>>The pressure line is off the tangent line by the pressure angle. I.e. rotating the analogue of the t1-t2 line by the pressure angle.
Yes, I agree. What Ive been referring to as the base circles, are those tangent points set at the same distance from pitch point ( cos of the elliptical radius) as the base circle normally is, so I still refer to them as a base circle points as they are an analog of the same thing in a circular gear. I think for the most point were speaking the same thing, in different languages. Originally, I used to use that (cos()*R) at pressure angle to determine each of the base points for any point in the rotation, then calculate the involution from that point. Now all that was ,to my mind ,necessary to pick a start point for the involution to occur, but
as you stated it, calculating a running contact point up that line sounds better and easier, with no need to
figure the involution angles involved.. maybe..
>>Enjoy your vacation! Try not to think about gears
I made an in depth mathematical analysis to understand how the constraints and degrees of freedom change at each point along the curve and made some intriguing discoveries.
Feel free to leave questions or feedback. The tool is not released to the public yet, but if anyone has a significant and immediate need for such a tool, we can discuss.
Looks good. Looks very similar as to how Gearotic does it, a digital subtraction where the involute evolves from the instantaneous rate of change during its construction? I find it works well until pressure angle drops too much , then the gears fall apart in real life running while they simulate fine. Backlash tends to be an issue
depending on contruction and elliptial coefficient. They look good though, generation seems smooth.
ArtF wrote:
Looks good. Looks very similar as to how Gearotic does it, a digital subtraction where the involute evolves from the instantaneous rate of change during its construction?
Right now, these curves are generated without any subtraction at all, just solving differential equations numerically in 2 dimensions. This produces "pure" involute curves that are exactly correct at every point. The only problem is that since these curves evolve, they may interfere with each other as they engage or disengage, so my last task is to find an appropriate solution to minimally clip, truncate or sheer off areas to prevent the teeth from jamming - that final piece may be done via a subtraction algorithm, but we'll see.
I just emailed Michael Valle wondering if he was still around, and he replied that he's "hard at work developing the next version and expect to announce some very exciting new features that are planned for it in the coming months."
Good to know. Its a nice program. I expect I'm several months away
from the next incarnation as well. Im finding the physics a bit harder to properly
simulate in 3d versus the Auggie simulator in 2d. ITs a great learning experience though.
As was mentioned above I am indeed hard at work on the next version of Gearify. I just released a 20ish minute video/talk on the status of the upcoming product.
Cool, looking forward to it. I sent you an email address, but it might be a good idea to post a link to your survey here too. I'll volunteer if you need any beta testers
