In the benthic imaging project we’ve been working on we’ve ended up using a tube mounted on the rear of the ROV for the camera and a second Raspberry Pi 4. It’s a little ungainly and driving around on the surface becomes a little sketchy, but it does work provided your camera + RP4 are compact enough.
I looked with interest at the new 1L rectangular enclosure and I’m wondering whether a hybrid box that utilised the lid as flange plate for a tube might open up a bunch of possibilities.
More compact than mounting RP4+camera end to end. Blocky, but only 175mm tall vs. 240mm+
Enclosure fits the RP4/5 easily, but is also big enough for the Nvidia Orin Nano! So that future proofs us when it comes time for doing any AI on the ROV itself.
Camera + Processor in same enclosure supports easy connection between the two. No pfaffing around with penetrators and cabling. Easily supports USB3 and MIPI/CSI options.
Questions for BR product folks, AEs (and others if interested):
Do you have a feel for whether this would work? What kind of depth rating do you think we could achieve with it? (we only need 100m or so).
That recessed indention in the lid of the rectangular enclosures? What’s it for? We’d need to mill a flat area for a 3" flange.
If we tried this, we’d have to re-anodise any machined surfaces. i.e. the Lid-Flange mating surface and a shortened 3" tube. Do you see any issues with that? Any alternatives to it?
If we stuff up can we get we purchase replacement lids?
Any other feedback appreciated. We’re investigating other mounting alternatives a the moment, including the payload skid, but rear mounting as a connectorised add-on works really well, and as above, having the camera+CPU in the same enclosure has real benefits.
It seems intuitively reasonable to me, although mounting the lid to the flange may require some ingenuity, especially if you want that connection to be separable[1]. The hole in the lid may also allow for some flex that rotates it away from the flange, so for extra strength you could add some internal supports that bridge from at/near the edge down to the bottom of the box.
From a rough comparison to our enclosure and end-cap dimensions I expect 100m would be fine as long as the mounting solution keeps the flange sufficiently pressed against the lid, without gaps or lateral sliding.
It’s not of structural significance:
Not sure on this one, although depending on your mounting and sealing solutions the machined surfaces may not be exposed.
From a quick look online it seems like re-anodisation is possible, but would typically be done for a whole part rather than just fresh surfaces (e.g. stripping the existing anodisation before applying a new anodisation layer).
That would be a question for sales@bluerobotics.com, although I imagine we’d allow that because if a support request came in for a lid with some kind of damage or manufacturing issue, we’d likely prefer to replace just the lid instead of the whole box.
Using screws would require adding un-sealed holes through the lid. ↩︎
Hi @PeterM -
Blue Robotics is not going to be able to provide you a lid that isn’t already anodized.
Would a general version of the lid (for either box) that incorporates a circular 3" or 4" port matching the standard end-caps of that size be an attractive alternative? You could use a dome or flatacrylic plate here to provide a window, and tighten 6 bolts into threads in the lid, sealing with an o-ring… just an idea that occurs to me!
Another alternative would be the same 3" or 4" geometry of an endcap, such that a water tight enclosure tube could be seated and secured in place with locking cord. Or maybe this is what you’re suggesting with your model?
Hi @tony-white,
Do you mean BR make available as a stock part?
We’re using an AVT alvium camera with the IMX304 chip (1.1”) and 8mm lens from CommonLands. The length of all that means we need a tube in addition to the box itself (see below). Ideally the tube needs to be detachable for lens adjustment, hence the idea of mounting a flange plate on a flat surface.
I’d initially had the flange plate mounted to the centre of the bottom of the 1L enclosure with the processing board (RP 4/5 or Orin Nano) mounted on the back of the lid. However, I had a rethink when I notice the hole pattern just touches the sealing o-ring (daw!), and also thought machining a lid might be easier. Matt at Bay Dynamics raised a good point though - the tube potentially might make the lid wiggle and move around a bit which would be bad for sealing. So might be back to hole for 3” flange plate in the bottom (2" too small and 4" feels too big).
I guess we could look to weld a flange plate and do away with the o-ring seal.
The anodising thing doesn’t look to be a big deal - looks like it can be removed and re-anodising done.
Hope that helps. Thanks for the interest!
I think the short answer is yes to this. I think a flat lid, without the label recess, with an option for a 3" port would be a useful add-on to the 1L product. In our case we need to add a tube because of the length of the camera + lens. So, if the 1L lid was machine for a port - i.e. had a goove for an O-ring, we’d need to adapt for a flange plate… But that might be as simple as cutting a large hole in a solid end plate. So you have two options:
1L Enclosure Lid + o-ring + view port. -OR- (for longer cameras)
1L Enclosure Lid + o-ring + flat plate (with hole cut in it) + o-ring + flange plate + tube + flange plate + view port.
Now, having said all that, Matt at Bay Dynamics had some concerns that lid mounting of a tube might cause the lid to be disturbed and that might effect the seal on it. So we’d flipped our design to modify the 1L enclosure not the lid. The challenge there is the 4 mounting holes in the bottom just touch the O-ring on the flange plate.
