I'm sorry I didn't reply in time. Thank you very much for your answer. Now I'm redesigning the reducer, using a mechanical double gear. The pressure is applied by the spring. Make two driving gears contact two surfaces of one tooth of output gear respectively. Back space can be eliminated. There is something wrong with the size of the aluminum parts being processed. I am changing it. I can do it next week. I hope it works.

It's a great job. I have a problem. I noticed that you used a sprocket reducer. But the sprocket reducer has a polygonal effect. The instantaneous deceleration ratio will be inaccurate. And the sprocket can not be tensioned, otherwise it will be stuck when rotating. I've tested it with a 1:4 4C sprocket and it does have this problem. and how do you solve this problem?

Thank you. It doesn't seem to be a problem . AR-Robot is a very good design

Hello. What a great job! But I always have a question. I found that the backlash of the normal planetary reducer with 50 reduction ratio is about 25arcmin, and the total length of ar2 and Ar3 is 620mm. As the second joint is used in this kind of reducer, it is found that the end positioning error should be about 4.39mm by calculation. Why do you make ar2 and Ar3 with an end positioning error of 0.75mm?