Cycloidal Speed Reducer

PM oil impregnated or Die cast gears 4 gears result in back to back cycloid gear pairs for full reduction, motor to output shaft. Step files of individual parts in ass. locations. FreeCad assembly models Note has difficulty producing AP214 assemblies failed. MathCAD analysis of gears to show speed reduction and stress using Lewis method. (Appears that made an error in the gear ratio. Suspect I flipped the gears so that 51 tooth gear is fixed. actual speed ratio is 586.5:1. ) PM material data part of public knowledge is insufficient to determine the cyclical yield strength of the gears. Particular manufacture data is needed. Where Zamak is a die cast material that has a more then 50 years used to cast gears. My Atlas lathe built in the 40's has Zamak gears as does those produced until ~ 1960's. This is why the information is available. A spring pre-loads all the ball bearings except for the compound gear. An alternate choice is for the compound gears is a cam roller bearing. These bearing are slightly different then a roller bearing with a separate sleeve for the shaft because it retains the position of the components due to lips between the races and the rollers. Further seals can be added. The second output bearing is a PTFE based material with wear and creep resistance material. The issue is the gap needed for the journal bearing which will increase the defection of the internal output gear more then the fixed internal gear. Alternate choice is a roller bearing that is special due to the small remaining space to the boundary. Some additional space maybe available by switching the output to the 50 tooth gear. Finally the freeCad gear model accuracy is unknown. For example no radius is present in the roots of the teeth or on the tips of the teeth. The stress levels are high enough and the effect of tolerences, gaps, need to be addressed by a specialist with gear design. For example the eccentricity could be adjusted to handle the clearances of the bearing. Stress an