Designfax weekly eMagazine

Subscribe Today!

Archives

View Archives

Partners

Manufacturing Center
Product Spotlight

Modern Applications News
Metalworking Ideas For
Today's Job Shops

Tooling and Production
Strategies for large
metalworking plants

Cool! New energy-efficient R290 enclosure air conditioners

Seifert Systems introduces PFAS-free SoliTherm^® SlimLine NEO air conditioners using eco-friendly R290 refrigerant. These units offer high energy efficiency (EER up to 3.6) and a compact, under-8-in. internal depth. Featuring maintenance-free design with external or recessed mounting options, they deliver up to 8,500 BTU/hr, providing flexible cooling solutions for varied industrial enclosure needs. Several models available based on size/cooling capacity needs.
Learn more and see all your options.

Surface inspection: From Army depot to factory floor

Born from U.S. Army requirements for rotorcraft inspection, the GelSight Modulus 3D surface measurement system has surpassed 100 units sold to commercial and Department of Defense customers. The handheld, micron-scale tool with interchangeable probe tips delivers fast, high-res measurements in places traditional tools can't reach.
Read the full article.

What is Cold Metal Fusion?

Cold Metal Fusion is an open industry standard for sinter-based metal additive manufacturing. It combines polymer SLS design freedom with reliable debinding and sintering workflows, enabling complex geometries, lightweighting, lattice structures, conformal cooling channels, and high-precision metal parts with predictable shrink behavior. Now available from TriMech Group, this process offers a faster, cost-effective way to produce strong, high-performance metal parts.
Learn more from TriMech Group.

Cool! Internal threading in hard materials now possible

INSACO has a new capability where they can machine an internal thread in ceramic, sapphire, quartz, and other very hard materials. This advance pushes the boundaries of what's possible to support advanced applications that demand high precision and complexity. Ultra-hard materials are alternatives for when metal can't do the job. Ideal for aerospace, medical, and industrial applications.
Learn more. Video available on right side of page.

ClampDisk micro fastener is new alternative for automotive and consumer electronics

Designed as a unique alternative in assemblies for the automotive and consumer electronics markets, the ClampDisk Press-on Fastener is a newer offering from PennEngineering that delivers a fast, simple way to achieve sheet-to-sheet clamped fastening while replacing the use of standard screws, nuts, and adhesives. ClampDisk eliminates over-installation, cross-threading, stripped screw heads, broken screws, and damaged product. This fastener can be removed easily with a sharp-edged tool.
See how ClampDisk works.

Simplify appliance glass assembly

Henkel's Technomelt PUR 9015 BV/WV is a polyurethane hotmelt adhesive providing high initial strength and long-term durability for glass and large-panel appliance assembly. It enables immediate handling, excellent substrate adhesion, and high thermal resistance, while supporting automated, cost-efficient production. It offers a flexible solution for high-reliability manufacturing.
Learn more.

Made-to-order stamped components for insert molding

Traditionally, OEMs source metal inserts and insert molding services separately. Not anymore. Plastics manufacturers and injection molders are now taking on more of the sourcing responsibility for insert molded parts, and they are partnering with Boker's, who has a long-term proven record for delivering precision stampings with quick turnaround times and ensuring metal inserts are mold-ready upon delivery. Boker's has immediate access to over 2,000 commonly specified and hard-to-find materials.
Learn more.

SDP/SI Shaftloc Fastening System

Shaftloc is a unique, reusable locking device for securely mounting mechanical components like gears and sprockets onto shafts without the need for keyways, set screws, or adhesives. Its simple, two-piece design offers a cost-effective alternative to traditional fasteners, providing high clamping force and vibration resistance. Installed with standard tools, Shaftloc is perfect for designers seeking flexible, hubless mounting solutions. Available in four styles.
Learn more from SDP/SI.

Epoxy engineered for heat-sink bonding

Master Bond EP54TC is a two-component epoxy engineered for heat-sink bonding and thermal management applications. Featuring the highest thermal conductivity in the Master Bond electrically insulating portfolio, it delivers exceptional heat dissipation while remaining electrically non-conductive and compliant with ASTM E595 NASA low outgassing requirements. It supports thin bond lines and efficient void filling to maximize thermal performance.
Learn more.

Metal 3D printing: Right at your desktop

From prototyping to tooling or batch production of end-use parts, the Studio System 2 from Desktop Metal brings metal 3D printing to any office, studio, or lab setting. This powder- and laser-free system consists of an easy-to-adopt two-step process: print using pre-bound metal rod feedstock and then sinter. It requires minimal training and operator intervention. Combined with next-gen Separable Supports and a software-controlled workflow, the Studio System makes metal 3D printing simpler than ever. This platform offers more materials than any other metal extrusion 3D-printing system on the market, including Inconel 625, titanium (Ti64), copper, tool steels, and stainless steels.
View the video and learn more.

Metal 3D printing: EOS adds four new materials

Industrial 3D-printing supplier EOS has added four new metal additive manufacturing materials to its portfolio: an iron-nickel alloy that boasts stability under fluctuating temps, a nickel alloy with high strength and extreme corrosion resistance, a low-alloyed steel prized for its high toughness and strength, and an industrial-grade stainless steel. Each has been optimized for EOS Laser Powder Bed Fusion systems.
Get all the details.

Application Note: Disc springs in mechanical braking system

Braking systems for off-highway equipment are commonly designed to be hydraulically actuated, but without an additional fail-safe system, this design alone has limited reliability. If a hydraulic seal is compromised, or the hydraulic cylinder loses pressure for any reason, the brakes fail. One solid mechanical back-up design uses SPIROL disc springs.
Read the full article.

Configurable welding platform for flexible manufacturing

Emerson's new Branson Polaris Ultrasonic Welding Platform offers a highly configurable, smart solution for advanced manufacturing. It features secure connectivity and real-time control to join diverse materials, from medical devices to food packaging. With adaptable power supplies and actuators, the system scales from benchtop lab trials to fully automated production lines, optimizing footprint and data storage to meet complex application needs.
Learn more.

SPIROL receives 2025 Supplier Excellence Recognition Award from Caterpillar

Kudos to SPIROL! The engineered fasteners manufacturer has received the 2025 Supplier Excellence Recognition Award from Caterpillar Inc. This prestigious award recognizes suppliers who demonstrate world-class performance and a sustained commitment to quality, delivery, and operational excellence.
Read the full article.

Eliminate cotters, bolts, nuts with SLIC Pin^®

The SLIC Pin (Self-Locking Implanted Cotter Pin) from Pivot Point is a pin and cotter all in one. This one-piece locking clevis pin is cost saving, fast, and secure. It functions as a quick locking pin wherever you need a fast-lock function. It features a spring-loaded plunger that functions as an easy insertion ramp. This revolutionary fastening pin is very popular and used successfully in a wide range of applications.
Learn more.

MIT researchers build world's largest collection of Olympiad-level math problems -- open to everyone

By Rachel Gordon, MIT CSAIL

Every year, the countries competing in the International Mathematical Olympiad arrive with a booklet of their best, most original problems. Those booklets get shared among delegations, then quietly disappear. No one had ever collected them systematically, cleaned them, and made them available, not for AI researchers testing the limits of mathematical reasoning, and not for the students around the world training for these competitions largely on their own.

Researchers at MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL), King Abdullah University of Science and Technology (KAUST), and HUMAIN have now done exactly that.

MathNet is the largest high-quality dataset of proof-based math problems ever created. Comprising more than 30,000 expert-authored problems and solutions spanning 47 countries, 17 languages, and 143 competitions, it is five times larger than the next biggest dataset of its kind.

What makes MathNet different is not only its size but its breadth. Previous Olympiad-level datasets draw almost exclusively from competitions in the United States and China. MathNet spans dozens of countries across six continents, covers 17 languages, includes both text and image-based problems and solutions, and spans four decades of competition mathematics. The goal is to capture the full range of mathematical perspectives and problem-solving traditions that exist across the global math community, not just the most visible ones.

"Every country brings a booklet of its most novel and most creative problems," said Shaden Alshammari, an MIT PhD student and lead author on the paper. "They share the booklets with each other, but no one had made the effort to collect them, clean them, and upload them online."

Building MathNet required tracking down 1,595 PDF volumes totaling more than 25,000 pages, spanning digital documents and decades-old scans in more than a dozen languages. A significant portion of that archive came from an unlikely source: Navid Safaei, a longtime IMO community figure and co-author who had been collecting and scanning those booklets by hand since 2006. His personal archive formed much of the backbone of the dataset.

The sourcing matters as much as the scale. Where most existing math datasets pull problems from community forums like Art of Problem Solving (AoPS), MathNet draws exclusively from official national competition booklets. The solutions in those booklets are expert-written and peer-reviewed, and they often run to multiple pages, with authors walking through several approaches to the same problem. That depth gives AI models a far richer signal for learning mathematical reasoning than the shorter, informal solutions typical of community-sourced datasets. It also means the dataset is genuinely useful for students: anyone preparing for the IMO or a national competition now has access to a centralized, searchable collection of high-quality problems and worked solutions from traditions around the world.

"The MathNet database has the potential to be an excellent resource for both students and leaders seeking new problems to work on or looking for the solution to a difficult question," says Tanish Patil, deputy leader of Switzerland's IMO. "Whilst other archives of Olympiad problems do exist (notably, the Contest Collections forums on AoPS), these resources lack a standardized formatting system, verified solutions, and important problem metadata that topics and theory require. It will also be interesting to see how this dataset is used to improve the performance of reasoning models, and if we will soon be able to reliably answer an important issue when creating novel Olympiad questions: determining if a problem is truly original."

MathNet also functions as a rigorous benchmark for AI performance, and the results reveal a more complicated picture than recent headlines about AI math prowess might suggest. Frontier models have made extraordinary progress: some have reportedly achieved gold-medal performance at the IMO, and on standard benchmarks, they now solve problems that would stump most humans. But MathNet shows that progress is uneven. Even GPT-5, the top-performing model tested, averaged around 69.3% on MathNet's main benchmark of 6,400 problems, failing nearly one in three Olympiad-level problems -- and when problems include figures, performance drops significantly across the board, exposing visual reasoning as a consistent weak point for even the most capable models.

Several open-source models scored 0% on Mongolian-language problems, highlighting another dimension where current AI systems fall short, despite their overall strength.

"GPT models are equally good in English and other languages," Alshammari said. "But many of the open-source models fail completely at less common languages such as Mongolian."

The diversity of MathNet is also designed to address a deeper limitation in how AI models learn mathematics. When training data skews toward English and Chinese problems, models absorb a narrow slice of mathematical culture. A Romanian combinatorics problem or a Brazilian number theory problem may approach the same underlying concept from a completely different angle. Exposure to that range, the researchers argue, makes both humans and AI systems better mathematical thinkers.

Beyond problem-solving, MathNet introduces a retrieval benchmark that asks whether models can recognize when two problems share the same underlying mathematical structure, a capability that matters both for AI development and for the math community itself. Near-duplicate problems have appeared in real IMO exams over the years because finding mathematical equivalences across different notations, languages, and formats is genuinely hard, even for expert human committees. Testing eight state-of-the-art embedding models, the researchers found that even the strongest identified the correct match only about 5% of the time on the first try, with models frequently ranking structurally unrelated problems as more similar than equivalent ones.

The dataset also includes a retrieval-augmented generation (RAG) benchmark, testing whether giving a model a structurally related problem before asking it to solve a new one improves performance. It does, but only when the retrieved problem is genuinely relevant. DeepSeek-V3.2-Speciale gained up to 12 percentage points with well-matched retrieval, while irrelevant retrieval degraded performance in roughly 22% of cases.

MathNet is publicly available at mathnet.csail.mit.edu.

Published May 2026

Rate this article