Loosening Torque is 30% Lower Than Tightening Torque 5

[filterengineer]

Hi Guys,

we're testing some tightening/loosening torques on a diesel fuel filter we designed here at my work. the can is 6061 billet aluminum and the head that it threads into is cast 6060 aluminum. they have stub ACME threads. when tightening the can onto the head it seats on an O-ring which is compressed axially between the head and the can. i'm wondering what would be the reason that the loosening torque that much lower than the tightening torque. we torqued the heads on up to 237Nm which is the customers specs. on average the loosening torque is about 205-210Nm. Any thoughts guys? Thanks!

 

[israelkk]

It is same as sliding uphill or sliding downhill. The thread helix angle when is in the tightening direction is in the positive direction (and resist the tightening) while in the opposite direction it is less resisting. You need to look at mechanical engineering equations for power screw jack equations for example.

 

[filterengineer]

ok, thanks for the info israelKK

 

[ornerynorsk]

Check your threads. The surface finish could be causing excess friction to tightening which would result in a markedly different loosening torque. Too coarse of a thread pitch will do the same IE the helix angle is approaching the point where it is no longer self-holding.

It is better to have enough ideas for some of them to be wrong, than to be always right by having no ideas at all.

 

[filterengineer]

yeah, we have a hard anodized coating on the billet can and bare threads on the filter head. the thread pitch is absolutely self locking as well. the fact that the loosening torque is lower than tightening is not necessarily a bad thing, I'm just curious as to the explanation.

 

[IRstuff]

Couldn't part of the reason be that your oring is now compressed and set? Or, you've managed to grind away burrs or imperfections in the tightening process.

What happens when you repeat the process?

TTFN
I can do absolutely anything. I'm an expert!
faq731-376 forum1529

 

[filterengineer]

IR, I could see the o-ring being already compressed. we did a number of tests with two different cans and the numbers were right in the same ballpark for every test

 

[tbuelna]

I agree with IRstuff. During installation, the axial force required to compress the o-ring plus the torque moment from high sliding friction at the opposing o-ring face contacts can add significantly to prevailing torque in the screw thread joint. Especially if the o-ring has a large circumference, is compressed a significant amount, or has high material stiffness characteristics. After installation, the o-ring can experience some compression set which will reduce the amount it contributes to prevailing torque in the screw thread connection.

Is there metal-to-metal contact between the filter can and head faces after the thread connection is torqued? Or is there a gap between them, with contact only at the o-ring?

 

[filterengineer]

tbuelna, there is a shoulder on the can which lands on a face of the filter head. the contact there was designed to stop the rotation on the can after a specified amount of compression was achieved on the O-ring. there were issues with a previous design where over compression of the O-ring seal created leaks. so the new design can have a much lower torque (for seating and breakaway). although the customer still needs to test the part to ensure that it won't come loose under vibration.

 

[IRstuff]

If the main issue is antivibration, then why not something more positive, like a stake or cotter pin arrangement? Just counting on friction might not be sufficient.

TTFN
I can do absolutely anything. I'm an expert!
faq731-376 forum1529

 

[filterengineer]

they've relied on friction to hold the can back for quite some time now and it hasn't been an issue. They just want to do a vibe test since the loosening torque is lower on this new design. I personally don't see any problems arising from it since its a fuel filter and shouldn't see much vibration at all. At least not compared to the transmission filter, which is of a very similar design and has no issues with loosening when exposed to vibration.

 

[Compositepro]

A simple free free-body diagram of a section through the thread contact surfaces will show that tightening torque not only applies a shear force at the contact surface but also a normal force that will add to friction. For loosening torque the normal force reduces reduces friction.
In this case, I believe that a good analogy is that it takes more force to drive a wedge into a crack than it does to drive one out of a crack. In fact, without friction, threads and wedges would always self loosen instantly.

 

[3DDave]

The oddity is that tightening friction is usually dynamic and loosening friction is static, which often makes items harder to loosen than tighten. Perhaps this is a really well lubricated item where static friction is not much different from the dynamic.

 

[tbuelna]

Good point. The peak wrenching torque should normally occur at the instant the preloaded threads break free when being loosened. Big difference between static and dynamic frictions.

 

[Tmoose]

Similar to what IRstuff said, What happens if the canister is tightened and loosened 6 times to burnish down the high spots before the final or test assembly?

What is the operating pressure?

Is it truly an o-ring? The previous design that could be accidentally (or by design!) overtorqued and damage the o-ring sounds kind of risky. Note conventional oil filters use an unrestrained flat elastomeric seal, and the tightening spec is some fraction of a turn after first contact. As there is no pressure balancing effect the physical strength of the elastomer and the initial compression must do all the work.
Under significant pressure o-rings achieve a deform and create a considerable amount of pressure balancing, so an 0-ring in a well finished, correctly dimensioned "gland" can seal 1000 psi+ pressures "automatically" with quite modest initial compression.
When under pressure I would expect an 0-ring to exert a bunch more force on the can and filter head and also provide a bunch of additional resistance to loosening for free.

 

[Spoonful]

The oddity is that tightening friction is usually dynamic and loosening friction is static, which often makes items harder to loosen than tighten

This is only true if the tightening process is in one continues motion.