No 4th axis: Will Helical Interpolation do this? - Page 3

# Thread: No 4th axis: Will Helical Interpolation do this?

1. Stainless
Join Date
Oct 2013
Location
New Jersey
Posts
1,115
Post Thanks / Like
Likes (Given)
0
949
I don't get it. Why in this day would you NOT be doing this in CAM? Please don't tell me you can't afford it because what you really can't afford is what you are doing now! For FREE for forty sessions and \$149 if you buy you can get CamBam. Yeah, I know, hobby program. I can draw a cylinder, extrude it, draw a rectangle, round the corners, project rectangle onto the cylinder to get a polyline. Use engrave operation to write toolpath following the polyline. I am a slow typist, I could just about finish the program in about the time it took to write this. Think of CamBam as training wheels until you can justify the big boy toys. Life is too short to write G-code by hand.

2. Banned
Join Date
Aug 2016
Country
ALAND ISLANDS
Posts
1,393
Post Thanks / Like
Likes (Given)
43
96
parametric core code for osp300

discretization available for user input as V2 inside OSC soubroutine

its based on simple geometry

Code:
```    LVL  =     ( window side 1, among circumference )
LVH  =     ( window side 2, among generator     )
LVR  =     ( window corner radius               )
LVCR =     ( cilinder radius                    )

( * )

CALL OINI    ( initializations )
( CALL OSUB1 ) ( variant 1 : rough corners   )
CALL OSUB2   ( variant 2 : smoothed        )

RTS

OINI

LVL = LVL / 2
LVH = LVH / 2

LAX = -LVL
LAY = +LVH
LBX = +LVL
LBY = +LVH
LCX = +LVL
LCY = -LVH
LDX = -LVL
LDY = -LVH

LA1X = LAX+0
LA1Y = LAY-LVR
LA2X = LAX+LVR
LA2Y = LAY-0

LB1X = LBX-LVR
LB1Y = LBY+0
LB2X = LBX+0
LB2Y = LBY-LVR

LC1X = LCX+0
LC1Y = LCY+LVR
LC2X = LCX-LVR
LC2Y = LCY+0

LD1X = LDX+LVR
LD1Y = LDY+0
LD2X = LDX+0
LD2Y = LDY+LVR

RTS

( . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . )

OSUB1

G00 X = LA1X+2.5 Y = LA1Y Z0
G01 X = LA1X

LV01 = SQRT [ LVCR * LVCR - LA2X * LA2X ]

*normal plane
G02 X = LA2X Y = LA2Y Z = +LV01 R = LVR
* shift plane
G02 X = LB1X Y = LB1Y           R = LVCR
*normal plane
G02 X = LB2X Y = LB2Y Z = -LV01 R = LVR
G01 X = LC1X Y = LC1Y
G02 X = LC2X Y = LC2Y Z = +LV01 R = LVR
* shift plane
G02 X = LD1X Y = LD1Y           R = LVCR
*normal plane
G02 X = LD2X Y = LD2Y Z = -LV01 R = LVR
G01 X = LA1X Y = LA1Y
G00 X = LA1X+2.5

RTS

( . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . )

OSUB2

LUH1 = ATAN2[ SQRT [ LVCR * LVCR - LB2X * LB2X ] , LB2X ]
LUH2 = ATAN2[ SQRT [ LVCR * LVCR - LB1X * LB1X ] , LB1X ]
LUH3 = LUH2 - LUH1

G00 X = LA1X+2.5 Y = LA1Y Z0
G01 X = LA1X

*normal plane
CALL OSC LV01 = 1 LV02 = LA2X LV03 = LA1Y
* shift plane
G02 X = LB1X Y = LB1Y           R = LVCR
*normal plane
CALL OSC LV01 = 2 LV02 = LB1X LV03 = LB2Y
G01 X = LC1X Y = LC1Y
CALL OSC LV01 = 3 LV02 = LC2X LV03 = LC1Y
* shift plane
G02 X = LD1X Y = LD1Y           R = LVCR
*normal plane
CALL OSC LV01 = 4 LV02 = LD1X LV03 = LD2Y
G01 X = LA1X Y = LA1Y
G00 X = LA1X+2.5

RTS

( . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . )

OSC ( quadrant Xc Yc )

IF [ LV01 EQ 1 ] N11
IF [ LV01 EQ 2 ] N12
IF [ LV01 EQ 3 ] N13
GOTO N14

N11 LUA  = 180
LUB  = 180 - LUH1
LUBD = - LUH3
GOTO NHALT
N12 LUA  =  90
LUB  = LUH2
LUBD = - LUH3
GOTO NHALT
N13 LUA  = 360
LUB  = LUH1
LUBD = + LUH3
GOTO NHALT
N14 LUA  = 270
LUB  = 180 - LUH2
LUBD = + LUH3
GOTO NHALT
NHALT

V1 = 0
V2 = 10
NAGAIN V1 = V1 + 1
Z+LVCR * [ 1 - LVCR * COS [ LUB + [ LUBD / V2 ] * V1] ]
IF [ V1 LT V2 ] NAGAIN

RTS

( . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . )```

3. Banned
Join Date
Aug 2016
Country
ALAND ISLANDS
Posts
1,393
Post Thanks / Like
Likes (Given)
43
96
output of shared structure is pretty similar to the code shared by Goo Proto in #11

just an observation : i aproximate an arch with 10 segments, while code from #11 uses 11

when you are near the cnc, a parametric structure has all the benefits : fast editable, faster trials, etc ...

4. Banned
Join Date
Mar 2017
Country
CHINA
Posts
1,576
Post Thanks / Like
Likes (Given)
0
1133
Originally Posted by Ox
I have no clue why you think that a 5x would be in any way quicker than the 3x on this?
Because he could use a corner-rounding or chamfer mill and make one pass around the perimeter. Contouring it with a ball end is gonna be a lot more passes.

That is, if it has to match the print and he isn't gonna just break the edge in a half-ass manner.

5. Banned
Join Date
Aug 2016
Country
ALAND ISLANDS
Posts
1,393
Post Thanks / Like
Likes (Given)
43
96
Originally Posted by SeaMoss
Because he could use a corner-rounding or chamfer mill and make one pass around the perimeter. Contouring it with a ball end is gonna be a lot more passes.

That is, if it has to match the print and he isn't gonna just break the edge in a half-ass manner.
3axis + reversed_chamfer_tool is more than enough

6. Banned
Join Date
Mar 2017
Country
CHINA
Posts
1,576
Post Thanks / Like
Likes (Given)
0
1133
3axis + reversed_chamfer_tool is more than enough
Won't make what's on the drawing ...

7. Titanium
Join Date
Nov 2011
Location
Missouri, USA
Posts
2,124
Post Thanks / Like
Likes (Given)
1354
2723
Originally Posted by Ox
I have no clue why you think that a 5x would be in any way quicker than the 3x on this?

Especially if you are thinking HMC with live 4th. For that to be any differ'n't than this guys 3x, the part would need to be in C/L of the B, and that would be no faster than a 4x vert, and likely slower than this guy tossing it in a vise.

A 4x would be easier to prog on this part, but his 3x will be quicker part-to-part, run at a cheaper rate, and he can skim his own off the top.

--------------------

Think Snow Eh!
Ox
First part that might be making a difference - I'm assuming the chamfer need to look even (equal leg length) and equal angle, and consistent all around.

With that said, keeping a standard chamfer tool at the CORRECT angle would absolutely accomplish it in a straight linear or arc movement at one level, no multi passes, no ball milling, no blending, no stepover, nothing. But this requires the part to be rotated so that the spindle is perpendicular to the vector of tangency of the feature being milled and the tube surface.

Would not need to be in CL of B. So long as the tool can reach the edge when the pallet kicks at an angle. Kick the tube vertical in fixtures, pack them as much as tool holder / chamfer tool allows, and go to town.

I only mention 5 axis because it could accomplish the same.

Meanwhile on a 3 axis machine, the chamfer is more and more imperfect as it gets away from top-dead-center of the tube. It'd be absolutely terribly off at the edges parallel to axial without making multiple passes.

Mill windows. Buzz chamfer around in single pass. Time for parts loading again.

Edit-to-Add: If my eyes suck and that's actually a corner round, and not a chamfer, this applies exactly the same, but with a corner rounding tool. Even more profitable since you're not 3d-ball-milling an aesthetic edge break that'd take forever.

8. Banned
Join Date
Aug 2016
Country
ALAND ISLANDS
Posts
1,393
Post Thanks / Like
Likes (Given)
43
96
Originally Posted by SeaMoss
Won't make what's on the drawing ...
... it is only a chamfer

9. Titanium
Join Date
Nov 2011
Location
Missouri, USA
Posts
2,124
Post Thanks / Like
Likes (Given)
1354
2723
... it is only a chamfer
Never underestimate how picky a customer can be when they introduce the word "aesthetic".

I've learned the hard way that the word "aesthetic" raises all sorts of flags unless my next question is met with "Oh, we just mean we don't want to see burrs or gouges in the edges. I don't care SPECIFICALLY how that's done." Then I'm ok.

10. Banned
Join Date
Aug 2016
Country
ALAND ISLANDS
Posts
1,393
Post Thanks / Like
Likes (Given)
43
96
Originally Posted by JNieman
Never underestimate how picky a customer can be when they introduce the word "aesthetic".

I've learned the hard way that the word "aesthetic" raises all sorts of flags unless my next question is met with "Oh, we just mean we don't want to see burrs or gouges in the edges. I don't care SPECIFICALLY how that's done." Then I'm ok.
hy i know what you mean ...

i am not afraid of such situations when i have to negociate such aspects

i am afraid when i know that i may be rejected because of them : some clients can refuse parts that we craft here, only to get them from another place where quality is much lower; those are heavy games

11. Originally Posted by JNieman
First part that might be making a difference - I'm assuming the chamfer need to look even (equal leg length) and equal angle, and consistent all around.

With that said, keeping a standard chamfer tool at the CORRECT angle would absolutely accomplish it in a straight linear or arc movement at one level, no multi passes, no ball milling, no blending, no stepover, nothing. But this requires the part to be rotated so that the spindle is perpendicular to the vector of tangency of the feature being milled and the tube surface.

Would not need to be in CL of B. So long as the tool can reach the edge when the pallet kicks at an angle. Kick the tube vertical in fixtures, pack them as much as tool holder / chamfer tool allows, and go to town.

I only mention 5 axis because it could accomplish the same.

Meanwhile on a 3 axis machine, the chamfer is more and more imperfect as it gets away from top-dead-center of the tube. It'd be absolutely terribly off at the edges parallel to axial without making multiple passes.

Mill windows. Buzz chamfer around in single pass. Time for parts loading again.

Edit-to-Add: If my eyes suck and that's actually a corner round, and not a chamfer, this applies exactly the same, but with a corner rounding tool. Even more profitable since you're not 3d-ball-milling an aesthetic edge break that'd take forever.

If it were a corner rounded - I'd tend to agree with you, but this being a 45, and with any good code of any sort will make the same part on 3x. I don't see at all why you think it's daffycult enough to put on a machine with 2x the costs.

I don't know why anyone would use a ball mill on this in the first place?

However - it does seem like the way things are these days - "let's throw money at it and see if it goes away".

-----------------------

Think Snow Eh!
Ox

12. Stainless
Join Date
Sep 2015
Country
UNITED STATES
State/Province
Ohio
Posts
1,173
Post Thanks / Like
Likes (Given)
734
1175
If he's putting on a big honkin chamfer, say .030 or so,I think it will look a little weird doing it in 3 axis. Rotary is the way to go to be more visually pleasing.

If its just to deburr, one pass around with a chamfer mill or ball mill to give a .005/.007 edge break should be good.

13. Titanium
Join Date
Nov 2011
Location
Missouri, USA
Posts
2,124
Post Thanks / Like
Likes (Given)
1354
2723
Originally Posted by Ox
If it were a corner rounded - I'd tend to agree with you, but this being a 45, and with any good code of any sort will make the same part on 3x. I don't see at all why you think it's daffycult enough to put on a machine with 2x the costs.

I don't know why anyone would use a ball mill on this in the first place?

However - it does seem like the way things are these days - "let's throw money at it and see if it goes away".

-----------------------

Think Snow Eh!
Ox
If it's just an edge break to deburr, like I mentioned, no big deal. If it's to put a more significant chamfer on it, for aesthetic purpose, the customer (IME) has nitpicky requirements for "looks" and it usually revolves around being consistent. I'm just saying that you can't get consistent geometry of chamfer around a cylindrical surface with a single pass of a standard chamfer tool in one axis.

A consistent chamfer size, and consistent angle at the edge .... you're not at the same angle to the spindle, in a 3 axis machine.

That's all I'm saying. In which case, a horizontal (yes, a more expensive machine) would make them faster than any damned thing, and cheaper (which is how expensive machines pay for themselves quickly)

14. Stainless
Join Date
Sep 2015
Country
UNITED STATES
State/Province
Ohio
Posts
1,173
Post Thanks / Like
Likes (Given)
734
1175
Not to argue, because maybe I don't understand what you're saying, but...

On a 3 axis chamfer you will always be 45 to the wall of the pocket.
On a 4 axis chamfer you will always be at 45 to the outside of the cylinder.

If a print called out .030 x 45 I think you could argue the point by measuring the straight wall that's left and saying "look it's .030 less than the wall thickness"

I have a super picky customer (medical) and this has come up before.
I usually catch it and ask which way they prefer.

15. Titanium
Join Date
Nov 2011
Location
Missouri, USA
Posts
2,124
Post Thanks / Like
Likes (Given)
1354
2723
I've encountered uber-picking quality people on this same topic regarding semiconductors and heat exchangers (note, it wasn't even the thermally critical areas...) as well as a special WRENCH we made... it was the handle... of the 'twisted' wrench. They didn't like the inconsistent appearance of the edge break. They accepted them but wanted to advise us to correct the "discrepancy" on future orders. *eye roll*

That's the only reason I'm picking that nit in this thread. "Aesthetic" raises flags with me depending on the customer, and sometimes a simple thing gets made non-simple for one process, and may require considering another.

....still think at the least, OP should consider getting someone to program it in CAM software instead of trying to do it by hand.

16. Stainless
Join Date
Sep 2015
Country
UNITED STATES
State/Province
Ohio
Posts
1,173
Post Thanks / Like
Likes (Given)
734
1175
Isn't Fusion 360 cheap/free for small shops?

If you have a CNC mill, even as a hobby machine, you're really limiting yourself not having some kind of CAM.

17. I use a ball mill to deburr odd shaped things. It's just one pass around the contours, not multiple passes. It took me a while to find out how to create the geometry for things like this. If you just create lines at the edges of the pocket on the solid and try and drive a 3d contour path with it, you usually won't be to happy with the results.

And if you aren't in a hurry and want to understand more I think doing it by hand can be a great learning experience. The computer is the easy part, knowing what you want out of it is usually the actual work part of it

And some things are actually easier by hand, at least until you can find out how to use the CAM better. Had a job the other day where there was a 3d arc-ramp cut, and I couldn't get MCAM to do G2/G3 in the code, all I could get was line segment output, which left facets on the part. Just programmed a 2d toolpath and added Z's in manually, and it did exactly what I wanted.

18. Hot Rolled
Join Date
Feb 2013
Country
UNITED STATES
State/Province
Idaho
Posts
698
Post Thanks / Like
Likes (Given)
151
749
Originally Posted by Ox
If it were a corner rounded - I'd tend to agree with you, but this being a 45, and with any good code of any sort will make the same part on 3x. I don't see at all why you think it's daffycult enough to put on a machine with 2x the costs.

I don't know why anyone would use a ball mill on this in the first place?

However - it does seem like the way things are these days - "let's throw money at it and see if it goes away".

-----------------------

Think Snow Eh!
Ox
In theory, using a ball endmill is a good way to approximate a 3D Chamfer. It isn't a surfacing routine, just one pass where the CAM package calculates which portion of the ball best approximates the 45 degree chamfer at that point. As the radius of the ball increases, the deviation decreases. In practicality, I have never been successful (patient enough) to do more than experiment with this. It generally requires such a large endmill that you can't get into any tight corners. I should probably revisit this, since I do tons of 3D chamfers, but in my experience as long as the change in slope of the chamfer is substantially less than the angle of the cutter, you can get a good looking chamfer that will pass the most stringent CMM QC check. As usual, your mileage may vary, and this is all contingent on the right CAM package.

19. This is where I learned how to create the geometry. I follow this process and the chamfers come out very nice and even

https://www.emastercam.com/forums/to...comment-906073

20. Aluminum
Join Date
Apr 2013
Country
UNITED STATES
State/Province
Texas
Posts
213
Post Thanks / Like
Likes (Given)
2
104
I use the trace tool path in InventorHSM to break 3d edges with a ball mill. Quick and easy. I suspect fusion has similar function.