By Alberto Schileo and Curtis D. Thomson
© April 2004 by the authors and ThePuristS.com
|
Part 2: Manufacturing Any watch, of course, starts with an idea. The design team takes this idea and transposes it into a few initial sketches, which in turn will become the first CAD renditions of the future watch. Then comes the phase of the technical choices, when problems are examined, solutions are devised, and issues are settled. Eventually, at the end of the process, the design team will produce a full set of CAD drawings which constitute the detailed blueprints of the future watch.
Once the watch schematics are complete, the design department will take the CAD drawings of each watch component they have been working on, and prepare these to be made through the equipment which is installed on the ground floor of the Daniel Roth/Gérald Genta factory in Le Sentier. In a nutshell, this involves breaking each piece down to the basic geometric shapes which make it up, so that the computer-controlled machines can process these one by one. ![]() Besides the usual assortment of Schaublin lathes, vertical drilling machines, etc., at Daniel Roth they basically use two different types of industrial machines, which are both computer-controlled: two are automated milling machines (which they acquired only two years ago), while the other two are spark-erosion machines. ![]()
The person responsible for the mechanical workshops then takes the break-downs of the CAD drawings and comes up with the most rational and optimized use of tools necessary to make each part. ![]() For a tourbillon main plate, for exemple, up to 30 different tools may be necessary (milling heads and/or drills), and for each, the computer will calculate a specific path, trying to minimize the need to switch back and forth among tools so as to save processing time. Despite this optimization, milling a single side of a tourbillon main plate still takes roughly 7 to 8 minutes, which also includes, however, the time necessary to switch tools (about 2 to 3 minutes of the total processing time).
Once the required tools are loaded onto the machine (a long process which requires first mounting each tool in a head which the machine can grip, then placing each tool in the machine, and then calibrating the machine to see, through the reflection of a laser beam, how far down from the tip of the drill head each tool actually protrudes), they let the machine make a single part.
This first test part is then checked against the CAD drawings to see if it conforms to specs. Initially, this will rarely be the case for two reasons: first, the tools themselves - though extremely rigid - have a tiny degree of flex when they are being used with lateral force applied to them; second, as they mill a piece on both sides, resulting in a thin layer of material being left between the two, the material itself will actually “bulge up” due to the pressure being applied by the machine during the milling process. ![]() The material normally used at Daniel Roth for the main plates is called Arcap, which is an improved version of “german silver” (a.k.a. “maillechort”, from the combined name of the two people who invented this alloy: Mr. Maille and Mr. Chorier). This version is not only harder but has better rigidity, thus making it more adapted to being worked on through computerized milling machines. ![]() Once the mechanical workshop team is happy with the corrections that have been done, and they are able to obtain a piece which meets the specs, they load the machine with 16 to 20 blanks, held in place through positioning holes, and lets the machine process each of them in turn. Each tool and each step is repeated for each plate before the tool is changed, so effectively the 16-20 plates advance simultaneously along the steps required to finish them. It takes about three hours for the machine to finish working on a full set of parts. ![]() After the machining is finished, each piece is checked through a computerized system which allows them to overlay the outlines of the CAD drawing onto a 176:1 scale picture of the actual piece, and a number of visual checks are then made to ensure that the piece actually corresponds to the drawing (see the thin red lines on the image below on the right). ![]()
As the very last step of the fabrication process, the machine will actually mill around the outline of the plate itself, occasionally leaving an aluminum-foil thin layer of material which keeps the plate attached to the rest of its blank. ![]() The plate is then detached from the disc it was made out of, mounted on a traditional lathe (a Schaublin 70), and “cleaned-up”. ![]()
The last steps consist of drilling the holes for the dial feet screws on the sides of the plate, and making the winding stem hole. This particular hole then requires the plate to go through a spark-erosion step, to open up the hole which accommodates the parts making up the time setting and winding gears operated through the stem.
For pieces made via spark erosion, the process is similar, meaning that they too start with the CAD drawing, which is then transposed into a “path” that the machine follows.
This process gives a very precise finish, but is only adapted to small productions due to the very long times involved in the processing. Often, because of this, plates are stacked 10 to 20mm deep, depending on the type of work being done, to allow several parts to be made simultaneously. ![]() The spark erosion process is very interesting because it operates with no direct contact between the part being worked on and the part of the machine which is doing the cutting. This actually consists of a very thin wire which is threaded through the part, and which is then discarded through the back of the machine. Current is applied to the part being cut and to the wire, and this creates a spark between the two which literally vaporizes the metal immediately around the wire. This, of course, also thins out the wire itself, which cannot therefore be used again and is then recycled.
For normal surfaces, the wire goes around the part three times, whereas where very precise surfaces are required (i.e. for working, contact parts), the wire goes back and forth seven different times. The work actually takes place underwater, which is necessary to provide better conductivity and to cool down the whole apparatus, so it is hard to take accurate pictures of the machine in action!
Whenever they are working on small, thin parts such as springs, etc., in between the CAD milling and the spark-erosion cutting, the plate is sent out to be hardened, so that when they get to the point of cutting the very thin blade of the spring the material will actually resist and not bend out of shape. Once again, all parts produced through spark erosion are also checked for tolerances against their respective CAD drawings after each production run, through the same equipment and process I described before.
Daniel Roth Discussion Board
|
ThePuristS.com Home Page
|