Version 3.51 online

[ArtF]

Hi Guys:

  Version 3.51 is online. It includes a new pdf help file created by Bob. We'll expand this help file as we go forward.
It can be reached from either the help/about license dialog ( press help button ), or fromt he main menu, ( press large gear icon on
top left of program and select "help file".

   Now that Ive cleared up many personal projects, work is restartin g on a few modules for GM. And we're workign again on our 3d printer
as well. New will follow as we go forward.

Thx
Art

[danmauch]

Looks good.
Do you have adding racks on the to do list ?
Dan Mauch

[ArtF]

Hi Dan:

   Yes, I have been consideri ng it.  I have a few things to finish up before I start coding it, but Im in the process of consideri ng the
math involved with creating teeth on an object based not on the usual formula's, but on the derivitiv e of curvature of an object.
 For example, the derivitiv e of curvature of a straigth rack ( instantan eous radius ) is infinite. It seems to me a tooth calculato r
using the radius from point to point would by its nature create normal V type teeth at pressure angle ( as a rack should have ). Now,
the same algorithm when faced with a straight line followed by a curve would blend from PA angular teeth to normal involute teeth
as the radius increased . (Picture a rack that is striaght along a line, then bends up or down to a new level.. ). Such an algortith m
coudl generate a rack from a straigth line, a rack with a radius in it , or..given a elliptica l curve woudl create proper PA involute teeth
on the ellipse..

   In other words, Im thinking of an algorithm that basically takes any 2d shape, circle, line, ellipse, or combinati on of shapes and
could then apply proper gear teeth to it ..within bounds of posabilit y of course.

   This would have advantage s to me as it would apply to almost all the gears and have better elliptica l repsponce than we currently have.
It woudl lend itself to more CAD like drawing of a gear. Anyway... its in the queue as something Im thinking of and planning roughly before I can get to a spot where Im free to implement such a thing and see how it works..

Art



     

[ArtF]

Dan:

 Hard to explain that last one properly. .

 Try this.. picture a circle. Cut it in half and separate the halves by a set distance.
Connect the two halves with straight lines.. You now have a shape that doesnt follow normal gearing.. but it could.
The teeth on the straight sections according to the math should have PA teeth like a rack, while the circular parts
should have involute teeth. Ive never seen such a beast but it should exist by the math. The gear you run it it, shoudl run well
along the inside ( or outside depending .. ) of the circles, and on the straigth rack section. It appears to me one algorithm shoudl
be able to tooth such a shape by using that derivitiv e of curvature instead of the current toothing formula's for a gear. Such shapes dont
allow themsleve s to be currently toothed properly. All current gear formula's assume a round blank shape. A conversio n of the formala's to
consider the shape at all points in the totality of the curve should work.. and on almost all shapes. I think...

   As I say, I havent begun coding such a thing, but when I walk my dog, its what I think about and form plans for at the moment.. :-)
My Dog isnt good at math at all, but he helps me think it through..

   

Art

[Mooselake]

How's the kitchen, Art?  Do you have pictures?  Hope you scored some honeydo points!

Good luck with racks.  Ageneral solution to tooth generatio n rather than specific algorithm s for each case would open a lot of possibili ties.  Could be a big help with more elliptica ls, and there's even a possibili ty of turning scanned pictures into gears, although the pinion generatio n could be a problem.  Hmm, rotating heads...

Kirk

[ArtF]

Hi Kirk

  I sometimes think that honeydo points get harder to come by the older you get.. lol , but yes, she's happy so far.
But its a 3 week wait for the granite counterto p she wanted.. so Ill post a pic after thats in.. I basically shoehorne d a kitchen
in the closet sized space we call a kitchen.. So instead of normal cabinets, they are all sliding racks to hold all her stuff, racks slide out
from the bases, or rotate out of the blind corner..t hen slide out for access. The whole point of all this was simply to take the requirmen t of
bending over out of the equation of cooking, so it aint fancy, just real functiona l..

 I agree, I think a general solution has facinatin g possabili ties.. I suspect its the error checking thats hard to ensure no bend in the given curve
violates the capabilit y of a gear of the prescribe d module from rolling in it.. To me, its all about the rolling.. current elliptica ls, though they work,
do not do pure rolling at times due to the varying pressure angle of the elliptica l motion. Using derivitiv es of curvature may make strange looking
teeth at times, but those teeth shoudl create a pure rolling mesh. Current elliptica ls see this as a discontin uity of the torque required to roll them.
A properly designed elliptica l gear would have a more continous or sinusoida l torque requireme nt, just as the halfcircl e thing I just posted about.. high frequency
spikes in the torque curve is symptomat ic of bad mesh caused by shape changes in the original curve.. Circles work great, but I think we need something new
if we wish to consider non-circular shapes properly. .  not that elliptica ls run bad.. but they could be better..


Art

[Jeff]

Bob - Art

Many thanks for Bob's help file - very useful for the occasiona l user like myself


Jeff

[ArtF]

Jeff:

  Glad you find it usefull, we'll expand it over time. We're also working to try to update thge main video as well so that ti matches the current gui.

Thx
Art

[BobL]

Hi Jeff;

 Thanks for the comment. PDF isn't perfect but good enough for now I hope. As Art mentioned earlier, we will try to update the tutorials as GM grows, if anyone finds something out of place or incorrect with the help file, please let us know and we will try to change it asap. BIg thanks to Art not only for writing such a great program but also for his input with this tutorial.


Cheers
Bob

[Monafly]

I'm all for your general solution!
It hopefully would allow one to 'project' gear teeth onto an arbitrary shape. Some that come to mind are heart or pear shaped gears.
A friend and I played with this some time ago in his CAD program (I believe it was ProE), and the two issues we ran into were having to expand or contract the starting shape so that an integer number of teeth came out, and then doing the same for a mating gear and the distance between the two. Took a lot of manual tweeking the way we did it, but would have lent itself to computing rather easily.

[BobL]

Hi Monafly;

  Thanks for the comment, your idea is a good one. I will include it on the list for evaluatio n, who knows, maybe someday you'll see something similar added to GM.

Cheers
Bob

[ArtF]

Hi Guys:

  Just a note on progress. The ToolBit wizard is almost ready for version 3.52 , it will generate a toolpath to turn a round stock in a 4th axis down to a
bar, and then shave a tooth into the bar with relief's so you can use it to cut involute gears and timing pulleys. I hope to have this completed within a week
as most of the work is done, final shaving of the relief is underway.  Ill post a video of a toolbit being cut when its ready.

   When this is released, I intend to redo the main GM video for the new interface and begin work on the general algorithm for toothing shapes.

Just a note as I know Ive been quiet lately. My computer is finally working well, many bugs removed after the upgrade. Weve build a trial water injector head for powder printing that we're playing with. ( hope it works, its easy and cheap for anyone to build.). More news soon hopefully, the tool cutting wizard turned out a bit harder than Id expect, but the data is looking very good so far.

Thx
Art

[ArtF]

Hi Guys:

 Just cut my first toolbit tonight. Pretty slick. I think it may be the worlds first wooden toolbit. ( Cut from 3/4" wooden dowel. ).

It was a module 5 involute. Seems to fit the mod 5 gear on my desktop quite well. Basically, you just right click on any gear or timing
pulley in the project screen, select toolbit wizrd and set the parameter s such as stock diameter, feedrate etc..  The code generated will
mill down the dowel to a bar, then tilt the bar to a user defined relief angle, then in multiple passes it will do the involute or timing tooth
shape on the edges of the relief angle. The profile is shifted to compensat e for the relief rotation. I would imagine users would then
harden the bit prior to cutting, but I dont think I can harden this hardwood. :-)

   I have a few tests to run tomorrow to make sure all tooth types work, Ive only tried the involute. . and if all is well I should have this module
in a release by weeks end. 

Art

[Mooselake]

http://en.wikipedia.org/wiki/Fire_hardening

Kirk

[ArtF]

OK..maybe I could harden it.. Suppose it's be OK for cutting those styrofoam gears I tend to cut...


Art