The Plastic output will not withstand the forces, I need a new way to mount things to the joint. Looking into cross roller bearings, but still attempting to keep the price low.
Added more ball bearings to the wave generator, removing the ball bearing spacer by filling it completely. This increases accuracy in the wave generator ellipse profile
Changed Tooth Profile
Changed Teeth and Circular Spline and FlexSpline to 62 and 60 respectively
Increased Thickness of Circular Spline outer plastic.
I tested the Joint up to 6.26KG at 33cm distance (20.26 Nm), any higher, motor stalls. I am assuming there is a lot of friction in the planetary gearset, I will be working on that.
Super happy with the fact the FlexSpline Teeth no longer skip!
After strengthening the Circular Spline, I found it helped prevent skipping, but it was not the solution as the skipping still persists, now just louder and more violently. After some research I am now thinking that it may also have to do with lack of tooth contact. I will be optimizing the tooth profiles and Wave Generator to see if the results change.
My plan before was to add steel around the circular spline to completely stop the flexing. I may still do this, but at this time I have no way to mill steel on my own and I would need to order custom cut pieces. Doing this contradicts the end goal of my project, so I will put a pause on this until either I obtain the tools to make these pieces or I find an accessible piece for purchase online.
It should not take long to change the Tooth Profiles and wave generator, hoping to get the next tests done this by tomorrow.
It seems with the stiffer Circular Spline, the strength does not change, rather if its not the teeth that skip, the motor just stalls. I am currently working on a fix. I may upgrade some parts to metal to prevent flexing in the Circular Spline as it is still present.
The reason the teeth are skipping under load is the Circular Spline is flexing enough to let the FlexSpline teeth pass through. I will be adding metal components into the Circular Spline to fix this.
I have tested the Joint in 1kg increments. The max torque it was able to achieve was 18 Nm, any greater mass tested causes the ABS FlexSpline teeth to skip. Still testing to find out how I can move forward.
Finally I found that the Joint sounds terrible, it might need some grease.
Testing out ABS as the material for the FlexSpline. It seems to be a fair middle ground between PLA and TPU with a good amount of flexibility while still keeping good rigidity.
So far I have used it for the FlexSpline, Circular Spline, and Wave Generator
After further testing the point where the mount failed last time, the teeth constantly skip to the point where its no longer lifting the weight, rather constantly skipping. Although the FlexSpline seems to have a lot of strain on it, the motor itself doesn't. While the teeth are skipping the motor is running at about 0.230 Amps. This is a very small amount, meaning its using a very small portion of its power. This is due the Planetary Gear acting as a multiplier on the Strain Wave Gear.
I am now going to take a step back in ratio in attempt to make the teeth of the FlexSpline Larger. Ideally this will stop the skipping from happening. Since there is little strain on the motor I am not worried on the lower gear ratio in the end. I may be looking into a 20:1 Strain Wave Gear which would leave the Gearbox with a 100:1 Ratio. With that Ratio the final design could still output close to 100NM of torque if everything goes to plan.
First Test was 2.28 KG at 10 CM, the Gearbox completed this with no problem.
Next was 2.28 KG at 20 CM, the Gearbox completed this with no problem
Finally I attempted 2.28 KG at 20 CM, the Gearbox did this fairly well. When doing this test at speed it was very noticeable that the outer casing of the Gearbox was flexing, I am assuming this was caused by the Circular Spline flexing. This caused the Teeth of the FlexSpline to skip occasionally.
Before I could further test the mount for the 2040 Aluminum snapped under the load.
I will be making this stronger and completing more tests very soon.
Jared Fenster
