Step 2: Added the appropriate contacts/connections. 

Bolt Joint Connection

The bolt head to plate = frictional contact. Made an imprint to face of plate where bolt head applies pressure using ansys spaceclaim. Augmented Lagrange was selected for formulation. It is best for frictional contacts. Small Sliding = off. 

Finite (large sliding is turned on because:

• Recomputes the closest-point projection every iteration; a contact node is free to “walk” across multiple target elements.

• Updates the contact normal and tangential directions as the bodies rotate/tilt.

• Allows open–close and re-contact anywhere within the pinball region.

• Costs a bit more, but it’s the physically correct choice for noticeable slip, edge contact, rocking, or separation.

• Step 2 and 3 adds an external load designed to partially unload and separate the plate from the magnet. That moves the contact patch toward an edge (rocking) and can produce re-contact at a different location.

Rule of thumb

• Slip distance $\ll$ 10% of the contact length scale (e.g., < 0.1× head diameter),

• Relative rotation is just a few degrees, and

• No separation/re-contact will occur.

Bolt thread to scanner hole = bonded. This is because the bolt shank fails before threads do in most scenarios. The shank is exposed to bending or shear whereas the thread portion is backed up by the “nut” material, which prevents bending. The break is just ahead of the first engaged thread in the nut, at the shank-to-thread transition. That’s where you get a combination of stress concentration and highest applied moment.

Plate clearance hole to bolt shank = frictionless, Augmented Lagrange, Pinball Radius = 0.4mm. This contact will introduce shear into the bolt shank once the bolt joint begins to seperate. As the clearance hole in the plate comes in contact with the bolt, shear will be added. 

Why a Pinball Radius of 0.4mm? 

When running the initial contact tool, I saw the original programmed control radius to be 1.7mm. This will cause the bolt to see a very large neighborhood, so nodes can flip into a near contact state long before the hole actually reaches the bolt. This will create artificial pressures/penetrations early. Setting the pinball ~ radial clearance (0.25) + ~one local element edge (0.1) keeps the pair open until the gap truly closes. When rerunning the contact tool, the initial status of this contact is OPEN. This is what I want to see. 

Plate to scanner (step 1) = No separation during preloading. Ensures gravity force does not affect movement of plate. (I was having this error in the beginning)

Plate to scanner (Step 2 and 3) = Frictional Contact with augmented lagrange, small sliding off for the same reasons. Pinball region is manually inputted to 6mm. 

Why a pinball region of 6mm? ***************************REDO

• In ANSYS contact, each contact node searches for nearby target faces inside a sphere—the pinball.
• If a target face is inside the sphere, that pair can become active contact; if not, the node is FAR.

• With finite (large) sliding, this search is repeated every equilibrium iteration, so the radius controls how far a node is allowed to migrate to a new target as the bodies separate/slide/rock

• pinball ≥ (initial gap + expected normal separation + a small safety margin).
• plate and magnet start touching (gap ≈ 0), and even in my “separation” step the measured openings are on the order of 10⁻⁴–10⁻³ m. So a 6 mm pinball is orders of magnitude larger than the actual opening—plenty to keep the contact “in range.”
• 6mm is a good choice because:
  

  If you leave the pinball Program Controlled, ANSYS often picks a generous value (frequently ≥ one element size or more). That can cause:

  • Spurious “jump” contact: when the plate peels, nodes may reattach to the back face of the magnet or to a neighboring feature that sits within a big pinball.

  • Chatter/overconstraint: multiple candidate faces inside the sphere → frequent switching → convergence trouble and ghost pressures.

  • Artificial stiffening: early contact with non-mating surfaces makes the joint look stronger than it is.

• Setting pinball = 6 mm (< 7.5 mm mesh element) narrows the neighborhood to “what we physically expect”. 6 mm pinball is intentionally a bit smaller than the 7.5 mm mesh size to exclude nearby, non-mating faces while still being much larger than any real separation you apply. ******************** REDO

Plate to frame = bonded. Tim contact = on, Trim tolerance of 5.5mm. Formulation = Pure Penalty. Pinball region = 6mm.

Why Formulation is Pure Penalty? 

A “bonded” contact in Mechanical can be enforced three main ways:

a) MPC (multi-point constraints).

• What it does: writes hard kinematic constraints (equal displacements) between coincident DOFs.

• Pros: exact tie (no relative slip/gap), very stiff; minimal penetration by construction.

• Cons: can create constraint loops (overconstraint) when many small faces meet one large face, or when the tied region winds around corners/holes; sensitive to topology mismatch; poorer robustness with large deformation and nonmatching meshes; can artificially over-stiffen the model.

b) Augmented Lagrange (AL).

• What it does: adds Lagrange multipliers with a penalty “assist”. It iteratively adjusts the contact traction so the kinematic error goes to (near) zero.

• Pros: accurate enforcement; sometimes more accurate pressure recovery than pure penalty.

• Cons: more nonlinear iterations; if the tied patch is big and the meshes are dissimilar, AL can still be touchy; more tuning (stiffness factors, augmentation).

c) Pure Penalty (what I am using).

• What it does: replaces the “bond” with very-stiff linear springs (normal & tangential) between the two surfaces. The solver tries to drive relative motion to ≈0 by using a large contact stiffness kn, kt,  computed from local $E$, area, and element size.

• Pros: very robust with many-to-one ties, dissimilar meshes, big areas, and large deformation; avoids MPC loops; converges faster; easy to run.

• Cons: it permits microscopic relative slip/gap, so the bond is “numerically compliant.” If you set stiffness too low you’ll see visible slip/penetration; too high and you re-introduce ill-conditioning.

• Why it is a good choice:  bonding many smaller frame faces to one split line on the plate (many-to-one). Penalty avoids the overconstraint/MPC loop issues and is typically the most robust & stable way to glue non-matching meshes.

Why Trim Contact = ON and Trim Tolerance = 5.5mm?

What “Trim Contact” does
When you define a bonded region by selecting faces, Mechanical builds an interaction patch by projecting contact nodes onto the target. Raw projection can (a) spill past edges, (b) include slivers around holes/chamfers, or (c) miss slightly misaligned areas because of tiny CAD/mesh gaps. Trim Contact = On tells Mechanical to clip the active bonded patch to the true overlap of the two surfaces (their intersection in projection) within a user tolerance. That prevents:

• Bonding across nearby but not intended faces, and

• Losing parts of the bond because of small initial offsets.

What the Trim Tolerance is

• It’s a geometric allowance for the projection/intersection, not a stiffness.

• A node on the contact face is considered part of the bonded intersection if its projected distance to the target is ≤ the Trim Tolerance.

• So it covers minor mismatches, mesh stair-stepping, or slight CAD gaps while avoiding bonding faces that are truly separate.

Why 5.5 mm in mymodel

• local element length at that interface is ~7.5 mm. With coarse, non-conformal meshes and flat-on-flat geometry, a few millimeters of trim tolerance keeps the bond continuous despite stair-stepped edges or tiny offsets.

• I chose 5.5 mm as a conservative but sub-pinball value: it is smaller than your nearby contact pinball (6 mm radius) and smaller than the typical hole/feature spacing, which helps avoid accidentally bridging to unintended faces while still capturing the intended overlap everywhere.

• Validated: turned on Contact Tool → Status and confirmed no “Over Constrained” regions appear and that the bond visually matches the intended footprint.

All other parts bonded. I learned that you cannot have multiple contacts with the same face involved. It causes an augmented lagrange error.