Can anyone comment as to the tolerance of the rails and disks within the enclosures?
I’m wanting to mount machine vision cameras which, once calibrated, absolutely can not move within the enclosures.
Do the rails and disks sit snug against the internal wall of the enclosure or would there need to be some extra component or a slightly larger made disk to stop any lateral movement?
I’m assuming the rails have some amount of clearance between the wall, so potentially a spacer block wuold be required to stop them from moving.
Has anyone managed to alter the disks in order to fit 4 rails for extra rigidity?
I’m leaning towards custom printing ,slightly modified disks, that would contact the enclosure radially at each axis and will accept a 4 rails setup.
The rails securely bolt to the flange. Neither the rails nor anything mounted to them touch the walls of the enclosure - this is because the ID along the length of the tube is larger than the ID of the sealing surface at either end!
By machine vision camera, do you mean stereo camera?
If the rails are bolted to one flange, and have a plate or disc between them, they are quite rigid. Using 4 rails could increase this further - attaching a 3D printed disc to this arrangement seems trivial. However you will not be able to make the rails/discs contact the inner diameter, as then you won’t be able to insert the assembly into the tube, past the sealing surface!
I think you’ll find 2 rails with a standard plate or disc plenty rigid enough, but without knowing what tolerances you can tolerate that’s impossible to say…
Yes it’s for a stereo pair in dome ports so the tolerance for movement is quite literally zero. I’ll have one camera per housing, so there’s more to go wrong than with a flat port stereo pair in one housing. They need to not move within the focal point of the domes and also not move relative to each other.
I think then, I’ll go with custom printed discs that will take 4 rails, and some means of locking them in place radially. 4 adjustable screws that extend to contact the ID of the housing, should suffice.
I’ll let you know how I get on once I’ve put them together.
I have no idea how you’d reach those screws that extend to contact the ID - I’d recommend against that as unnecessary or requiring of insanely tiny hands!
From my experience with stereo cameras, any misalignments calibrate out pretty well, but the orientation does need to stay consistent. I’d be more concerned with holding the enclosures externally securely than the internal components! 4 rails seems like overkill - they really are quite stiff on their own, and rock-solid once tied together with a mounting plate.
Having gone down this road several times, for building custom MV stereo cameras, my recommendation would be to mount the camera to the flange of the domed port.
This will allow you to restrict any movement resulting from mating clearances on the far endcap. You will see movement on the flanges to the tubes, resulting from external pressure at depth and internal pressure as the system heats while on-deck, pushing on the locking cord from each direction. Additionally, the flanges still have rotational play as the indexing tab of the flange is smaller than the notch in the tube to allow for ease of assembly.
All of that movement will translate directly to the camera and potentially add additional error to the resulting stereo data, if mounting on the back endcap.
When mounting directly to the dome flange, you can also index that endcap directly to the stereo mounting bar to eliminate any movement between the two cameras.
Yes, that’s what I have planned. Four rails on the dome cap flange should hopefully give enough stability.
Out of interest, have you come across any decent glass domes that fit these enclosures? I’m going to run with the standard acrylic initially but I will be using high res cameras, so I want to get the most out of them.
I think you’re missing what @RyanC is saying - rather than use RAILS at all, mount the camera to the end cap flagne that has the dome port or flat acrylic window, and don’t use rails to support it at all! This locks all the optical components together, and lets you mount whatever other electronics separately to RAILS?
Also worth noting that the domes are no longer acrylic, but polycarbonate! I think you’d be hard pressed to distinguish the difference between polycarbonate, glass, and acrylic in any case… assuming no scratches are present on any of them and the camera is correctly positioned!
For photogrammetry, assuring class level sub mm measurements, I think there is a difference between the materials. Optically, I’ll concede there likely isn’t a noticeable difference, but whilst the poly/acrylic might be robust, it will scratch where glass wont. I’ll definitely use them, but the moment we see scratches they’ll have to be replaced. That said, the robustness is a definite plus and at the cost, it may be beneficial to simply carry an inventory.
On the mounting, I see what Ryan is getting at.. the camera needs to be secure to the dome end not the rear, which is what I want to do, but the rails were how I was intending to do that. short rails on the dome end solely carrying the cameras, and short rails at the back carrying the switch.
If you have an idea how to mount the camera to the flange directly, I’m open to suggestion. The lens is fairly long, so the camera will need to sit back inside the housing. My plan is rails and mount it to a circular plate. It gives easy(ish) adjustment fore and aft for swapping out lenses.
The other option is a custom printed block bolted directly on the flange, but these would need to be lens specific to set the centre of aperture. Although.. as I type this, im realising that’s simply a case of adding millimetres to the depth of the block and that actually.. it would totally remove the possibility of error that you would get from lining it up on rails.
I think I’ve changed my mind whilst writing this. You’re entirely right. Print a custom mount!