GearHeads Forum

Hi Sky:

  Just a note, for the next version both troubles have been fixed and the involutes and epicycloidals now
do a radial rooting step as well to shave the side walls once to remove any aliasing artifacts..

Thx
Art

Interesting.. Ill check that as well, Ive already fixed the previous trouble. This almost looks like an even/odd miscalculation... either way, Ill fix it up..
thx for the bug..

Art

When you do gears with many teeth, the toolpath shows on top of the tooth, doesn't seem to hurt anything.  (See screenshot).

lol.. Ok, Ok...You wouldnt be complaining if it were multipage gears, it really helps align the pages..but you have a point if its one page gears.
They should not have the alignment circles..only an idiot cannot align a 1 page gear.

I will have that removed by next iteration, it was already on the list.

Art

One last thing [ for now  ;D ]

Printing gears out in 2D.

You get the gear with all the details on the page but also circular rings that don't seem to relate to the gear PD, OD etc ?

Do you realise just how much printing ink costs ?  ???

John:

>>In GM the screen shot gave the max size of the cutter needed, is this in GT ? Can't seem to find it.

  As yet..no. The toolpath display shows pretty clearly if the tool is too large though, watch the center of the tooth, if the path from either side crosses it, the tool is too large.
To see what I mean, set a tool of 0, then 1, then 2 etc.. and see the effect as it gets larger, once seen its pretty obvious if a tool is too large. I will look into codifying it
though, but GM was often wrong, so Im thinking of better ways on that one.

>>Zoom, laptop here no scroll wheel or 3rd button, how do you zoom?

  Press shift, and left mouse for zoom. Ctrl and left mouse for rotate.

>>Also seem to be getting lost between screens and although I eventuall y get back to the tools screen that shows the two gears I often don't know how i got there?

  It does take getting used to. Heres the basic rule. If you select the project tab on the left, youll be placed o the project screen. If you select the tool tab, youll be placed in your most
recently opened tool ( round gears, escapements..whatever..).  The program CAN and does at times switch you around automatically. This will actually become intuitive, I've gotten used to it and it works pretty well. Just think of those two left tabs as your selection. In reality, the bottom tabs also function that way. If you select a tool's bottom tab, youll switch to that tool, if you select properties, youll go back to the project.

This is further modified by the need of some tools ( liek DXF or STL output tools) to have you look at the project screen, so they
  can also autoswitch about to ensure your looking at what you need to. In the case of CNC output, when you select CNC, youll beon the project
  screen so that you can select an object to machine, but will automatically switch to the cnc screen once having done so.
 
    I know this can seem confusing, but Ive been using it for months and have found it soooo much better than GM's confusion. If you just consider
generally that the two left tabs ( project) and (tools) will switch you back and forth ( almost always ) youll be fine.


  Sorry about the radius issue, radius is actually the standard in gear terminology and calcuations, but I do know too dia is pretty typically prefered by
  machinists, so it will be shown shortly for outside measurements .Radius will still be prefered for pitch , base etc. Tools of course will also be in dia.
 
 
  Nice to see GT is now being used in anger..we'll clear all the bugs out in short order that way.
 
 
  Art
 

OK clear as mud !  ;D
I'll digest it later.

Doing a 24 T gear and using 4th axis normal milling cutter.
In GM the screen shot gave the max size of the cutter needed, is this in GT ? Can't seem to find it.

Zoom, laptop here no scroll wheel or 3rd button, how do you zoom?

Also seem to be getting lost between screens and although I eventually get back to the tools screen that shows the two gears I often don't know how i got there?

Bit like driving into Birmingham, been there at least 500 times but never drove out on the same road i drove in on !!  ::)  ::)

Please don't forget diameters, all my bar stock, 6 tonnes of it is labelled in diameter, hell of a job to go round and relabel it all in radii.  :'(

John:

Ahh, now I see. For some reason I sometimes dont get notified on some posts..

So your question is basically why GM's gears dont look like GT's when you do a profile shift.
 Good question..(I figured it would come from you! lol )

     The reason is that GM did not do a proper profile shift. In GM , the shift was a percentage
of the addendum. So a shift of 1.0 ( never really correct to shift by 1.0 ) was a shift by however much the addendum
was by a factor of 100%, a shift of .5 woudl be 50% of the addendum. this was the result of a misunderstanding by me.
When I investigated for GT, ( Ive investigated a LOT for each module of GT's recode.), the concensus of the papers available
that a proper profile shift is a percentage of that gears Module. ( All calc's are in module/metric then converted to imperial
afterwards in order to ensure both modes are identical in nature...also different from GM.

  So lets say you need a profile shift of .5, in reality thats a shift of .5 times the module of the gear, so a Mod5 gear
would shift by about 2.5mm in that case. In the case of GM, you also woudl find that the addendum on a mod5 is 5mm, so a .5 shift also
shifted by 2.5mm. But the similarity stops there. In reality, GM didnt go far enough. Specifications for gears call for the profile
shift to affect more than just the pitchline, though often simple explanations only list the profile shift  as shifting the pitchline.

  True specification calls for the profile shift to move the pitchline, but ALSO it affects the pressure angle of the gear, as in
 
  PressureAngle = PressureAngle *  ((WheelShift + PinionShift) / (WheelTeeth + PinionTeeth));
 
This means the addendum will indeed change..but the pressure angle also changes..something GM did not do. When you think about it,
the pressure angle DOES change. A shift should induce the same effect as shifting a cutting rack deeper into the material ( or shallower).
The actual effect of that moved rack is one on Pitchline as well as the effective pressure angle. It also inpinges on the centering distance.
(Thx for the note, I will add center distance to the text output.).

  In addition to that , (should the gear be a planetary gear), the math has an obligation to take profile shift into account in a different
way for correction of clearance. In planetaries the pinion is of greater importance in the calculation of actual physical effects
of the racks motion. GT also recomputes the centers, pressures, center corrections and dedendums/addendum calculations based on
shift. Youll notice the shift of only one gear affects both wheel and pinion in GT, and this too is by proper specification. The
pitchline and pressure angle having been effectively changed by the shift necessitates a coresponding , though different, change in the
pinion. The results are vastly different than GM's, but follow industry specification as closely as I could determine them. If youd like
to understand the full mathmatics of each modification, yell and Ill send you a paper full of the proper equations. I tried also to verify
GT' ouput by doing size and pressure comparisons to industrial gears sold by various vendors and then checked agains MIT's MITCALC for gears
which seemed to validate all the end numbers of shifted gears. None of this is a perfect science though, MITCALC is somewhat hard to run,
variosu vendors agree, others diagree about how exactly a profile shift should affect an end gear. So for my part I analysed mathmatically
what would happen numerically to the numbers in a gear affected by motion of the rack, and nmear as I could tell the new equations are much
closer if not an exact match to reality of the rack.

    I have verified each type numerically for the shift affects as well as the effects gleason equations have on bevels, and also the transverse
and axial effects the cals have on helicals. I CANNOT ensure that Im correct in all cases though, nor can I be sure what "correct" is as the
specification of one standard are not alwys in agreement with another source. When this occurs I ruminate on the actual math effects I can understand
from the phyical realities of a rack cutter metaphore and go with my gut feeling as to effective end results.

   Its my hope then that the profile shift, stubbing and such will all effect the gear as a cutter rack would in reality even though the gears are
calculated from equation algorithms not phyical modeling algorithms.

 But theres another way to check isnt there. If you use the GTA ( which is a physical modeler that really uses a virtual rack to cut with ), ask for
a gear with the same tooth count and shift it -.5 ( as the GTAS is actually backwards from spec..). The gear youll get if you tooth it is the
physical gear one could expect from cutting with a rack. Youll notice two things. First..the roots of the GTA gear more closely resemble the roots
of the GT shifted gear. Youll also note the gear in GTA is smaller in radius..so small in fact you cannot see the pointed tips. This is because
the GTA is faulty. It didnt increase the size of the blank for the profile shift. It is for reasons such as this that GT is being written, poor
conformance to spec/reality. When GT has its own GTA, its my hope that one can double check by simply creating a round GTA gear from a rack..if Im
correct such a gear should match the math model gears from the round gear module. One will prove the other. At the moment the GTA looks to be
numerically matching GT's output. When I do the GT-GTA module, I'll use that for conformance checking..

Sorry for the verbosity, this is one topic I think needed a more full explanation to explain my philosophy on accuracy ..
ARt

In GM draw a gear 1.5 Mod 24 teeth 20 degree PA, Alter the shift to 0.5, leave stub at 1, leave width at 0.5 regen and look at the drawing.
Looks fine to me.

Now repeat in GT and the screen shot is wrong.

Again in GT sizes are given in radii.
If two gears selected there is no centre distance.