I've got a SeaFrogs 4-inch (100mm) dome that I haven't used in years. It doesn't have any extension, and is meant to be used with the old Sony 16mm f/2.8 pancake lens and VCL-ECF fisheye adapter. That combination is widely panned for its image quality, so I never spent any money on it. I tried it with a cheap manual fisheye, but the manual focus and fixed aperture was too much of a pain to manage, so it's been on a shelf since then. I've considered getting a Tokina 10-17mm, as I already have a Metabones IV, but that would require an extension, which SeaFrogs only sell for their newer 90mm port mount, as opposed to the 80mm that my housing and port use.

 

Well, now I've got an almost-new a6700, and a new SeaFrogs housing on the way, so I'm revisiting this idea. The new housing uses the 90mm port mount, so I figure it should be possible to 3d print a straight-walled cylinder that has the larger-diameter bayonet on the outside of one side, and the smaller-diameter one on the inside of the other. 

 

Question is, can a 3d printed part of reasonable thickness hold up to the pressure of 30-odd meter depth with some safety margin? I don't own a 3d printer myself, so I would have to use a service for it - what kind of settings should I request they use? I'm guessing that 100% infill is a given, but I have little experience with 3d printing, so I don't know what I don't know. 

 

Also, the bayonet lugs and o-ring grooves present surfaces that are perpendicular to the direction of printing - is this even possible to print? Do I need to include support structures in my design that are removed after printing? Or is it something that should be machined? Or maybe omit the lugs altogether? Just have o-ring friction hold the parts together on the camera table, and then vacuum (and water pressure) should apply much more force than the little plastic tabs would've been able to withstand anyway. 

 

Finally, what's the best way to estimate the required extension length? I understand that the entrance pupil of the lens should be aligned with the geometric center of the dome, and the dome appears to be a 180 degree hemisphere so its center should be at the base of the transparent element, but where is the entrance pupil? By how much should the lens extend into the dome? 

Probably not a good idea. It could possible be done, but port rings are generally aluminum for a good reason. Weight bearing and strength of filament are not going to be great for this. 

 

It might function but i would not trust it. The effort would out weigh the minimal cost of the ring.

 

Now gears are a different story...

Metal extension rings are typically used with metal housings. The extension rings for Ikelite, AOI and SeaFrogs are plastic. Of course this is, I'm guessing, injection-molded plastic rather than 3D printed, but still - how thick should the 3d printed ring be in order to resist the pressure? The symmetrical cylinder shape should give it additional strength, as opposed to a flat-sided box.

I use 3D printed ports and have done some pressure testing. At one point I was trying to get one to fail so I could better understand their limits, so I made one with intentionally thin walls of only 2.5mm (I normally use more like 5-6mm). Surprisingly, it withstood the limit of my pressure chamber which equated to 225ft/69m. So 3D printed ports/extensions can certainly be a viable option. However, FDM 3D prints are not watertight using standard printing settings, and figuring out the right settings can be difficult. There are other challenges too including lots of CAD work, dimensional tolerances, and as you mention the horizontal overhangs. It can all be figured out, but it will probably require quite a bit of work/experience before one gets really good results. 

Edited February 24, 20241 yr by Isaac Szabo

Designing things for external pressure is more problematic as the stresses are different.  Issac's thin port withstood the pressure however a thin walled cylinder like that is not inherently strong under external pressure forces and may have collapsed if you grabbed the cylinder for example.  Cylinders are not inherently strong under external pressure, only spheres are.  Similar to the trick  where you can stand on an aluminium drink can and it supports your weight until you flick the side walls and it collapses or the titan submersible where it was the cylinder that failed.

 

I'm sure it can be done but I think one of the issues is probably going to be getting reproducible strength and water tightness from your print and couple that with the need to get smooth o-ring sealing surfaces.  Properties of injection moulded plastics and metals are predictable, but I would suggest that similar to composites, 3D printed parts are not.

 

If you want something custom I would think that machining it out of aluminium would be a better proposition.  You can also machine acrylic or polycarbonate, if you could find tubing with the right dimensions to allow you to machine the part you need. 

 

regarding the entrance pupil, the 10-17mm is listed in this table of entrance pupil locations: 

 

https://wiki.panotools.org/Entrance_Pupil_Database#Third_party_lenses

1 hour ago, Chris Ross said:

Designing things for external pressure is more problematic as the stresses are different.  Issac's thin port withstood the pressure however a thin walled cylinder like that is not inherently strong under external pressure forces and may have collapsed if you grabbed the cylinder for example.  Cylinders are not inherently strong under external pressure, only spheres are.  Similar to the trick  where you can stand on an aluminium drink can and it supports your weight until you flick the side walls and it collapses or the titan submersible where it was the cylinder that failed.

 

Standing on an aluminum can is not similar to a pressure differential. Instead you can put your mouth over the opening of an empty can and suck the air out. The can walls will collapse with little effort.

 

1 hour ago, Chris Ross said:

I'm sure it can be done but I think one of the issues is probably going to be getting reproducible strength and water tightness from your print and couple that with the need to get smooth o-ring sealing surfaces.

 

Those are issues when trying to dial in the ideal print settings, but you only have to go through that process once. After you've figured them out you can apply them to all future projects.

 

1 hour ago, Chris Ross said:

If you want something custom I would think that machining it out of aluminium would be a better proposition.  You can also machine acrylic or polycarbonate, if you could find tubing with the right dimensions to allow you to machine the part you need. 

 

Of course those are viable options too. They're just more expensive/difficult. I'm guessing you might be able to buy a 3D printer for the cost of having a port machined (unless you can do the machining yourself).

Yeah, I don't have ready access to machining equipment, and I have no idea how to even design a part for manufacture with it - I'm sure there are significant differences between designing for 3D printing in plastic and designing for milling on a lathe or whatever.

 

Isaac, if you don't mind sharing, how thick did you make your RS 13mm adapter for SeaFrogs A9, and what did you print it from? I figure, with the 90mm port opening diameter, and 70mm lens outer diameter, I can do 5-6mm wall thickness on the bayonet itself and leave enough clearance for a zoom gear, then flare out to ~1cm walls on the outside portion of the cylinder - would that be enough? Although, thinking about it, I need to have an o-ring groove right at the thinnest part, thinning it further, but it will be supported by thicker material around it.

2 hours ago, Isaac Szabo said:

 

Standing on an aluminum can is not similar to a pressure differential. Instead you can put your mouth over the opening of an empty can and suck the air out. The can walls will collapse with little effort.

 

 

Those are issues when trying to dial in the ideal print settings, but you only have to go through that process once. After you've figured them out you can apply them to all future projects.

 

 

Of course those are viable options too. They're just more expensive/difficult. I'm guessing you might be able to buy a 3D printer for the cost of having a port machined (unless you can do the machining yourself).

Yes sucking the air out is probably a better representation, however as I understand it one of the issues is stability when materials are subject to compression, so for example if you press in lightly on the sides of the can it collapses even more easily.

 

I have never done any 3D printing but it seems to me that the properties of the material are quite dependent upon the way the individual beads of plastic are laid down - does that vary with material and with the type of printer used?  I've handled quite a few 3D printed items and they never struck me as something that would readily make an smooth surface to get an o-ring to seal upon.  I assume different materials and print setting get around that? 

 

On the o-ring groove issue I would think you could make the walls thicker around the groove if needed just by reducing the ID of the port. 

 

The Tokina is 70mm OD and your port opening is 90mm - so the you could make the port 100mm OD at the base which would give a 5mm ledge where you insert port into the housing.  My o-rings are 2mm OD with a groove ~1.5mm deep, so if you have 5mm wall thickness behind that, the ID would 100-2x(5+2+5) = 76mm ID and the top would be 80mm ID to take your port.  You could taper the port out to that ID up at the top potentially. 

 

That ID is about equal to the N85 port of the Nauticam ports.  Whether 76 mm ID is too small would depend on whether the zoom gear needs to fit through that hole or not.  If you don't use mounting lugs and rely upon vacuum to hold the port in you can make the bottom of the o-ring groove a little thicker as well. 

Actually, giving it some further thought, I could reduce my risks by using a stock extension ring for the barrel part and just making a very short ring plug to reduce its port-side opening diameter to 80mm. If either the 23mm or the 37mm ring is within a few mm of the needed extension length, it can even sit flush with the top of the extension ring, with very little surface exposed to water. This ring itself would have to take very little load, as it would be supported radially by the dome port bayonet, and axially by an internal ledge on the extension ring. Hmm... just gotta wait for the housing to get here so that I can take a proper measurement of how deep in the housing the camera is going to sit.

 

Edit: Basically something like this:

Edited February 25, 20241 yr by Barmaglot

1 hour ago, Barmaglot said:

Actually, giving it some further thought, I could reduce my risks by using a stock extension ring for the barrel part and just making a very short ring plug to reduce its port-side opening diameter to 80mm. If either the 23mm or the 37mm ring is within a few mm of the needed extension length, it can even sit flush with the top of the extension ring, with very little surface exposed to water. This ring itself would have to take very little load, as it would be supported radially by the dome port bayonet, and axially by an internal ledge on the extension ring. Hmm... just gotta wait for the housing to get here so that I can take a proper measurement of how deep in the housing the camera is going to sit.

 

Edit: Basically something like this:

That should work, do you need a lip to limit how far it will slide into the extension ring?

14 minutes ago, Chris Ross said:

That should work, do you need a lip to limit how far it will slide into the extension ring?

Photos of the rings on their website seem to show an internal ledge that would limit how far it could slide in:

 

 

Although if I end up needing something like 30mm of extension, might as well make it wider on the outside so that it gets supported by the rim of the extension ring.

7 hours ago, Barmaglot said:

Isaac, if you don't mind sharing, how thick did you make your RS 13mm adapter for SeaFrogs A9, and what did you print it from? I figure, with the 90mm port opening diameter, and 70mm lens outer diameter, I can do 5-6mm wall thickness on the bayonet itself and leave enough clearance for a zoom gear, then flare out to ~1cm walls on the outside portion of the cylinder - would that be enough? Although, thinking about it, I need to have an o-ring groove right at the thinnest part, thinning it further, but it will be supported by thicker material around it.

 

My 13mm port has 6mm thick walls. I had planned to pressure test different wall thicknesses to failure and plot the results so I could have actual data on how thick the walls needed to be, but even the thinnest port I tried survived the max pressure of the chamber. So I didn't have a meaningful reason for going with 6mm other than that it seemed like it was much more than adequate. I recently got a stronger pressure chamber and tested the port down to 145psi/100m/330ft with no issues. 

 

I use PETG. I've tried a number of other materials (including some more expensive/exotic), but PETG has a number of advantages over the others I tried including good print quality, water resistance, layer adhesion, toughness, price, etc.

 

I'm very busy at the moment fulfilling 13mm conversion orders, but I might be willing to print your design for you if you want since it might be difficult to find someone else who has the waterproof settings dialed in.

Edited February 25, 20241 yr by Isaac Szabo

4 hours ago, Chris Ross said:

I have never done any 3D printing but it seems to me that the properties of the material are quite dependent upon the way the individual beads of plastic are laid down - does that vary with material and with the type of printer used?  I've handled quite a few 3D printed items and they never struck me as something that would readily make an smooth surface to get an o-ring to seal upon.  I assume different materials and print setting get around that? 

 

Yes, there are many variables at play including material type, printer model, nozzle size, and many different print settings.

 

One important aspect of 3D printed o-ring groove walls is a small layer height. I use 0.10mm. Obviously, the walls won't be completely smooth due to the tiny ridges of the layers, but the compressed o-ring will span enough layers to make it watertight. For example, a compressed o-ring with a 1.5mm flat surface will span 15 layers. Using large/soft o-rings helps with this.

2 hours ago, Barmaglot said:

Photos of the rings on their website seem to show an internal ledge that would limit how far it could slide in:

 

 

Although if I end up needing something like 30mm of extension, might as well make it wider on the outside so that it gets supported by the rim of the extension ring.

The bottom ledge looks a little narrow, I would think having it sit flush on top of the ring would be best if the design allows.  What is your o-ring cross section?

 

The force pushing the dome and adapter combination into a 90mm ID port is equal to the differential pressure multiplied by the cross sectional area.  I calculate the inward force at 40m to be about 250kg.  This places the adapter material in compression pushing it inwards so you want to spread this over a reasonable cross sectional area.

 

3 minutes ago, Isaac Szabo said:

 

Yes, there are many variables at play including material type, printer model, nozzle size, and many different print settings.

 

One important aspect of 3D printed o-ring groove walls is a small layer height. I use 0.10mm. Obviously, the walls won't be completely smooth due to the tiny ridges of the layers, but the compressed o-ring will span enough layers to make it watertight. For example, a compressed o-ring with a 1.5mm flat surface will span 15 layers. Using large/soft o-rings helps with this.

Right, i see how that would help - do you need extra care placing the o-ring compared to the smooth surface of something like an aluminium port?  Seems like it might drag a little on the grooves?  or does it work out OK in practice? 

16 minutes ago, Chris Ross said:

Right, i see how that would help - do you need extra care placing the o-ring compared to the smooth surface of something like an aluminium port?  Seems like it might drag a little on the grooves?  or does it work out OK in practice? 

 

I'm not sure I'm completely following you here, but no I haven't noticed that any extra care is needed. The ridges from 0.10mm layers are not very significant. 

1 hour ago, Isaac Szabo said:

 

Yes, there are many variables at play including material type, printer model, nozzle size, and many different print settings.

 

One important aspect of 3D printed o-ring groove walls is a small layer height. I use 0.10mm. Obviously, the walls won't be completely smooth due to the tiny ridges of the layers, but the compressed o-ring will span enough layers to make it watertight. For example, a compressed o-ring with a 1.5mm flat surface will span 15 layers. Using large/soft o-rings helps with this.

Did you manage to get the settings dialed in to have the parts waterproof on their own? I was under the impression that covering the parts with epoxy was the way to go for proper waterproofing. 

 

Do you smooth out the mating surfaces for the O rings in some post-processing, or do you fully rely on the soft O-ring to fill the gaps? 

59 minutes ago, Chris Ross said:

The bottom ledge looks a little narrow, I would think having it sit flush on top of the ring would be best if the design allows.  What is your o-ring cross section?

The SeaFrogs o-ring for 90mm ports is specified as 90x3mm, so I'm guessing a 2mm deep x 3mm wide groove should do it.

 

1 hour ago, Isaac Szabo said:

 

My 13mm port has 6mm thick walls. I had planned to pressure test different wall thicknesses to failure and plot the results so I could have actual data on how thick the walls needed to be, but even the thinnest port I tried survived the max pressure of the chamber. So I didn't have a meaningful reason for going with 6mm other than that it seemed like it was much more than adequate. I recently got a stronger pressure chamber and tested the port down to 145psi/100m/330ft with no issues. 

 

I use PETG. I've tried a number of other materials (including some more expensive/exotic), but PETG has a number of advantages over the others I tried including good print quality, water resistance, layer adhesion, toughness, price, etc.

 

I'm very busy at the moment fulfilling 13mm conversion orders, but I might be willing to print your design for you if you want since it might be difficult to find someone else who has the waterproof settings dialed in.

Thank you; this is very helpful. The fact that 6mm thick walls easily stood up to 10 bars of pressure is quite reassuring. If and when I go forward with this, I think I will use a local service for draft prints until I get the fit perfect, then get in touch for a final print.

17 hours ago, Barmaglot said:

I've got a SeaFrogs 4-inch (100mm) dome that I haven't used in years. It doesn't have any extension, and is meant to be used with the old Sony 16mm f/2.8 pancake lens and VCL-ECF fisheye adapter. That combination is widely panned for its image quality, so I never spent any money on it. I tried it with a cheap manual fisheye, but the manual focus and fixed aperture was too much of a pain to manage, so it's been on a shelf since then. I've considered getting a Tokina 10-17mm, as I already have a Metabones IV, but that would require an extension, which SeaFrogs only sell for their newer 90mm port mount, as opposed to the 80mm that my housing and port use.

 

Well, now I've got an almost-new a6700, and a new SeaFrogs housing on the way, so I'm revisiting this idea. The new housing uses the 90mm port mount, so I figure it should be possible to 3d print a straight-walled cylinder that has the larger-diameter bayonet on the outside of one side, and the smaller-diameter one on the inside of the other. 

 

Question is, can a 3d printed part of reasonable thickness hold up to the pressure of 30-odd meter depth with some safety margin? I don't own a 3d printer myself, so I would have to use a service for it - what kind of settings should I request they use? I'm guessing that 100% infill is a given, but I have little experience with 3d printing, so I don't know what I don't know. 

 

Also, the bayonet lugs and o-ring grooves present surfaces that are perpendicular to the direction of printing - is this even possible to print? Do I need to include support structures in my design that are removed after printing? Or is it something that should be machined? Or maybe omit the lugs altogether? Just have o-ring friction hold the parts together on the camera table, and then vacuum (and water pressure) should apply much more force than the little plastic tabs would've been able to withstand anyway. 

 

Finally, what's the best way to estimate the required extension length? I understand that the entrance pupil of the lens should be aligned with the geometric center of the dome, and the dome appears to be a 180 degree hemisphere so its center should be at the base of the transparent element, but where is the entrance pupil? By how much should the lens extend into the dome? 

Before you embark on this DIY journey: The Tokina 10-17 works well in the standard salted line 6" dome without any modifications.  That's a relatively cheap solution.

Seafrogs might sell you the base part of the dome for your newer 90mm housing if you contact them. Don't know if there is one that has similar length but it might be much easier.

I have done on myself several dome ports, zoom gears, flip brackets for macro lenses....What I have found about 3D printing is no piece which have to be waterproff must be done on 3D printing, I don't trust on it. So when I am planning to make for example a new dome port I use to 3D printing as a prototipe provider. I build the whole piece in 3D printing and make all the tests needed in dry. After that, I divide the dome port in 3 pieces, bayonet fitting, cylinder and dome. I order the dome to a workshop which makes it the way I need and really cheap for my surprise, then I take the prototype to the turner and ask him to make an exact model in Delrin (way cheaper than alluminium) and I finally print in 3D de bayonet fitting for the housing. Then I put it all together, tipically the bayonet screwed to the body and the dome glued with polyurethane sealant. The last 4" dome port I did this way it's now 2 years that I'm diving with. And it costed me about 150 euros.

17 minutes ago, Robin.snapshots said:

Before you embark on this DIY journey: The Tokina 10-17 works well in the standard salted line 6" dome without any modifications.  That's a relatively cheap solution.

I know, and I actually have this port for the A6xxx Salted Line housing, but I am interested in the Tokina 10-17mm for its CFWA capabilities, and this requires a small port.

 

18 minutes ago, Robin.snapshots said:

Seafrogs might sell you the base part of the dome for your newer 90mm housing if you contact them. Don't know if there is one that has similar length but it might be much easier.

Unlike the 6-inch and 8-inch domes, the 4-inch dome does not have a removable extension barrel that can be swapped; one would need to replace the entire port base as it is molded as a single piece with the bayonet mount. It is also out of production - I recall reading mentions that it was produced for SeaFrogs by another manufacturer who went out of business.

17 hours ago, Robin.snapshots said:

Did you manage to get the settings dialed in to have the parts waterproof on their own? I was under the impression that covering the parts with epoxy was the way to go for proper waterproofing. 

 

Yes, the parts are waterproof on their own. If you were to use normal print settings the parts wouldn't be waterproof, and you would have to do something like cover them in epoxy to make them waterproof. 

 

17 hours ago, Robin.snapshots said:

Do you smooth out the mating surfaces for the O rings in some post-processing, or do you fully rely on the soft O-ring to fill the gaps? 

 

I used to smooth out the o-ring surfaces with epoxy or by turning them on a lathe, but then I discovered that it wasn't necessary. 

20 hours ago, Barmaglot said:

The SeaFrogs o-ring for 90mm ports is specified as 90x3mm, so I'm guessing a 2mm deep x 3mm wide groove should do it.

 

 

Looking at my Nauticam ports the O-ring groove seems to be 1.5mm deep and the o-ring 2.0 mm OD.  Getting the gap and groove depth right is the key to compressing the o-ring enough to ensure it seals and also so it won't extrude if the support below the o-ring is insufficient.  I would measure and copy what Sea Frogs does with their o-rings.  If you have a set of calipers you can measure your existing ports to replicate this on your extension.

Yeah, that's probably the best way to do it. I do have a set of calipers, so I can measure both width and depth of the existing o-ring groove and replicate it. Fingers crossed that the housing gets here before I leave for a Myanmar liveaboard trip on March 4th - I ordered it on February 7th, but due to Chinese New Year it didn't ship until the 20th, and right now tracking shows it as 'Leaving from departure country/region' yesterday.

I personally would not trust any FDM part underwater. I would only be trusting an SLA part, which once you get your part designed, you should be able to get printed at most any print house. 

 

Fusion 360 is probably the simplest free modeling tool, and with youtube and some calipers, you might be able to model both connections.