CNC Dry Run and First Cut Debugging: The Process That Keeps Your Machine Alive
There is a reason every experienced operator runs a dry run before every single job. Not because they are cautious. Because they have crashed a machine before and they never want to do it again. The dry run and first cut sequence is not a formality. It is the last line of defense between your program and a five-thousand-dollar mistake.
Skip it and you are gambling. Run it properly and you catch almost everything that can go wrong before metal ever meets metal.
What Dry Run Actually Does and What It Does Not Do
The Mechanics of a Dry Run
When you hit the dry run button, the controller executes every line of code exactly as written. The spindle spins. The axes move. Tool changes happen. Coolant turns on. But the feed override is locked at zero, so the tool never engages the material. It traces the entire program in the air.
This sounds simple. But dry run catches things that simulation cannot. Simulation works with a mathematical model of the tool path. Dry run works with the actual controller logic, the actual offsets, and the actual machine kinematics. If your work offset is wrong by 0.05mm, simulation might not flag it because the tool path looks clean on screen. Dry run will show you exactly where the tool goes, and if that wrong offset sends it into the vise, you see it happen in real time.
What Dry Run Cannot Catch
Dry run does not show you chip formation. It does not show you surface finish. It does not show you deflection under cutting load. A tool that looks perfect in dry run can chatter violently under real cutting forces. A program that runs clean at rapid speed can stall when the feed engages because the controller cannot keep up with the lookahead calculations.
Dry run is a geometry check and a logic check. It is not a cutting check. That is why you always follow dry run with a first cut on scrap, not on the production part.
Setting Up the Machine Before You Press Start
Raising the Z Zero: The Step Everyone Forgets
Before any dry run, raise the Z work offset by at least 50mm above the part. This gives the tool clearance to move freely above the stock without any risk of plunging into the vise or the fixture.
If you forget this step, the first Z move in the program will drive the tool straight down into whatever is sitting on the table. I have seen operators lose a vise jaw over this. It takes five seconds to adjust the offset and it saves you from a two-hour cleanup.
Set the override switches to zero. Set the feed hold to ready. Make sure the E-stop is within arm’s reach. Then start the dry run in single-block mode.
Single Block vs Continuous Mode for Dry Run
Run dry run in single block, not continuous. Single block executes one line at a time and waits for you to press cycle start again. This lets you watch each move, confirm the coordinates on the DRO, and catch errors before they cascade.
Continuous mode runs the whole program without stopping. It is faster but dangerous for a first-time run. If there is a wrong coordinate on line 200, continuous mode will execute it before you even notice. Single block gives you time to read, react, and stop.
Once you have run the program once in single block with no issues, you can switch to continuous for the second dry run to verify timing and sequence. But the first pass is always single block.
The First Cut on Scrap: Where the Real Debugging Happens
Choosing the Right Scrap Piece
The scrap piece for your first cut should be the same material as the production part, or at least the same hardness. Cutting aluminum and cutting steel produce completely different tool behavior. If you test a steel program on an aluminum scrap block, you will not see the chatter or the deflection that would appear on the real part.
The scrap should also be the same thickness as the production stock. A thinner scrap piece lets the tool deflect more, which hides problems that would show up on a thicker part. Match the stock as closely as possible.
The Feed Rate Strategy for First Cut
Do not run the first cut at full program feed. Start at twenty-five percent feed and fifty percent spindle speed. Watch the cut. Listen to the sound. A good cut sounds like a steady buzz. A bad cut sounds like a grinding or squealing noise. If you hear anything other than a clean cut, stop immediately.
If the first cut at twenty-five percent looks clean, bump to fifty percent feed. Watch again. If it is still clean, go to seventy-five. Then full speed. This gradual ramp-up lets you catch problems at low severity before they become expensive.
Measuring After the First Cut
After the first cut, stop the program and measure the part. Do not assume it is correct because the program ran without crashing. Measure the critical dimensions first. If the holes are the right diameter and the pockets are the right depth, the program is probably good. If anything is off by more than 0.02mm, check your offsets before running again.
Most first-cut errors are offset errors, not program errors. The G-code is usually fine. The work offset or the tool length offset is wrong. Measure, adjust, re-run. Two or three iterations is normal for a new setup.
Common Dry Run Crashes and How to Prevent Them
Rapid Moves That Plunge Through the Part
The most common dry run crash is a rapid move that goes straight through the part because the Z clearance was never set. This happens when the programmer assumes the tool starts above the part but the work offset puts Z zero on the top surface. The first move is G00 Z-10.0 and the tool slams into the vise.
Prevent this by always checking the first Z move in single-block mode before you let the program run. Look at the DRO. Confirm the Z position before you press cycle start for the next block.
Tool Changes That Hit the Fixture
On machines with automatic tool changers, the turret or magazine move can collide with the part or the fixture if the tool change position is not set correctly. This is especially common on lathes with live tooling and on mills with tight vise setups.
Set the tool change position to a safe location, usually high above the part and clear of any fixtures. Verify this position in dry run by watching the turret move. If it comes anywhere near the vise, adjust the tool change coordinates before running again.
Overtravel on the Axes
Dry run can push an axis beyond its physical limit if the program contains a coordinate that exceeds the machine travel. The controller will throw an overtravel alarm and the machine will stop. This is not a crash, but it stops production and it scares the operator.
Check the program coordinates against the machine travel limits before dry run. If your program calls for X500 and the machine only travels to X400, you will get an alarm. Fix the coordinate or adjust the work offset so the program stays within the machine envelope.
Debugging Specific Program Errors During Dry Run
Wrong Feed Direction on Contours
When you dry run a contour operation, watch the direction of travel. If the tool is cutting on the wrong side of the profile, your compensation direction is flipped. G41 and G42 are relative to the direction of travel. If you programmed a clockwise contour but the tool is offset to the outside instead of the inside, the part will be wrong by exactly two times the tool radius.
Catch this in dry run by watching the tool path on the graphic. If the offset looks wrong, stop and check your G41/G42 call. It is almost always a direction error, not a value error.
Peck Drilling That Retracts Too Far or Not Far Enough
Peck drilling cycles retract between pecks to clear chips. If the retract distance is too small, chips pack in the hole and the drill breaks. If the retract distance is too large, you waste time on every peck.
During dry run, watch the Z motion on a peck cycle. The tool should peck down, retract to the R-plane or clear, then peck again. If the retract goes all the way to the initial plane on every peck, your cycle is inefficient. If the retract barely moves, your chip clearance is insufficient. Adjust the Q value or the peck increment before the first cut.
Missing Dwell Times at the Bottom of Holes
G82 drilling cycles include a dwell at the bottom of the hole. If you programmed G82 but forgot the P value, the cycle runs as G81 — no dwell, no pause. On deep holes, this can cause the drill to bottom out with full force and break the tip.
During dry run, check that the spindle actually pauses at the bottom of each hole. On most controllers, you can see the spindle speed drop to zero during the dwell. If the spindle keeps spinning, the P value is missing or set to zero. Fix it before you cut.
Building a Dry Run Checklist That Saves Time
Write down the steps and follow them every time. Not because you are forgetful. Because under pressure, even experienced operators skip steps. The checklist keeps you honest.
Check the work offsets against the last known good values. Check the tool lengths. Raise Z zero by 50mm. Set all overrides to zero. Load the program. Run in single block. Watch every move. Confirm the first five moves manually on the DRO. Then run continuous. Raise the override to twenty-five percent. Cut scrap. Measure. Adjust. Repeat.
This routine takes an extra fifteen minutes at the start of every job. It eliminates the two-hour panic that comes from finding a crash after the tool has already hit the vise.
When Dry Run Is Not Enough: High-Speed and Five-Axis Work
High-Speed Machining Requires Tighter Checks
On high-speed machines, the lookahead function calculates tool paths several blocks ahead. If your program has sharp corners or sudden direction changes, the controller may not decelerate in time and the tool will overshoot the corner. Dry run at reduced speed catches most of these. But for critical high-speed work, run the program at the actual programmed feed rate on scrap, not at a reduced override.
The reduced-override test does not replicate the actual servo behavior. What looks smooth at twenty-five percent feed can jerk at full speed because the axis motors cannot accelerate fast enough.
Five-Axis Dry Run Needs Kinematic Simulation
A five-axis program that looks fine in three-axis dry run can still crash because the A or B axis rotation drives the tool holder into a clamp. The tool tip might be clear, but the tool body is not.
If your controller supports kinematic simulation, run it. This mode simulates the actual rotation of the tool axis and checks for collisions between the tool holder and the machine structure. If kinematic simulation is not available, run the program at very slow feed with the spindle off and watch the axis movements carefully. Any axis that approaches its limit should be verified manually.
The Mindset That Separates Good Operators From Lucky Ones
Every operator who has never crashed a machine is either very new or very lucky. The ones who stay lucky are the ones who run dry run every single time, even for programs they have run a hundred times before.
A program that worked yesterday can crash today if the work offset shifted, the tool wore, or the stock moved in the vise. Dry run takes minutes. A crash takes hours. The math is simple. The discipline is the hard part.
Run the dry run. Watch every move. Measure after the first cut. Adjust. Run again. That is the process. There is no shortcut, and there is no substitute.
