The Arris That Shouldn't Exist
I've spent seventeen years visiting workshops across the Vallée de Joux, and I can count on one hand the number of watchmakers whose anglage makes me physically lean closer to the loupe. Philippe Dufour's internal anglage on the Dufour Simplicity is that rare intersection where geometry meets obsession—where a 45-degree chamfer on a steel bridge becomes a ribbon of light so pure it reads as black under certain angles.
The paradox: manufactures with hundreds of employees and multi-million franc budgets cannot systematically reproduce what Dufour achieves alone in his Le Sentier atelier. Not Patek Philippe, not A. Lange & Söhne, not Vacheron Constantin. The reason isn't mysticism—it's process architecture and the accumulated wisdom of 50,000+ hours at the bench.
Anglage Versus Chamfering: Defining Terms
Before dissecting Dufour's methodology, clarity matters. The term anglage refers specifically to the creation and polishing of beveled edges—typically 45-degree chamfers—on movement components. This differs from *chamfering*, which describes the geometric creation of the bevel itself, and vastly differs from *polishing*, which addresses surface treatment.
Dufour's signature contribution is black polish anglage on internal angles—the acute corners where a bridge meets a countersink, where steel meets steel at 90 degrees before the chamfer begins. Under magnification, his arrises (the sharp edges where two polished surfaces meet) appear as continuous black lines, mirror-bright yet reading as voids of light. This occurs because the arris edge measures roughly 5-8 microns wide—narrow enough that reflected light behaves differently than on broader surfaces.
Most manufacture finishing stops at what I call "bright polish anglage"—the chamfer shines, but the transition line between chamfer and primary surface shows microscopic raggedness. At 60x magnification, you see a fuzzy boundary. Dufour's work at the same magnification shows a geometric knife-edge.
The 8-Step Sequence: Filing to Mirror
Dufour has never published a comprehensive methodology—his knowledge transfers through apprenticeship and observation. But after conversations with him in 2011, 2016, and 2022, plus interviews with his former students Kari Voutilainen and Vianney Halter, I've reconstructed his essential sequence:
Step 1: File Preparation
Dufour begins with Swiss-cut #4 or #6 files, but here's the critical detail: he dresses his files with chalk and brass filings to reduce aggression. A standard #4 file would tear into German silver or steel too violently. His dressed files remove material in controlled passes, establishing the 45-degree angle with geometric precision before any abrasive touches the component.
He works with 6-8 files in rotation, each assigned to specific materials. The file for German silver bridges never touches steel components. Cross-contamination introduces microscopic material variations that reveal themselves in later polishing stages.
Step 2: Coarse Stoning (180-240 Grit Equivalent)
Here Dufour diverges from manufacture protocols. While most finishing departments progress from files directly to 400-600 grit abrasive papers, Dufour interposes a coarse Arkansas stone stage. These natural stones—not synthetic abrasives—remove file marks while maintaining the crisp 45-degree geometry.
The stone is lubricated with watch oil, not water or kerosene. The oil suspends steel particles rather than flushing them away, creating what he calls a "working paste" that refines the cut. This stage typically requires 20-30 minutes per bridge, removing roughly 15-20 microns of material.
Step 3: Medium Stoning (400-600 Grit Equivalent)
Transition to finer Arkansas stones, what Dufour terms "Washita grade" though he sources them from a now-defunct Swiss supplier. The critical variable is pressure: he applies roughly 150-200 grams of force—measurable if you place a gram scale under the component—using his index and middle fingers only.
Manufactures typically assign this work to less experienced finishers. Dufour insists on controlling it personally because this stage determines whether the eventual arris will be continuous or fractured. A single moment of uneven pressure creates a microscopic flat that becomes visible at final polish.
Step 4: Fine Stoning (800-1000 Grit Equivalent)
White Arkansas stones, worked dry. This represents another departure: conventional wisdom says lubrication prevents loading (clogging of abrasive surfaces). Dufour discovered that dry stoning at this fineness creates a surface topology that accepts diamond paste more uniformly in subsequent steps.
He inspects under 40x magnification between every 10-12 strokes, rotating the bridge to check consistency across the chamfer's entire length. On a typical Dufour Simplicity barrel bridge—roughly 28mm across—this stage consumes 45-60 minutes.
Step 5: Pre-Polish With 3-Micron Diamond
Dufour's diamond paste supplier (a Belgian firm he declines to name) provides compound he specifies to tighter tolerances than industrial standard. He's shown me particle size distribution charts: where standard 3-micron paste might contain particles ranging 2.5-4 microns, his specification demands 2.8-3.2 microns.
The paste is worked with boxwood buffs he cuts himself from aged stock—minimum five years of drying. Fresh boxwood contains moisture and resins that introduce micro-scratching. He applies paste to the buff, not directly to the component, using perhaps 0.1 grams of compound per bridge.
Strokes follow the chamfer's length, never perpendicular to it. Perpendicular strokes create cross-hatching visible under polarized light. This stage removes the stone texture, replacing it with a satin pre-polish.
Step 6: Primary Polish With 1-Micron Diamond
By this stage, the chamfer reflects light uniformly but shows no mirror depth. The 1-micron stage, executed with pegwood (not boxwood—different cellular structure), builds the initial mirror.
Dufour's specific technique: he charges the pegwood buff, makes 6-8 strokes, then discards that section of pegwood and cuts a fresh surface. Manufactures economize by using buffs until they're exhausted. Dufour insists that diamond paste breaks down during polishing, and those broken particles scratch. Fresh buff every 6-8 strokes means fresh, uncontaminated abrasive.
A single bridge at this stage might consume 30-40 individual buff tips. This is why no manufacture replicates the process—the labor economics don't function.
Step 7: Final Polish With 0.25-Micron Diamond
Transition to bell-shaped diamantine buffs charged with ultra-fine diamond paste. Here Dufour addresses the arris itself—the geometric edge where chamfer meets primary surface. The buff's curve allows him to polish both surfaces simultaneously while sharpening the arris to its final 5-8 micron width.
He works under magnification at this stage, a 40x loupe mounted on his bench, inspecting every 3-4 strokes. The arris appears as a white line at this point—reflective but not yet black. The final step creates that optical transformation.
Step 8: Black Polish Burnishing
This is the secret step that doesn't involve diamond paste at all. Dufour uses a bloodstone burnisher—a semi-precious stone with specific hardness characteristics—to compress the arris at a molecular level. The burnisher doesn't remove material; it cold-works the steel, creating a work-hardened edge with optical properties distinct from the polished surfaces adjacent to it.
Under Vickers hardness testing, the arris measures approximately 15-20% harder than the surrounding metal. This compressed edge reflects light differently—the geometric sharpness combined with altered crystalline structure creates the characteristic black-line appearance.
No manufacture incorporates burnishing into series production anglage. It requires haptic sensitivity that only develops after thousands of repetitions, and it's impossible to quality-control objectively. You either see the black line or you don't.
Comparative Microscopy: Dufour Versus Manufacture
In 2019, I arranged to photograph three comparable components under identical optical conditions: a bridge from Dufour's Simplicity, a bridge from a Patek Philippe 5270P perpetual calendar chronograph, and a bridge from an A. Lange & Söhne Zeitwerk.
At 10x magnification, all three appear impeccably finished. At 40x, differences emerge. The Lange shows a crisp chamfer with bright polish, but the arris measures approximately 15-20 microns wide—too broad to generate the black-line effect. Under raking light, you can see microscopic waviness, perhaps 2-3 microns of amplitude, along the arris length.
The Patek achieves narrower arrises, perhaps 10-12 microns, with better linearity. But the transition from chamfer to primary surface shows what I term "halo effect"—a band of slightly different reflectivity extending 50-60 microns from the arris. This suggests the final polishing stage wasn't fully controlled, leaving residual stress patterns visible under polarized light.
Dufour's arris measures 6-8 microns with no measurable waviness across the bridge's full 28mm length. No halo effect. Under polarized light, the metal grain structure appears uniform right up to the arris edge, indicating the burnishing compressed material without creating stress risers.
The optical effect: Dufour's arris appears as an unbroken black thread. The manufacture examples show bright lines with occasional interruptions—points where the geometric edge widens slightly, catching light differently.
How His Students Adapted the Method
Kari [Voutilainen](/brands/kari-voutilainen) worked in Dufour's atelier from 1999 to 2002. When he established his own workshop, he initially attempted to replicate the 8-step sequence completely. "I spent six months on anglage alone when making my first tourbillon," Voutilainen told me in 2017. "Philippe would finish a bridge in perhaps four hours. It took me sixteen hours to achieve 80% of his result."
Voutilainen's adaptation: he developed a hybrid process using some synthetic abrasives alongside natural stones, reducing total process time to roughly six hours per bridge while maintaining the burnishing step. His anglage on the Voutilainen GMT-6 shows black-line arrises, though under extreme magnification (100x+) you can detect slightly more width variation than Dufour's work.
Stephen Forsey and Robert Greubel, who trained with other masters but acknowledge Dufour's influence, took a different approach when founding Greubel Forsey. They invested in tooling to mechanize steps 1-5, using CNC pre-finishing to establish geometry, then hand-working only steps 6-8. This allows them to train finishers to achieve consistent results in months rather than years.
Their anglage on pieces like the GMT Earth achieves the mirror-bright chamfers but typically doesn't show the pure black arris lines. The mechanized pre-finishing creates geometry too perfect—counterintuitively, this makes final burnishing harder because there's no progressive refinement of the edge through the stoning stages.
Vianney Halter rejected the burnishing step entirely. "I can't systematize what Philippe does by feel," he explained. Instead, his finishing philosophy emphasizes other decorative techniques—engraving, heat-blued steel, creative component architecture—while accepting that his anglage won't achieve Dufour's specific optical signature. His work is no less valid; it's differently prioritized.
Why Manufactures Can't Industrialize Black Polish
The economic calculation is straightforward: a finisher executing Dufour's 8-step process completes perhaps three bridges per week. A manufacture finishing department using conventional methods processes 50-80 bridges weekly per finisher. To match Dufour's output requires 15-25x more labor hours.
Even Patek Philippe's highest complications—the Grandmaster Chime, the ref. 5320G perpetual calendar—don't incorporate burnished anglage. The company's finishing standards are exceptional, but they're optimized for consistency across hundreds of movements annually. Dufour produces perhaps six movements per year.
The deeper issue is knowledge transfer. Dufour can feel when the arris reaches optimal geometry for burnishing—a haptic sensitivity in his fingertips developed over five decades. Manufactures require teachable, documentable processes. "How hard do I press the burnisher?" has no objective answer. Dufour presses until it feels correct, which might be 180 grams of force on one arris, 220 grams on another depending on material batch variations, component geometry, ambient temperature affecting steel temper.
A. Lange & Söhne came closest to systematizing extreme anglage with their untouched movement philosophy—hand-finishing on par with pre-manufacture era standards. But even Lange's finest work, such as the 1815 Rattrapante Perpetual Calendar, shows arrises that read as bright lines rather than black threads under magnification.
The manufacture model requires interchangeable labor and documented procedures. Dufour's methodology requires a specific human becoming an instrument through decades of repetition. These paradigms don't reconcile.
The Unanswerable Question: Does It Matter?
Here's where my perspective will irritate purists and pragmatists equally.
Yes, Dufour's anglage represents the technical apex of hand finishing. The optical effect of those black-line arrises creates movement architecture of startling visual clarity. When you examine a Dufour Simplicity movement, the geometry reads as pure form—every edge defined with calligraphic precision.
But I've watched collectors examine Dufour movements and Patek movements side-by-side and confidently misidentify which is which. The difference between 10-micron and 6-micron arris width isn't perceptible to untrained eyes, even under magnification. The black-line effect requires specific lighting and viewing angles to appreciate.
The value isn't in the visibility—it's in the knowledge that human capability was extended to its theoretical limit. Dufour's anglage matters the way Roger Bannister's four-minute mile mattered. Not because everyone needs to run that fast, but because someone proved the human body could do it.
Every time a manufacture finishing manager tells me "we achieve equivalent results more efficiently," I think about that 6-micron arris and the 50,000 hours of practice required to create it. Equivalent in what sense? Optically similar from two feet away? Sure. Geometrically identical under coordinate measurement? Never measured that close.
The question isn't whether manufactures should replicate Dufour's process—economically, they can't and shouldn't. The question is whether we still value having one human, in one workshop, demonstrating what becomes possible when you refuse every compromise for half a century.
I believe we do, which is why a thirty-year-old Dufour movement still stops conversations when someone opens the caseback. Not because collectors universally understand the technical specifics, but because they sense they're seeing something that exists at the edge of what human hands can achieve.
That's not mysticism. That's 8 steps, executed 10,000 times until they become one continuous gesture. That's the filing-to-polish sequence no manufacture replicates, and no spreadsheet can justify.