Cobots and Collaboration Combine to Make Some of America’s Best Bagels

This entrepreneur leveraged a mechanical engineering past to automate her renowned California bagel bakery.

Emily Winston, owner of Boichik Bagels (left), and Russ Bowman, vice president of equipment manufacturing at BakTek (right), with BakTek’s custom two-lane dough former. (Image: Boichik Bagels)

Emily Winston, owner of Boichik Bagels (left), and Russ Bowman, vice president of equipment manufacturing at BakTek (right), with BakTek’s custom two-lane dough former. (Image: Boichik Bagels)

Manufacturing custom automation is a complex process that requires discussion, evaluation and coordination between all the players. This process doesn’t change, regardless of the industry for which the automation will ultimately be used. A great example of this process is the development of a custom robot and dough-forming machine for Boichik Bagels, a Berkeley, Calif.-based bagel company.

The final product took 18 months to develop and required strong partnerships, especially considering the solution was engineered during the height of the COVID-19 pandemic, with one partner located half a continent away. 

This was the case for Emily Winston, owner of Boichik Bagels. Winston opened Boichik Bagels in 2019, using a retail-size bagel rolling machine to avoid tired hands. When The New York Times in 2021 proclaimed Winston’s bagels some of the best in the U.S., she decided it was time to get serious and automate production.

Winston designed her new 18,000 square-foot bakery specifically to accommodate the new automated equipment, and her background as an engineer worked to her advantage. Before founding Boichik Bagels, Winston earned a B.S. in mechanical engineering from Cornell University and an M.S. in transportation technology and policy from UC Davis. She also worked as an automotive engineer for General Motors.

Winston reached out to Apex Motion Control, a Canadian robotics manufacturer based in Surrey, B.C. Apex designs Baker-Bots, an industrial cobot work cell that handles all sorts of tasks associated with food preparation and the associated material handling tasks. The company makes other robotic and automation systems for the food, dairy and pharmaceutical industries.

Winston also contacted BakTek, an original equipment manufacturer with machining and fabricating capability in Livermore, Calif., about designing a dual-lane dough forming machine. She then coordinated a discussion with the two manufacturers to integrate the machines. She wanted Apex to modify a standard Baker-Bot with an end-of-arm tool that would load boards of bagels onto racks. She also wanted to showcase the Baker-Bot and the dough former—it would be an interesting attraction to her shop and allow customers to see the bagels as they were being made.

Modifying the originals

First, a machine called a bowl-lift hoists a mixer bowl with 100 pounds of dough and places the dough into a hopper attached to the BakTek dough roller.

The BakTek line forms the dough into a thick log and then draws the dough thinner. Next it cuts it into portions and rolls each piece of dough first into a “snake,” then into a circle to create the bagel’s shape. The line lays out the bagels in a set pattern onto a cornmeal-dusted board, which is then moved over to the Baker-Bot.

The Baker-Bot’s end-of-arm tool—essentially a three-fingered robot hand—grips the board and places it on a rack. These racks go into a retarder to proof the bagels at a cold temperature.

The BakTek machine can make up to 12,000 bagels per hour—about 10 times the amount of bagels that Boichik could make with its previous machine. But the new system needed some customization to do the job, which included replacing some of the standard guarding found on industrial machines with a light curtain.

“The enhancement of guarding lets customers, who look through a glass window at the robot, see the bagel-making process better. We also added light curtains to the machine. That way customers can see bagels being delivered onto the pans,” says Russ Bowman, vice president of equipment manufacturing at BakTek.

Customizing the Baker-Bot required adding an articulated end-of-arm tool with six joints, a photo sensor and software programmed to direct the robot to rack boards. The Baker-Bot uses the photo sensor to detect the board. If the machine doesn’t see the board after three seconds, it goes into “stand-by” mode, stopping work and remaining still.

Apex builds a standard Baker-Bot with a food-grade stainless steel enclosure, a touch screen tablet and proprietary software.

“This specific Baker-Bot was programmed to pick the bagel board from the conveyor and quickly place the boards onto bakery racks. It was important that, during the movement of these boards, the raw bagels did not move on the board. The end-of-arm tooling is specific to each application,” says Doug Henderson, Regional Sales Manager for Apex Motion Control.

Henderson adds that Apex trained the machine to turn its robotic hand 180 degrees and move from left to right, “waving” at customers.

“This was Emily’s idea, to have the Baker-Bot greet the customers. She’s all about the experience,” says Henderson.

Henderson says the basic tasks that Winston asked to be programmed in were not especially difficult. Yet the grabbing of the bagel boards from the conveyor, the specific movement of the boards to the racks and doing these tasks at the required speed presented a welcome challenge for Apex’s engineers.

 “For over 25 years, we’ve been designing and manufacturing collaborative robots (cobots) which work alongside people for many industries. They perform tasks such as robotic decorating, palletizing, tray handling, and primary and secondary packaging. One of the challenges with this project was to decide which end-of-arm tooling would do the job best. We also rigorously tested the software program for this specific application to ensure the cobot performed optimally,” says Henderson.

After delivery, Apex sent a team to finalize and calibrate the settings for the Baker-Bot to ensure the required precision.

Take the time to get it right

There are multiple lessons of Winston’s success story: start with available technology, determine what new tasks machines should achieve, make modifications, and apply acquired knowledge to future goals.

Winston also recommends reviewing each step of the process that will be automated and talking with the robotics manufacturer about what components can perform these tasks well.

With this automation implementation under her belt, Winston now hopes to automate baking.

“That’ll take a new type of robot and different partners. I want our next robot to boil and bake the dough in a way that truly replicates the artisan process,” she says.