Vacuum Pumpdown Calculator

Model assumptions — read before trusting absolute numbers

• Sizing & comparison tool, not a predictor. Verify base pressure against real pumpdown data, vendor pump curves, and leak tests.
• Topologies: Series = one daisy chain (far-chamber penalty ~n²). Strings = parallel identical chains off the hub; worst case is the deepest chamber. Banks = massless junctions; your chamber V/A/material applies to the leaf chambers, not the junction node.
• Pump speed is constant except backing-pump ultimate, modelled as S(P)=S_nom·(1−P_ult/P). Virtual leaks and turbo speed-vs-pressure rolloff are not modelled.
• Real leak = constant throughput on the hub, sets a flat floor at Q_leak/S_eff. Early flatline = leak; slow creep-down = outgassing.
• Outgassing (no bake): q(t)=q₁ₕ·(1h/t)^α. α≈1 unbaked metal, ≈0 baked (flat H₂ floor).
• Bakeout is qualitative. Identical to no-bake until heaters fire, then thermal desorption (burst, then collapse to an H₂ floor). Shape is meaningful; absolute peak and baked floor are order-of-magnitude only — only a real bake / RGA settles it.
• Turbo spin-up ramps backing→full over ~45 s (smooth handoff). Steep drops on small chambers are real, not artefacts.
• Pumping modes: Direct = short fat throat, lowest base pressure. Remote = foreline pipe sets base pressure (a narrow line throttles a big pump). Backing only = single stage, cannot beat its ultimate — need a turbo to go lower.
• Foreline: bellows derated by your conductance factor (~0.3–0.6); 90° valve added in series as bore aperture + short tube (L_eff≈3d). Roughdown from atmosphere is approximate — use vendor roughing curves for that region.
• Conductance (air, 20°C): molecular 12.1·d³/L + aperture 9.11·d², viscous 135·d⁴·P/L, Knudsen sum. Solver: coupled backward Euler, stable for any chamber count.