Hi Hari,

A simple test we do is pump the unit up slightly (maybe 1 psi), and check for bubbles.  It's like fixing a flat on a bicycle.  The positive internal pressure opens up and exposes any questionable seals.  Most of the leakage problems tend to be of the shallow water variety, which this test finds.  Deep water leaks are more rare, since the pressure tends to close up any cracks.

Otherwise the SOP is to lightly grease all O-rings and seals, and assemble the unit in the lab to prevent sand etc. from sticking to the greasy O-rings.

Thanks

Is it possible to remove the compressible air from your design further? One option is to fill the cavities with a dielectric fluid such as mineral oil. Another option is to have a non-rigid shell, or a section of membrane that can expand or compress depending on depth/pressure. 

While googling this I came across a nice open source project from South Australia incorporating both of these methods into an underwater camera design. They might have more ideas and experiences to share.

Abhijit and Hari joined us at our Variety Hour show to talk in more detail about this project - check it out to learn more about how it's progressing.