Load Sharing and Effective Number of Planets in Epicyclic 2

[NuclearNerd]

Hi all

I'm designing and epicyclic gearbox under tight space constraints. In order to increase the load capacity, I'm thinking about increasing the number of planets from 3 to 4, 5 or 6. I know that without flexible planet pins the load won't be taken up evenly by all 6 planets. Does anyone know a good rule of thumb for the "effective number" of planets under these circumstances (ie, a 6 planet, precisely machined carrier might act as if it had 5.2 planets)?

Thanks in advance,
Brendan

 

[NuclearNerd]

A few more hours of googling, and I'm closer to answering myself. The old AGMA standard 420.04 said "to compensate for unequal loading, the total power capacity of all pinions should be the calculated capacity of one pinion plus a maximum of 0.9 times the calculated capacity for each additional pinion"

This factor seems to be superseded by more conservative numbers in AGMAs newer standard A6123, but I don't have access to it. Timken includes a plot in their flexpin paper:

http://www.timken.com/en-us/solutions/windenergy/Documents/FlexpinBearingPaper.pdf

I would still like some independent confirmation though. What factors do the gearbox designers here use for 4,5 and 6 planets?

 

[tbuelna]

NuclearNerd,

Whether you use 3, 4, 5, or 6 planets in your gearset depends first and foremost on what gear ratio you want to achieve. A single stage, simple planetary with 3 planets can get a ratio of about 12.5 max; 4 planets about 5.7 max; 5 planets about 4.1 max; and 6 planets about 3.4 max.

Mounting the planet gears on carrier flexpins is also not the only method that will produce load sharing among the planet gears.

A 2 or 3 planet configuration will naturally always tend to load share. But even 4, 5, or 6 planet configurations can be made to load share adequately by using flexure mounted ring gears and floating the sun gear. And if money is no object, close tolerance machining and select fit assembly will produce load sharing.

Finally, if your gears have a wide face, using spherical roller bearings for the planets will eliminate tooth edge loading due to torsional wind-up in the carrier structure.

Best of luck to you.
Terry

 

[gearguru]

tbuelna - you probably mean stages or sets of pinions, not a number of pinions in a simple planetary; the ratio of a planetary is determined by numbers of teeth of sun gear and annulus (ring gear) only. It is possible to design the planetary as "heavy duty" using 6 planets, while "regular" planetary has just 3 (or 4) planets. Obviously the carriers are different and the numbers of teeth for sun and annulus are selected to enable such configurations.

 

[MikeyP]

Gearguru,

As the ratio of the number of teeth between ring and sun increases, so does the required diameter of the planet.

If the sun and ring are nearly the same size, then I can (theoretically) fit loads of tiny circles in the gap between. If the sun is nearly zero diameter, I can only fit 2 planets in.

M

--
Dr Michael F Platten

 

[TVP]

The following is an excerpt from ANSI-AGMA 6123-B06:

9 Load sharing
Design of epicyclic gear drives should address distribution of transmitted power between each of the parallel power paths if the configuration contains more than one planet gear in the system. Concerning multiple path transmissions, the total tangential load is not quite evenly distributed between the various load paths (irrespective of design, tangential velocity, or accuracy of manufacture). Allowance is made for this by means of the mesh load factor, K_?.

K_? = T_Branch N_CP
----------------
T_Nom

where
T_Branch is torque in branch with heaviest load;
T_Nom is total nominal torque.

If possible, K_? should be determined by measurement. Alternatively, its value may be estimated from
table 8.

K_? is equal to or greater than 1.0. K_? equals 1.0 when all planets are assumed to equally share the load. K_? is greater than 1.0 when it is assumed that one planet will carry more than its equal share of the total load.
(end excerpt)

K_? varies from 1.0 to 1.61 in Table 8 depending on the application level, AGMA accuracy grade, and whether or not the mounts are flexible. There are 4 application levels:

1 slow speed gears, mining mill drives, etc.
2 moderate quality, i.e., commercial marine, non--military
3 & 4 high quality, high speed, gas turbine/generator drives, military marine, wind turbines.

Section 9 also has more information on float, load share, and load imbalance. I suggest you spend the $140 for this standard if you need to design an epicyclic gear drive system.

 

[gearguru]

MikeyP,
I know that. But just to have enough space to locate more planets does not mean, that I have to, especially if the carried torque is low. The ratio is determined by the numbers of teeth in sun and annulus. Loads and numbers of planets are different story.

 

[MikeyP]

Apologies, gearguru (grandmothers, eggs, sucking etc.)

I was merely pointing out that tbuelna did not say anything about number of planets affecting ratios, only that the number of planets set a limit on the achieveable ratio.

M



--
Dr Michael F Platten

 

[gearguru]

No problem. We talk about the same from different points of view. (by the way - aren't the planetaries amazing part of gearing? Surprisingly a lot of math hidden behind those few gears, it's worth to study it, one can learn a lot here).

 

[tbuelna]

gearguru,

Sorry if my post wasn't clear. I agree with you in that a planetary gear ratio is dependent upon the relative numbers of gear teeth and what elements are fixed or rotating. But the maximum number of planets may be limited by the gear ratio you seek to achieve. NuclearNerd has limited space avaiable, so one would naturally assume that he would choose to employ as many planets as possible, given his ratio requirements. And thus his (valid) concerns regarding load sharing.

Best regards,
Terry

 

[plasgears]

If size constraints will allow, a plastic planetary gearbox system could produce the desired flexibilities to assure load sharing. You may end up with a metal output shaft based on stress calcs.