how to validate that a unit is sealed 4

[dgowans]

Our products generally consist of a PCB assembly that gets sealed in a plastic enclosure (roughly 8"x4"x1.5") via a hot-plate welding operation. Our current method of verifying that the welding operation produced a fully sealed unit is to heat the sealed product for a period of time sufficient to build internal pressure that causes the product to expand - the 1.5" dimension increases. This is a time consuming process that we're trying to revise or eliminate. Additionally, our new products are continually getting smaller so even if we wanted to stick with the same test method the amount of "puff" seen in the new enclosures is pretty minimal, making evaluation difficult.

We have proven that we can perform the same test by evaluating the units in a vacuum bell, but the equipment is quite expensive and there is resistance to going this route.

Does anyone have any other suggestions as to how we can perform this testing? Obviously fast and inexpensive are preferable, but at least we don't need a solution immediately.

Thanks in advance.

 

[StewPad]

Helium (don’t have to be pure) and a group test

 

[btrueblood]

The problem with a group test is twofold.

1. a single bad unit may not register enough change in the group reading to reject the lot.

2. if small numbers of bad units do register within the lot test, you need a way to single out the bad units to find the leakers, or else reject the whole lot. This can be much more expensive.

Better to find a fast, efficient, single-item test. It can then be applied on a sampling basis...

 

[ornerynorsk]

Can a liquid sealer be applied post-weld as insurance to the process and just forego the leak testing once it has been validated, save for the occasional spot inspection?

 

[PratsA]

In my former job manufacturing mil-spec hermetic packages we used both the hot-liquid/bubble test and helium testing.

Helium testing is not effective with larger leaks, as most of the helium will escape before you transfer the part to the detector. Hence the use of the bubble test, using Galden to test it at 125°C.

So I'll add another voice suggesting acceptance criteria; this will make clear what testing is needed. You can check the mil-specs to help come up with one.

http://www.dscc.dla.mil/Programs/MilSpec/ListDocs.asp?BasicDoc=MIL-DTL-24308