Ask any machine builder what keeps a non-standard machine accurate, and most will talk about the servo and its encoder. I will tell you something else: on the machines I have built and rebuilt, accuracy in the loaded state comes from the mechanical lock, not the servo loop. A servo holds position while the loop is alive and the load is still, but the moment a press pushes in or a tool pulls out, the screw winds back and the coupling compresses. That is where a proper mechanical locking and positioning mechanism earns its keep. This article is about the mechanical side of holding position: the devices, the engineering choices, and the mistakes I have made and watched others make.
1. Positioning and Locking Are Two Different Jobs
First, keep the two jobs separate in your head. Positioning moves the element to a defined location. Locking holds it there against a force and a time. On a stroking head or a pallet stop, those are different mechanisms doing different work, and conflating them is how machines drift.
Positioning is usually done by a cam, a hard stop, a tapered dowel, or the servo itself. Locking is done by a detent, a clamp, a wedge, or a brake. A good design sequence is: position on hard datum first, then lock with a mechanism that adds stiffness instead of fighting the positioner. If the lock has to both move and hold, you have given it two jobs and it will do neither well.
My working rule: never let the locking mechanism be the positioning mechanism. The lock holds what the datum placed.
A classic example is a rotary index table. The servo or the cam indexes the plate, but the plate only holds its angle against an indexing or clamping device that seats into a shot pin or taper. On the first machine I designed this way, I relied on the servo to hold the index for a drilling operation, and the holes walked 0.3 mm by the end of the shift. Once I added a tapered locating pin behind the drive, the drift disappeared.
2. Shot Pins and Locating Pins: The Workhorse Datum
For repeatability, nothing beats a pin going into a precisely machined hole or bushing. Shot pins are the highest-force, highest-rigidity way to fix a slide or a turret in a known spot, and they are the natural first choice for index tables, pallet stops, and tool changer arms.
The engineering is all in the fit. A straight shot pin locating into a bushing gives you a datum repeatable to a few microns if the pin and bushing are ground and the bushing is pressed into the plate with a controlled interference. A taper pin or a conical seat drives the work home and removes backlash by commanding the element into the seat. The trade is that taper pins need a controlled seating force and can bind if the taper angle is too shallow.
Choose the pin size from the load, not the habit of the previous machine. The pin has to carry the shear and the bending from the applied force, and the bushing has to survive the life of the machine. I usually start with a 12 mm hardened pin for a small index table and scale from there, but the honest method is to compute the shear area against the worst-case stop force with a safety factor of at least three for a device that people depend on.
| Pin Style | Repeatability | Backlash | Seating | Best Use |
|---|---|---|---|---|
| Straight pin in bushing | Micron class | Fit dependent | Push-in | Index, pallet, turret |
| Taper/cone pin | Excellent | Eliminated by seat | Force commanded | Heavy-duty stopping |
| Ball detent | Good | Small, spring | Spring loaded | Light indexing |
| Spring-loaded plunger | Good | Small | Self locating | Tooling, light duty |
The spring-loaded plunger is the underrated cousin. It is cheap, it self-locates, and for light pallet stops or swept-lowering guards it is plenty. People dismiss it because it feels crude, but for repeatable light holding it can outlive every fancy brake on the machine.
3. Wedges, Clamps and the Mechanical Advantage Game
When the load is coming in sideways or the pin cannot reach, a wedge or cam lock takes over. A wedge converts a force into a much larger normal force across the contact by the tangent of the wedge angle, and it does it without any complex links. A typical 10 degree wedge multiplies force by a factor approaching six before friction, and even with a realistic friction coefficient of 0.1, you are still getting a solid mechanical advantage that holds the load when the driving force is removed, as long as the angle stays below the self-locking limit.
Self-locking is the property that makes wedges so useful and so dangerous at the same time. Below the friction angle, the wedge stays put even if you release the actuator. That is what you want for a holding device. But it also means the wedge will not release by gravity or a light spring, so you have to design the release stroke with enough force to back the wedge out. I have watched a designer wedge a slide tight, then spend a week fighting a release that was too weak because the self-lock was too good.
tan(alpha) < mu gives a self-locking wedge, and the holding force roughly multiplies as 1/tan(alpha) with a safety factor for friction variability.
Cam clamps work on the same principle with a shaped profile instead of a straight taper. The advantage is a controlled motion, lock at the designed position, and a profile that can be made to release just past top dead center. I reach for cam clamps on quick-change pallets and on covers where the operator expects one lever motion to clamp and another to release, because the operator feel is better than a mystery wedge you have to hammer.
4. Brakes and Clamps for Slides and Rotaries
Friction brakes and diaphragm clamps are the workhorses when the moving element cannot accept a drilled hole for a pin. A spring-applied, air-released brake on a linear slide is the classic fail-safe hold: power or air loss stops the slide, which is exactly the behavior a safety study wants. The stiffness of a friction brake is lower than a pin, but the convenience of clamping anywhere along the stroke is worth it when datum holes are impractical.
I have a personal preference here that contradicts some training material. I use friction brakes only where the load is moderate and the position tolerance is forgiving, above a tenth of a millimeter. For anything tighter, I cut the datum and use a pin. You can sometimes get away with a brake a few microns off on a continuous path, but on a drilling operation a couple of microns of wind-up shows up as an out-of-round hole you cannot explain to the customer.
For rotary tables the same logic applies. A pneumatic or hydraulic brake on the periphery of the plate holds it for machining, and the shot pin provides the actual datum for repeated stations. Combining a brake for damping and a pin for positioning gives you the best of both, and it is the configuration I put on every index table I have responsibility for. The brake takes the chatter load, the pin carries the accuracy.
Use the brake to absorb the energy and stiffen the loop; use the pin to define where the station is. Let each mechanism do the one job it is good at.
5. Detents and Plungers for Operator Convenience
Not every lock is fighting a machining force. A lot of them are just there so the operator feels the position and nothing moves while the machine idles. Manual detents, spring plungers, and ball locks are perfect there, and oversizing them is a common, harmless mistake. The real sin is under-sizing the pin on a forced stop and over-sizing the clamp on a clearance operation, so you carry more mass and cost than you need.
Measure the detent load. A guard that is only held against vibration by a 6 mm plunger is fine. That same plunger trying to hold a slide against a 2 kN process force will shear or pop. I keep the load audit on the drawing: write next to each lock the force it must resist and the direction, and the table becomes self-checking when someone tries to reuse the wrong device from the library.
6. Design Rules I Have Learned the Hard Way
The first rule is datum before lock. Never design the lock as the thing that also finds the position. On one turret I designed, the clamp was supposed to both align and hold the tool, and the repeatability died every time a different tool weight changed the seating. We stopped aligning with the clamp, added a taper pin, and the problem vanished in an afternoon.
Second, keep the datum clean and guarded. A locating pin that gathers swarf is a pin that stops locating. I always protect the seat area with a guard or a wiper, and I make sure the blow-off air or the guide covers that area because a machinist will happily blow chips sideways into the bushing otherwise.
Third, design the release with the same care as the lock. If locking takes 20 N and releasing needs 200 N, the operator will hammer it or the mechanism will see early wear. Count the release force on the drawing and provide the lever, cam, or pilot pressure to make it a one-hand operation. Self-locking wedges and cams are the usual culprits, so the release stroke has to be engineered, not hoped for.
Fourth, think about what happens when the machine power dies in the middle of a cycle. A spring-applied lock is fail-safe; an air-applied lock probably is not. Where the tool could drop or the pallet could move under gravity, the requirement is a spring or a mechanical stop, and you need to write it on the risk assessment, not discover it after the incident.
7. A Practical Selection Sequence
Here is the order I work through, and it saves me from impulse-buying a brake when a plunger would have done:
- Define the worst-case force and direction the lock must resist.
- Define the repeatability tolerance of the position.
- Choose a datum (pin, taper, hard stop) that meets the tolerance.
- Add a brake or clamp only if the load needs stiffness or damping beyond the datum.
- Choose the actuator style (spring, air, manual cam) to match the access and fail-safe needs.
- Verify the release force and document it on the assembly drawing.
Follow that order and you end up with a lock that is no bigger than it needs to be, a datum that actually locates, and a mechanism the operator can undo with one hand. It is not glamorous engineering, but it is the engineering that keeps a non-standard machine accurate at noon on a Thursday, which is when the customer actually measures it.
One more habit worth stealing: standardize the locking hardware across the machine family. I keep a short list of proven pins, plungers, wedges and brakes in the corporate library, and I refuse to let a designer invent a new lock for a standard application. The maintenance team learns a handful of devices cold, the spares shelf stays small, and the failure modes become familiar instead of novel. When a non-standard machine finally needs a truly new lock, that is the moment to slow down and do the load audit properly, because novelty is where the field service calls come from.
