Vacuum Leak Testing 5

[mechintern]

I am performing a leak test where an encapsulated electronic device is submerged about 10 cm under water. A vacuum pump then pumps air out of the seal chamber to 20inHG. I was wondering if this is a better test than submerging the unit under 2 meters of water. Basically we are dealing with a pressure issue because water pressure is a function of depth. Is there any correlation between water pressure and air pressure. Is the suction force significant enough on the unit to induce leaking at 20inHG or am I better off throwing it in a dunk tank? Any comments are greatly apreciated.

 

[IFRs]

I would pump helium into it and use a helium detector. The one I used was made by Varian.

 

[IRstuff]

I'd agree with IFRs; helium leak testing is the standard method for testing hermetical seals.

The device is "bombed" by high pressure helium. If there is a leak, the helium will get inside the device. The device is then placed under vacuum and the mass-spectrometer detects the presence of helium and what the leak rate is. This is all done in a single chamber and is automatic.

TTFN

 

[ChrisSvitak]

One thing to keep in mind when doing MSLT is the "time" issue. You don't want to keep parts out of the bombing chamber too long or you willl miss a gross leak. In most instances or at least per MIL-STD-202G you should do a bubble test in addition to the MSLT.
I do agree with IRstuff that helium leak testing is the easiest and quickest way to go.

 

[IFRs]

You can also pump helium into the object, seal the port and use a detector on the outside. They are extremenly sensitive and quick acting instruments

 

[scooter911]

From my days in scuba I believe it's 33feet of salt water and 34 feet of fresh water are equal to 14.7 psi. Looked it up it's actually 33.9 feet, or 1 foot of water is equal to .4335 psi. One inch of mercury is equal to .4912 psi, with 1 atm equal to 29.92 inches of mercury. All this is on the web, type in conversions. By the way, a cheap He detector will run you in the $15K range, but if there's a leak the detector can find it.

Good Luck

 

[IFRs]

scooter911 - Varian makes (or used to) a hand held He detector for about $7000. We used one for a while a few years ago. I think they are used to detect leaks in underground piping, and things like that.

 

[scooter911]

IFR's correct, it looks like Varian sell one called a Helitest most likely in the $7K range.

 

[profengmen]

It looks to me like 34 feet of plastic tubing and some water is less costly than $7,000.

 

[IRstuff]

A water submersion test will only find gross leaks, either by bubble observation or actual detection of moisture within the enclosure after the duration of the test.

Hence the problem. You have only a finite amount of time to run the test. A slow leak, such as those caught by a helium leak test, may not fully develop into a problem until months or years or submersion have elapsed.

The helium leak test is the only way to determine if you have subtle process problems that will bite you later. If you pass a brute force submersion test, you won't know if there are latent problems. If you do pass a helium leak test, you know that you have a solid construction with no hidden problems.

TTFN

 

[profengmen]

helium leak testing is the standard method for testing hermetical seals.

If this is in fact a standard, then it should be followed.

I'd be curious to know the application; Is this device to prevent leaking from the outside or the inside?

If it is to prevent leaking from the outside pressure, then the test needs to simulate outside forces; the geometry of the device is to be in compression, and simulating an internal pressure will exert forces that the device might not be designed for and will give false results.

If the device is to hold pressure from the inside, then exerting external forces on it is not a realistic simulation.

 

[IRstuff]

The normal helium "bomb" essentially simulates an external overpressure.

Again, the issue is one of test time. If you have 10 years or so and several dozen parts to run the leak test, then you can do it with the existing test that the poster is using. Naturally, you'd need to cycle temperature/voltage to simulate actual operating conditions. This is a VERY expensive approach.

Alternately, you need to do it differently and make some simplifying assumptions. A helium leak test will determine if the fundamental, as-built, encapsulation is hermetic.

If so, you can then move on to determine its performance over life. This can usually be simulated with temp-humidity test similar to those run for MIL-STD-883 components. A helium leak test after a temp-humidity test will allow you to verify that the encapsulation will protect the interior over some lifetime. Additionally, since a temp-humidity test will often cause corrosion if the moisture gets in, so dissection and inspection can also tell you what's happened.

TTFN