Watertight prints for high pressure environments
 
Notifications
Clear all

Watertight prints for high pressure environments  

  RSS
Justinas
(@justinas)
Eminent Member
Watertight prints for high pressure environments

Hi,

I am testing to see if it is feasible to 3D print parts for an underwater ROV that would dive to ~50 meter depth, and therefore be subjected up to 5 bars of water pressure. The parts I would print might be interfaces and adapters between acrylic tubes and domes, back plates with holes for potting wires, sealing caps, motor housings, etc. The following pic is an example of one of the simpler parts I would like to 3D print - an end cap for a acrylic tube with some holes for potting in cables, and relying on three 65mm rubber gaskets to make a watertight interface between the cap and the cylinder. Nevertheless, such a part should be watertight by itself, because if water can make it's way inside the part through one end, it will make it out the other.

Among the other, more complex parts would be replicating something like the WetLink Penetrator plug from Blue Robotics, using PLA/PETG for hard parts, and TPA for making these custom-shaped seals:

First, I want to establish if any of this is at all realistic. For testing 3D print resistance to high water pressures, I have built a very cheap hyperbaric water chamber using a water filter housing, a pressure gauge, a ball valve, and a thin syringe to purge air and pressurize the housing  with extra water. This housing is too small for a full ROV, but it is enough to check the watertightness of various plastics and print settings.

For these tests, I have printed several 30mm diameter, 50mm tall hollow cylinders (zero infill). I would then insert them into the chamber, pressurize it to 5 bars, and leave for an hour or two. Then, I would remove them, dry them with a towel, and shake to see if any water made it inside.

So far, I have tried PLA and PETG. Print settings were:

Profile: 0.20 Quality, with following overrides:

Infill: 0%
Perimeters: 4
Top layers: 7
Bottom layers: 7
Linear advance: 0.00
Extrusion multiplier: 1.1

Explanation: top and bottom layer count was increased to account for sag due to no infill support for the roof, and just to make sure that these end caps of the cylinder have a lesser chance to leak than the walls, which is what I want to test primarily in order to reduce the possible failure points. Linear advance was disabled to avoid gaps at the start and end of each layer - better have too much extrusion than too little. Extrusion multiplier was increased by 10% to really squish that plastic and fill any gaps in between the lines. I think that 10% is pretty much the limit - the extruder on my MK3S sometimes slips when filling in the top and bottom layers.

For PLA, I used 225°/60° for the first layer, 220°/60° for the rest. For PETG, I used 250°/90° for all layers. These values are somewhat higher than the norm, again, to make sure the plastic bonds really well.

Results:

PLA didn't fare well at all. It leaked both through the walls and the top/bottom, regardless of all that plastic squishing and numerous layers. As I pressurized the chamber to 5 bars, I could see bubbles forming on the PLA cylinder, and it sank to the bottom in less than a minute. It was 90% full of water.

PETG fared a bit better, but not great either. It leaked too, but at a much slower rate, and did not sink. After 2 hours, it was 75% full of water.

I am wondering if these results can be improved. Ideally, I want no leaks whatsoever, but I am not sure if the inherent porosity of 3D printed plastics can be overcome with the right settings. I have already reviewed this article, and multiple similar posts on this and other forums - that is where I got the idea to try the print settings I mentioned above. However, it seems these settings are only good for submerging parts just below the water surface, not 50 meters down.

I am aware that I could seal my 3D printed parts perfectly by applying a few thin coats of epoxy. However, that would throw the tolerances off, and it is damned difficult to apply epoxy in a even-thickness, predictable coating that I could account for in the CAD software, especially on delicate features like threads or other mating surfaces. I might have to use epoxy in the end, but I would really like to explore pure-printing options first.

Oh, and yes, I have already explored printing the parts hollow, and filling them with epoxy. I though that would retain the dimensions. Unfortunately, epoxy too leaks through the microscopic holes in the plastic, making it's way outside, and throwing off the geometry. Besides, it is difficult as heck to fill a 3D printed part with epoxy without leaving air bubbles inside.

I am wondering if flexibles, such as rubber-like TPA would fare better, and perhaps allow me to print custom-sized gaskets, so that I could avoid having to use epoxy on delicate geometry, and rely on these TPA gaskets to seal the critical interfaces, while epoxying everything else. 

 

I would very much appreciate your comments and suggestions. Thank you! 🙂

 

 

Posted : 13/10/2021 2:04 pm
Dan Rogers
(@dan-rogers)
Noble Member
I make lures waterproof(

I have some plug models that typically are trolled at 100 to 150 feet in salt water.   Making them so they don't fill with water and subsequently leach salt all over your tackle box as they dry out is important.

There are some techniques to make your prints waterproof.  First, the salt environment is harsh - so PETG or ABS are a must.  Then you need to turn off  the K value in custom G code (it makes prints smoother, but adds gaps into the print layers) .  Set it to K0 and mind the if statements so you get it everywhere.

Then you have to (absolute must) seal your part.  I use acrylic spray, or Epoxy coating.  The spray is easier to apply, but you need to hit every surface.  Epoxy is thicker and wears better, but harder and messier to apply.

If your containers are at local pressure you won't have to worry about structural pressures at these depths, but otherwise, you will need to make sure you ENGINEER the part to withstand the bathyspheric pressures - think super thick, with internal structure such as 3D honeycomb, or Cubic.  These have additional leak containment advantages - but once a leak starts at high pressure, it's going to inundate your part.

Posted : 13/10/2021 2:52 pm
Justinas
(@justinas)
Eminent Member
Topic starter answered:
I really want to avoid coating

 

Posted by: @dan-rogers

I have some plug models that typically are trolled at 100 to 150 feet in salt water.   Making them so they don't fill with water and subsequently leach salt all over your tackle box as they dry out is important.

There are some techniques to make your prints waterproof.  First, the salt environment is harsh - so PETG or ABS are a must.  Then you need to turn off  the K value in custom G code (it makes prints smoother, but adds gaps into the print layers) .  Set it to K0 and mind the if statements so you get it everywhere.

Then you have to (absolute must) seal your part.  I use acrylic spray, or Epoxy coating.  The spray is easier to apply, but you need to hit every surface.  Epoxy is thicker and wears better, but harder and messier to apply.

If your containers are at local pressure you won't have to worry about structural pressures at these depths, but otherwise, you will need to make sure you ENGINEER the part to withstand the bathyspheric pressures - think super thick, with internal structure such as 3D honeycomb, or Cubic.  These have additional leak containment advantages - but once a leak starts at high pressure, it's going to inundate your part.

Hi, Dan, thanks for your input. As I wrote in my post, I have already set K value to 0 in these tests. As for sealing, I would see this as the absolutely last resort - I want to explore the options to make 3D prints truly waterproof without additional coating, because they would throw off the tolerances of the mating parts.

As for structural strength - yes, that is very much on my mind. It should be easily doable. In these tests, I also made a 1 perimeter print, and even that didn't get crushed in the pressure chamber. Of course, there might be fatigue in the long run, but right now, leaks worry me much more.

Posted : 13/10/2021 3:26 pm
Dan Rogers
(@dan-rogers)
Noble Member
More perimeters - over extrude a little.

Up your perimeters a LOT - like 10.   Think about compartmentalized infill - the 3D shapes will help if one springs a leak.  At 4 perimeters at .2mm, I can get down to about 50 feet before I see salt encroachment on a round (bullet shaped) plug.  Acrylic spray helps a ton because it soaks into the layers - I give the lures 2 to 3 coats.  For structural fit areas I then remove as little acrylic as needed to get a good fit - the acrylic is still soaked into the layers in those spots - and with this approach I can run a lure for an hour at 150 feet in salt water without any leakage.

Posted : 13/10/2021 4:10 pm
BogdanH
(@bogdanh)
Honorable Member

@justinas
Thank you for sharing this! If someone would ask me if PETG is watertight (up to 5bar), I would probably say "yes" -I mean, 5bar isn't that much.

Could you observe where drops start to collect? Maybe at seams? Is kinda hard to believe that bond between layers would be that weak -just thinking aloud here.

[Mini+] [MK3S+BEAR]...
Posted : 13/10/2021 4:19 pm
Dan Rogers
(@dan-rogers)
Noble Member
Print yourself a PETG cup and fill it with water

It'll leak even sitting on the counter.

Posted : 13/10/2021 4:25 pm
Justinas
(@justinas)
Eminent Member
Topic starter answered:
Another idea
Posted by: @dan-rogers

Up your perimeters a LOT - like 10.   Think about compartmentalized infill - the 3D shapes will help if one springs a leak.  At 4 perimeters at .2mm, I can get down to about 50 feet before I see salt encroachment on a round (bullet shaped) plug.  Acrylic spray helps a ton because it soaks into the layers - I give the lures 2 to 3 coats.  For structural fit areas I then remove as little acrylic as needed to get a good fit - the acrylic is still soaked into the layers in those spots - and with this approach I can run a lure for an hour at 150 feet in salt water without any leakage.

 Alright, I will try 10 layers, both at the walls, and bottom/top, and will let you know of the results. Partially removing the coating might be very difficult with subtle geometry like threads... Then again, I am not sure how much does that acrylic spray add to the thickness (meaning, what is the minimum coat thickness that makes the part waterproof). Perhaps it is easier to use than brushing on epoxy.

Posted by: @bogdanh

@justinas
Thank you for sharing this! If someone would ask me if PETG is watertight (up to 5bar), I would probably say "yes" -I mean, 5bar isn't that much.

Could you observe where drops start to collect? Maybe at seams? Is kinda hard to believe that bond between layers would be that weak -just thinking aloud here.

The drops seem to be collecting at the seams in the walls. Then again, it is difficult to tell, because the drops (and air bubbles on the outside) are very large.

Posted by: @dan-rogers

It'll leak even sitting on the counter.

I have printed several flower vases in PLA with a vase mode - these don't seem to leak when filled... Are you saying PETG is even more leaky that PLA?

 

Another idea - what about ABS with acetone vapor smoothing? Is there a chance that would work?

Posted : 13/10/2021 4:33 pm
BogdanH
(@bogdanh)
Honorable Member

The drops seem to be collecting at the seams in the walls. Then again, it is difficult to tell

In Slicer (Print settings) set seam position to aligned. I know that's not good for keeping stuff watertight, but at least you'll maybe see if seam is weak point... is just an idea.

[Mini+] [MK3S+BEAR]...
Posted : 13/10/2021 5:20 pm
Stefan
(@stefan-3)
Eminent Member
alternative material

i have no own experience with this material, but if you don't have success with PETG, try this PP (Polypropylene).

This company claims "Watertight printing possible with only single wall prints" with their material:

https://formfutura.com/shop/product/centaur-pp-2789?category=460

 

Posted : 14/10/2021 5:45 pm
Justinas
(@justinas)
Eminent Member
Topic starter answered:
ABS doesn't work... Only PP left to try

So, I have tried ABS with vapor smoothing. The results were disappointing. Even after smoothing to a gloss finish (which really throws off the dimensions), and leaving it to dry for 24 hours, the ABS part still leaked. I printed another one, with 5 perimeters, 15 top/bottom layers and extra extrusion, then left it in acetone vapors for even longer. Then let it dry for 48 hours. And when I put that piece into the pressure chamber at 5 bars... It imploded 😀 It appears that acetone severely lowers the yield strength of ABS. Or perhaps I should have let it dry for a week or something. In any case, ABS is off.

Posted by: @stefan-3

i have no own experience with this material, but if you don't have success with PETG, try this PP (Polypropylene).

This company claims "Watertight printing possible with only single wall prints" with their material:

https://formfutura.com/shop/product/centaur-pp-2789?category=460

 

I have ordered Centaur PP as per your recommendation. It is now my last hope of achieving watertight prints.

Posted : 17/10/2021 11:47 am
towlerg
(@towlerg)
Noble Member
Reheat in sand

I seem to remember a YouTube video of a guy who was making 3d printed high pressure do hickies. He was embedding the parts in sand and heating, I assume the sand to retain the shape and the heat to recrystallize. Sorry can't remember details or even if it was YouTube.

Posted : 17/10/2021 12:22 pm
Justinas
(@justinas)
Eminent Member
Topic starter answered:
Yeah, tried it
Posted by: @towlerg

I seem to remember a YouTube video of a guy who was making 3d printed high pressure do hickies. He was embedding the parts in sand and heating, I assume the sand to retain the shape and the heat to recrystallize. Sorry can't remember details or even if it was YouTube.

I know this technique, I used it some time ago with very fine salt dust. It is difficult to get right, but even under ideal conditions, the parts shrink and deform quite unpredictably - it is much more difficult to keep accurate dimensions with this method than it would be with epoxy coating. 

Posted : 17/10/2021 12:36 pm
sergey
(@sergey)
Trusted Member
RE: watertight parts

i was able to make few watertight prints . the secret is to use high layer width and high extrusion multiplier . 

for 0.4 nozzle try 0.6 layer width or larger , start with 1.05 multiplier and go from there .

another thing is not to align the seam location . I'm not sure if prusaslicer improved in this regard , i remember when i tried to randomize the seam position with it , it would do the randomize thing , but the location of the seam was the same for all outer perimeters of the specific layer , which defeated the purpose . The seam should be different for each outer perimeter . if you set outer perimeters to 3 for example , the seam should be at different location for all 3 on the same layer . 

update : just checked , it's doing the same thing , aligning all outer perimeter seam at the same location . it seems the're nothing you can do about it , use a different slicer .

 

for material types i think anything other then pla should be good enough . 

This post was modified 2 months ago by sergey
...
Posted : 17/10/2021 1:08 pm
Justinas
(@justinas)
Eminent Member
Topic starter answered:
PP test results

I have finally received that Centaur PP and conducted some tests. I printed the same hollow cylinder, and put it under 7 bars of water pressure.

With default settings (Verbatim PP profile in PrusaSlicer), the cylinder leaked like a sieve through walls and top/bottom.

However, with these modifications:

Hotend temperature 230°C

Extrusion multiplier: 1.1

Bottom ant top layers: 7

Perimeters: 4

Seam position: random

Extrusion width: 0.6 (all)

Infil/perimeters overlap: 45%

Enable ironing: true

Retraction - Extra length on retract: 0.2

 

The resulting cylinder was actually watertight! It stayed under 7 bars of pressure overnight, and did not leak at all - it was the same weight in the morning, and shaking it, I couldn't hear any water inside.

However... The cylinder imploded from the pressure. It got squished. Even 4 perimeters didn't make it rigid enough to withstand the pressure. The deformation was plastic - the test piece did not return to it's original shape after I took it out of the chamber. However, I was able to restore it's shape with pliers. It remains questionable whether the air inside the cylinder was compressed, or leaked out during the implosion. In any case, there was no water inside.

So this means that PP is quite a viable plastic to achieve true watertightness, with 2 caveats:

1. It is very elastic, and it would be difficult to make structural components from it - like that bulkhead in my first post on this topic. The water would simply deform it and break the seal. Perhaps there is some way to engineer such a component to take advantage of the deformation (meaning that the component would expand and push harder on the O rings as it deforms in the middle), but this requires testing.

2. Epoxy doesn't stick to PP, at all. Meaning that there would be no way to further waterproof such a component in the critical areas. This also means that I couldn't pot the wires through the holes if the bulkhead was made from PP using epoxy. Perhaps the same could be done by melting a piece of PP filament with a small torch (lighter), and dropping it over the holes with wires, like a wax.

I will also investigate if it is possible to make a multi-material print, like, print most of the test piece from PLA (for structural strength), and do the several outer layers in PP (for water barrier). But I don't expect much - it is said that PP doesn't stick to anything but itself. It would probably delaminate immediately.

Posted by: @sergey

i was able to make few watertight prints . the secret is to use high layer width and high extrusion multiplier . 

for 0.4 nozzle try 0.6 layer width or larger , start with 1.05 multiplier and go from there .

another thing is not to align the seam location . I'm not sure if prusaslicer improved in this regard , i remember when i tried to randomize the seam position with it , it would do the randomize thing , but the location of the seam was the same for all outer perimeters of the specific layer , which defeated the purpose . The seam should be different for each outer perimeter . if you set outer perimeters to 3 for example , the seam should be at different location for all 3 on the same layer . 

update : just checked , it's doing the same thing , aligning all outer perimeter seam at the same location . it seems the're nothing you can do about it , use a different slicer .

 

for material types i think anything other then pla should be good enough . 

I have created a Github issue about this, and someone replied, saying that PrusaSlicer 2.4.0 beta 1 has this implemented - the seams on the same layer are staggered a little bit. Check it out: https://github.com/prusa3d/PrusaSlicer/issues/7225

Also, I'd like to share a link to great article by Jakub Koči, he did quite a lot of testing on 3D print waterproofing: https://blog.prusaprinters.org/watertight-3d-printing-part-2_53638/

It was very useful to me, perhaps someone else might benefit from it was well.

Posted : 02/11/2021 10:35 am
Share: