Full recess planetary gear set 3

[lamieristoff]

Hello all,

I have read lots of informations on books and on internet on simple planetary gears set, and I don't understand the "design rule" on the approach and recess action.
In the standard AGMA 6023 B06, I can read that the approach action has to be equal to the recess action in all meshes in order to be the most efficient, And in this forum, I found some topics on Full recess action in planetary gears.
Someone can explain to me what is the most efficient rule that I have to follow?

For info I am designing an epicyclic gear set with
ratio : 4.3
sun gear : 18
planetary : 21
ring : 60
modulus : 0.6
Input : sun
Output : planet carrier
input speed : 15000 RPM
input torque : 0.1Nm

thank you for your help

 

[dinjin]

I would try 50 percent long on the sun, 25 percent long on the planet,
and 25 percent short on the ring gear. The center distances are not
standard. Experiment with other percentage to see different results.

 

[lamieristoff]

Hi Dinjin,

I'm not sure to understand your post, Are you talking about the percentage of the recess length in each meshes?

 

[dinjin]

It is the percentage of increased addendums on the sun and planet and decreased addendum on the ring gear. You do not have recess action until parts are mated together.
Some folks do not modify the planets and simply modify the sun and ring using the
long and short addendum on the two respectively. You can use higher percentages but make certain you do not have a pointed tooth on the sun. This limits the percentage amount that you can apply.

 

[tbuelna]

Full recess action at the ring/planet mesh is mostly beneficial when scoring is a concern. Based on the limited information provided in the OP, it would not seem that scoring is a concern with this epicyclic gearset. The net efficiency gains from modifying the geometries to produce full recess action at the planet/ring mesh would be quite modest. With such small gears and high speeds the viscous losses from the planet bearings can be a far greater problem than losses from mesh contact sliding. Thus it would be wise to focus efforts on optimizing the planet bearing design.

 

[lamieristoff]

Thank you for your answers.

I will try your ideas and i will let you know the results.

the last test with 50%/50% appraoch/recess at each meshes work quite well