CNC Multi-Station Fixture Programming: Setting Up Work Offsets and Tool Paths That Switch Cleanly
Running a multi-station fixture on a CNC mill means one program has to hit four, six, or eight different work offsets without crashing, without losing tolerance, and without the operator having to baby it. The programming is not hard. What makes it hard is the coordination — making sure every station’s tool path knows where it is, when it starts, and when it hands off to the next station without overlapping into the wrong zone.
Most shops treat multi-station work as a single program with a bunch of subroutines. That works until something shifts by half a millimeter on station three and the whole part is scrap. The fix is not better fixtures. It is better offset management and smarter tool path sequencing.
How Work Offsets Work Across Multiple Stations
Every station on a multi-station fixture has its own work offset — typically G54, G55, G56, and so on. The controller stores the X, Y, Z zero point for each station independently. When the program calls G55, the machine instantly switches to that station’s zero without moving anything. The tool stays exactly where it is in machine space, but all subsequent coordinates are now referenced to station two’s datum instead of station one’s.
This sounds simple until you realize the tool has to travel from station one to station two. That travel move crosses open space — no part, no fixture, just air. If you program that move wrong, the tool smashes into a clamp, a vise jaw, or the part itself.
Setting Up Offsets Without Guessing
The most reliable way to set multi-station offsets is to use an edge finder or a probe on each station individually. Do not try to calculate offsets from the fixture drawing. The drawing does not account for the actual position of the part after it is loaded.
Load the part into station one. Touch off on a known datum — usually a precision locating pin or a machined surface on the part. Set G54. Move to station two. Touch off again on the same datum feature. Set G55. Repeat for every station.
The critical step: after setting all offsets, run a dry pass that moves the tool to each station’s safe Z height without cutting. Watch the tool clearance visually or through simulation. If the tool gets within 2 mm of any clamp or fixture element on the travel move between stations, adjust the offset or add a clearance move.
One mistake that shows up constantly: operators set the Z offset on each station but forget that the part height might vary slightly between stations. If station three’s part is sitting 0.1 mm higher because of a chip under the locator, the tool will plunge 0.1 mm too deep. Always verify Z on every station, not just X and Y.
Tool Path Sequencing: The Order Matters More Than You Think
A multi-station program is not just a list of operations. It is a sequence. The order in which stations are machined affects clamp deflection, heat buildup, and cumulative tolerance drift.
Machining Station One First Is Usually Wrong
Most programmers start with station one because it is the first one they set up. But station one is often the one with the least rigid clamping, because the operator has not yet tightened all the clamps. If you rough station one first, the part shifts when you go to tighten the clamps for station two, and now station one is out of tolerance.
The better sequence: start with the station that has the tightest, most repeatable clamping. Usually that is the last station on the fixture — the one farthest from the load/unload side. Machine that station first while the part is still fresh in the fixture. Then work backward through the stations. Each subsequent station benefits from the part being fully seated and all clamps engaged.
Using Subprograms to Isolate Each Station
Each station should have its own subprogram. The main program calls the subprogram, switches to the correct work offset, runs the operations, then returns to a safe position before switching to the next offset.
The subprogram structure looks like this: the main program positions the tool at a safe height above station one, calls O1001 (station one program), the subprogram runs all operations for station one using G54, then returns with M99. The main program then moves to the safe height above station two, calls O1002, and so on.
Keep the subprograms independent. Do not let one subprogram reference another subprogram’s offset. If station three needs to reuse an operation from station one, copy the code into station three’s subprogram rather than calling station one’s subprogram from station three’s offset. Crossing offset boundaries inside a subprogram call is a recipe for crashes.
Travel Moves Between Stations: Where Most Errors Hide
The actual cutting on each station is usually fine. The travel moves between stations are where things go wrong. The tool has to go from the last cut position on station one to the first cut position on station two. That move crosses the fixture, and the fixture is full of clamps, locators, and hardware that the tool does not care about.
Programming Clearance Heights That Actually Clear
A safe Z clearance between stations is not a fixed number. It depends on the tallest clamp on the fixture. Measure the tallest fixture element between station one and station two. Add 5 mm to that measurement. That is your minimum safe Z height for the travel move.
Most programmers set a blanket Z clearance of 50 mm or 100 mm and assume it is safe. It is safe on an empty fixture. But once the part is loaded, the clamps extend upward, and a 50 mm clearance might not be enough. The tool does not care about your assumption. It cares about the actual geometry.
Use G43 (tool length compensation) on every travel move. If you switch tools between stations, the tool length changes. Without G43, the Z position on the travel move is calculated from the wrong tool length, and the tool can dip into the fixture.
Avoiding Rapid Moves Through the Fixture Body
Rapid traverses (G00) between stations are fast, but they are also uncontrollable in terms of exact path. The controller takes the shortest path between two points, which might cut diagonally through a clamp that you thought was out of the way.
For travel moves between stations, use linear feeds (G01) at a moderate rate instead of rapid. This forces the controller to follow the exact path you programmed. You can route the move around the clamps deliberately. A G01 move at 2000 mm/min between stations adds maybe two seconds to the cycle. A crash adds two hours.
If you must use rapid, program an intermediate waypoint that lifts the tool high enough to clear everything, then moves laterally, then drops to the next station. Do not let the controller figure out the path on its own.
Part Zero and How It Shifts Across Stations
The part zero is the datum from which all dimensions on the part are measured. On a multi-station fixture, the part zero should be identical on every station. In reality, it never is. The locators have wear, the parts have slight variation, and the clamping force shifts the part by a few microns each time you open and close the vise.
Using a Common Datum Feature Across All Stations
The best way to control part zero across stations is to machine a common datum feature on every station — usually a pocket, a hole, or a step that all other dimensions reference. Machine this datum first on each station, before any other operations.
When you machine the datum on station one, you establish the part zero for that station. When you move to station two, you re-machine the same datum feature using station two’s work offset. This re-establishes the part zero for station two relative to the same physical feature on the part.
The result: even if the work offsets drift by 0.02 mm between stations, the datum feature is fresh on each station, so all dimensions that reference it stay accurate. This is far more reliable than trying to keep six work offsets perfectly aligned.
Compensating for Locator Wear Over Time
Locators wear. After a few hundred cycles, the precision pins that position the part in each station develop play. The play is small — maybe 0.01 mm — but it adds up across multiple stations.
Check locator wear monthly. Run a test part through all stations, measure the part zero on each station, and compare. If station four has drifted by 0.03 mm relative to station one, adjust the G58 offset by that amount. Do not adjust the fixture. Adjust the offset. The fixture is hard to change. The offset is one number in the controller.
Cycle Time Optimization Without Sacrificing Accuracy
Multi-station fixtures exist to save time. One setup, multiple parts, no re-clamping. But the programming can eat all that time back if you are not careful.
Overlapping Non-Cutting Moves With Spindle Deceleration
The spindle does not have to wait for the tool to finish moving to the next station before it starts slowing down. Most controllers support background spindle speed changes. While the tool is traveling from station one to station two at G01, the spindle can already be decelerating to the next station’s RPM.
Program the spindle speed change at least 50 mm before the tool reaches the next station’s first cut position. This way, the spindle is already at the correct speed when the tool engages. On a program with eight stations, this saves 10 to 15 seconds per cycle — not huge on one part, but significant over a production run of 500 pieces.
Grouping Similar Operations Across Stations
If station one and station three both need a drill operation, do not drill station one, then move to station two and mill something, then move to station three and drill again. Group the operations. Drill station one, then drill station three, then move to station two and mill. Then mill station four, mill station six, and so on.
This reduces tool changes and keeps the spindle at a consistent speed for longer stretches. The travel moves between stations of the same operation type are shorter because you are moving between adjacent stations rather than jumping around the fixture. The program is also easier to read and debug when operations are grouped by type rather than by station order.
Common Mistakes That Wreck Multi-Station Programs
Forgetting to Reset Modal Codes Between Stations
Modal codes carry over from one subprogram to the next. If station one’s subprogram ends with G91 (incremental mode) and you forget to switch back to G90 (absolute mode) at the start of station two’s subprogram, every coordinate on station two will be wrong.
The fix: always start every subprogram with a modal code reset block. G90 G40 G49 G80 — absolute positioning, cancel cutter compensation, cancel tool length compensation, cancel canned cycle. This clears everything and puts the controller in a known state before the first move of the new station.
Not Accounting for Fixture Asymmetry
Many multi-station fixtures are not symmetric. Station one might be at the front left, station two at the front right, station three at the back left. The travel distances between stations are not equal. A program that assumes equal spacing will have the tool arriving at the wrong Z height on the stations that are farther apart.
Measure every travel distance on the actual fixture. Do not assume. Program each travel move with the exact coordinates for that specific station pair. A multi-station fixture program should have a travel move table — even if it is just comments in the code — that lists the from-station, to-station, clearance height, and feed rate for every transition.
