Månadens Tips

TIP OF THE MONTH 18 : Tests chambers

How do Leak Detectors respond in test chambers containing steam?

TIP OF THE MONTH 18 : Tests chambers

QUESTION: Prior to the Helium leak test all our parts to be tested are cleaned by washing. The parts are dried by flushing with compressed air. Then our products are leak tested. Leak testing in vacuum also acts as a final drying step. So we are spending a lot of time and effort for pre-treatment and leak testing. Nevertheless a customer has claimed a leaky part recently. Is our leak detector corrupted?

ANSWER: This is a frequent question which covers two aspects of leak detection. For that reason we shall respond in two tips of the month. In this tip we shall treat the time consumption needed for a leak test. Next month we shall discuss the sensitivity limits after a drying step.

Coming back to your question: Your leak detector is most probably working OK. Generally the time constant of a vacuum leak detection system is a function of vessel volume and the effective pumping speed for the tracer gas. The gases in the test chamber are pumped away with the effective pumping speed of your pumping equipment. However, moisture in the chamber does release only a comparatively small fraction of vapour to the gaseous state. This is a very slow process which is slower than any pumping process. For that reason a constant pressure level can be observed in the test chamber which is determined by the vapour pressure of the liquid inside. The vapour emmision transports energy away from the liquid. So the remaining liquid will cool down and can even transform to ice. The vapour pressure of ice is even lower than the vapour pressure of the corresponding liquid (water). A second pressure level can be observed and the pump down process gets even slower.

ADDITIONAL INFORMATION: A blind-flanged leak detector can pump down the gases in its own vacuum system quite fast. A tracer gas background level is reached which is specific for the respective unit. This pump down behavior is shown in picture 1 for the leakdetector ASMGraph.

Picture 1: Pressure and leak rate pump down curve on a blind-flanged leak detector

The pump down behaviour is dependant on the volume of the sample or the test chamber, respectively. The bigger the volume connected to the leak detector, the longer the pump down time. In addition to volume effects the growing surface of the object can also emit gases. The pump down process can be prolonged by conductance or moist surfaces which are an infinite source for gas emission. For our experiment we have connected a “dry” container from our stock to the leak detector. The volume of the container is roughly 1 L.

Picture 2: Pressure and leak rate pump down curve on a “dry” container

The slope of the pressure pump down curve (right axis) now is less steep compared to our first experiment. The broad peak at low pressure reveals that our container definitely was not dry – residual humidity is responsible for the prolonged pump down time although no moisture was visible on the sample. Helium leak rate drops within few seconds over a range of four decades from 10-5 mbar l/s to 2 x 10-9 mbar l/s. However, for reduction of this leak rate by a factor of 2 the same time interval again is needed. The achievable Helium background level is bigger by a factor of 10 compared to the dry blind-flanged leak detector.

In a third experiment we have introduced only a small droplet of water into our container. The red pressure pump down curve now shows two distinct levels. The first level at a level of roughly 1 mbar (right axis) can be assigned to water vapour. The second level between 0,1 and 0,01 mbar can be assigned to ice vapour pressure. Going along with the pressure curve the leak rate signal also shows a flat regime at a constant water vapour pressure. At the crossover point of 0.5 mbar the leak detector switches from gross leak to normal mode. In this experiment it takes 35 seconds from cycle start until the leak detector reaches a background signal. This leak rate level is reached within 3 seconds with a blind-flanged leak detector.

When the detector switches from gross leak to normal mode the gases are guided through other channels of the valve block. In addition the Helium pumping speed of the unit may change. In our example the change of the gas path results in a sudden jump of the leak rate signal to a level which is one decade lower. Freezing of water is also reflected in the leak rate signal. How can you interpret such a leak rate “measurement” and meet a reliable OK/NOK decision during serial production?

Picture 3: Pressure and leak rate pump down curve on a “moist” container

The data in our examples are demonstrate conclusevly that moisture in samples can extend the pump down time massively. An influence on the leak rate reading can also be shown.

So our recommendation clearly is to invest additional time and effort in ex-situ pre-treatment and drying processes. Your results will be easier to interpret and the OK/NOK decision is much clearer. In addition the service intervals of your leak detector will be extended.

If you still need to test moist samples we shall explain to you in our next tip which leak rates you will get and where your detection limits are.

Överst på sidan

Kontakta oss Företaget Nyheter Event Site map Nyhetsbrev  

Välj hemsida :

Adixen Worldwide | Adixen China | Adixen Germany | Adixen France | Adixen India | Adixen Italy | Adixen Japan
Adixen Korea | Adixen Netherlands | Adixen Taiwan | Adixen UK | Adixen USA | Adixen Singapore
Copyright © 2010 Alcatel Vacuum Technology. All right reserved Legal Notices