter]

I've updated the title of this issue to reflect that, In addition to the saturation and bleeding that is already included in the simulation, the sensors themselves also exhibit non-linearity as they approach saturation (in addition to any readout chain electronics effects).

These parameters are being recorded as part of the I&T (integration and testing) process and the parameters are being recorded in the e-traveler database at SLAC.

This issue is relevant for the SN group who worry about understanding how well they can understand the efficiency of measuring the light curve of close very bright SN.

[jchiang87]

The parameters being used by I&T to characterize non-linearity are not a very complete description of the non-linear response of the sensors. Currently, the info is just the maximum deviation from a linear fit over some range of e-/pixel values, e.g., 1e3-9e4 e-/pixels, and the "blooming full-well" value (as defined in LCA-10103). If we want a complete description of the non-linearity suitable for use with the Stack, I think it needs to be measured from TS8 data as part of the cp_pipe tasks.

[cwwalter]

@jchiang87 and @rhl What's the status of non-linearities in the camera model these days?

Was this measured in I&T and incorporated into the model?

(I think same question for x-talk)

[cwwalter]

We are waiting for data from the camera ream before we can implement this.