Site: Denken Engineering Company, Ltd.
Technology and Information Systems Development Division
2-1-40 Sikiden-cho
Oita 879, Japan

Date Visited: 12 December 1995

JTEC/WTEC Attendees: A. Lightman (report author), D. Shelton, L. Weiss


Syoji Hiura

Hiroshi Kibe

Toshiyuki Akamine


The RP concept was formulated circa 1985 by Akamine (Autostrade) in his desire to have 3D models of CAD designs he was creating. Akamine is a software engineer, and Autostrade does CAD design for clients. He approached the local prefecture for development guidance and assistance, and he was put in contact with Denken Engineering.

Denken was founded in 1976, and its first products supported the local steel industries in the areas of automation, hardware and software, robotics, and mechatronics. It has expanded into LSI fabrication and testing. The Solid Laser Plotter (SLP) series of resin-based rapid prototyping machines is a joint effort of Denken Engineering and Autostrade Co., Ltd., assisted by Nippon Kayaku with resin development and by Shimadzu with laser development. Current size is 300 employees, ¥3 billion per year. Autostrade is a 5-person shop, and the SLP work is a joint effort with a staff of 5-6. Denken financed the SLP product development from internal R&D funds. It filed its initial patent application in 1990, and it was anticipated to become profitable starting in 1996.

Denken participates in Professor Nakagawa's (University of Tokyo) program to coordinate vendors. Vendors and academics relevant to the topic areas meet monthly for each of these subjects: materials, interfacing, standards, and promotion (one meeting per topic area). Expenses are covered by MITI.


The SLP series is based on laser photolithography. It bears strong resemblance to the Mitsui Zosen COLAMM system, where the model is built inverted, attached to a platform that rises as each successive layer is attached to the bottom-most face. The liquid resin layer is deposited on a specially prepared plate that is transparent to the laser. The platform and previously built structure are lowered into the resin, leaving between the part and the plate a liquid film that has the correct thickness for the next layer. The new layer is "written" from beneath the plate, using conventional photoexposure, moving the exposing beam via an XY plotter assembly. After the layer is written, the new structure is raised, separating the layer from the plate, and the process is repeated until all layers are fabricated. The final assembly is removed from the support plate, and postcure, if needed, is performed. The part can be finished as required for the end application.



Denken's main goal is to provide inexpensive desktop modeling systems for both network and personal use.


Denken's current priorities are (1) reduce cost, (2) heighten resolution, (3) increase accuracy, and (4) improve reliability.

Improved reliability is ranked only fourth at this time because of the high reliability already achieved. Denken representatives also acknowledge that their capability still lags behind U.S. standards.

Future Goals

Future goals include engineering models (the SLP-5000 addresses this need) and a microfabricator.


Denken currently uses one photoacrylate. It has restrictions as a result of the exposure wavelength (680 nm). Denken's new Nd:YAG SHG (second harmonic generation) laser (946 nm fundamental, 473 nm SHG) will permit alternative materials and Denken representatives indicated that they are looking to use an epoxy. They have a very small research staff and they have not performed extensive literature searches. (They were quite interested to learn of the DeSoto resins formulated for Quadrax, which also operates in the visible range. They said they will make contact with those firms to obtain more information.) Their current resin is a two-part system, mixed at time of use.


The current SLP is directed to the modeling environment. Engineering applications are a future goal. Most SLP installations are in design or academic environments. The Ministry of Education subsidizes academic purchases up to 50%.


Denken has produced three machines: SLP-3000 (discontinued), SLP-4000 (current model), and SLP-5000 (first delivery was scheduled for 2/96). Sales at the time of the team's visit were 3, 24, and 1, respectively, for the 3 machines. Anticipated sales for calendar 1996 were 50 units total.

The SLP-3000 was discontinued because of its size and weight; it did not fit the desktop office environment concept for the personal modeler. The newer models are closer to the desired size. (It should be noted that this system is based on photopolymerization of a liquid resin. Use of this material in an office environment may present a health risk.) Denken claims that the technology factor limiting overall performance is the interaction of the laser with the photopolymer.

Build Chamber

The part is built attached to an elevator that rises from the exposure window (build is inverted compared to the common stereolithography systems). The liquid polymer is exposed from beneath by laser illumination through a specially prepared window. The window is held in a frame that acts as a spillover gate to define the film thickness on the window (0.2 mm). Resin is pumped from the surrounding trough onto the window, where it is evenly spread over the surface by a roller recoater. The excess resin spills over the gate back into the trough. Sufficient resin is initially loaded into the trough to build the part. The resin is a two-part formulation, mixed at the time of use.

Once the resin film is prepared, the build elevator is lowered into the liquid film, leaving the desired next layer thickness (usually 0.1-0.2 mm; experiments have been tried at 0.03 mm). The new layer is polymerized. The elevator then rises to allow a new layer of liquid to be spread, and the process is repeated until all the layers have been formed. All the formed material, except for the layer being created, are in air during the build process rather than in the vat of liquid resin, as in the standard systems.

The window is coated with a proprietary material for easy release of the most recently formed layer. The window tint appears milky, and there may be some scattering that could limit the minimum line width. The elevator platen is acrylic, which enhances adhesion. A support structure is created to provide sufficient bonding to the elevator (so that separation is at the window when the elevator rises and not within the part) and to support overhangs. Part removal is the same as for other systems (cut away the supports). In cases of large surface contact, layer damage could occur. This may limit the geometries that can be built.


Laser680 nm laser diode supplied by Audio Technica (there is a joint patent for this application). The laser power is 10 mW and the laser footprint is 0.1 mm in diameter at the part surface. (Diode lasers are inherently elliptical. Anamorphic optics would have to be used to form a circular beam. This is common in compact disk player applications, the source of this laser.)
Work envelope150 mm (X) x 150 mm (Y) x 200 mm (Z)
Cost¥4.9 million, machine alone
¥8 million with software and control computer
Annual maintenance with laser replacement
Cost driversXY translator is the most expensive subsystem, the laser is the next most expensive
Mechanical accuracy X,Y: 0.04 mm, Z: 0.1 mm
SpeedA 3-blade propeller, 25 mm high and 75-100 mm diameter, took 10 hours to build


Laser Shimadzu Nd:YAG (946 nm) with SHG (473 nm) 20 mW Work envelope 220 mm (X) x 200 mm (Y) x 225 mm (Z) Cost ¥15 million, complete Rationale The shorter wavelength allows use of photoinitiator-resin combinations, which permit more accurate model fabrication, providing potential to build models for engineering applications. Cost driver Laser


The SLP series can accept CAD data in a variety of formats: SLP, STL, CT, MR, 3D digitizer data, and FE mesh. Autostrade's background CAD programming provides experience for easy modification of data to meet specific customer requirements.

Akamine is considering applying finite element analysis (FEA) to determine optimal scanning strategies in order to minimize distortion for specific geometries through modeling the laser-polymer interaction and the resultant stress formation.

Published: September 1996; WTEC Hyper-Librarian