Posted on Leave a comment

French fries a la Doosan

French fries a la Doosan


DHR Engineering was created as a company for developing new tools and solutions, and the spirit of innovation still thrives in our team. It sounds entirely realistic for us to dedicate a whole weekend, outside of working hours, to have fun with something related to the company’s daily tasks. On that account, we decided to organise a homemade hackathon and make some french fries with the help of the DOOSAN six-axis robot.


Project Goals

  • Program a six-axis robot to fry potatoes in a standard deep fryer and serve them.
  • Record a video of the robot in action.

Why We Did It

  • To demonstrate the capabilities of DOOSAN and its applications in various fields.
  • For fun! The opportunity to solve a new, interesting problem with a 24-hour dealine tested the team’s creativity and the capabilities of the machines we have at our disposal.

Action Plan

  • Brainstorm and outline the different actions the robot will need to perform for successful cooking (e.g., opening the refrigerator, transporting the fries, frying, etc.).
  • Purchase and modify various materials and tools to be used (e.g. a box for transporting raw fries, a deep fryer, etc.); for each item, design interface parts to facilitate the connection between the robot’s gripper and the respective object.
  • Set up the “kitchen” – positioning of the robot, deep fryer, refrigerator; prepare for video recording.
  • Program the robot – determine the sequence of movements, find the right settings and speeds.
  • Test, debug, and record the video.

Ready! Set! Go!

With great enthusiasm, we began discussing exactly what we wanted to achieve and how to do it. Some steps were suitable for the robot, while others were left for ourselves to save time. Here is the final result of our brainstorming:

Preparation (human tasks)

  1. Arrange and anchor all objects included in the process (the robot, the fryers, various containers).
  2. Pour the oil in the fryers and heat it up.
  3. Prepare the boxes with raw french fries in the refrigerator.
  4. Place a bowl for the finished products.

Frying the French fries (robot movements)

  1. Open the refrigerator.
  2. Grab the first box with raw fries.
  3. Close the door.
  4. Remove the lid from the box.
  5. Drop the fries into the deep fryer net.
  6. Leave the box aside.
  7. Grab the fryer net and immerse it in the hot oil.
  8. After a certain time, repeat steps 1-7 for the next box of raw fries.
  9. After a certain time, grab the fryer nets with the cooked fries and empty it into the bowl.
  10. Take the salt shaker and season the potatoes.
  11. Serve the potatoes in a special basket.

From the list, it is evident that a special interface was needed for each of the following items: box, lid, deep fryer, salt, serving basket, refrigerator. We started designing and printing the parts.

Design and Fabrication of 3D Parts


In order to perform all the necessary operations for frying the potatoes, we first needed to determine how the robot would manipulate each item in the process. To ensure repeatability and a secure grip, we decided to use a dovetail clamp assembly wherever possible – it allowed for fixed positioning in two planes and provided a secure grip during the robot’s movements. To each item that the robot needed to grasp, we attached (bolted or glued) a 3D-printed complementary dovetail component – this way, the robot was able to grasp most of the items with the same interface.

Jaw Design

As we mentioned in the first article from our 3D printing series, the jaws of the gripper can combine multiple operations in one. In our case, we used the front of the jaws to grasp the dovetail clamps, while the middle cylindrical part was designed for the salt. By chance, the refrigerator door had a very convenient geometry and a simple bolt, sticking out of the jaws, was enough for the robot to open the fridge. Not the prettiest solution in the world, but hey, we needed it to work ASAP!

Since the gripper had only two positions for the jaws – open and closed – we needed to model a slight overlap when the jaws were fully closed to ensure that the robot wouldn’t drop the item it carried.

Manufacturing of 3D Printed Parts

We had limited time and several printers at our disposal. To accelerate the process, each part was printed separately, as soon as its design was complete. This, of course, led to a few unsuccessful designs that needed to be adjusted and reprinted – an iterative process that is common in the engineering field. The most complex parts were ready in less than 90 minutes.

Kitchen setup and programming the robot

Once we started the 3D printers, we went on to set up the “kitchen”- half of our workshop was rearranged, we moved the robot and brought a few tables. Various parameters were taken into account, such as the robot’s reach, order of movements and framing and light for the video.

The programming of the robot was a tricky task as well. Like most six-axis robots, the Doosan has two movement modes – “linear” and “joint”. The linear movement allows for better control over the trajectory and was used for accurate positioning and gripping the objects we manipulated. The joint movement, on the other hand, is designed for shortest and fastest operation between two predefined locations. That mode was used for transition between the different steps in the process. We started with a few steps, outlining the workflow and added or adjusted the list with each test iteration. Speed of movements (directly related to inertia of the carried objects) was also something we had to figure out by trial and error.

Testing, debugging and video

The most intensive part of the whole adventure was when we started testing the kitchen setup. As expected, various problems popped up – some of the printed parts didn’t fit the containers, some weren’t strong enough and we had to enhance and reprint them.

The most important lesson we learned though, was that it is crucial to fix everything very well to the ground. There were several occasions where somebody would accidently displace a table by a few centimetres and all the robot positions would run off. That was a hard lesson to learn but a valuable one.

Finally, after a long and intensive day, we were ready for the real test: recording a one-shot video of the robot, going through all the steps and delivering freshly made French fries … After three long minutes it was done and we sat down to enjoy some delicious, effortlessly-made French fries!


To sum it all up, it was a great weekend! Three ingredients were necessary for it to happen: a dedicated team, the powerful Doosan robot and the incredible 3D printing technology!

As always, we will be more than happy to help you with anything related to automation, robots or 3D printing (or cooking)! Let us know in the comments!

Posted on Leave a comment

Case study 1: Plastic jaws for a pneumatic gripper

Case study 1: Plastic jaws for a pneumatic gripper


Automated CNC machines have long been the standard equipment in the metal working industry but connecting all the different operations in a fully automated flow still requires human intervention. Companies like DoosanRobotics have developed six-axis robots that, with the help of various end-tools (grippers, suction cups, etc), can complete many complex tasks, reducing the need for human labour in repetitive tasks. However, can these robots cope with the most common operation – loading and unloading parts from the CNC mill?

What are the barriers ahead of a fully automated workshop?

For most parts the manufacturing process is divided into two steps – processing the upper and side surfaces (1) and processing the bottom surface (2). For a robot to be able to bridge the gap between the two operations, the robot needs to be equipped with a suitable tool which can handle the blank and the processed component precisely and securely (one such tool is the Schunk gripper). Usually two sets of gripper jaws are used – standard parallel jaws to hold the blank and custom made jaws that fit onto the partly processed part (for example, a cylindrical detail on one side of the part created with the first operation in the mill). 

The custom jaws are often made of aluminum blocks with the exact shape that needs to fit in cut inside. The preparation of those jaws requires investment of time and money on its own. Once manufactured, the aluminum jaws allow almost no changes to the design of the part they hold, and if changes are made, the jaw set is no longer usable. Also, in some cases, the process cannot be automated because the standard and custom jaws need to be manually swapped between operations. The modern solution to all of these problems is 3D-printing.

Figure 1 – forces acting on the jaws during use


Quick, easy, cheap

In the first article of the series we described the requirements that a metal part should fulfil in order to be replaced with a 3D printed one. In the case of the gripper jaws, all of the conditions are met:

  • Gripping and moving small to medium sized parts exerts forces less than 1500 N and the plastic jaws will have no problem withstanding those. That includes cyclic loading as well – the set of jaws we currently have at the workshop have been used for more than 20 000 cycles.
  • The ambient temperature in the workshop is about 25 degrees Celsius.
  • We need just a few sets of jaws so no large batches are produced.

But why would we want plastic parts after all? Here are some good reasons: 

  • 3D printing allows for very complex designs without manufacturing overhead. This gives us the capacity to create jaws with multiple interfaces that will be able to grip the processed part in all stages of its production. As can be seen on Figure 2, interfaces can be located on both sides of the jaws (flat interface on one side, curved interface on the other side), maximising the number of shapes that can be manipulated with the set.
  • The manufacturing time for a set of jaws with dimensions of 100 х 40 х 15 mm is only 2 hours, with zero additional processing required. According to the calculator on our website, the price is 35 BGN as of March 2023.
  • In case the design of the processed part is altered in any way, the plastic jaws can be easily modified in the 3D model and reprinted in just a few hours. Similarly, a new set can be created if the jaws are damaged or worn out
  • The plastics are way softer than metal and the printed jaws can never scratch or damage the processed part. The part appearance is the first sign of a good (or bad) quality work

Figure 2 – interfaces on both sides of the jaws – flat (inside) and curved (outside)


The secrets of good design

Making a good tool requires careful design in various areas – manufacturing, use, reliability, maintenance. Here is what we took into account during the creation of the plastic jaws:

  • To reduce manufacturing time, warehouse management complexity and downtime lost in replacing jaws between different operations and parts, we wanted to include as many interfaces as possible in the jaw set. On Figures 2 and 3 you can see the three interfaces we managed to combine – cylindrical grip for the shape after the first operation; flat surface on the inside for the blank; a second flat gripper with different size available if the outside interface is used. The combined jaw set lets us automate the whole manufacturing process for the part on Figure 3 – loading the blank, swapping orientation after the first operation and unloading the finished product after the second operation.

Figure 3 – the same jaw set can be used for multiple operations


  • A very important design consideration is the calculation of cyclic and peak forces that act on the jaws. Both the geometric design and 3D printer parameters are based on that calculation. On Figure 1 you can see the main forces acting on the jaws – vertical forces that appear during lifting (red) and horizontal forces due to gripping the part (blue). Taking into account the force calculations and following the best design practices, we developed the appropriate geometry. Regarding the print itself, we used 80 % infill density and the reliable PETG plastic. Also, the filament layers are parallel to the horizontal forces, where the most strength is required.
  • Doosan collaborative robots can achieve repeatability of 0.1 mm. We needed a secure and precise way to mount the custom printed jaws to the robot in order to maximise its capabilities. As the 3D printing technology still cannot support such small tolerances, we used a well known trick to get the job done – metal inserts (Figure 4). A dedicated slot was left at the back of each jaw for tight tolerance (h7) inserts – they guarantee that the relative position between the robot and the jaws is always the same.

Figure 4 – precise jaws mounting thanks to the locating inserts



The 3D printed gripper jaws are definitely a success – we have been using them for months in our workshop with zero problems.

Next week’s topic is related to a seemingly insignificant activity – cleaning the mill table. Even though many operators overlook that part of the job, the presence of metal chips restricts automation and may cause misalignments and scraped parts. We have a simple yet effective solution to offer. 

See you next week!