Machining Bevels.

[ArtF]

Hi Guys:

  One question we havent dealt with is blanking. .

  What do you think about the proceedur e of making the blanks for a bevel, how much
should be left to the maker.. Shoudl GM put out code to create the bevel of the blank, or should it assume the blank is properly sized and beveled for the process? Makeing the bevel can be time consuming on the 4th axis.. there is no real fast way for me to do that. Shoudl I worry about it all all? Or should the program assume you have a beveled blank available of the correct face  width?

  How about zeroing? Some get a bit tiffed that I zero Helicals to the top of the blank front for example.
The reason I do this , rather than zeroing the Z to center of axis, is to reduce the number of bad collision s due to improper centering or blank size. Running into the blank is much harder when one is zeroed to the blank rather than the arbitrary Z zero of axis center. While somewhat non-standard it seems to work as Ive
heard of very few ""oops" remarks as to collision s. If the blank isnt proper size you simply get a gear thats wrong, but didnt hurt anything mechanica l...

   Your call really, codewise it doesnt matter. SO what do you think about blanking. .and zeroing..?

Im about to do the output..

Art

[tweakie]

I would be happy to turn the blank to the correct size / shape.
Not fussy about Z zero position but as you say there is less risk with it at the top of the blank. (why is it that the expensive cutter breaks in a heartbeat and the old dull cheapo lasts forever ?).

Tweakie.

[Dan]

Art,

I agree with Tweakie. If you output the blank dimension s, it would be very easy to make it on a lathe. May be consider even outputtin g a DXF of the blank in addition - isn't really necessary as Mach Turn has the required tapper wizard, but just as an option.

I agree with the point you made on zeroing.

Dan

[John S]

Same as Dan.
2D drawing of blank would be nice.

Zero from wherever suits you. We as operators will adapt.

[John S]

Just an add on.

Talking of blanks, just had a look at the on screen informati on for bevels.
It gives the pitch cone angle and root angle but no face angle which is nessesary for turning the blank up.

It's also necessary to define what the angle is to, a 45 degree mitre gear will probably have a 47 - to 49 degree face angle depending on size but it gets quite confusing when turning the blank whether the 47 degrees is from the X axis or Z axis.

Large ratio's won't suffer from this but mitre gears have such a small differenc e it's easy to get the wrong angle.

John S.

[ArtF]

John:
 Noted.. Ill put out the face angle as well. ( which is basically the pitch angle with the addendum angle added to it. The reason I hadnt till now was if we decided to use parallel cutting, the angles woudl stay the same as the pitch cone angle.

  Looking at the math of all this, Ive decided that it doesnt make sense NOT to tilt the table. A tilted 4th can do pretty much any bevel while any other method will fail on large angles and likely on all helical bevels or zerols.

  This being the case Im thinking zeroing IS difficult as we need a point readily identifia ble on the blank. Zeroing to center of a tilted table can be difficult with the mounted blank.  Any thought on an easily definable zero point?  Perhaps inside top edge of the mounted blank? ( Thats assuming a bevel is already applied? )

  Gone to take a good look at my 4th to see what I can do to make it easily set to an angle..

Art


   

[Dan]

Art,

When we tilt the 4th axis, we want the bevel face to be parallel to the XY plane (at its top end), right? Then why not zero the tool on top of the face there and on the OD... Does it makes sense...?

Dan

[John S]

OK so tilt is agreed.
Following on to what Dan says and it makes perfect sense, we need a point that you can define easily.

Y is easy as you can touch off on both sides of the blank and take half the measureme nt to get on the centre line.

Z is also easy, touch off on the top outer edge as this is the highest point, rememberi ng that the pitch angle will be the one you set up on.

X, again easy as it the outer edge.

All points easily dome with wiggler and / or cigarette paper.

[Chuck]

Art,

I expect you will be forced to come up with several machining options.  While setting the 4th axis to the pitch angle will help a bunch I think you should consider other setups.

Using standard endmills and 4 axis should be the first option for larger pitch.

If you create a Gcode generator and a simple setup you can make formed cutters  that will allow just about anything.   Don't get stuck on only using endmills.

You can do most strait cut gears on a 2.5D/3D if you create a cosine stretched formed endmill.  This would allow most users to make a cutter with one easy setup and the gear with another easy setup.

A non stretched involute form will do most of your gears with the 4th axis set at the pitch angle.  This can be a fly cutter for strait teeth or a formed endmill.

I do have a question for you Art,  is the involute curve the same at the small pitch end of a tooth as it is at the large end?  By this I am asking if one profile form can make the correct shape for the entire tooth with varying pitch?

If a form cutter can be made to cut the small pitch end and the angle of the gear set to cut the widening tooth any pitch strait tooth could be made with a formed fly cutter or a formed endmill.  The spiral teeth would require a formed endmill.

Soft material (soft compared to the machine rigidity) can be quickly done with a non rotating formed shaper tool with a 4 axis setup.

As for tool zero, if you specify points on the gear blank appropria te for the tool used I think it is up to the machinist to set up the machine and blank correctly or adjust his zero points as needed to put the zero where you want it. With 4 axis at any angle a Y centering touch on each side of the gear using a fixed X and Z then a point at the top of the gear blank for X and Z makes the most sense.

Chuck in Wyoming

[ArtF]

Hi Chuck:


>>If you create a Gcode generator and a simple setup you can make formed cutters  that will allow >>just about anything.   Don't get stuck on only using endmills.


   We wont be, I plan on using tilted 4th axis at first, as it will cut most any bevel angle..
I then plan to put out code for cutters for normal spurs and helicals. . The tilted 4th axis is just the first step to prove the concept and get bevels cutting..


>>You can do most strait cut gears on a 2.5D/3D if you create a cosine stretched formed endmill.  This would allow most users to make a cutter with one easy setup and the gear with another easy setup.


   Im not sure what a cos stretched endmill is usefull for.. can you explain that one. If I stretch an endmill by cos, the involute would be lost woudlnt it? At least on spurs it would..


>>A non stretched involute form will do most of your gears with the 4th axis set at the pitch     angle.  This can be a fly cutter for strait teeth or a formed endmill.

    YOu mean bevels? I think the invcoltes woudl be lost.. (I think.. )


>>I do have a question for you Art,  is the involute curve the same at the small pitch end of a tooth as it is at the large end?  By this I am asking if one profile form can make the correct shape for the entire tooth with varying pitch?

   No. Thats the problem, since the diamter of the pitch changes, ( and the base circle), a proper involute changes over the length of the face. Unless I use a parrallel machined bevel.. which is possible but weakens the teeth..


>>If a form cutter can be made to cut the small pitch end and the angle of the gear set to cut the widening tooth any pitch strait tooth could be made with a formed fly cutter or a formed endmill.  The spiral teeth would require a formed endmill.

      A formed cutter doesnt help much with a proper bevel.. as far as I can see.. The involutes would be deformed across the length of the face..

 
    Zeroing and such I sont think will be a problem.. its more of a question of intuitive ness to ensure peopel can zero simply and easily.. I guess thats one that can be worked on when we're cutting to make sure people are happy.. pretty easily changed anyway I think..


 Thx

ARt

[BobL]

Hi Guy's;

 So is this the new standard method we will be using?

Y - touch off on both sides of the blank and take half the measureme nt to get on the centre line.

Z - touch off on the top outer edge as this is the highest point, rememberi ng that the pitch angle will be the one you set up on.

X, - touch off on outer edge


Cheers
Bob

[Chuck]

Art,

Sorry but I am thinking in terms of the non generatin g bevel gear shappers.  But if the involute curve changes from small pitch to large pitch then ALL my formed cutter ideas are crap.  As I knew, bevels are a whole different beast than strait spurs.... .with endless variation s!

How about a tapered endmill and 4 axis? or do we call it 4.5 axis setup for the pitch angel.  I have a nice Gcode program to make tapered endmills with a specific angle and tip width.


While trying to educate myself I found a few links to industry standard processes and equipment:

Straight Tooth Bevel Gears
Google books:
http://books.google.com/books?id=6PE0jZZ-nr0C&pg=PA445&dq=bevel+gear+shaper&hl=en&ei=OOIUTeKhNcOB8gay6cSyDg&sa=X&oi=book_result&ct=result&resnum=1&ved=0CCkQ6AEwADge#v=onepage&q=bevel%20gear%20shaper&f=false

http://books.google.com/books?id=U7ZKAAAAMAAJ&pg=PA328&dq=bevel+gear+shaper&hl=en&ei=6d8UTYqFE8P88Aa434T7DA&sa=X&oi=book_result&ct=result&resnum=1&ved=0CC8Q6AEwADgU#v=onepage&q=bevel%20gear%20shaper&f=false

http://www.youtube.com/watch?v=Adi0GgUc2Z4
http://www.youtube.com/watch?v=hvDO7t8h_sI
http://www.youtube.com/watch?v=LEvpy66AqWA
http://www.youtube.com/watch?v=8H-FDtDMfMg
http://www.youtube.com/watch?v=ul9wCexWlXM

Spirals
http://www.youtube.com/watch?v=0whwtairKOI

And nothing to do with gears but still cool: (state of the art 1860)
http://www.youtube.com/watch?v=SW1JFaFWtKI
http://www.youtube.com/watch?v=VLZHV3v_FIo

[ArtF]

Hi Chuck:

    Thx fer the links, I'll browse them over the holidays.

 Im not sure a taper adds much capabilit y, sounds like it'd perhaps make the math worse..
but Ill dwell on it. I think an endmill should suffice for the first iteration s.. we'll see how they look
and go from there, there are several ways to skin this cat I think, obviously though its a complex
thing, you dont see homemade bevels very often..un less their basically engraved, and thats pretty slow.

Have a good XMas..

Art

[Chuck]

Art,

The tapered endmill would allow you to cut the small pitch teeth.  It would let you get to the tooth root without undercutt ing the sides.  The taper makes the mill MUCH stronger for the size of cut.  I don't think the math will get much worse.  You wouldn't be able to use G42/43 offsets, but for computed paths would you use G42/43 anyway.

The tapered endmill would allow CNC to replicate  the generated octoical tooth form of bevel gear machines for strait and spiral bevels.

Have you looked at octoical bevel tooth forms?  This is the common tooth for for generated bevel gear shaper machines like Fellows and Bilgram.  This form should be much easier to generate than involute for bevels.

More REALLY GOOD light reading ( a few thousand pages) from Google scanned machine shop books:

The last 1/2 of this book is really good.
http://books.google.com/books?id=icxKAAAAMAAJ

Good theory and descripti on of octoical bevel tooth forms
"AMERICAN MACHINIST GEAR BOOK"
http://books.google.com/books?id=339syGI6JW0C&pg=PA9

http://books.google.com/books?id=O7dKAAAAMAAJ

http://books.google.com/books?id=Y1ZJAAAAMAAJ

http://books.google.com/books?id=U7ZKAAAAMAAJ

Lots of theory and math:
http://books.google.com/books?id=0KVBAAAAIAAJ

http://books.google.com/books?id=aLVKAAAAMAAJ

There are more but these are all I have looked through and selected as relevant over the last 2 days of speed reading.

Chuck

[ArtF]

Hi Guys:

  Chuck just had a great idea offlist.. How about we agree to machine bevels with the 4th axis tilted only to 45 degrees. This effective ly splits the differenc e of troubleso me angles bringing them to a level where we can probably do them all.. and its a much easier thing to arrange.

   This means one could find a setup or adjustmen t that puts their tabel to 45 degrees and know the code is OK for that gear. Otherwise we'll be trying to angle to things liek 23.1254 degrees ..the numbers will rarely be integer..

   By tilting to 45, the worst bevels at 89.99 degrees become 45 degree to the code, much easier to cut.. and lower bevels like 45 degrees, become straight cuts..  lower become upward cuts.. downside being a probable
code modificat ion to cut from the other direction to eliminate plunge problems with an endmill..

  This is handy because statistic ally most bevels will fall between 20 and 70 degrees meaning the normal cut angle will range from upcutting at 25 degrees to downcutti ng at 25 degrees. Its very likley we can do all types with that small a variation in angle... at least in my head at the moment, Im still visualizi ng.... ..

    It DOES also mean that I can do the fractiona l angling internall y so youd only have to worry about making your 4th angle to exactly 45 degrees for any bevel gear.

  Anyone see a probalem with this compromis e solution?  John? Anyone? :-)

  I think Ill start to do the code on this .. so yell if you see a pitfall..

Art