I need to design a planet carrier for a planetary gearbox to which I'm gonna be doing some simple vibration experiments. It basically consists of the ring, sun and 3 planets and the ring it's no more than 20cm diameter. I've seen many types of designs for planet carriers surfing the web, mainly star type or triangle. I only need the carriers to keep the planets in place but eventually I will need it as the input or the output, so a shaft is needed.

Also, I don't know what kind of bearings I should use on each end of the carrier for the shaft that will go into the planet gears. I'm not a mechanical designer so I don't know about the tolerances either. I was thinking of just measuring the inner diameter of the planets, and design a small shaft to fit very tightly into the gear and the bearing, and have the bearing fit into a bore in the carrier by pressure also, but I don't know if that's also the correct design to go with. I will send everything to be manufactured so I won't have opportunity to mend any mistakes.

Sun = 32 teeth (1.5 module), 48 PD, Ring = 80T 120 PD & Planets = 24T 36 PD

  • $\begingroup$ First, could be more specific as to what you are asking. I'm not sure if I understand the exact question. $\endgroup$ – Daniel K Jun 7 '17 at 1:57
  • $\begingroup$ @DanielKiracofe you are right, my question is kind of unclear, but I was trying to ask for general recommendations for the design of the carrier, mechanically speaking. $\endgroup$ – spe4ker Jun 7 '17 at 22:32

The carrier can be pretty much any shape - if you don't have the manufacturing capacity, a circle will do just fine. The fancy shapes are to reduce weight (and as result friction and moment of inertia) by reducing the amount of material used, but they are in no way necessary. Only layout of the axes is important: central plus three planets.

You'll need axles and bearings on the carrier for the planets (the usual approach is just friction bearings, e.g. brass cylinders, but if your loads are higher, you may go with small ball bearings) - but you don't need bearings for the central axis.

If your carrier is where you output the torque (ring gear is fixed) you just fix the central axis to it firmly. If it's dual-sided (say, axles with bolts, edges extending over the sides of the ring - self-stabilizing against axial direction forces) - the other side just requires a hole larger than the axis propelling the sun gear. (you can use a bearing there, but it's better to fix the bearing to the chassis than to the carrier. It's also not necessary at all; the planets acting as balls of a ball bearing, the gearbox acting as bearing for both the sun and ring.)

If you use a variant with fixed carrier, moving sun and ring (axial reversal of rotation) then you just fix the carrier to the chassis, no bearings for anything except the planets.

  • 1
    $\begingroup$ In addition to load, you should also consider speed in the choice between a sliding bearing and a rolling element bearing, where speed is the linear velocity at contact = radius * rpm. e.g. 1/4" shaft at 100 rpm, bronze bushing is just fine. 2" shaft at 6000 rpm, you're going to need a ball or roller bearing. $\endgroup$ – Daniel K Jun 7 '17 at 10:56
  • $\begingroup$ @SF thanks, I have come into the first problem, the gears I have are Sun = 32 teeth, 48 PD, Ring = 80T 120 PD & Planets = 24T 36 PD, how do I work out the spacing of the three planet gears to get them as evenly spaced (i.e. as close to 120° ) as possible? $\endgroup$ – spe4ker Jun 7 '17 at 22:49
  • $\begingroup$ @quo: is it one-off or for a production line? Because for one-off, I'd go with the easy approach: experimental. Assemble ring, sun and planets inserting them as close to 120° as you can, where they fit, then use their center holes to mark drilling points for axles on the carrier. $\endgroup$ – SF. Jun 8 '17 at 0:54
  • $\begingroup$ @SF: it's actually for performing experiments in a lab, not for production line, but I would still like to know the mathematical analytical method to determine this. I haven't been able to find some explicit example in the Web, otherwise I'll do it the experimental way as you said. $\endgroup$ – spe4ker Jun 8 '17 at 1:01
  • $\begingroup$ @quo: I don't know such method myself. If the sun and ring teeth number is divisible by 3, or ring's is a 3n multiple of sun's it's a trivial 120 degrees, but I'm really not sure how to approach analytical solution for arbitrary number of teeth... or even if one always exists. $\endgroup$ – SF. Jun 8 '17 at 8:20

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.