Hello, we’ve been using an a blue robotics 4" enclosure (11.8" length), acrylic tube, acrylic dome 500m, and an aluminum end cap with 10 x M10 holes. We found that the enclosure seems to leak when the ROV descends past 15 ft, enough to fill up the dome within ~8 seconds. All provided O-rings and locking cords are used on both endcaps and we vacuumed checked every WLP including the PRV in a separate enclosure that we know seals well and found that all of them seal fine. When vacuum checking the enclosure on the surface it shows a good seal. We have been putting sealing tape on the outside near the endcaps and it had seemed to be working up until recently. We were running the ROV at a depth of ~20 ft for 13 minutes when the rear endcap just popped off the enclosure. Maybe the inside pressure became greater than the outside pressure and the cap came off, but we have a PRV installed so it should’ve vented the pressure right? Also all voltage and temps on the inside of the enclosure were normal at the time of failure so we dont see how the inside pressure could have been greater than the outside pressure. Could the endcap have come off because of the slow leak?
Hi @Rhys_Sellahewa -
That’s concerning! Can you share more details on your setup?
- Are the enclosures locking or non-locking?
- Are you using cable penetrators in the 10x M10 holes? What type, and what cables / loads are present during your failures? (If you’re using potted penetrators, this is almost certainly the source of your leak.)
- Have you done a positive pressure test with soapy liquid to find leak points?
- Have you removed the dome and inspected it for cracks in the area hidden by the dome retaining ring?
15-20 ft is really very little pressure, so something is definitely wrong! If water leaks into the enclosure, the internal pressure will increase until it is equivalent to the external pressure, or just below it - depending on the leak source…However if this occurs and you bring the enclosure to the surface, internal pressure may be higher internally - this is when a PRV should trigger (if the enclosures have locking cords properly installed.)
The enclosure is locking and we have been running it with the locking cords installed, all of the penetrators are wet link except for the PRV and they range in size from 5.5 mm to 8.5 mm. We have done a positive pressure test with soapy water and found the seal between the flange and the rear endcap to be leaking, after re seating the o-ring and re tightening the flange and endcap the positive pressure test showed no leak from there. We haven’t inspected the dome that much (I was just using it as a reference point in my previous post) but there are no visible cracks on the acrylic. Also we assume the the endcap popped off at the ~20 ft depth because we lost connection with it at that point and when we bought it up the endcap was already off.
Hi @Rhys_Sellahewa -
If the endcap popped off with the locking cord installed, that must have been some serious internal pressure! To confirm, you have a PRV and not a vent plug installed?
Are you carrying significant power through any of the WetLink penetrators, and did you have to modify any of them to accommodate your cable jacket (drilling them out to make them pass-through penetrators?)
It’s a bit concerning that you had a leak on the flange - this was coming from the face-seal o-ring located between the endcap plate and the flange? After fixing this, does the system pass vacuum tests?
What is on the other end of your cable penetrators, and is their any chance for water to enter the cable at this point? If it does, it will leak along the internal length of the cable and past the penetrator into your enclosure…
Hi @tony-white -
Yes it is a PRV, not a vent plug. There is one WLP which we had to drill out to accommodate the cable jacket to our 12 volt power line, after doing so that WLP passed vacuum tests. The leak on the flange was coming from the face seal o-ring, after inspecting it nothing looked out of the ordinary, all we did was re seat the o-ring and the system passed vacuum tests which was weird. We have had issues in the past where water would leak through the insulation of our Ethernet cable and enter the enclosure that way, we switched to a heavier duty weather proof cable hoping that would fix the issue (maybe its still leaking from the ethernet insulation?). All the other wires have waterproof cable jackets so it shouldn’t leak from there. It is concerning how much pressure was needed to pop off the endcap, it doesn’t seem like a leak would do it, how else could that much pressure build up?
Hi @Rhys_Sellahewa -
The face seal o-ring typically only can fail from not being in place when sealed against the end-cap plate, or using waaaay to much silicone grease.
If the o-ring was in this state, a fair amount of water may have entered the enclosure - but this should only cause the endcap flange to pop off if the system ascended rapidly? Do you have any pictures of the damaged locking cord?
Now that WetLink jacket-passthrough penetrators are available, this may be a better choice in the future…
If you had quite a lot of heat generated in the enclosure, this could also increase the internal pressure - try running the numbers to predict!
Hi @tony-white -
It is possible that water could have entered from the face seal o-ring but would that raise the pressure enough to pop off the endcap? We’ve kept everything preserved since the failure so I would check the status of that o-ring to see if that was the issue but I’m currently out of town. The ROV definitely didn’t ascend rapidly as it was stationary during the time of failure.
We did have 2 slightly corroded ESCs on board but the heat needed to raise the pressure enough to pop off the endcap is huge. After looking at the numbers I don’t think it is possible for those 2 ESCs to generate the amount of heat needed to pop off the endcap, especially when the entire system runs on 12 volt. The next plausible idea would be rapid gas production? Could corrosion in the ESCs cause an electrical fault where a large amount of gas is produced? Or is that only possible with an electrical fire? Maybe if the enclosure was leaking at the time and water got to those corroded ESC’s it would be enough to trigger an electrical fault where gas is produced?
Definitely will be using some jacket pass throughs in the future.