GearHeads Corner

Hi Art,
Looked at the video on collars and that's exactly what I'm trying to do, but when I downloaded 2.0 tonight, when I go to put on a collar of any percentage, the diameter is all the way to the root of the tooth.  Same at 1% or 100%.  Thickness seems to work ok, but the dfx doesn't really work for this because DFX is 2D and I think what this needs is an STL file.  Ran the DFX into Bob-CAD and it doesn't show the collar at all.

Steve

Hi Steve:

  That is weird. What gear type were you using? Im still having trouble with a couple of pulley types, but all gears Ive tested seem to work.. Im thinking it "may" be an inch issue so Ill do a diagnostic run on imperial's today.
  The collars do not appear in dxf's, a 2d dxf wouldnt show them in any event ( other than a circle perhaps), but I havent yet attached the collar contours to the 3d dxf output so Ill see if I can do that as well..

Art

Hi Steve:

  I just uploaded a new version which should work better on collars on most of the objects
it screwed up on previously. Still looking for troubles...

Art

Yes, that revision seemed to fix the problem.  I was doing an 8-tooth sprocket with a .25 shaft and a thickness of .11.  I wanted a collar at least 3x the thickness and probably 75% of the width (to give the set screw some decent threads).  I can see it in the program now, but not coming out as something that can be milled as of yet. 

I think anything much bigger and the amount of waste would be pretty significant but well worth it if it could be left to mill without attention vs. the 4 hours it took me to make and fit collars to my combination wheel.  Right now that has:

1.  3/8" shaft collar
2.  Nylon washer, 3/8" hole 1" diameter
3.  3 x 6-32 socket head cap screws to hold the:
  3a.  Winding sprocket (4")
  3b.  1/4 inch spacer
  3c.  ratchet wheel (3")
  3d.  1/4 inch spacer w/threads
This assembly rotates on shaft.  Chain pulls down against the click attached to ouput sprocket to provide power.
4.  Output sprocket (6 7/8") with attached click, click spacer, click spring,(2 x 4-40 shcs w/lock nuts to hold click and click spring) threaded x3 for 6-32 shcs.
5.  5/16" shaft collar screwed to output sprocket w/set screw to hold it to shaft

So right now, the gears would be too large to economically use the collar idea, BUT on the pinion-sized sprockets, its very necessary since they're only about 1" diameter.  Alternative is to make them solid, make collars and the screw them together which is what I'll probably do for right now.  This clock idea is still a work in progress.

Thanks for all the work on this.  I'm sure other people will start thinking about collars now too!

Steve

Glad it fixed the problem Steve, Im aware of the cause and still looking for the final error
prone ones.

Keep up the good work, :)

Art

Decided to do a multiple step solution for the collars. 

    Use Gearotic to cnc the sprocket.
        1.  Use 4 jaw independent chuck to prep round stock (face, drill 1/4" hole)
        2.  Move chuck to mill
        3.  Mill sprocket to .125 depth
        4.  Move chuck to lathe and cut away everything that doesn't look like a collar
        5.  Thin sprocket to .11
        6.  Part off collar/sprocket
       
This workaround may not be as elegant as doing the whole thing on the mill (probably in 4th axis), but it worked fine once I got the rythym going.  1/4 inch hole is used to center the chuck under the mill spindle.  I have it on my rotary table until I get that chuck-to-mill table adapter from Sherline.  Once the rotary table is centered, the only thing you have to adjust is the z-axis.  I did a half dozen sprockets -12, 10, 8, 7 tooth and the CNC worked like a dream.

Thanks Art!

Steve

Steve:

  Sounds like a lot of processes but not really. Its really what I envisioned Gearotic being
used for. I see gearotic as a helper in operations though not a program to do all of them. While we may grow it eventually into a more full featured CAM program, I always truly meant it as an adjunct to cad/cam , a way of making some things faster than they might otherwise have been.
  When I redo the 4th axis programs we'll consider what else can be done to help . :)

Art

New issue has arisen.  The teeth on the 7 and 8 tooth sprocket are nicely pointed, but when I checked the 10 and 12 tooth sprockets, the teeth are fatter (wider along the face)) overall and have lost their pointedness.  Same tooling, same fixture, same process as before.  The result is the chain will only go down part way onto the tooth, and not all the way to the root.

Any ideas?

Steve

Hi Steve:

  Probably my fault, I changed a few things in there while working on collars, likely a
typo. Which sprocket type was it..?

Art

I'm doing all my sprockets to fit a #25 chain.  The larger sprockets seem to work pretty well, though the teeth aren't as pointed as I expected.  The 12, 10, 8 and 7 tooth have sharper points.

I think I may have a perception problem.  Picture the sprocket with the chair wrapped so that only 2 teeth are full dept, and the rest of the chain must have an angle away from the sprocket.  Like connecting a 8 tooth with an 84 tooth.  There will be some kind of requirement that the angle to the sprockets must have to work.  I'm thinking it would be called something like a minimum circumference.  At 7 teeth, it might be only 1 or two teeth engaged at a time, where a 40 tooth would have the chain wrapped half way around it engaging 20 teeth.  I'm not an engineer or mathematician, only a humble biologist, which means I tend to do things a see what the result is instead of working out the why's and wherefore's and then making the drawing and part.  I tend to reverse engineer (make the part work and then draw it out).

So, I think my next experiment will be to establish some kind of arbor to arbor distance, make up a chain and see if it does what I think it will.  I can envision testing a sprocket/sprocket almost touching and then move the arbors out by one diameter of the driving sprocket and testing again.

Steve

Do you want pics of the different sprockets I'm talking about?

Steve:

  I understand what your saying, ( you explained it quite well for a non engineering type.. :) )

  Reviewing them here the teeth look near identical for all tooth counts. BUT, I did do
multiple changes in there.. so by all means take a photo so I can see if it looks different than mine. We'll go from there..

  As to the arbor issue, typically even with large ratio you'll get several teeth engaged, the math just works out that way, but in extreme cases an idler would be used to press the chain into a further engagement  on either one side or both...

Art

This photo shows the difference between the two types of teeth that I've been generating.  Both were DFX exported to BobCAD/CAM.  Generally only did rough profiling on the larger gear with 3/16" end mill.  I would think that doing a final profile with a smaller diameter end mill might change the profile of the tooth.  I also had to "reverse the start point", otherwise the cutter tried to work within the part and not the waste area.  BobCAD doesn't seem to think this is an issue.  Changing the tool offset doesn't seem to work.

Hi:

  Well those small ones look odd..but the large one just looks...wrong..

Does the DXF show those sharp points? I cant find a type or size that makes that...
It looks very much like the tool size was off or something.. At a minimum the large one should match its dxf exactly.. Were these cut 2.5D?

Art

I didn't like that one either.  Here's the profile I'm probably OK with.  Its my 7 inch sprocket and the weight chain sprocket.  Wrapping the chain around these seems to tell me they're ok.

Yes, those look much better. In the end the best test is as you noted if the chain fits them, if it rides well the sprocket is fine generally..

Art