A vacuum level that will not come up is a common problem on site, and the first reaction for many people is to pull the pump apart. There is another move that can be made before that: stop the vacuum pump and watch how fast the reading falls back. As long as the starting reading and the moment the pump stops are the same, the rate of the fall corresponds directly to how well the system is sealed, and it is a clue that costs almost nothing.
Why the reading must change after the pump stops
Once the vacuum pump stops extracting, the oil and gas mixture left inside the system keeps expanding, and the pressure balance that existed in the pipework, the filter housings and the separation tank is broken. The spaces that had been drawn down to low pressure are now open to the outside, air finds its way in through the places that are not sealed properly, and the vacuum reading falls back from -0.0999 MPa towards the lower edge of the -0.08 MPa working range, eventually approaching atmospheric pressure.
So what matters is not the reading at one moment but the slope of the reading over time. A well sealed system admits very little air per unit of time and the reading comes back down gently; a system with a gap, an untightened flange or a cracked weld keeps taking air in and the fall is clearly faster. The slope separates these two states better than the reading does, because the reading itself is affected a lot by the oil temperature, the oil level and the amount of residual gas at that moment.
What each rate of fall points to
Match the rate of fall to the symptom and the direction of the search narrows quickly. The table below uses the working vacuum of -0.08 ~ -0.0999 MPa and the limit vacuum of ≤5 Pa of the ZYD-100 to show what each decay state usually means and which part of the pipework to look at first.
| Decay state after the pump stops | Probable sealing condition | Where to look first |
|---|---|---|
| The reading hardly moves and stays high for a long time | Sealing is sound, the vacuum system is in a normal state | Record the baseline data as a reference for later comparison |
| The reading falls slowly and is still inside the working range after several hours | A slight leak exists, smaller than the pumping capacity of the system | Sealing rings, thread joints, welds on the vacuum pipework |
| The reading falls quickly and returns close to atmospheric pressure | The leak is clearly larger than the pumping capacity of the system | Flange faces, pump inlet and outlet, flexible pipe connections |
| The rate of fall changes from fast to slow without a pattern | A leak path inside the system opens and closes with the pressure | Isolate section by section, suspect the tank connections and the valves |
How to run a test that can be compared
A rate of fall can only be compared when the conditions are the same, so several preconditions have to be fixed before the test, otherwise the number obtained has nothing to refer to. What is compared is always the same machine under the same conditions, not absolute values between different machines, because the volume, the length of the pipework and the number of connections differ from one machine to the next.
- Run the test after the machine has stopped taking oil and after the heating has been off for a while, so the oil temperature is stable and the vapour coming off does not disturb the reading.
- Record the starting vacuum reading and the moment the pump stops before stopping it; both values are needed to work out the fall per unit of time.
- Read from the same gauge for the whole test and do not swap gauges halfway through, because the scale and the damping differ from one to the other and the readings are not comparable.
- Record at least three points in time, keeping the interval between neighbouring readings the same.
- Do not open or close valves, do not add oil and do not carry out other maintenance work while the test is running.
Keep this set of data as the baseline curve for the machine. Running the same test on the same machine at a later date often says more through a small change in the rate of fall than through a single reading: a sealing ring ageing slowly, a weld on the vacuum pipework opening up under vibration, these changes cannot be seen in one reading but they do show up in a curve.
Where this method belongs
The value of this method is that it separates the leak rate from the capability of the pump. A vacuum pump in good condition with a vacuum level that will not come up, and a vacuum pump that has lost output and therefore will not reach the level, look the same on site but differ a lot in the cost of pulling the pump apart. Doing one decay test first, ruling the sealing problem in or out, and only then deciding whether to open the pump makes the order of the search far more economical.
It should be said that this is a qualitative to semi-quantitative judgement and it does not replace a leak detector for locating a fault. What it does is split the problem into sealing and pump, so that time is not spent on the wrong branch. The limit vacuum of the ZYD-100 is ≤5 Pa, and that figure only means something when the system has been drawn down properly, and the decay test is exactly what confirms whether sealing is holding the level back.
That curve of the reading falling after the pump stops says more about the sealing condition than any single vacuum reading ever will.
