Hi,
I would like to know some reference to tighten end caps and penetrators to my ROV, has anyone used a torque wrench or something similar to determine empirically how much should things be tightened on the ROV?
Are they O-ring sealed parts? I think the correct way to tighten O-ring fittings is to tighten it until it’s snug, and then another little bit (1/16 - 1/8 of a turn). I have been criticized a few times in the past since I have a tendency to overtighten O-ring parts and then the O-ring gets extruded or is crushed.
Hi @tguajard,
There are various screw tightness indications in our BlueROV2 assembly guide, although they’re based on the specific parts, or the tools that come with the ROV kit. As @btrue mentioned, the advice there for penetrators is
It’s worth noting that our new WetLink penetrators do actually have tightening-torque specifications, and since the part that mates with the end caps is the same (material + size + o-ring) then that specification (3.5Nm) likely applies for our M10 potted penetrators as well.
I brought this question up internally and was told the end-cap bolt tightness is likely just the relevant one for stainless M3 bolts, but the relevant design engineer will be able to confirm next week. The best resource I could find for that is this tightening torques table. The closest match there was DIN 7984 - Hexagon socket thin head cap screws, which applies to the A4-70 bolts that we use, so that suggests a max tightening torque of 0.6Nm. I’ll try to get back with confirmation next week on whether that’s appropriate.
Note that the screw tightness for the various frame components is determined by things like whether the screws are indenting into the frame (as covered in the guide I linked to), so those will likely have different torque specifications that are more difficult to estimate effectively.
Thanks for the answer, if you could confirm those numbers it will be great.
Regards
I personally wouldn’t worry too much about the torque specs as long as your enclosure(s) pass your vacuum test. In my limited experience, It’s not going to leak if it passes the vacuum test that is roughly comparable to the operational depth.
We had a leak some weeks ago, and we don’t know why as everything seemed tight enough, we think it may have been o-ring damage from too much force. That’s why we want to make some kind of manual with references and values for everything.
1 Like
I’ve spoken with the relevant engineer and been told that those recommendations (3.5Nm for the penetrators, 0.6Nm for the end-cap bolts) should be fine.
On this front,
It’s a face seal o-ring, so too much compression is impossible
You’d break the screws as the aluminium bottoms out
My understanding is that o-ring over-compression is more of a design consideration than an installation one - even more so in face-seal cases, where the components being connected are meant to be solidly pressed against each other, rather than having a gap.
When installing it’s important to ensure the o-rings and surfaces are clean and correctly lubricated, and in the right position (it shouldn’t pinch on a corner - gradually tightening the bolts in a star pattern can help avoid this). If a leak occurs it’s very likely that it’s from some dust or hair or a small cut or gouge in the o-ring that has compromised the seal, rather than an issue with the amount of compression (which has been designed for).
If it’s any further peace of mind, there’s a blog post here which covers o-ring squeeze ratios, and recommends 25% squeeze for static applications of NBR70 o-ring seals. Our face o-ring is listed as AS568 size 154, which according to here has a cross-sectional diameter (CS) of 2.62mm \pm0.08 mm. In the provided CAD model our 4" Flange has a gland depth of 2.06mm. That gives
\begin{align}
\%_{squeeze} &= \frac{CS - d_{gland}}{d_{gland}}\cdot 100\%\\
&= \frac{2.62\pm0.08 - 2.06}{2.06}\cdot 100\%\\
&= (27.2\pm3.9)\%
\end{align}
which is very close to the recommended 25%, and also fits well within the ‘rule of thumb’ presented here for 10-40% compression of static seals.
Hello,
As instructed in the operation manual (here), we would like to check these following screws:
- “Tighten the M5x16 screws that hold the frame to the center and bottom panels using the short part of the M3 hex key as the handle or an M3 hex driver.
- Tighten the M3x12 screws that hold the back-end caps to the flange seals using the M2.5 hex driver.
- Tighten the M3x16 screws that hold the clips to the electronics enclosure.
- Tighten the M3x12 screws that hold the dome and the front battery end cap to the flange seals using the M2.5 hex driver.”
Is it possible to have a full table of the torque recommendations?
Thanks
Fiona
Hi @Eloi,
These are all M3 A4-70 (316-stainless) screws, so would have a max tightening torque of 0.6Nm, as discussed above.
The resource I linked to in my original comment specifies a maximum torque of 2.5Nm for M5 screws, so that at least provides an upper limit. Qualitatively, our BlueROV2 Assembly Guide recommends
I expect that would occur at a bit less than the maximum, given the HDPE frame material isn’t as rigid as the metal end cap and flange components, and it’s sufficient with just a slight indentation.
We don’t have recommended torques for many of the screws, particularly given the internal screws are mostly for positioning components rather than bearing much load. I expect that a table of torque recommendations for each screw would provide minimal benefit to users, and take quite some time to compile and document.
That said, if you’re very keen and have some good reasons for it I can raise it for you internally to see what the others think 
Many thanks for your reply.
We are always very keen to get accurate information. It would be great to have an exhaustive table of torques such as this one from Nortek: Torque procedures for screws in threaded plastic – Nortek Support Center (nortekgroup.com)
Thanks for the reference. I’ve passed on your request internally and will let you know what the response is
If we are able to add it to our testing schedule, it will likely not be until at least a few months away.
Update on this,