Provide a detailed summary of the following web content, including what type of content it is (e.g. news article, essay, technical report, blog post, product documentation, content marketing, etc). If the content looks like an error message, respond 'content unavailable'. If there is anything controversial please highlight the controversy. If there is something surprising, unique, or clever, please highlight that as well: Title: Building a legged robot in 80 days: Lessons learned, tears shed, blood spilled Site: 107-systems.org This is the heroic tale of how a team of five dedicated individuals designed, assembled and programmed a six-legged hexapod robot within 80 days starting from scratch. The article takes the format of dated diary entries. Since the work was completed while keeping up with family, work and community obligations the diary does not show a neat linear progress from start to finish – real life is rarely so simple. Instead there are large leaps forward followed by weeks of little to no progress. This article attempts to reflect reality and not to be some fairy tale with a convenient story arch. However, if you, dear reader, accept those preconditions then there are many gems to be found. Introduction The European Land Robot Trial (ELROB) is an annual event with a focus on providing a real-world testing ground for various robotic systems. The tasks required of the participants roughly stay the same although the competition alternates between civilian applications such as search and rescue and military applications such as reconnaissance. The competition is hosted by a different country each year. In 2020 a ELROB with a focus on military applications was supposed to be hosted by Austria. Unfortunately the event had to be postponed to 2021 due to the outbreak of the Corona virus. In 2022 the competition was once again cancelled due to the Corona virus and was rescheduled to 2023. Due to the repeated pattern of an hopeful announcement followed by a crushing cancellation I lost sight of participating at ELROB while working on more pressing concerns. At the begin of March 2022 I once again took a look at the ELROB homepage and discovered that this year, xapod robots to thefor once, ELROB was actually going to take place. Alas, the deadline for application had already passed and there were only 87 days left to the official start of the competition, for which I had no robot. I took a couple of days mulling the situation over but ultimately the desire to have a go at the competition won out over concerns of an extremely tight deadline. The first order of business was to assemble a team. Thankfully I could recruit Bernhard Mayer , Jonas Wühr and Pavel Kirienko to the effort. Robert Hahn, OE3BHC , a ham radio friend of mine later joined the fray to precision-manufacture the many shafts required for the robot. I’ve always been fond of legged robots, since they have an edge over wheeled or even tracked vehicles when traversing rough terrain. UAV’s don’t need to bother with the underlying terrain at all, however there are severe energy constraints on how long they can operate. Loitering or simply staying on a spot and observing with little to no energy consumption is next to impossible for flying robot. Adding a long-time fascination with six-legged hexapod robots to the mix it became rapidly clear what kind of robot I intended to build. The project was christened L3X-Z which is a phonetic wordplay on „legs“ as well containing the company names LX Robotics and Z ubax Robotics. T – 114 days: Work space simulation and linkage design for a potential legged robot Although I had not even found out yet about ELROB taking place this year I had been constantly playing with various concepts for building a hexapod robot. And while there is a considerable number of hexapod robots with fully electric joints those are unfortunately limited in terms of joint durability and maximum robot weight. I wanted to go a different route and build a large-scale hexapod capable of carrying a considerable payload. In order to generate the required power to lift the robot’s legs I decided to power the force-intensive femur („thigh“) and tibia („lower leg“) with miniature hydraulic cylinders instead of electric servo motors. A primary concern when designing a legged robot (or any manipulator) is the achievable work space. A large workspace enables you to step over obstacles and crouch below overhangs. A tiny work space means that you can perform only very small steps, even when you have very long limbs. When you are working only with pen and paper this task can become daunting very quickly. Thankfully I found a free software tool called Linkage Mechanism Designer and Simulator which I used to prototype various different lengths for femur and tibia, different kinds of cylinders (varying in their length when fully retracted or extended) and different positions of the rotational axis and cylinder mounting. Using the Linkage Mechanism Design and Simulator to design a leg with a suitable range-of-motion smartly compromising between various competing constraints T – 87 days: ELROB 2022 – to go or not to go? After finding out that ELROB would in fact take place this year and despite a strong initial desire to participate in the event the actual decision to go was not made lightly. The deadline for registration had already passed, I had no robot and the event was less than 90 days away. After thinking hard about the situation for a couple of days I came to the conclusion that it could be done but only with the help of an experienced team. Before officially trying to obtain a place on the competitions roster I first went out to recruit a team: Bernhard Mayer and I have a decade-old relationship of jointly working on robotic projects. Jonas Wühr was brought in by Bernhard Mayer, as they are members of the same robotics club in Bavaria. Pavel Kirienko – whom I happened to know via our collaboration on the OpenCyphal project – was swayed to come to Austria by promising delicious Austrian cuisine. T – 85 days: Designing a hexapod leg with FreeCAD In parallel with assembling the team I designed the first version of a hexapod leg using FreeCAD. The tool was chosen for multiple reasons, the most important of which are: personal familiarity with the tool, cost (or lack thereof) and the ability to export parts as DXF files (it’s not straight forward but it can be done). All custom milled plates in the final robot have been manufactured by Schaeffer AG . The company is specialized on creating customized front plates but its tools and processes can be diverted to mill any custom parts up to a height of 10 mm. L3X-Z’s femur was designed in FreeCAD and exported to DXF … Schaffer is providing a multi-platform tool called Front Panel Designer which allows direct import of DXF files to define the outer contour of your front panel (or hexapod robot leg). The holes or cut-outs can also be exported as a DXF file from FreeCAD and re-imported into Front Panel Designer. A word auf caution: One has to be very careful in aligning the contour DXF file with the hole DXF file as otherwise the holes in the manufactured part are not where you’d expect them to be. … and then re-imported into Schaeffer AG’s Front Panel Designer. T – 80 days: Official application for participation at ELROB 2022 After successfully recruiting the team and confirming everyone’s availability I filled and sent a ton of paperwork (i.e. the events application forms) to ELROB’s organizing body, the Fraunhofer Institute for Communication, Information Processing and Ergonomics (FKIE). Despite considerably overshooting the official registration deadline initial feedback was positive so we decided to go ahead full steam breathing life into a robot that so far only existed in our imagination. T – 74 days: Assembly of the first leg prototype After arrival of the custom milled plates I immediately assembled the first leg prototype using screws as substitute for the as of yet not available shafts. Upon completion it rapidly became clear that there’s a problem with the design: The tibia cylinder is blocked by one of the screws connecting both femur plates before it could reach full extension. In the worst case this could have led to leg damage should software control have failed to stop cylinder expansion before reaching the maximum allowed extension length. Not achieving the designed range of motion for the tibia joint also limits the work space of the leg. At any rate a design should be fail-safe to prevent physical damage even in case of software failure. Back to the drawing board … L3XZ leg v1 – full expansion of tibia cylinder is prohibited by an ill-placed screw connecting both femur plates (red) T – 65 days: Assembly of the hydraulic valve blocks While L3X-Z’s hip joints are actuated by Dynamixel MX-28AR servo motors femur („thigh“) and tibia („lower leg“) are powered using hydraulic cylinders as those limbs bear the brunt of the robot’s weight. As L3X-Z has six legs and each leg contains two hydraulic cylinders this means there is a sum total of 12 cylinders to be controlled. Cylinder control is done via suitable valves which can be used to control the cylinders state (expand, retract and hold) as well as the speed with which the cylinder extends or retracts. Note: The speed of extension or retraction does not only depend on the valve but also on the flow rate provided by the hydraulic pump as well as how many cylinders on the same hydraulic circuit are actuated at the same time. L3X-Z uses two Leimbach Modellbau 0H506 valve blocks – and two separate pressure circuits – each consisting of six independently servo-operated valves. 2 x Leimbach 0H506 6-port valve blocks are used for controlling L3X-Z’s 12 cylinders Each hydraulic valve is actuated by a FUTABA S3107 micro servo which in turn is controlled via the standard 50 Hz/1-2 ms PWM signal commonly used in RC model building. On the right side of each valve block the inlet for the pressure circuit as well as the return outlet leading to the hydraulic reservoir can be seen. Each 6-port valve block is actuated by 6 FUTABA S3107 micro servos T – 63 days: Assembly of the second leg prototype Seven days after detecting and rectifying the issues of the first leg prototype the parts for the second leg prototype arrived. There were still no custom made shafts available so M4 and M6 screws were used for its assembly. L3X-Z leg v2: The new version is free of the issues uncovered in the first prototype Writing this article I found only picture where you can see both the first and the second version of the leg. It’s not perfect for side-by-side comparison but if you look closely you’ll be able to spot some differences: A (nearly) side-by-side comparison of L3X-Z leg v1 vs v2 T – 61 days: Dynamixel MX-28AR hip joint installation With only 61 days left to the official start of ELROB 2022 we had to move forward speedily. L3X-Z’s base plate was designed and sent off to manufacturing only a couple of days after the plates for the legs had arrived. L3X-Z’s base plate as designed in FreeCAD After arrival of the base plate the first order of business was to check if the servos really would fit into the openings designed for them. Thankfully this proved to be true so I went ahead by upgrading the firmware of all Dynamixel MX-28AR used for hip rotation to the 2.0 firmware using the Dynamixel Wizard 2.0 tool. Furthermore a unique id was set for each servo where each leg is assigned a number from 1 to 6 which corresponds to the hip servos network id. L3X-Z’s base plate after installation of 6 x Dynamixel MX-28AR A bit of background on the servo selection: The Robotis Dynamixel MX-28AR was chosen because of its high-resolution 12-bit absolute position encoder, a high stall torque of 2.5 Nm and its usage of RS485 (as opposed to the standard TTL interface) for communication. RS-485, as opposed to default TTL, is a differential bus and has therefore a higher resilience against common mode electromagnetic