How to: Leak detection 3

[Perd]

We are at a bit of a dead end here needing some ideas on which direction to pursue.

We are building a unit that is required to be water tight. It should be able to withstand being submerged in a meter of water for 30 minutes without water passing any of the seals.

The problem we are currently experiencing is this:

There are 26 seals in the unit with several having the potential to leak. The best method we currently have to detect where a leak is occurring is to apply positive pressure to the unit with helium at about 0.5 PSI. We then use a handheld helium detector and sniff around the unit to determine where the helium is leaking. Obvious problems here are if we set the detector to detect even the smallest amount of helium it detects false fails as very small amounts of helium pass through the seals. The other problem is the test is very time consuming and often does not catch all of the leaks the first or second time so we are spending manufacturing time pulling the units apart several times before they eventually seal.

The only other method we have used is applying positive pressure to the unit and dunking it in a tank of water to see where the leaks come from. This is not what we want to be doing at $3000+ per unit. The risk of damaging the internal electronics is high.

Anyone have any experience or ideas on this subject. We have been kicking around the idea of trying to use some type of thermal imaging camera but I am no chemist so we haven't really figured if any type of gas would work well to inject and use the thermal camera to detect leaks.

Thanks for any help/suggestions.

 

[GTstartup]

What about snoop?

 

[hydtools]

You are assuming that the seals are effective in both directions of pressure application. That pressurizing the interior of the box is the same as pressurizing the outside of the box, the real application. This theory may or may not be true.

I would establish a submersion test in 3 meters of water to verify what you actually want to know. Inspect the sample. Then establish testing samples of boxes to verify seal integrity on a production basis before installing the electronics. Maybe dye the water so that leaks are visible when the test box is opened.

Ted

 

[TheTick]

Perhaps a vacuum test? 1m H₂O << 1 atmosphere.

 

[Twoballcane]

Can you add a pressure gage to the unit? If you pressurize it and then cap it. Wait for 24hrs. If the pressure is the same, you have no leaks, if pressure is different, you have a leak some where.

To find leak, spray soap water on the seals and if you see bubbles, you found the leak.

good luck


Tobalcane
"If you avoid failure, you also avoid success."

 

[DennisP]

This sounds like an application for a helium leak detector. The device will pull a vacuum on the container and then helium is sprayed around the seals. The mass spectrometer in the leak detector will detect any helium passing the seals and also give you an idea of the severity of the leak. The unit we use is sold by Alcatel although there are other manufacturers available. They are generally capable of detecting a leak as small as 10e-7 to 10e-8 torr liters/second.

Dennis

 

[MintJulep]

26 seals and you can't put it together in the shop without leaks.

Is this thing supposed to be serviceable?

 

[Twoballcane]

"Obvious problems here are if we set the detector to detect even the smallest amount of helium it detects false fails as very small amounts of helium pass through the seals."

How do you know you are getting false fails?

The best way to find a leak if you are using helium is to first find a possible leak, then add a dab of alchole to the suspected area, then run the wond over the suspected area again. If the spectrometer did not pick anything up again, you got a leak.

What is the material used for the seals? If it is not Buena N, helum will pass right thru it.

Tobalcane
"If you avoid failure, you also avoid success."

 

[Perd]

I'll try to post more info to help with my question.

GTstartup: Not sure what you are referring to when you say "What about snoop"

hydtools: The tests have been done to determine what amount of pressure leak down is acceptable per the customer. The allowable leak down over 7 minutes is 0.2 PSI. Anything more and the unit is said to fail and subsequently have to be opened and reworked. Unfortunately the unity cannot be built without the electronics to determine if the seals are all good as some of the electronics make up the seals too.

TheTick and Twoballcane: We currently pull vacuum for the initial test. If the unity holds vacuum for 7 minutes it's said to be a good unit and no leak discovery is necessary. The problem is when the unity does not hold vacuum. It's not as easy as putting soap on the unit to see where it leaks. There are MANY points that can leak on the unit and a soap test will not work for several of them.

What I am kind of envisioning is in a perfect world a Thermal Camera would be able to detect helium and we could just pressurize the unit and see where the helium is leaking as it would show a nice difference compared to ambient temperature. I also kicked around the idea of potentially positive pressurizing the unit and the leak being able to displace something (material/dense gas/I don’t know) to where we could have the unit sitting on the bench in a case or whatever and the introduction of positive pressure into the unit would then displace whatever was around the unit so we could physically see where the thing is leaking.

 

[IRstuff]

"snoop" is, I think, a soap-like solution that's used to find gross leaks, not unlike the stuff the tire guy puts on the tire to find pressure leaks.

TTFN

FAQ731-376

 

[MiketheEngineer]

Pressurize it with hot air and use a thermal imager??

 

[IRstuff]

Unless it's rip-roaring leak, there won't be enough flow to cause a larger temperature change.

TTFN

FAQ731-376

 

[Perd]

"Unless it's rip-roaring leak, there won't be enough flow to cause a larger temperature change. " Which is what I kind of anticipated. I figured the delta would have to be pretty large between the inlet air and the ambient temperature.

Guess I’ll keep throwing darts until we can come up with a better method for leak detection.

 

[hacksaw]

clean dry air with ultrasound detectors works great

now if your spec calls for heliem testing that is a different story.

 

[Perd]

hacksaw - Nope, there is no requirement for method used to find the leaks. We have just been using the helium to this point as it is the best solution we know of. It's marginal at best for our application though which is why we are looking to improve the process.

Any good information you could link about the ultrasound detectors you mentioned?

Thanks

 

[MiketheEngineer]

Guess I better stick to Structural... Seemed like a good idea at the time...

 

[Compositepro]

0.2 psi in 7 minutes is a gross leak rate if you are trying to protect the electronics. Helium leak detectors are thousands of times more sensitive.

I can't think of any simpler test than dunking the box in a transparent tank of warm water. You do not need to pressurize your box. The air will expand as it gets warmer. You will test all the seals simultaneously and instantly see where the leak is.

 

[hydtools]

snoop:

http://www.krackeler.com/products/1182-Supplies/21500-Snoop-Leak-Detector.htm

Ted

 

[MikeHalloran]

Lubricate the seals with silicone vacuum grease.

Better, reduce the number of seals to ~<2.



Mike Halloran
Pembroke Pines, FL, USA

 

[handleman]

What is your production volume of these $3000 units? If you can determine what the acceptable leak rate is (everything leaks, as you have proven with your helium detector), you can do helium leak testing with a master. Steps:

1. Build vacuum chamber to enclose your unit
2. Pull vacuum in chamber
3. Evacuate the unit
4. Pressurize unit with helium
5. Sample gas from the chamber for helium.

You need to create a reference leak that leaks at the maximum allowed amount. Compare your detector reading on the reference leak to the detector reading on your unit. That will give you a quantifiable comparison, and it will check the sum of all leaks in the system. The downside is that it won't tell you which seal is leaking.

A similar process can be done with hydrogen (yes, hydrogen). The gas used is actually nitrogen/hydrogen mixture at a level below the LEL, so it is completely safe and less costly than helium gas. Check out

http://urltea.com/2wis