Superfluid Helium-4 as a Test Bench for Spontaneous Collapse: Heating Signatures, Model Discrimination and Quantum-Evaporation Readout

White-noise collapse models are now excluded at their natural parameters by X-ray searches, so bulk heating constrains spontaneous collapse only if the noise is colored. In that regime the ratio of the heating per unit mass of superfluid He-4 to that of a solid distinguishes the main model classes. This paper merges and corrects four of the author's earlier He-4 studies (colored-noise calorimetry, the collapse wind, Diósi–Penrose heating and quantum-evaporation readout). Every number was re-derived with independent code and every bound checked against its primary source. Main results:(i) Rest-frame colored CSL: He-4 is heated more than a solid by up to (c_s/c_He)², i.e. 85–2170 for Pb to sapphire. A helium calorimeter with 10 pW/kg residual heating improves the best existing bound at r_C = 10⁻⁷ m by a factor of 5.4, and by about 72 for r_C ≲ 10⁻⁸ m.(ii) Collapse wind: if the colored noise has a frozen spatial pattern at rest in the CMB frame, heating becomes material independent, suppressed by F ≃ (2√π/3)Ωr_C/v, with a 7.9% annual modulation peaking on 12–13 June.(iii) Diósi–Penrose: with R0 set to the zero-point width, the predicted heating is 12.5 pW/kg for Cu, 66–90 pW/kg for W and 7.4 fW/kg for solid He-4. Colored DP probes single-atom recoil, where helium is less sensitive than heavy solids.(iv) Discriminant: the He/solid heating ratio is 1 for white CSL, white DP and the collapse wind, up to (c_s/c_He)² for rest-frame colored CSL, and below 1 for colored DP.(v) Readout: single collapse events lie far below quantum-evaporation thresholds, so readout is calorimetric. A multi-bounce Monte Carlo shows that the evaporation efficiency is set by the cell walls (η ≈ 0.07 at 30% roton reflectivity, 0.72 with low-loss diffuse walls). A background budget for surface and underground operation and a list of corrections to the earlier preprints are included.