Carbon fiber composites are incredibly strong for their weight; that’s why they’re key to the newest aircraft designs. However, they’re only strong in one direction, so they’re generally layered or woven in grid patterns before being shaped into structures. That means one set of fibers carries the load some of the time, and another set carries it at other times—which is not the most efficient use of the material.
In 2014, Hemant Bheda was CEO of Quantum Polymers, a company that makes extruded plastic rods, plates, and other shapes for machined parts. The company used chopped up carbon fiber in some of its materials, but a potential customer asked for a material which would require continuous carbon fiber to be embedded in a polymer material in carefully laid paths that would give the material super mechanical properties
“I said that we couldn’t do it,” Bheda recalls, but he kept thinking about that request. As he says was typical for him, an EE whose career has focused on software, and particularly on image compression and video, he kept wondering if the issue was a software problem instead of a hardware problem.
Then, he says, he came across a paper about 3D printing. Why, he thought, “can’t we come up with an algorithm for optimal orientation of the fiber in 3-D space and only use the fiber in the directions you need it?”
Bheda joined with Wiener Mondesir to start a company to explore this possibility. They named their company Arevo for “a revolution in manufacturing.”
With backing from an aircraft company, they spent two years trying to build a 3D-printing system and software that could make production parts that met aerospace requirements. The system was based on traditional fused deposition 3D printers, which are the kind that people use at home. The conductive heating process used to melt the material, which was a composite of a polymer and continuous carbon fiber, was slow. And it didn’t meet the mechanical and quality requirements of the aerospace industry, Bheda said. “We knew we needed to move to a laser-based system [that could give us faster,] contact-less heat transfer without fraying the fibers.”
That, however, would take more time, more people, and more money. In late 2016, Bheda and Mondesir pitched venture capitalist Vinod Khosla on the concept. “He got it immediately,” Bheda said, as did Bill Gates, who Khosla quickly brought in to meet with the Arevo team. They “immediately comprehended the potential, that our approach was a radical departure from the current way fibers are stacked, that it would optimize how we used the fibers, [and] lead to less material” being needed, Bheda said.
Khosla Ventures put US $7 million into the company. After raising additional rounds from investors including Airbus, Arevo’s funding now totals $34 million. The company has 35 employees, with an engineering team of 25, including 14 engineers or scientists with Ph.D.s.
Arevo’s goal has always been to get its process working well enough to produce aircraft parts. And that’s still on the agenda. But like so many Silicon Valley startups, Arevo wasn’t opposed to making a little pivot along the way.
In March of 2018, the team began planning a demo for the company’s Series A investors. They had software—which uses calculations of the direction of stress on a particular part to inform the optimal layout of the carbon fiber filament—ready to go, and had built a prototype laser-based 3D printer. As the engineers brainstormed possibilities for what to produce, Bheda recalls, “Someone said, ‘Why don’t we print a bike?’”
Why not? But they needed a bike design. They got it from experienced bike design firm Studimowest, where, Bheda says, the designers “were excited by how the technology could push the limits of design.”
In addition to allowing new shapes, printing a bike frame would vastly simplify its construction. Traditional bike frames contain dozens of parts that have to be assembled by hand, a complex process with lots of room for error.
The designers at Studiowest “sent us a design, provided us with the design requirements including the load, and we ran it through our software,” Bheda said. “The software told us how to orient the fibers. We then printed the bike and sent it to a local bike shop” to add components like wheels and pedals and brakes, he said.
“Then all of us rode it,” Bheda said. “It worked, it felt like a bike, but it didn’t seem like any big deal.”
Then, the designers came over and rode the bike. ““They were extremely surprised that it rode exactly how they had specified it to be,” Bheda said. “We wondered why they were surprised; they explained that they usually go through multiple iterations of design, ride, make some changes, ride again. It usually takes a long time, as much as 18 months, but in less than 14 days from the time that they sent the design, we had the bike.”
The bicycle designers urged Arevo to show the product to the CEOs of large commercial bike companies. They did so in April 2018, at the Sea Otter Classic in Monterey, the largest consumer bike show in North America.
“The CEOs we met with were intrigued, but skeptical,” Bheda said. “They thought it was too good to be true.” But the showcase, along with a little help from some personal connections, led Arevo to electric bike startup Emery. And this April, the Arevo team was back at the Sea Otter show, watching the introduction of Emery’s electric bike with a 3D-printed carbon fiber composite frame.
Since the show, Emery has been taking orders for the Emery One e-bike and will start shipping to customers this quarter. When I visited Arevo in June, the company had two of its five printers laying down bike frames; it expected to have all five printing bike frames within a matter of weeks, which will allow it to build 10 bike frames per week. Then, the company expects to increase the speed of the process until production hits 75 frames per week.
The system looks like the kind of desktop printersthat use coils of PLA as feedstock, with a few significant differences. Arevo’s version is a lot bigger and uses an industrial robotic arm as its core. And there’s an extra step in the graceful choreography of laying down the strings of material—the cut—because the carbon fibers in the composite don’t just break neatly on their own when the print head stops heating and pulls away. Finally, there is the printer bed—instead of a flat platform, it’s a 3-D structure, printed from plastic, that allows the carbon fibers to be draped at angles, instead of requiring those angles to be built up out of layers. This is an essential part of making sure fibers are oriented in the right direction.
New bike designs that take further advantage of 3D printing are on the drawing board. Arevo is already working with another yet-to-be-announced bike manufacturer on a radically different design, with just one fork for the front wheel and other unusual features. And Arevo will be in the bike manufacturing business for the foreseeable future, because the plan, at least for now, is not to sell printers to others, but to market its software and printing capabilities as a service, keeping the manufacture of any products in house. It also has a lab, in which it frequently checks samples of the carbon fiber composite as it comes in from a supplier, and comes out of the printers, priding itself on avoiding so-called “voids,” areas in which gaps form between the layers of material.
“The technology is new,” Bheda says, “so we don’t want to lose contact with the end customer; if we don’t make the parts ourselves, we won’t learn about issues and be able to advance the technology.”
Eventually, Bheda says, the company will get back to those aircraft parts, after the bike business proves out its technology. Arevo is also eyeing wind energy and building construction as potential future markets.
And then there are flying cars. That’s not a joke. Bheda says the flying car market could turn out to be Arevo’s sweet spot. “They will be manufactured in a larger volume than airplanes, the manufacturing technology being used for current aircraft won’t scale to that, and they want to use thermoplastic. Our technology is at least three years ahead of any other thermoplastic technology, so we will be ready