How much pull-off force does a ball socket hold?
Between 30 and 100 N, depending on size. The table shows the minimum pull-off force that JW Winco (the Ganter catalogue) lists for its DIN 71802-size steel joints.
| Ball diameter | Thread | Catalogue min. pull-off (N) |
|---|---|---|
| 8 mm | M5 | 30 |
| 10 mm | M6 | 40 |
| 13 mm | M8 | 60 |
| 16 mm | M10 | 80 |
| 16 mm | M12 | 80 |
| 19 mm | M14×1.5 | 100 |
Axial joints from another maker list the same range: a minimum of 30 N to a minimum of 100 N, depending on size. Treat these as catalogue minimums for a new part, pulled along the stud axis. They are maker values, not DIN requirements.
What does that number mean?
It shows the strength of the grip between ball and socket. It does not show how much load the joint can carry. A catalogue pull-test figure is not an allowable working load, and a hand tug is not proof of load capacity. The number is for a new part and a straight pull. It says nothing about what happens under vibration, repeated shock or wear.
It also depends on which way the load acts, and that depends on the joint (Q1-07). On an angle joint the push and pull of the rod act across the stud, so pull-off is only a retention check against stray pulls. On an axial joint, whose stud is in line with the rod, a pull on the rod is a pull on the socket, so pull-off becomes the figure that matters.
Do not mix this up with a different problem. A rod end can lose its ball when the race around it wears and the ball is squeezed out. That is a different failure, covered in the syzmachine article What Is "Ball Pull-Out". This article is about the socket coming off the stud, which can happen to a new part.
What decides how much retention you need?
The largest force that tries to pull the socket off, multiplied by a safety factor.
The pulling force can come from several things added together: the weight of the link if the joint hangs, the inertia of the link as it speeds up and slows down (mass × acceleration), shock peaks at the end of a stroke, and prying, if a long part tips the socket sideways at the tilt limit.
Example: a 0.3 kg link hangs below its joint. Its weight is about 3 N. A peak acceleration of 20 m/s² adds about 6 N. The total is about 9 N. If your design review sets a factor of 3 (an example, not a rule), you need at least 27 N of retention. A 30 N joint just clears it; a clip would add margin. The right factor depends on what happens if the joint comes apart, so agree it with whoever is responsible for the design. This is a retention check; the strength check in Q1-07 uses a different, larger factor.
How do you test retention?
Test it in the direction your linkage actually pulls. Clamp the socket firmly, pull the stud away along the worst direction at a steady rate, and record the peak force and how the joint let go: sprang off, cracked or deformed. Test several parts and use the lowest value, and repeat after the joint has been swung through many cycles if it works that way.
Pull, push and sideways prying are different loads, and the joint is not equally strong in all three, so test each one separately.
Why do ball joints come apart in vibration or shock?
| Cause | What to check | Fix |
|---|---|---|
| Joint hangs with no safety clip | Is the joint below its stud? The abstract of DIN 71802:1992-04 recommends that clip-less angle joints not be mounted hanging | Use a clipped form, or change the layout |
| Pulling force above the catalogue retention | Shock peaks, not only steady load | Add a clip; it adds to the retaining force |
| Joint tilted past its limit | Does the stud neck touch the socket edge? Over-articulating the joint like this is a condition to avoid | Reduce the tilt (Q1-09) |
| Long part prying the socket | Is the joint carrying a long, loaded component? | Support the part, or use a larger joint |
| Loose thread | Does the whole joint come off, stud and all? | Lock the thread. This is not socket pull-off (Q1-23) |
After assembly, check that the socket is fully seated and any clip is in place and undamaged. A hand pull confirms only that the ball is engaged.
What can you do if the retention is too low?
In order of preference: change the layout so the load does not pull the socket off the stud; add a safety clip (C to CS, B to BS), which costs about 3° of tilt (Q1-02); move up a ball size, which raises the catalogue pull-off figure; or switch to a rod end with a bolt through the ball (Q1-26). Do not switch to an axial ball joint to fix this: its pull-off direction is the direction the rod pulls, so it makes the problem worse.
How much retention does a vehicle linkage socket need?
Vibration and shock are continuous there, and it is the shock peaks, not the steady weight, that pull a socket off. Redo the earlier example with vehicle-style numbers, as an illustration only. Take a 0.15 kg shift or throttle link whose inertia acts along the stud axis, and assume a 10 g shock peak (an assumed figure, not a measurement). The weight is about 1.5 N, the shock adds about 14.7 N, and the total is about 16.2 N. With a factor of 3 (again an example, not a rule) the retention needed is about 49 N. A 30 N or 40 N joint falls short, a 60 N joint clears it, and a clip adds margin.
The number is still a retention check. It says how hard a new socket holds the ball, not how much load the joint may carry; the strength check is the one in Q1-07. Tilt matters here too: a joint worked near its limit pries the socket sideways, and engine movement can use up tilt without anyone noticing (Q1-09).
A shift or throttle link is a control link, so the factor and the choice of joint are the vehicle maker's to set. SYZ's small ball joint page is the place to start on the product side.