How do gear mates work?

onshaper

Hi all,

I've been trying to implement a gear mate to connect the crank pulley to the timing pulley, but I can't get it to work.

Also, is there a way to reset part orientation in an assembly? I drew everything correct but I haven't been able to line it up after I dragged the pulley to see if the gear mate was working.

Thanks
Luke

https://cad.onshape.com/documents/fab7f27232d3460081425531/w/0f689583090f433b8719c49b

andrew_troup

I had a quick look at your model: the main problem I see is that you have picked a Mate Connector at one end of the big-end journal. Your mate connectors for positioning the crankshaft should be at the midplane of that journal, corresponding to the midplane of the bore

The con rod and piston and crankcase should also use connectors lying rigorously on that plane, regardless of future dimensional changes. 

As a rough guide to good practice, dimensional changes to parts should not cause problems to any robust mate scheme.

andrew_troup

@Onshaper  You first need to add revolute mates to each of the pulleys, then the gear mate requires picking those.
The orientation of the part will be determined by the orientations of the connectors.
(ON EDIT: I got that last statement wrong: that's true in the general case but NOT in the case of gear, or rack and pinion mates, which are "mates of mates", not mates of connectors.
The linked clip does explain the synchronisation process nicely)

Here's a clip explaining how it all works:
https://www.onshape.com/videos/whats-new-gear-mates

Good luck!

onshaper

Hi Andrew,

Here's where I got lost. One of my gears is fixed to a shaft which is on a cylindrical mate. Seems reasonable. I have it as a cylinder mate because that way the model is more robust.

In the end, I had to delete the cylinder mate and replace it with an offset revolute. While it does work, a change in one of many dimensions will lock up the mate solver.

philip_thomas

Guys - I am glad someone posted the video. I cannot emphasize the importance of 'resetting' the mates before you add the gear relation. It will drive you nuts if you dont. The reason is that when you apply the gear mate, the first thing it does is reset the mates. If the teeth dont mesh to begin with (in the reset state), it will not work. Tip, get the teeth to mesh first by editing the one of the mate connectors in one of the revolute mates. Once you understand whats going on, this gets a lot easier.

andrew_troup

@Onshaper Ideally, the selection of mates is governed by the degrees of freedom each part would need to have in real life. Unless you have encountered a bug, it should not be the case that you have to select an unrealistic mate type, with extra degrees of freedom (which the cylindrical mate has - it permits axial translation as well as the rotation you would normally expect to be appropriate to a gear)

If you need your gear to slide (say along a splined shaft) then you cannot use a gear mate, unless you use a workaround. For example, you could model a dummy disc (hidden in the assembly) to act as an intermediary. I'm thinking this disc could have the necessary revolute mate to the mating gear, and a slider mate to your sliding gear. You could then create a gear mate between the mating gear (provided it used a revolute mate) and the dummy disc.

andrew_troup

@Onshaper@ 
Without seeing your model it's hard to know why your mate solver is locking up; unless it's a bug (it probably isn't) then you must have overconstrained the assembly with your mate scheme.

Check that you have not removed the same degree of freedom with several mates, for instance.

onshaper

Hi, perhaps I can be more specific.

First of all, a link to my file is in my original post! You can see the working model there.

The axial translation of the crankshaft is constrained by the cylindrical mate used on the piston and all other mates in between. I ORIGINALLY chose to use a cylindrical mate on my crankshaft>crankcase because that would allow for the whole assembly to automatically center itself making for a robust assembly. It worked well.

The problem WAS that I cannot use a cylindrical mate in a gear relationship. It appears to only work with revolute and slider. I see this as an area where gear relationship falls short.  

Is it because gear relationship can only handle 1 DOF? Seems to me that "gear" should work fine with any two mates that have only 1 rotational DOF.

By changing the mate to a revolute, I have lost some robustness in my assembly, but the gear assembly works! A trade off.

Thanks for the good discussion!

andrew_troup

@Onshaper
Sorry about my selective blindness. When I got to the signature of your original post, I thought everything below it was none of my business :-)

It seems to me that you might have picked the way you modelled the assembly in an unrealistic way:
A real engine does not use the piston's inability to walk sideways to control the end-float of the crankshaft!

But you could choose to get similar convenience, plus what I would call (from a solid modelling perspective) robustness, if you modelled the crankshaft big-end journal midplane so that it was constrained to lie on the same construction plane as the midplane of the piston, regardless of changes to the lengthwise dimensions of the crankshaft.

Then you would not require a cylindrical mate for the crankshaft (I presume).  You seem to be exploiting that extra degree of freedom for expediency (the sworn enemy of true robustness in a solid model), rather than to portray the real assembly motion which would happen in a real engine. 

When I get time, I'll look at your model, (I'll grab a copy now so you can carry on changing it if you want) but I hope I at least gave you some things to think about.

andrew_troup

I had a quick look at your model: the main problem I see is that you have picked a Mate Connector at one end of the big-end journal. Your mate connectors for positioning the crankshaft should be at the midplane of that journal, corresponding to the midplane of the bore

The con rod and piston and crankcase should also use connectors lying rigorously on that plane, regardless of future dimensional changes. 

As a rough guide to good practice, dimensional changes to parts should not cause problems to any robust mate scheme.

onshaper

Makes sense, thanks Andrew!

jon_woellhaf

andrew_troup said:

@Onshaper  You first need to add revolute mates to each of the pulleys, then the gear mate requires picking those.
The orientation of the part will be determined by the orientations of the connectors.
(ON EDIT: I got that last statement wrong: that's true in the general case but NOT in the case of gear, or rack and pinion mates, which are "mates of mates", not mates of connectors.
The linked clip does explain the synchronisation process nicely)

Here's a clip explaining how it all works:
https://www.onshape.com/videos/whats-new-gear-mates

Good luck!

Just what I want to see, but the video has been removed!

brucebartlett

@jon_woellhaf try this

https://www.youtube.com/watch?v=GWfAXqQMkEY

jon_woellhaf

Excellent! Just what I needed, Bruce. I didn't think to expand the revolute so I couldn't find the control.