Why is a rising vacuum pressure not yet a diagnosis?
When a vessel, chamber or header is isolated from its pump, the pressure always rises. The question is why. Gas can enter through a real leak in a flange, weld, seal or valve; it can come off internal surfaces as adsorbed water and gases desorb; it can bleed out of blind threads, porous welds or trapped pockets; it can permeate through elastomer seals and hoses; it can come from the product or process; or the gauge may simply be reading wrongly.
Each source behaves differently over time and responds differently to temperature, pumping history and gas type. A good diagnosis uses those differences deliberately instead of assuming the first plausible explanation. Replacing seals because “it must be a leak” wastes shutdown time when the real cause is moisture in a vessel that was open to humid air for a week during the monsoon.
What should be checked before any test?
Most failed vacuum diagnoses go wrong before the first reading. Confirm these basics first:
- The gauge suits the pressure range and has been checked recently. Pirani and other thermal-conductivity gauges read differently for water vapour, helium and solvent than for air, so their reading under a vapour load can mislead.
- The gauge is mounted where it measures the volume of interest, not at the pump inlet or behind a long, thin line.
- The isolated volume is truly isolated: valves are closed and holding, and any line between the valve and gauge is included in the volume.
- The volume is known, including internal pipework and any connected equipment.
- Temperature is stable, or at least recorded. A vessel warming in the afternoon sun will show a pressure rise of its own.
- The process is in a defined state: empty and dry, or loaded with known product.
How does a rate-of-rise test work?
Pump the volume down to a stable pressure, close the isolation valve and record pressure against time. The gas load entering the volume is then estimated as:
Q = V × Δp ÷ Δt
where V is the isolated volume in litres, Δp the pressure rise in mbar and Δt the time in seconds, giving Q in mbar·l/s.
Worked example (illustrative only): a 500 litre vessel rises from 0.10 mbar to 0.60 mbar in 100 seconds after isolation. Q = 500 × 0.50 ÷ 100 = 2.5 mbar·l/s. This number is the total gas load at that moment; it does not yet say whether the load is a leak or outgassing.
The shape of the curve is the key evidence. Record the rise over a long enough period, ideally several times the first interval, and plot it.
| Observed behaviour | Most likely mechanism | Why |
|---|---|---|
| Straight-line rise that continues well past normal working pressure | Real leak from atmosphere | Air keeps entering at a near-constant rate while the internal pressure is far below atmospheric |
| Fast initial rise that slows and levels off | Outgassing or vapour reaching equilibrium | Surface or product release falls as pressure rises towards the vapour pressure |
| Levelling near about 20–30 mbar at room temperature | Water present in the volume | The vapour pressure of water at typical room temperature is in that range |
| Rise that falls with each repeat pump-down | Outgassing, gradual conditioning | Adsorbed gas is being removed with each cycle |
| Rise that tracks vessel temperature | Temperature effect or temperature-driven outgassing | Warmer surfaces release more gas; gas expands |
| Erratic jumps or a reading that differs from a second gauge | Measurement fault | Gauge contamination, wiring or wrong gas calibration |
Real systems often show a mixture: a levelling component from outgassing on top of a steady leak. Repeating the test after more pumping time, or after mild warming and cooling, helps separate the two.
How do you locate a real leak once you suspect one?
Once the evidence points to a leak, the next job is finding it. The method depends on the size of the leak and the vacuum level.
- Divide and isolate. Close intermediate valves and test sections separately to narrow the search to one part of the system.
- Inspect the usual suspects: flange gaskets, O-rings, valve glands, sight glasses, instrument fittings, threaded joints, flexible hoses, shaft seals and recently disturbed joints.
- For coarse leaks, pressurise the isolated system slightly with dry air or nitrogen, within its safe rating, and use leak-detection fluid or an ultrasonic detector. Never over-pressurise a vacuum vessel that is not rated for positive pressure.
- For finer leaks under vacuum, spray a tracer gas such as helium around suspect joints while watching a helium leak detector or mass spectrometer connected to the system. A response pinpoints the leak.
- Listen and look while vacuum is applied; large leaks hiss, and some seals visibly draw in.
- After repair, repeat the rate-of-rise test under the same conditions to confirm the improvement.
Solvent sprays are sometimes used to find leaks by watching a gauge change, but they introduce flammable and process-contaminating vapour and should only be used where the plant’s safety rules allow.
How do you reduce outgassing and vapour load?
If the problem is gas released from inside the system, sealing work will not help. Instead:
- Keep vessels closed or blanketed with dry nitrogen when not in use, especially in humid weather, so surfaces do not take up moisture.
- Allow more pumping time, or warm the vessel within its design limits, to drive off adsorbed water before production.
- Clean surfaces of oil, grease, product residues and fingerprints, which release vapour under vacuum.
- Replace aged or unsuitable elastomer seals and hoses that permeate or release gas.
- Remove trapped volumes such as blind tapped holes, double-sided welds and unvented screws where the duty requires it.
- Where product moisture is the load, fit a condenser or cold trap before the pump.
What else can look like a leak?
Several non-leak conditions produce the same symptom of poor vacuum:
| Condition | How it shows | How to confirm |
|---|---|---|
| Worn or contaminated vacuum pump | Poor ultimate vacuum even with vessel isolated | Blank the pump inlet and test the pump alone |
| Contaminated pump oil | Vacuum limited near the vapour pressure of the contaminant | Oil condition; improvement after oil change |
| Restricted line or filter | Good vacuum at the pump, poor at the vessel | Compare gauges at both ends |
| Process gas or product load | Rise only when product is present | Test empty and loaded |
| Passing internal valve | Gas from another part of the system or a purge line | Isolate upstream connections individually |
| Gauge error | Readings disagree with a reference gauge | Cross-check with a second, suitable gauge |
How should results be recorded and accepted?
Keep a simple record for every test: date, plant state, volume, gauge type and position, starting pressure, temperature, pressure readings with time, the calculated gas load, the curve shape, what was changed between tests and the conclusion. The conclusion should be one of three honest outcomes: a mechanism confirmed by more than one piece of evidence, an unresolved load that needs further work, or an invalid test that must be repeated.
Acceptance limits for a production system, such as a maximum allowable rate of rise, should come from the process requirement and the equipment specification, not from a general rule. Once set, the same test at the same conditions becomes a valuable routine check after every shutdown or seal change.
How Reddy Compressors helps
Reddy Compressors is an independent multi-brand compressed-air, vacuum and blower specialist operating from Tarnaka, Secunderabad, since 2002, and not an authorised dealer of any brand. We have tie-ups with all major and reputed brands. With over 20 service engineers and technicians, we cover the whole of Telangana and Andhra Pradesh, and we supply equipment and spares across India.
- Testing vacuum pumps on their own to separate pump condition from system leakage.
- Helping plant teams carry out isolation and rate-of-rise tests, and interpreting the results.
- Servicing and overhauling rotary-vane, liquid-ring, screw and claw vacuum pumps and Roots boosters of many makes.
- Supplying seals, gaskets, filters, oils and pump spares, identified from the nameplate and confirmed before quoting.
- Supplying compressors and dryers for clean, dry air used in pressure leak tests, and carrying out compressed-air leak detection on the plant side.
Call +91 99590 60718 or +91 92468 88812, or write to reddycompressors@gmail.com. Office: Srinilayam, 12-13-200, First Floor, Street No. 1, Tarnaka, Secunderabad 500017.
What to send us
- A sketch or photo of the system, with vessel volume and pipe sizes.
- Vacuum pump make, model and nameplate photo.
- Gauge type and where it is fitted.
- Pressure readings against time after isolation, with temperature if known.
- The required working pressure and what changed before the problem started.
Frequently asked questions
How do I know if my vacuum system has a leak or outgassing?
Isolate the volume from the pump and record pressure against time. A real leak from atmosphere gives a steady, roughly straight-line rise that continues well past working pressure. Outgassing and vapour give a fast initial rise that slows and levels off. Repeat after more pumping or a temperature change: outgassing usually reduces with conditioning, while a genuine leak stays much the same.
What is a rate-of-rise test in vacuum systems?
It is a test where the system is pumped down, isolated, and its pressure rise is recorded over time. Multiplying the isolated volume by the pressure rise and dividing by the time gives the gas load in mbar·l/s. The value shows how much gas is entering, and the shape of the pressure curve over a longer period helps show whether that gas comes from a leak, outgassing or vapour.
Why does my vacuum stop at around 20 to 30 mbar?
A vacuum that stalls in that range at normal room temperature very often indicates water. The vapour pressure of water at typical room temperatures lies there, so as long as liquid water is present the pump is mainly removing vapour. Look for wet product, condensate in lines or the vessel, or water in the pump oil. A condenser, cold trap or gas ballast may be needed.
How do you find a leak in a vacuum vessel?
Narrow the search by isolating sections and testing each, then inspect flanges, seals, valve glands, fittings and recently disturbed joints. Coarse leaks can be found by gently pressurising with dry air or nitrogen within the vessel rating and using leak-detection fluid or ultrasound. Fine leaks are best found with helium sprayed around joints while a helium leak detector watches the system.
Can a vacuum gauge give a false reading?
Yes. Thermal-conductivity gauges such as Pirani types are calibrated for air and read differently for water vapour, helium or solvents. Contamination, damaged sensors, wrong mounting position or long thin connecting lines also distort readings. Cross-check with a second suitable gauge, ideally a gas-independent type, before deciding on a leak or pump fault based on a single instrument.
Why does my vacuum get better after pumping for longer?
That pattern usually points to outgassing. Water and gases adsorbed on internal surfaces, seals and product are gradually removed, so the gas load falls with time and the achievable pressure improves. It is common after a vessel has been opened, cleaned or exposed to humid air. Keeping the vessel closed or blanketed with dry nitrogen between batches reduces the effect.
Related pages
Calculators and worked examples
These tools give indicative figures for screening, comparison and for preparing an enquiry. They run in your browser and send nothing. No prices or tariffs are assumed: enter your own to see costs. Efficiencies and other figures shown are example values: replace them with your own. Results are not a performance guarantee or a design certificate; final selection is confirmed by a service engineer against the machine data sheet and the site.
Vacuum rate-of-rise gas load
Turn an isolation test (pressure rise over time with the pump valved off) into the gas load entering the system, compare a later interval to see whether it is a leak or outgassing, and find the speed needed to hold a working pressure.
The worked example below shows the method with fixed numbers. Turn on JavaScript to change the inputs.
Worked example
Inputs
- Isolated volume incl. pipework (litres): 500
- Pressure at isolation (mbar a): 0.1
- Pressure at the end of the interval (mbar a): 0.6
- Interval (s): 100
- Later interval of the same test: pressure rise (mbar; 0 = not measured): 0
- Later interval: duration (s): 100
- Working pressure for the pumping-speed check (mbar a; 0 = skip): 0
Result
- Rate of rise: 0.005 mbar/s = 0.3 mbar/min
- Gas load Q = V × Δp ÷ Δt: 2.5 mbar·L/s = 9 mbar·m³/h = 0.25 Pa·m³/s
- As air at 20 °C: 10.7 g/h = 0.148 litres per minute of atmospheric air
Method. Q = V × Δp ÷ Δt, with V in litres, Δp in mbar and Δt in s, gives Q in mbar·L/s (× 3.6 = mbar·m³/h; × 0.1 = Pa·m³/s). Mass of air at 20 °C: 1 mbar·L/s ≈ 4.28 g/h. Curve shape (screening values used by this tool): a later interval rising at 80% or more of the first rate points to a real leak; 50% or less points to outgassing or vapour levelling off; between, a mixture. Required effective speed to hold a working pressure p: S = Q ÷ p. The isolated volume must include every pipe up to the closed valve; record temperature, because a warming vessel shows a rise of its own. A leak from atmosphere stays roughly constant only while the inside pressure is well below about half an atmosphere.
What to take from it. The number gives the size of the gas load, the curve shape gives its cause: test empty and loaded, and repeat after more pumping, before replacing seals.
Technical content reviewed by D. Srinivas Reddy, Managing Partner and service engineer, Reddy Compressors. Last updated 17 October 2026. How our engineering content is reviewed.