Calling all "Gear Heads",
I'm having trouble finding an old Creo planetary model, and recalling, exactly how to define the gear-pair to establish proper planetary gear motion in MDX/MDO. My system is a simple epicyclic set with 3 planets. My trouble is defining the ring gear. I seem to recall there is a trick to how to define the ring gear's assignment so MDX understands it's an inward facing involute set and provides proper movement and torque & speed transfer. I don't want to use a "generic" gear type. I can't seem to find anything on-line, within PTC support, or inside this community - what's the trick I'm not able to recall?
Trick#1 Make a dummy "cage" to hold the three planets.
Center gear pin to ground.
Cage pin to ground.
three planets pin to cage.
Trick#2 Ring gear pin to cage NOT ground.
Trick#3 Tuning ring gear ratio user-defined ratio until pitch diameters are close.
Thanks for your suggestion. Unfortunately, this will not solve my issue. I have attached a representative model set for review. The challenge is to have a proper planetary gear train that will allow control over any of the main elements - sun, carrier and ring. For example, in a typical automotive planetary transmission, we might "gang" 2 or more simple planets together to be able to select a number of different output speeds. This is done by selecting the sun, ring or carrier to be held while simultaneously locking any two of these members together to change how the torque is passed through one planetary set and on to the next (those are your "clutches" and "bands" in your transmission). The model I have shared allows for this with basic pin joints all going to a "ground". In my actual system, the ring gear is part of "ground", but I kept it open to either fix it or allow rotation based on a servo motor lock (on or off). I have done the same with the carrier. The sun is always the input to my/this system.
Kinematically, I can get the set to function properly and provide speeds that make sense, but I'm trying to get gear forces and loads out as well - that's where it doesn't play nice. I do have an issue with the kinematic analysis where I lock the carrier and let the ring gear become the output - this doesn't even solve. I have a few analyses created to play with - if anybody can figure out how to arrange the gear pairs to get these to work, that would be great. I know there will be redundancies for the dynamics, but that's what you get with more than 1 planet gear (although MDO seems to split the load evenly, as it should).
Way back in the days of Wildfire, we had Mechanica Motion to use. I actually provided input for the original gear specification/capability to Brian Shepherd when he was in charge of developing their simulation tools in the 90's. The code they produced was fantastic and even properly calculated an uncommon phenomenon called torque feedback in certain planetary transmission system arrangements...something tricky to do by hand...at least 20 years ago. Over time PTC has completely lost track of that older fantastic code that allowed us to arrange mechanisms that would handle very involved situations. I'm seeing this issue here - I used to be able to put together complex, linked planet systems, but now I can't get this simple, stand-alone one to work - SERENITY NOW!
"BEST PRACTICE" NOTE - for those of you interested enough to look at my model, please take note of the way I've documented the joints and gear pairs - I've provide proper names to them all for very easy recognition - a practice I seem to find very seldom among users - I highly recommend it.
For what it is worth I was able to get your model to simulate with these changes. I could not lock the ring gear movement using its pin joint because it has to reference the carrier. To find the torque applied to the ring gear under load (locked carrier), you will have to use some other mechanism to stop rotation such as a pinned link-torsion-bar, or a cam interface against one of the splines on the ring, rather than the planar constraint I used. The gear ratio at the ringplanets is not intuitive as I will explain.
1. Change Grounded Ring to pin to the carrier instead of ground.
2. Add to the Grounded Ring a general constraint to stop rotation (Y plane of grounded ring to Y plane of ground)
3. Ringplanets Change GEAR 1 to 0.3125 instead of 0.625 pitch diameter and flip (upper flip) GEAR 2 so purple direction arrows are in opposite directions. The icons will look like you made another planet to sun gear ratio.
Well folks, I have received a reply from PTC Technical Support concerning planetary gears:
Seriously PTC? This is shameful...time for any of us who needs a decent kinematic/dynamic mechanism tool to find a real one.
This is yet another example of how PTC has let their previously well-recognized and cutting-edge engineering tools directly embedded into a decent CAD package go to straight to hell, right beside Behavioral Modeling Extension (BMX) and Mathcad.
Attached is a document created by Dr S. Shome, in response to your Planetary gears query. This deals with the kinematic side of things. Regarding the forces we will need more information.
It would certainly seem to...I will read-through the shared document, try to apply it to my model and write a response as soon as I can.
Thank you for providing some feedback. I have taken a look at the PDF file shared and find it only provides a solution to operate the gear system properly when the carrier is fixed to ground and the ring is allowed to rotate and act as the output of the system. I made the change as documented by page 10, by making the planet-to-ring gear pair as "generic" for this arrangement to function and checking the box so as not to create an internal carrier body. The only detail not shared in the PDF file on page 10 is how the direction of "Gear2" needs to be set. The instruction simply says "needs to be flipped" - should they be set in the same direction, or opposite? My planet-to-ring gear-pair definition have both axes pointing in the same direction. My model modified based on this document has been added to this posting.
When I run this same gear-pair configuration for the case where we have the ring as the fixed member and the carrier as the output, it does not provide a proper solution of the system - the planet gears do not rotate. I have two analysis definitions setup that both have the sun as input with either the ring or carrier fixed and the other as the output member. The only difference between them is including a servo motor that "locks" either member to allow the other to become the output. If MDX can handle a planetary, we should be able to have these two situations solve properly without making changes to the definition of the gear-pairs. Somebody at PTC must re-visit the old Mechanica Motion code for gears and resurrect it - it was fantastic!
I do have a working model - it is attached. I did need to make an adjustment from the one most recently shared. Here are my observations:
For now I am satisfied with the operation, but can't explain why PTC did not recognize this approach...at least by their support desk. I would like to see this model used as a "suggested technique" by PTC so others can apply it - of course at their own risk. Maybe PTC can recognize the importance of this application and formally release MDX/MDO with planetary capability...better yet, modernize the code with an idealization to have validated calculations of all forces, speeds, torques and loading that can occur in a helical geared planet system by all elements.
Thanks for the confirmation. I'll record a tips&tricks video for this so it might get some more attention. We'll update the TS article and add your model as example or have it for TS to share. We'll review the Planetary option for the Mechanism roadmap.