OT/General Ball Jointing 101 (Pic heavy!)

I have been trying to work out the fundamentals of WHY ball joints work the way they do and how the jointing systems can be improved and modified in logical ways. I thought I would share the information with you guys - it's useful for me, and is really helping me wrap my own mind around all of the angles and options!

I will continue updating this thread with info as I explore different options and layouts - feel free to add your own breakthroughs too! I would love for folks to share photos of existing joints, especially interesting ones, so we can all get a better idea of the mechanics involved!

Part 1: The Basic Ball
Part 2: Double Joints
Part 3: Off-Center Pivots & Uneven Cuts
Part 4: Depth & Roundness
Part 5: Shims
Part 6: Cut Angle
Part 7: Friction & Surface Area
Part 8: Stringing Channels
Part 9: Stops/Locks
Part 10: Guides
Part 11: Non-Ball Joints
Part 12: Range of Motion in Human Joints

Note that any listed above that are not linked are not yet complete - I'm mainly listing this here to keep notes for myself so I can remember all the things I want to go over. I'll link each part as it's complete to make the individual entries easy to browse if you don't want to scroll through the whole thread - but please don't let any plans I have to cover these things stop you from adding your own information!

Other really helpful info can be found in twigling's workshop notes - including many illustrations and really good advice.

 

Part 1: The Basic Ball

This circle represents a balljoint, as seen from the side. This simple sphere is what all the joints are built on, to provide smooth rotation. Even hinge-type joints really are balls - just flattened, to provide movement in only one direction.

In the center of the ball is a small blue dot. This is the ball's pivot point. If you rotate the ball, it will rotate around this pivot point. As long as the ball joint is perfectly rounded along the angle of rotation (i.e. not oval, egg, or rounded rectangle shaped), this point will always be in the same relative location in the ball.

The more I learn about jointing, the more I realize this pivot point, though imaginary when working in clay, controls the way the joint moves in a lot of interesting ways.

I will go into a lot more detail about pivot points, and note where the pivot point is especially important. For now, just pay attention to the pivot points and their locations.




Upon the ball, we add the rest of the limb. This could be for a knee, elbow, ankle - pretty much any part. Right now, the joint fully encloses the ball, and thus, this joint cannot flex, except if pulled out of socket. It doesn't have anywhere to go.




Flexing each half of the limb by 45 degrees results in a collision of the two limbs in the red area. Of course, we can do this easily in a diagram, but in real life, two objects cannot inhabit the same space.




In order to make the joint flex to this 90 degree angle, the red area needs to be cut away. Cutting away half of the joint from each side results in this - a perfect meeting of the joint at 90 degrees, allowing 90 degree rotation - but no further.




Unrotating the joint to its straight position, with these new cuts. The total angle of the cut is 90 degrees, with a 45 degree cut on each side.




A single ball joint can be pushed further than 90 degrees, but the cuts remove more and more material. For example, 140 degrees is roughly the flexibility of the human elbow - and is the amount of rotation shown here. But the further you go, the greater the area of the collision between the two parts, and the more material that will have to be cut away.




And the more you cut away, the stranger and more unnatural the joint looks when straightened, exposing more and more of the ball and creating a large gaping angle, equal to the amount of rotation (so the cut here is 140 degrees, with 70 degrees cut away from each side).





So single ball joints can only be used to go further than 90 degrees in a limited manner, without creating a big old gaping cut.

How can we get around this?

 

Part 2: Double Joints

In order to reduce the amount of cut away material, we'll need to add another ball. That is, a double joint. The balls for the joint can overlap, creating sort of an oval or pill shape. But even though it may appear like an elongated single ball, it really is two balls working separately, each with their own pivot point in the center.

The top of the limb will rotate around the pivot point of the top ball, and the lower limb will rotate around the pivot point of the bottom ball.




For now, the cut is placed in the center, equally far from both the top and bottom pivot point.




Each part of the limb is rotated 70 degrees, for a total of a 140 degree rotation. There's still a pretty significant amount of material that must be cut away due to the two halves of the limb colliding.





With the cuts made, the joint doesn't look too bad at this 140 degree angle - you could almost believe it's an elbow.




When unrotated, there's still a fairly large cut...




BUT! If you compare this 140 degree rotation with this double joint to the 140 degree rotation with the single joint, the difference is clear, though not that dramatic. I've flipped the single joint, so the comparison is easier.

Note the red lines, showing where the cuts end vertically on the limb. See how the cut for the single joint has to extend further up and down than with the double joint? The double joint has a cut that's a bit less open than the single joint. The red shows the difference in the two.

The angle of both cuts is still the same - 140 degrees. But the position of the angle's corner changes, moving further to the left as the distance between the pivot points increases, making the total area of the cut smaller.




This becomes more dramatic as you increase the distance between the pivot points even further...




And rotate both pieces 70 degrees (again, 140 degrees total) then the amount of collision area is decreased...




With the cuts made...




And then straightened... This version has a much shallower cut, extending quite a bit less far up and down, and the total size of the cut is quite a bit smaller.




When compared against the previous double joint with a smaller distance between the pivot points you can see, again, the cut has gotten shallower, covering a smaller total area.

There is probably a way to figure out exactly the areas and ratios and whatnot here - I'm no Pythagoras, but it's still useful to know that the further you move the pivot points, the shallower the cut will be!




So what conclusions can we draw based on this information?

• For a simple, centered joint like this, whether it's a single or double joint, the angle of a cut's opening is equal to the amount the joint can flex. If the cut is 90 degrees, the joint can flex 90 degrees.
  
• The greater the angle of rotation, the more collision between the parts, and the more material will have to be removed, resulting in a larger cut.
  
• Creating a double joint allows for a shallower cut, even though the actual angle of the cut will remain the same.
  
• The greater the distance between the pivot points of the double joint, the smaller and shallower the cut.

Please note: The diagrams above are somewhat approximate - I've made them as accurate as I can without driving myself nuts in Photoshop - but that everything may not fit together quite -perfectly-. The corners of the cuts down by the pivot point may need to be rounded due to a small amount of collision there for these to actually work.

 

Part 3: Off Center Pivots & Uneven Cuts

Double-jointing is not the only way to reduce the size of the cut.

What happens when we move the position of the cut, -and- make a double joint? Here, the cut is moved up, so that it is no longer centered between the two pivot points. Instead, it is level with the top pivot point. This means that the top ball will be set fairly shallow into the limb, while the bottom ball will be set quite deeply. This will completely change the way the parts interact with each other.

The pivot points in this example are set the same distance apart from each other as in the last example in Part 2.




Moving both parts of the limb 70 degrees with the cut off-center like this results in a bit of a strange and more complex shape than before. The area where the two pieces collide is quite a bit smaller though.




Previously, the cuts made have all been half from each side of the limb, along the obvious middle line. But in this case, there is no obvious middle - and no obvious line of where to cut.

So in this case, we'll make the cut as shown above - removing all of the collision area from only the top part of the limb, leaving the bottom intact.




Now, I mentioned above in a "Please note" that you may have to round the corners a little bit compared to the diagrams to make things rotate properly.

The more "exotic" an arrangement you do (like un-centering the placement of the cut, or cutting more away from one side than the other), the more you'll have to be careful about your corners. They will tend to collide more in rotation than a basic boring joint.

Here, the top of the limb has been rotated back only 35 degrees (halfway between its full rotation of 70 and its unrotated position of 0), which results in another area colliding on the corners. So it must be trimmed away further, checking in various amounts of rotation until the joint can move freely to any desired angle without colliding.

For most of my diagrams (including this one) I am not going to bother rotating to various positions and trimming away (it takes long enough to make the diagrams as it is!). But just be aware that these are approximations and I'm trying to illustrate the overall angles and principles involved - the exact tiny details may be a little bit off!




This results in a bit of an uneven cut which can look a little wonky, but much less material has been removed than with the cut centered between the pivot points.






By the way - moving the cut off-center can be done with single joints too:




Small, slim area of collision, with a fairly obvious spot to make a cut:




And slightly awkward when straightened - but it can always be given a better contour. This creates a very small cut.




Now, I don't know how useful this information is, but it's an interesting observation anyway...

In the examples with the centered cut, the total angle of rotation for the joint was equal to the total angle of the cut.

And it sort of still is, if you know where to look. For the off-center double jointing example above...

If you measure the angle of the cut from the corner of the top limb across its cut, that's 50 degrees. And if you measure the angle of the bottom limb, from the corner and down (it has no cut), then that's 90 degrees. 50+90 = 140... the total rotation this setup is capable of.




It works with the single joint too. Remember, it could rotate 90 degrees total.

If you measure from the corner of the upper part of the limb, up its cut, that's 20 degrees. And from the corner down the cut for the bottom part of the limb, that's 70 degrees. 20+70 = 90, the total rotation this setup can do.




Like I said, dunno how useful that particular bit is, just an interesting observation that appears to hold true for every way I've tried.

Remember, knowing how things will move and being able to predict and change which parts move is key to having a really nice, solid, poseable doll!

 

Part 4: Depth & Roundness

Touched on briefly in Part 3 was the concept of how deep a joint is set into the socket. This becomes even more important when you are jointing the areas where the limbs attach to the body: the shoulder and hip.

This concept really works for all limbs, but is easiest to demonstrate using a torso-like shape with the cup, and a limb-like shape with the ball.

Note that something else has changed - the sides of the limb are now flush with the sides of the ball. Why this is important should become clear later in Part 4.

Right now, this limb can't really rotate at all - it might be able to wiggle a bit, but that's it.




But set the ball more shallowly - about a quarter of its diameter, and now there is about 25 degrees of the ball exposed on each side.




Which means the limb can now rotate about 25 degrees in either direction before it has to stop - when the limb's side collides with the body.




Decrease the depth even further, so that 40 degrees of the ball is exposed...




And the limb can rotate 40 degrees in each direction.




But the shallower the cup, the less friction and grip it's going to have on the ball. This means the limb will be more likely to slip out of place.

Remember, what's important is how much of the ball is exposed, allowing it to rotate before parts will collide and prevent further rotation. So how to expose more of the ball?

One way: set the outer edge of the limb back from the ball.

Here, the limb is set halfway into the cup as it was in the first example. Except now, one side of the limb is set back from the ball. This is a common arrangement on the upper arm. This exposes about 60 degrees of the ball.




Which will allow for 60 degrees of rotation.




But setting the edge back is not the only way. It can just be cut away to the angle of rotation necessary. Cutting away 70 degrees...




... Also gives you a 70 degree rotation:




Again, what matters is how much of the ball is exposed. More ball = more mobility, and the exact amount of mobility is perfectly predictable.

 

Part 5: Shims

I'm not sure if there's a better name for this, but I believe twigling referred to them as "shims" in her Workshop Notes, so that's what I'll call them - it sounds technical. :P

This is any part which acts as a small ball-and-cup, and fits over the end of an existing joint to allow greater mobility.

Notice that there are two pivot points. Shims act as sort of a half-double joint, with their own separate pivot point. The top pivot point here is for the shim. The bottom pivot point is for the ball-and-limb.




This shim is pretty small, and is set at a moderate depth. But as in Part 4, what really matters is how many degrees of the ball are able to move in rotation. Here, there is approximately 25 degrees of the shim's ball exposed.




...Which means that there is 25 degrees of rotation possible for the shim. Because the limb is set into the shim, it rotates along too.

Notice that the pivot point of the limb's ball changes position here. This complicates some of the angles and so forth - with the previous examples of double joints, the pivot points stayed in the same position.




Of course, right now, the limb can't really flex at all where it fits into the shim - none of that ball is exposed, so this is really doing basically the same thing as adjusting the depth of the ball without a shim.

In order to make the shim really work, more of the limb's ball has to be exposed too, to allow the limb to rotate in the shim's cup.

But just by trimming away a little bit, the limb can be rotated an additional 55 degrees, for a total of 80 degrees.




Straighten everything back out, and the small size of the cut is obvious. Using a very small shim, a large degree of rotation can be easily achieved with just the slightest bit of trimming away.




Shims are used on a lot of different dolls. Here are some examples...

Martha Armstrong-Hand's porcelain jointed dolls use shoulder shims to increase the amount of rotation, and allow the dolls to bring their arms across their chest.




Fairyland's resin BJDs use shims for greater hip mobility and smoothness in appearance when bent:




My first sculpted doll also uses shims for hip mobility and smoothness of appearance:




Fairyland also uses shims on the knees of some of their dolls. The shims are visible from the front when the leg is straightened, and helps create a smooth appearance to the knee when bent, as well as increased mobility. The back edge of the knee shim is thicker than the front, raising the pivot point for the shim and thus allowing for a greater rotational angle - they can't quite sit on their legs, but they can bend them back more than 90 degrees with this half-shim, half-double-joint setup.

 

Part 6: Cut Angle

Looking at company BJDs, you often see a variety of interesting angles to the cuts, especially in areas like the knees - the kneecap may be part of the top of the leg, or the bottom - or it may be a part of a double joint, with various contours to make it seem more natural.

Designing the contours of the cuts and angles is not just a matter of flexibility, but aesthetics. And it's made easier once you understand this:

Past the point where the two surfaces of the part collide when fully bent, the shape of the cut doesn't really matter.

This means that for a simple ball joint, as shown in Part 1...




All that really matters is that the area of collision allows for a 45 degree rotation for each part (90 degrees total). This is the part on the left, highlighted in pink. Its contour doesn't have to be exactly straight across like this, but it does have to match up so that the joint can rotate to this angle.

However, the other side of the limb, on the right, highlighted in blue... Its shape really doesn't matter very much and will not affect how much this joint can rotate.

So, for example, the shape of the cut on the right side can be modified to follow the contour of the cut-out portion:




When bent, this creates an interesting contour, which may be desirable for something like an elbow, that does look quite sharp when flexing.




But just modify the contour a little...




... To get a pretty convincing knee that would look nice both straight and when flexed.




As long as there are no undercuts or strange angles that would interfere with the joint opening and closing, you can have any sort of contour to the cut beyond the point where the two pieces meet when fully flexed - without worrying about it altering the way the joint will work.




Even very strange cuts for, I dunno, the Swamp Thing's elbows would work just fine - there may be minor points of collision that need smoothing, but it doesn't affect how far the limb can flex or straighten in any way. It's the 90 degree angled cut in the back, and where the two pieces come together (as shown in Part 1) that really matters.

 

Part 7: Friction & Surface Area

This part is not complete!

There's one other reason this off-center cut/pivot arrangement may be preferrable to one where everything's centered...

When the limb is straight, you can see this best. Notice the pink and green, and grey shaded areas.

The pink shading here indicates the surface area of the top ball, in contact with the cup of the top part of the limb.

The green shading indicates the surface area of the bottom ball, in contact with the cup of the bottom part of the limb.

The grey area is shaded to indicate the part that would be cut away to allow the joint to flex.

You can see that there's more surface touching the bottom ball (green) than the top ball (pink). This means there's less friction on the top ball. Friction is what will prevent the joint from moving - greater surface area = greater friction.

This would make it more likely for the top joint to move than the bottom joint - it will tend to want to "let go" easier and flex with less pressure. You can, of course, encourage this effect by adding some sort of lock/stop system on the bottom ball, so that it fits into the bottom cup and makes it even less likely to move vs. the top ball.

This can be used for areas such as double jointed knees and elbows, where you may want to encourage one part of the limb to always flex to its full limit before the other part begins to flex at all.

 

Part 8: Stringing Channels

This part is not finished!

Of course, every change you make has consequences - by contouring the cut to look more natural when fully extended, it means that the top and bottom of the limb won't contact each other when flexing. The top part flexed to 70 degrees has a gap, and it has to be dealt with properly.



If it isn't, it can result in the top part of the limb over-flexing. This may not seem like a problem, but say you bend the limb, flexing only the top part first, and it goes past the proper 70 degrees to something like 80 degrees. Then, if you bend the bottom part to its full 70 degrees, the top part has to un-flex by 10 degrees to accomodate it. It's not the end of the world, but it is a little sloppy.

So how to fix this? Well, there's a couple ways... Locks/stops could be used to limit the top limb's movement to a maximum of 70 degrees.

Or, perhaps a bit simpler, is to handle it in the way the stringing channels are cut. I'll cover stringing channels a little bit more later on, but for now...

The blue shaded area indicates where the hole would be drilled through the center, to handle the elastic stringing.




In order to make a path for the elastic to travel when the joint is flexed, a slit must be cut into the surface of the ball joint. These kinda things:




So when this off-center joint is flexed, the stringing channel for the elastic would bend too. There would need to be a slit extending from the straight position all the way through the fully-bent position - the pink shaded area indicates where the slit would go.

As long as the slit is only cut this far and no further, it will not be able to move past 70 degrees without popping out of joint. If the slit was cut a bit more (into the green shaded area) then the top of the limb could continue rotating past 70 degrees.

 

I've been updating the existing done parts with new pics (the ones with the dots for the pivot point instead of the blue lines) - I figured out a more accurate way of making the diagrams.

I'll be adding some more soon - it may take a while for me to complete this whole guide as it's sort of being done as I feel inspired, as a bit of a helper thingie for my own brain. I can't work too fast on it or I can't assimilate the info myself, and I get burnt out and start making leaps of faulty logic and whatnot. But rest assured, it's still very much on my to do list.

 

Gawd it's been a long time since I updated this - so sorry for the delay. I can't do a proper update right now but there are a few things that I've learned recently that I want to note down before I forget:

1) The most "comfortable" position for a doll should be the standing position. This means that it should be easiest for each joint to sit in standing position, and any other positions are more difficult. Making a joint's positioning equally difficult for standing vs. not standing (say, a knee being bent) may work, but will be less stable for standing.

2) A hinge joint (like a knee) can be made slightly oval rather than round on the ball part of the joint, so that there is a slight resistance to the joint rotating. This makes the joint more "comfortable" non-rotated and thus more stable for standing. This functions as a "soft lock." The cup part of the joint can remain precisely round - only the ball of the joint needs to be modified.

View attachment 587

3) The length of the stringing channel matters, and if you're getting a joint that is kicky, try measuring the depth of the stringing channel at different points along its length - if it's significantly deeper at one point than another, the elastic has to stretch more there and it will resist going into that position. You can use this principle to make a joint resist bending by making the shallowest part of the stringing channel the one the elastic uses when the joint is in standing position.