Electronics > Mechanical & Automation Engineering

3D printed compound planetary gears, any ideas why efficiency is so rotten?

(1/2) > >>

Infraviolet:
I've 3d printed normal planetary gearsets before, they've worked pretty well, efficiencies better than 80% * have been achieved with no difficulty so long as I left reasonable clearances in terms of shrinking the outward pointing gear teeth slightly (0.08mm or so).

But I had a go at trying a 12:1 compound planetary gearset, I only got a torque increase of 6-fold, efficiency was therefore about 50%. This same gearbox could not be backdriven reliabily** even with no load on the input shaft. But turning the input by hand had the output moving pretty smoothly.

Any idea why the efficiency of this mechanism was so bad compared to reasonably effective normal planetary sets? It seems that increasing clearances for the teeth here actually made the gearbox worse.

If you haven't heard of it before, a compound planetary type gear is also known as a perpetual wedge. The carrier is driven as the input, this turns and as it does the planets roll around the inside of a fixed ring gear. Each planet is double height, with the upper portion of the planet being a different gear to the lower portion, namely having 1 less teeth. These upper portions mesh with an output ring gear. The output ring makes one turn for every Num_of_teeth_on_lower_part_of_planet rotations of the input shaft.

Thanks

*efficiency being given by (output_torque/input_torque)/gear_ratio, so giving an indication of how closely the torque increase factor in a reduction gearbox matches the factor by which speed is reduced. A 100% efficient system would give an increase in torque equal to the decrease in speed.

**sometimes applying force to the output would let it move the input jerkily for a short while, but it would always lock up at some point during turning.

P.S. This is the eficiency achieved with lubricant grease applied, before adding the grease the effficiency wasn't vastly difference, so I think that surface finish effects are not the main factor here, rather it is something to do with the gear geometry

coppercone2:
I recommend you take it apart and compare it with a good gear that you printed by using mechanical gear measurement techniques. Gear metrology is pretty hard core

Infraviolet:
I'm thinking trial and error looks like my best option here. I'll compress the design's "z" dimensions for quicker printing (less strength) and try some alternative tooth counts with the same overall reduction ratio, as well as trying a variety of different scripts to generate the gear tooth profiles (though I'd swear they are already proper involutes). Varying the pressure angles of the teeth is another thing I'll try. Once I print one which can backdrive I'll know the efficiency has improved. I'll also be able to have a normalplanetary gear using exactly the same ring and planets as in the compound version, so can make exact comparisons of efficiency rather than having to compare this compound planetary against normal planetary's I've previously done at diferent tooth counts and module scales.

Doctorandus_P:
My first guess is the tooth geometry. 3D printing just isn't all that great for gears because lack of accuracy.
Second guess is overall stiffness of the whole thing. When you put some force on it, it deforms a bit and things get out of alignment.
Third guess is lack of proper bearings. Put some ball bearings in the planet wheels.

Note that these are all guesses. Posting some pictures of your gearbox will help a lot.

MarkT:
Involute gears have significant sliding friction so need to be made of low friction materials or lubricated with very good surface smoothness, whereas cycloidal gears have only rolling friction.  Try 3D printing cycloidal for comparison.  Cycloidal gearing is used universally in timepieces due to the much lower wear rates and ability to have very few teeth on a pinion (large gear ratios).

Navigation

[0] Message Index

[#] Next page