sdack underwater rig (was Re: Ultimate Intervalometer - a script for shooting over a long duration -

So it's the USB data connection (the CHDKPTP connection), between the Pi and the Ixus that I'm unsure of.

There are also USB extenders that are actually protocol converters and use ethernet cable like https://www.amazon.com/WEme-Extender-Connection-Ethernet-Extention/dp/B01FHTCH68/
These are likely less than 100% compatible, so you'd definitely want to test with chdkptp and pi before committing. There are also appear to be some cheap ones that are just cable converters, which obviously wouldn't help.

You might be able to combine POE with something like this.

Also FWIW, these links are just meant to be examples of the type of thing I'm talking about, not a specific recommendation. I have no direct experience with any of them.

Thanks Reyalp, for splitting the thread.
Fortunately the depth we're looking at is less than 5mtrs.  So pressure is not an issue.

My main concern is choppy waters bashing the rigs around and possibly pulling on the cable glands.

The cameras want to have a fixed frame on a seed rack, to document plant growth over @ 30 day period. 

Two cameras, one on the "test rack", over which nutrient rich water will be pumped, from 450mtrs below and one on the 'control rack" with no additional nutrients.
Sdack

Fortunately the depth we're looking at is less than 5mtrs.  So pressure is not an issue.

That will definitely make it easier to seal.  But the pressure at a depth of 5 meters is about 1.5 atmospheres. So a leak in the bottom of your enclosure would allow the internal water level to go 1/3 of the way up the inside by the time the pressure equalized.

Post some pics of your rig when you get it all together?  It would be very cool to see.

Also, I agree with @reyalp 's comment :

FWIW, those black frames look suspiciously like hardware failure to me. 1/8 at f3.2, ISO800 should show *something* at 9:00 AM outdoors under almost any weather conditions at aus latitudes.

but I did not notice a response from you? If one of your cameras is indeed flakey then sending it 5m down where you can't get at it does not seem like a good idea?

Somewhat off topic, but you might find the openrov community has useful resources for this project. Even if you don't need an ROV, the materials and techniques may be helpful.

Regarding the black image, it would be interesting to see one of the original files if possible.

That will definitely make it easier to seal.  But the pressure at a depth of 5 meters is about 1.5 atmospheres. So a leak in the bottom of your enclosure would allow the internal water level to go 1/3 of the way up the inside by the time the pressure equalized.

Post some pics of your rig when you get it all together?  It would be very cool to see.

Will do

but I did not notice a response from you? If one of your cameras is indeed flakey then sending it 5m down where you can't get at it does not seem like a good idea?

Sorry for slow reply, I've been configuring and reconfiguring the underwater system and ordering parts like mad.  The project is also something of a moving target.
Current configuration looks like this

• One Raspberry Pi with 3/4G mobile data connection and Ethernet cable attached to a Router above surface on the base raft
• Two housings underwater each containing an Ixus 160 and an Ethernet cabled Raspberry Pi with an umbilical Power and Data over Etherne connetion to the router above

I'm currently looking at ways to pierce the cable through the housings without invoking floods

Regarding the black image, it would be interesting to see one of the original files if possible.

here's a link to the last, black frame which, on closer inspection, isn't 100% black..



and the first exposed frame



curious!

here's a link to the last, black frame which, on closer inspection, isn't 100% black..
...
curious!

Yeah, you can see the same scene if you crank the brightness enough. Attached +5 ev in raw therapee. Curious indeed.

I still don't think this can be explained by lighting if the exif exposure values are correct, but it's less obviously suggestive of hardware failure.

Thanks for looking at this...
I've had the camera running all afternoon but haven't had a black frame, despite a gloomy late afternoon light slipping into dusk.
Sdack