raw · notes · ingested 2026-07-02
DIY vacuum, leak-tightness & non-condensables for sorption cooling (research note)
confidence: high
Source: https://measurequick.com/identifying-true-leaks-in-a-vacuum/
Consolidates: MeasureQuick (real-vs-virtual leak), HVAC decay-test practice, Wikipedia/refrigerant vapor-pressure, vacuum-vessel sealing guidance. Credibility: High (authoritative practitioner + engineering sources).
Why it’s the make-or-break
Water boils near 6 mmHg (~0.8 kPa) at chilling temps → deep vacuum; any tiny leak accumulates air that blankets HX surfaces (8% loss @60% load, 14% @100%).
Real vs virtual leak (the #1 skill)
Read decay on a linear scale: true leak = fast rise then steady climbing slope (visible even at 10,000 µm); outgassing/moisture (O-rings, threads, grooves, the water) = rise then flat tail; tight vessel = flatline. Don’t chase phantom leaks.
Sealing
Threaded + PTFE-tape leaks under real vacuum → use O-ring seals. DIY vessels: welded stainless pipe + ISO/KF flanges with O-ring centering rings; double-O-ring grooves for critical joints.
Non-condensables
Nitrogen sweep 2–3× then deep-evacuate; add a purge/re-evacuation port for maintenance.
Refrigerant difficulty ladder
- Water — best latent heat; hardest (deep vacuum, 0 °C freeze floor, big vessel, degassing). Safe.
- Methanol — moderate sub-atmospheric, sub-zero ice, easier seal; toxic + flammable, unstable >120 °C.
- Ammonia — above atmospheric (pressure tubing, no vacuum, easy leak detection); toxic + corrosive.