Humanoid robot chest shell processing: thin-walled parts deformed, can only rely on rework?

Robot body processing: Why is the chest shell the most prone component to tipping over?

I have been working on robot structural component processing for almost 8 years. In the past six months, I have handled no less than 15 sets of chest shell samples for humanoid robots. My biggest feeling is that the robot body processing of core structural components for humanoid robots is much more difficult compared to ordinary industrial robot parts.

Especially the chest cavity component, which can be said to be one of the most representative parts in the entire humanoid robot processing - it needs to be both lightweight and rigid, and also needs to accommodate a pile of core parts with tight tolerances, which can cause problems if not paid attention to. Today, I will take the actual processing of a 60cm humanoid robot chest shell in our factory as an example and share with my colleagues the experience we have stepped on and debugged on site.

This order is from a domestic humanoid robot entrepreneurial team. The design requirements are: using 6061 aluminum alloy integrated processing, the net weight of the finished product is controlled within 1.2kg, the overall external dimensions are 320 × 210 × 85mm, the thinnest part of the side wall thickness is only 2.8mm, there are more than 20 positioning holes of different sizes, the positional accuracy requirement is not more than 0.05mm, and the flatness requirement is within 0.03mm. At that time, when we first tried processing, we stepped on the most common pit: deformation.

The dimensions of thin-walled aluminum parts float after processing - is your clamping method correct?

For the first processing, we used the conventional process route: cutting → rough milling of the outer shape → fine milling of the inner and outer shapes → drilling, and the clamping used was a regular vise clamp on the side. After machining and measuring, the flatness of the top surface exceeded 0.08mm, and the two side walls were concave inward by almost 0.1mm. The positional accuracy of the holes was also significantly off center, making it impossible to assemble.

Our team analyzed the workpiece for a long time and found two core issues: firstly, the clamping method is incorrect. The vise clamps the side wall, and the clamping force directly causes elastic deformation of the side wall. When processing, the size is correct, but it rebounds when released; Secondly, there is not enough time left for stress relief after rough machining. 6061 aluminum material rolling itself has internal stress, and rough machining cutting breaks the stress balance. After machining, it slowly deforms and can still drift by 0.05mm after a day.

After finding the problem, we adjusted the process: first, we changed the clamping and made a customized vacuum suction cup fixture, which can suck the entire bottom surface of the workpiece, evenly distribute the clamping force, and prevent deformation caused by local stress; Secondly, the machining allowance and process were adjusted. After rough machining, a 1mm allowance was left on one side. Then, the workpiece was placed in a constant temperature storage area for 24 hours to release stress, followed by semi precision machining. After semi precision machining, it was left to stand for another 6 hours, and finally precision machining was performed.

After this round of modification, the deformation problem has been effectively controlled, the flatness has stabilized within 0.02mm, and the side wall size deviation has also been controlled within the tolerance range. The deepest feeling when processing thin-walled aluminum alloy parts is that the thinner and more complex the parts, the less urgent it is to rush to meet the deadline. The step of stress release saves time, which is a waste of time.

How to ensure that the positional accuracy does not exceed the tolerance when there are more than 20 positioning holes distributed on all sides?

The deformation problem has been solved, and the second problem has arisen: the positioning holes of the chest shell are distributed on four different surfaces: the front, back, left, and right. The original process was to process the front surface, flip it over to clamp the back surface, and then change the direction to process the side surface. Each time the direction is changed, it needs to be re aligned, and there will be an error of 0.01-0.02mm each time. After processing the four surfaces, the accumulated error of the hole position can reach up to 0.06mm, which is exactly beyond the tolerance given by the customer.

The customer requested that the positional accuracy of over 20 holes be controlled within 0.05mm, which is still unacceptable for us. How to solve it? We ultimately used a five axis linkage machining center to complete all surface processing in one clamping scheme: the billet was clamped onto a five axis turntable, and after one alignment, the turntable automatically rotated the angle. All holes and surfaces on the front, back, and four sides could be processed in one go without the need for disassembly or tooling changes, avoiding the accumulation of errors from multiple clamping at the root.

Here is also a reminder for friends who have just started working with humanoid robots: don't think about using three-axis machining to save costs. For complex robot structural parts machining, it is difficult to ensure a stable pass rate without five axes. Our factory currently has over 200 processing equipment, including multiple five axis machining centers dedicated to producing core structural components for humanoid robots. This is because many complex structures are not suitable for three-axis machining. In addition, we also optimized the programming process by using the "drilling bottom holes first and then reaming holes" technique for all holes, which is more suitable for small batch samples than directly boring holes. The dimensional stability is better, and the surface roughness can also meet the requirements. After adjustment, the positional accuracy of all hole positions remained stable within 0.04mm, meeting the customer's requirements.

What should I do if the inner cavity weight reduction groove machining tool cannot be extended? Can we split the welding?

This client's design aims to reduce weight by creating a continuous weight reduction groove inside the chest shell, located at the angle between the side wall and the bottom surface of the shell. After the conventional right angle milling head is inserted, the tool holder will interfere with the inner wall of the shell, making it impossible to process. This is also the third pitfall we have encountered.

At first, we thought about splitting and processing before welding, but the customer requested an integrated structure, which would cause deformation during welding and not meet the required strength. This plan was rejected. Finally, the problem was solved by the five axis swing head function: we used an extended rod rounded milling cutter to adjust the tool angle through the five axis swing head, allowing the tool holder to avoid interference with the inner wall and tilt into the machining of the weight reducing groove. The entire groove was machined without the need for disassembly, ensuring the integrity of the structure.

Here is also a small suggestion for structural designers to optimize the process: when designing the inner cavity groove, try to leave more than 10 degrees of swing angle space for the tool, and do not make it a 90 degree right angle closed structure, which can greatly reduce the difficulty of processing and also reduce processing costs. When we review drawings, we usually point out such issues in advance, making small changes on the design side and saving a lot of trouble on the processing side. This is also the reason why Huiwen insists on intervening from the design stage in robot body processing, which can help customers avoid many detours.

What are the easiest pits for the robot to step on during the processing of 15 sets of chest shells?

After making so many sets of chest shells and CNC machining of thin-walled parts, I have summarized several key points and shared them with colleagues in robot body machining:

① Thin walled aluminum processing, cutting parameters are not necessarily the harder the better: we are currently making 2-3mm thick aluminum alloy sidewalls with a speed of 8000r/min, a feed rate of 1200mm/min, and a cutting amount of 0.2mm/blade, which results in much less deformation and better surface quality compared to high feed and large cutting blades;

② Constant temperature processing is very necessary: Our precision machining workshop has implemented constant temperature control, with temperature fluctuations controlled within ± 2 ℃, and the thermal deformation caused by temperature changes controlled within a very small range, which is also the basis for ensuring dimensional stability;

③ Testing should wait until the temperature is consistent before testing: During the machining process, the workpiece will expand due to cutting heat, and the size will be measured accurately immediately after machining. After cooling, it will change. We usually leave it at room temperature for 2 hours after machining before testing to avoid misjudgment;

④ If the design can be integrated, do not split it: although integrated processing is difficult, it reduces the splicing process, increases overall strength, and reduces assembly errors. Nowadays, humanoid robots are moving towards lightweight, and integrated structures are the trend. The processing difficulties will have to be overcome sooner or later.

As a hardware design and manufacturing service provider for robot bodies, our factory relies on the technical background of the Chinese Academy of Sciences and has passed ISO9001 and IATF16949 quality management system certifications. We have been recognized as a specialized and innovative enterprise with more than 20 related patents. Our business covers the entire process of design (mechanical structure/electrical solution design/process optimization), manufacturing (CNC machining/mold/surface treatment), assembly (component trial assembly/product assembly and debugging), and supports a set of processes from prototype samples to small and medium-sized batch production, running from scratch.

In terms of humanoid robots, Huiwen Intelligent Manufacturing Innovation adopts a combination process of "new materials+new molds+precision machining", relying on integrated design, intelligent manufacturing, and assembly full chain services to complete the entire process from research and development to batch delivery in one stop, with a stable batch delivery qualification rate of over 98%.

In addition to five axis machining of humanoid robots, we can provide comprehensive OEM services for mechanical dogs, flexible robotic arms, bionic robots, medical robots, and more. If you have the need for processing humanoid robot bodies, please feel free to send drawings for consultation on process plans and quotations.