weekly schedule.

Time block	Wed	Thu	Fri	Sat	Sun	Mon	Tue	Wed
Global class	3 h
Local class		1,5 h
Research			1 h
Design		2 h	2,5 h
Fabrication		1 h	1 h			3 h
Documentation		2 h	2 h	3 h		1 h
Review

Week 07 · Computer-Controlled Machining.

overview.

This week is about large-format CNC machining — designing parts for real-scale fabrication, generating toolpaths, setting up the machine, and physically building something that a person can actually use. It’s the most physically demanding week so far, and the one where things are most likely to go wrong in interesting ways.

The machine at Fab Lab León is a TEC-CAM 500 (Pérez Camps), a large-format CNC router with a vacuum table and a RichAuto A11 controller. My instructor Pablo has been working with this machine for years, so I was in good hands — even when things didn’t go as planned.

---

group assignment.

The group assignment for this week covers: operating the machine safely, documenting the test for feeds and speeds, and characterising the joint behaviour for the material we’re using.

→ [Link to Fab Lab León group assignment page — to be added]

The machine: TEC-CAM 500.

The TEC-CAM 500 is a large-format CNC router manufactured by Pérez Camps. Its nominal footprint is 1800 × 3000 mm, with a working area of 1220 × 2440 mm (the difference accounts for the machine structure and mechanical limits). Key specs from the lab documentation:

Parameter	Value
Nominal dimensions	1800 × 3000 mm
Working area	1220 × 2440 mm
Spindle	2.2 kW, water-cooled
Gantry height	120 mm
Repeatability	0.025 mm
Fixturing	Vacuum pump, 3 independent channels
Weight	900 kg
Max material thickness	~40 mm (tool-dependent)
Accepted file formats	STL, DXF, and others
Compatible materials	Plywood, MDF, bakelite, PP, solid wood, PVC, brass, foam, acrylic

Safety.

Before touching anything in the CNC room, everyone signs a safety document provided by the lab. The TEC-CAM 500 is the most dangerous machine in the lab — the spindle runs at up to 18,000 RPM, the gantry moves fast, and a mistake can be irreversible. The rules are not suggestions.

The rules split into two categories:

Personal safety:

• One person operates the machine at a time.
• Never disconnect cables by pulling, or with wet hands.
• Hair tied back; no loose items — necklaces, scarves, bracelets — that could catch a spinning tool.
• Maintain safe distance from the gantry during movement and cutting.
• PPE inside the CNC room: ear protection and safety glasses at all times; dust mask when needed.

Machine rules:

• Any malfunction or emergency: press the emergency stop or disconnect the machine; notify the Fab Lab Manager (Nuria, 661 802 774).
• Do not use if there is burning smell, abnormal noise, or smoke.
• Cleaning and unclogging only with the machine disconnected.
• Repair and maintenance under Fab Lab Manager supervision only.
• No metals or liquids in contact with internal components.
• Never use compressed air to clean the electrical panel.
• Never power on with the pendrive already connected to the controller.
• Always verify material thickness before machining.
• Format the pendrive before use; save each job with a new filename — never overwrite.
• Always activate the vacuum fixturing system; check for obstructions before use.
• With materials that produce fine chips, activate the dust extraction system.
• Clean the machine and workspace after every session.

The vacuum table.

The table uses a vacuum pump to hold the material down — no clamps needed. It has three independent suction channels, each controlled by its own ABIERTO/CERRADO switch on the front of the machine. For a full-size sheet, all three zones are open. For smaller boards, you can close unused zones to concentrate suction where the material actually sits — but you need to surround the board with scrap pieces to maintain the sealed area and prevent air leaks.

Underneath the working material sits the martyr board — a 3 mm thick sacrificial layer that absorbs the final pass of the cutter when it goes slightly past the material thickness. It wears down over time, which is why the surface looks like a map of every job that has ever run on the machine.

Material.

The board I used is poplar-core plywood with birch face veneers. The stamp on the edge reads:

EFFICIENCY POPLAR 1508 Y+V2V3 PR2 · 2500×1220×18mm · CE EN 13.986 EN 636-1 E1

Nominal thickness is 18 mm, but I always measure the actual thickness twice with calipers before entering any value in Aspire. In CNC work, measuring twice before cutting once is not a cliché — it’s the difference between a clean job and a ruined board.

Measure twice, cut once.

The real thickness of this board was 18.15 mm. There is a label on the gantry that says — in capital letters — “ALWAYS CHECK THE MATERIAL THICKNESS”. It is there for a reason.

The end mill.

The standard mill at Fab Lab León for plywood cutting is a 6 mm single-flute straight end mill (labio recto in the tool library). Single-flute because it evacuates chips more efficiently than multi-flute tools on wood, which prevents heat build-up and recutting.

Installing the end mill (done before powering on the machine):

1. Snap the collet into the collet nut — it clicks into place.
2. Insert the end mill into the collet.
3. Thread the nut onto the spindle: rotate backwards until you feel a click, then forward to tighten.
4. Tighten with two flat spanners: 21 mm on the collet nut, 30 mm on the spindle flat — one-hand force only.

Electrical startup.

The machine is powered by raising the differential switches in the electrical panel on the wall — not from a button on the machine itself. The dust extractor is a separate unit that gets switched on just before pressing RUN.

Feeds and speeds test.

Before cutting the actual project, Pablo ran a test to confirm that 9,000 mm/min — the lab standard for 6 mm single-flute on 18 mm plywood at 18,000 RPM — produces acceptable results. We cut three pockets at 3,000 / 6,000 / 9,000 mm/min and compared the wall finish.

The chip load formula for reference:

Feed rate (mm/min) = Chip load × flutes × RPM

For a 6 mm single-flute end mill on plywood, chip load range is 0.28–0.30 mm:

Chip load	Calculated feed rate
0.28 mm	5,040 mm/min
0.30 mm	5,400 mm/min

The lab default of 9,000 mm/min sits above the calculated range, but in practice the wall quality at all three speeds was indistinguishable on this material. No visible difference in surface finish, no burning, no chatter. 9,000 mm/min confirmed as the lab standard.

Full confirmed parameters:

Parameter	Value
End mill	6 mm, 1 flute, straight
Spindle speed	18,000 RPM
Feed rate	9,000 mm/min
Plunge rate	1,500 mm/min
Pass depth	10 mm max
Ramp angle	30° (always enabled)
Cut depth	material thickness + 0.15 mm

Joint test.

T-bone reliefs are required at every interior corner because a round end mill cannot cut a sharp 90° corner. The T-bone diameter is set equal to the end mill diameter — 6 mm in this case. I cut a test joint in the same material to verify fit before committing to the stool geometry. Slot width:

slot_width = measured_thickness + tolerance = 18.15 + 0.20 = 18.35 mm

---

individual assignment.

A note on CAD.

I came into Fab Academy with no CAD background. None. My professional experience is in industrial automation which involves a lot of structured thinking but almost no 3D modelling. The closest I’d been to parametric design before this was laying out control panel schematics in AutoCAD.

Fusion 360 is, for me right now, all three things at once: a technical challenge I’m committed to overcoming, a tool I genuinely need for the final project (a height-adjustable standing desk with four synchronized actuators), and a source of real frustration when the software behaves in ways I don’t yet understand. The parametric workflow, the sketch constraints, the CAM environment — each of these is its own learning curve on top of the others.

This week was the first time I used Fusion for something that ended up physically cut in 18 mm plywood. That gap between the screen and the machine — where a wrong number becomes a miscut joint that can’t be undone — is what makes CAD feel high-stakes in a way that code doesn’t. I expect Fusion to be one of the hardest ongoing challenges of this programme. I’m not going to pretend otherwise.

Design concept — spiral 0.

For spiral 0, Fab Lab León provided a reference Fusion 360 file — a flat-pack stool using a cross-slot joint — as a starting point to gain familiarity with the CNC workflow without spending the first session fighting the CAD tool from scratch. The concept: two identical pieces of plywood that cross-lock at 90° to form a stable seat at approximately 50 cm height. No fasteners. No glue. Assembly by hand. The cross-slot joint is the only connection mechanism.

Before committing to CNC time, I prototyped the concept at scale on the laser cutter. The laser doesn’t replicate the T-bone geometry — the kerf is essentially zero — but it was enough to validate the proportions and confirm the slot position and assembly logic.

This is a workflow I’ll repeat: prototype fast on the laser, then move to CNC once the geometry is confirmed. It saves material, machine time, and avoids forcing oversized joints in 18 mm plywood.

Design in Fusion 360 — spiral 0.

The Fusion 360 file provided by the lab contained the stool geometry as a set of sketches. Key parameters documented from the file:

Parameter	Value	Notes
board_thickness	18.15 mm	Measured with calipers on the day
slot_tolerance	0.20 mm	Added to board_thickness for slot width
slot_width	18.35 mm	= board_thickness + slot_tolerance
tbone_radius	3 mm	= end_mill_diameter / 2
piece_height	270 mm	Leg height
piece_width	300 mm	Seat width

The design is two identical pieces. Each piece has: a rectangular cross-slot at mid-height, T-bone reliefs (⌀ 6 mm) at all four interior slot corners, curved cutouts (r = 60 mm) at the inner leg corners to reduce stress concentration, and 4 tabs per piece (8 × 3 mm) to keep pieces attached to the board during cutting.

Note: Spiral 0 works — you can sit on it — but the slot tolerance was insufficient and the seat discs mismatched due to a scaling error in Aspire. Both issues fed directly into the spiral 1 redesign.

CAM setup in VCarve Aspire.

Key things about Aspire that Pablo taught me during this session:

• Postprocessor: Always use “Fab León G-Code” — never change this.
• Material origin: Always bottom-left corner.
• Tool library: Never edit the base library entries — use “Edit” to adjust per-job.
• Toolpath order: Pockets → interior profiles → exterior profiles — cutting in any other order risks losing fixturing mid-job.
• Ramps: Always enable, 30° — prevents direct plunge, extends tool life.
• Critical bug: Aspire forgets the selected tool every time you reopen a toolpath dialog — always reselect before saving, even if it looks correct.
• File naming: Max 8 characters visible on the RichAuto display.
• Save workflow: Save to local disk first, then copy to pendrive — never save directly to pendrive.

Vector troubleshooting: When importing DXF from Fusion 360, open vectors prevent toolpath calculation. Fix: Join Vectors. If vectors are still open after joining, there are overlapping line segments — use the Trim tool to remove duplicates, then Join again. This cost about 20 minutes during my setup.

Final CAM parameters:

Parameter	Value
Material	Plywood, 18.15 mm measured
End mill	6 mm, 1 flute, straight
RPM	18,000
Feed rate	9,000 mm/min
Plunge rate	1,500 mm/min
Pass depth	10 mm
Cut depth	18.30 mm (18.15 + 0.15)
Tabs	4 per piece, 8 × 3 mm
Output format	.TAP
Postprocessor	Fab León G-Code
Output files	PATAS.TAP, ASSUP.TAP, ASINF.TAP
Estimated cut time	4 min 56 sec

Machine operation.

Full startup sequence:

1.  PPE: ear protection + safety glasses.
2.  Install end mill (machine off).
3.  Raise differential switches in the electrical panel.
4.  Controller: HOME → OK  (machine homes to reference position).
5.  Jog to XY origin (board bottom-left corner).
6.  Set 0 for Xand Y axis (button XY→0).
7.  Activate vacuum pump.
8.  Z-probe: press first MENU → and once ON/OFF → probe lowers until contact → Z0 auto-set.
9. Activate dust extractor.
10.  Load file: RUN/PAUSE DELETE → U DISK → select filename → OK.
11. RUN/PAUSE → verify spindle spinning → OK → job starts.
12. Stay at machine, hand near STOP/CANCEL.

Sound is the main feedback during cutting. A consistent tone means everything is normal. Any sudden change — a crack, a pitch shift, a grinding noise — means stop immediately with STOP/CANCEL and assess before continuing.

Milling.

Assembly — spiral 0.

Spiral 1 — tea table.

The problems from spiral 0 were clear enough to act on the same week. Rather than re-cutting the stool with adjusted tolerances, I took the spiral 0 file as a starting point and redesigned it into something new: a low tea table with three legs in a Y configuration, a square top with rounded corners, and an engraved ñ on the surface.

Design in Fusion 360 — spiral 1.

Working from the spiral 0 file, I modified the geometry in Fusion 360 to create the tea table. The key changes were: replacing the two-piece cross-slot with a three-leg Y system, resizing the overall proportions for a lower table height, and adjusting the slot width to fix the tolerance problem from spiral 0.

The slot adjustment was done using Fusion’s Scale tool on the specific slot elements — since the file is a sketch-based DXF import rather than a fully parametric model, there are no parameters to edit directly. The scale factor was calculated as:

scale_factor = target_dimension / current_dimension
36.3 / 36.6 = 0.99180

Laser prototype first.

Before committing to CNC, I cut a quick laser prototype of the new joint geometry to validate the slot fit with the updated tolerance. The laser kerf is negligible compared to the CNC end mill, so the prototype doesn’t replicate the exact fit — but it’s fast and confirms the assembly logic and proportions before spending machine time on 18 mm plywood.

CAM setup for spiral 1.

The same Aspire workflow as spiral 0, with the toolpath order strictly maintained: pockets first (the ñ engraving), then interior profiles (slot cutouts), then exterior profiles (leg and seat outlines).

The cut.

The joint fit was clean. Assembly by hand, no mallet needed. The table is stable under load. The ñ engraving came out sharp.

---

problems and solutions.

Problem	What happened	Root cause	Resolution in spiral 1
Slot too tight	Second piece would not enter the slot cleanly at 90°	Tolerance of 0.20 mm was insufficient; cut edge roughness and slight board variation add friction	Applied scale factor (0.99180) to slot geometry in Fusion; validated with laser prototype before CNC cut; assembly fit confirmed clean
Seat discs mismatched	The two circular discs forming the seat do not align — one overhangs the other	A scaling error during DXF manipulation in Aspire changed the dimensions of one piece relative to the other	Redesigned as tea table in spiral 1; verified all dimensions in Aspire before placing toolpaths
Open vectors from DXF import	Toolpath calculation failed on first import	Overlapping line segments at corners in the Fusion 360 DXF export	Trim overlapping lines in Aspire with the Trim tool, then Join Vectors; verify all contours are closed before assigning toolpaths

---

what I learned.

The CNC is the machine that requires the most preparation before you touch it. By the time you press RUN, most decisions are already locked in — material thickness, tool selection, feed rates, joint tolerances, toolpath order. If any of those is wrong, the error shows up physically in wood, and you cannot undo a cut.

A few things I’m taking forward:

Prototype fast, prototype cheap. The 10-minute laser test was worth more than an hour of simulation in Aspire. I could hold the geometry in my hand and decide whether the concept was worth committing to CNC time. I’ll do this every time.

Tolerances are not guesses. The 0.20 mm tolerance I used was based on the joint test, not intuition. Even so, it wasn’t enough. Spiral 1 tests 0.35 mm and validates with a sample joint before cutting full pieces.

Sound is data. The machine communicates through the cutting sound. A consistent tone means everything is fine. An unexpected change — pitch, rhythm, crack — means stop and assess. I didn’t need this rule this week, but I know I will.

Aspire forgets things. The tool selection bug is real and non-obvious. Every time you reopen a toolpath dialog, verify the tool is still selected before saving.

---

files.

Spiral 0 — stool

• taburete_igf.f3d — Fusion 360 source file (spiral 0)

Spiral 1 — tea table

• mesa_te_igf.f3d — Fusion 360 source file (spiral 1)