What's new
What's new

Best approach for curved surface engraving

drcoelho

Stainless
Joined
Feb 19, 2017
Location
Los Altos
Assume I have a cylinder and want to engrave some letters/marks around the curved edge of the cylinder and I have a 4-axis CNC mill. Which is the better approach:
- have the spindle/cutter stay on the x/y plane, and use the 4-axis to move around the cylinder curve
- OR use ball cutter and move the spindle in x/y/z plane with only occassional 4-axis movement when getting too close to vertical on the cutter
 
I would rotate the 4th axis as your going, and don't forget to "wrap" the text along the arc or things will look funky real fast depending on the curvature.
 
IMO, Probably easiest to cut on centerline in Y (or X) and movement in X(or Y) + 4th if you need to make any depth of cut or tool length adjustments at the machine.

If most of movement is 3d with occasional 4th reposition, depth adjustments become a trig problem - either need to be done in cam, unless you're using 3d comp? I bet most people don't though.

Shouldn't be that much slower using the 4th than 3d milling. Unless you're using a spindle speeder and going super fast...

If you're using a ball endmill to do it, got a nice cylindrical part, rotary on center, know exactly what depth of cut you need, I suppose doing it 3d wont be much different.

With an engraving bit you'd want to stay on centerline and use 4th motion, otherwise once you move off centerline you aren't really cutting with the tip anymore, instead more to the side of the cutter... results wont look right i'd imagine since the tool radius varies.


Peronsally, I'd use a 1/32 or .8mm ball, and now that I think about it I'd program the tool with a slight tilt to cut more on the side, isntead of the tool tip. Use TCPC, so rotary motion on the 3+1vmc or 5axis machine.
 
Yes, the quality of the engraving will be better if you can cut slightly off-center on your cylinder (allows you to use a portion of the cutting edge that has more surface speed than when using nearer the tip). Provided you can program for it, you will see superior results.

PM
 
Thanks to all who offered responses to my question, I REALLY appreciate all the very valuable help....

To plastikdreams regarding your comment "When are you going to get a machine and stop with all the hypothetical bullshit here?", you really are unfair, don't see why I can't learn ahead of my machines arriving via reading and queries on this forum, gotta tell you that I have learned a huge amount with all my questions that have directly driven decisions I'm making about machine choice, tooling,etc...and I have already purchased and received most of the tooling and have on order a Weiler Praktikant lathe new from germany (sitting in midwest ready to be delivered), a new FPS 300M (Deckel) manual milling machine (to be shipped from germany in about 4-6 weeks) and a new Brother Speedio sitting at Yamazen ready to be delivered.

The questions I'm asking are directly related to projects I'm planning and as I sit here wondering how one would go about solving various problems. So again, please be gentle....thx.
 
Keeping the tool normal and rotating the 4th is kind of optimal, BUT:

1- Programming it is tricky. HSM Works/Fusion might be able to do it, but it shits the bed a lot on simultaneous 4th axis stuff. In talking with CAM developers, the simple fact is that 4th axis simultaneous motion programming is sort of niche, and all their focus is in the (rapidly growing) 5 axis motion side. 4th axis work is sketchy and bugs galore (even in NX).

2- You need to dial-in the 4th axis. These are really optimal as indexing tables, and doing simultaneous motion can produce some bizarre results. You get 90% of the complexities of full 5 axis motion, but none of the control tools associated with doing something about it (tool center-point control, glass scales, axis calibration routines, etc). Absolutely dial-in the backlash compensation on your control (the Sankyo doesn't have traditional backlash, but all rotary tables have some lost motion in the system that needs to be dialed out via the control's backlash comp parameters).
 
I'll be using NX. I'm also not adverse to doing some code development myself to generate the g-code as needed. I do understand the world is leaving me behind with regard to simultaneious 4th axis CAD/CAM tools. My future direction will involve doing some mathematical models to generate g-code, so again not adverse to going that direction.
 
Thanks to all who offered responses to my question, I REALLY appreciate all the very valuable help....

To plastikdreams regarding your comment "When are you going to get a machine and stop with all the hypothetical bullshit here?", you really are unfair, don't see why I can't learn ahead of my machines arriving via reading and queries on this forum, gotta tell you that I have learned a huge amount with all my questions that have directly driven decisions I'm making about machine choice, tooling,etc...and I have already purchased and received most of the tooling and have on order a Weiler Praktikant lathe new from germany (sitting in midwest ready to be delivered), a new FPS 300M (Deckel) manual milling machine (to be shipped from germany in about 4-6 weeks) and a new Brother Speedio sitting at Yamazen ready to be delivered.

The questions I'm asking are directly related to projects I'm planning and as I sit here wondering how one would go about solving various problems. So again, please be gentle....thx.

I jumped the gun and have to suffer the consequences, for that I am sorry.
 
My thought process is biased as I only have a Haas indexer not a 4th axis, but visually for simple text I think scribe letter, rotate, scribe letter would be indistinguishable from actual 4 axis engraving. If you were wrapping a logo you would have to go 4 axis obviously, but for simple text and numbers one could avoid any weirdness in the post
 
Keeping the tool normal and rotating the 4th is kind of optimal, BUT:

1- Programming it is tricky. HSM Works/Fusion might be able to do it, but it shits the bed a lot on simultaneous 4th axis stuff. In talking with CAM developers, the simple fact is that 4th axis simultaneous motion programming is sort of niche, and all their focus is in the (rapidly growing) 5 axis motion side. 4th axis work is sketchy and bugs galore (even in NX).

Tricky???

Most of what I read online about surfcam is people complaining about or shitting on it. It's the only mill&lathe CAM software I've used, however it really couldn't be easier to program this. Simple 2d contour of your text or pattern, then translate it with wrap cylinder. There. Simple 4th axis motion on centerline. If another cam package can't handle that... pretty pathetic.

4th simultaneous a niche? Plenty of vmc's with 4th's and horizontals out there. Sure this has a lot to do with the type of work and parts you make, but there's plenty of times where these toolpaths can be an optimal strategy. Not that difficult to use 5axis toolpaths, restricting motion to only 4axes either...


2- You need to dial-in the 4th axis. These are really optimal as indexing tables, and doing simultaneous motion can produce some bizarre results. You get 90% of the complexities of full 5 axis motion, but none of the control tools associated with doing something about it (tool center-point control, glass scales, axis calibration routines, etc). Absolutely dial-in the backlash compensation on your control (the Sankyo doesn't have traditional backlash, but all rotary tables have some lost motion in the system that needs to be dialed out via the control's backlash comp parameters).

Bullshit. 90% of complexities of 5axis to setup a 4th and engrave on a cylinder? LOL. How?
Bolt rotary to table, align platter, use probe or indicator/haimer/whatever to find center. Program from centerline.

Glass scales? backlash comp? OP is asking about engraving here. A few arc-seconds of lash aint gonna affect engraving logos or text one bit. WTF you smoking?
 
I often run 3 axis simultaneous when I come across a task like this. If I had an A axis I would run A and X, leave Y on center and let z be "positional". I guess in the end it really depends how far around the outside of the cylinder you need to get.
 
....... 4th axis work is sketchy and bugs galore (even in NX)....

Is there something unique about 4th programming or the code required on a Brother that makes it not play well or more related to engraving specifically?

Shop I retired from was an NX user and did a fair amount of 4th axis work. Not engraving but some really weird 4th contouring that worked really well.
 
Most of what I read online about surfcam is people complaining about or shitting on it. It's the only mill&lathe CAM software I've used, however it really couldn't be easier to program this. Simple 2d contour of your text or pattern, then translate it with wrap cylinder. There. Simple 4th axis motion on centerline. If another cam package can't handle that... pretty pathetic.

Sure, you can wrap it. But that depends on the pattern or toolpath. Seriously intense engraving with a lot of Z axis motion to get fine details can get wonky when wrapped. Simple text? Not an issue.

It seems obvious to me that what the OP is after is not simple...

4th simultaneous a niche? Plenty of vmc's with 4th's and horizontals out there. Sure this has a lot to do with the type of work and parts you make, but there's plenty of times where these toolpaths can be an optimal strategy. Not that difficult to use 5axis toolpaths, restricting motion to only 4axes either...

Yes, it is a niche. Most HMCs offer full 4th axis motion as an expensive option (at least on all the Mori and Makino HMCs I've seen). While there are applications where simultaneous 4th axis motion are really efficient, they are a tiny fraction of what most 4th axis machines do... hence niche.


Glass scales? backlash comp? OP is asking about engraving here. A few arc-seconds of lash aint gonna affect engraving logos or text one bit. WTF you smoking?

Oh really? Do you want actually amazing looking text to print? Or are you happy with a shit show engraving where every axis direction change has a gouge in the text? Again, given the OP's aims with this machine, he is not going to be happy chasing down motion errors. Engraving is just the beginning of what he wants to do (hint: most folks aren't talking about writing their own routines to generate G-code).
 
Is there something unique about 4th programming or the code required on a Brother that makes it not play well or more related to engraving specifically?

Shop I retired from was an NX user and did a fair amount of 4th axis work. Not engraving but some really weird 4th contouring that worked really well.

An issue with the Brother machines is that they have 3rd party tables and require a bit more parameter setup to work really well, depending on the rotary table used. A direct drive Yukiwa table is less finicky than my Sankyo table, for example (less gubbins between the encoders and the workpiece with direct drive; no gear train). Yamazen and I went down this road and we have it figured out, but very few applications in production machining require simultaneous 4th axis motion, so a learning curve for all is involved. Andy and Yamazen were (as always) fantastic about this and will help the OP get set up, but it isn't as simple as plugging one of these things in and getting it running.

Brother is also committed to it, hence their introduction of inverse time feed, but this is all still very new on the control and I understand some enhancements to it are inbound (it currently runs out of digits required to make it go at programmed feed rates, this is being corrected, and Mode B path smoothing doesn't work in inverse time yet).

As far as NX; wrapping text is easy. Contouring cuts with a 4th are easy (if the underlying CAD data is clean). Detailed engraving is where this gets difficult, NX really doesn't have great engraving tools being the core issue. I mentioned Fusion because it has the opposite problem - great (V Carve-like) engraving tools, no good way to wrap them around a 4th axis reliably.
 
Oh really? Do you want actually amazing looking text to print? Or are you happy with a shit show engraving where every axis direction change has a gouge in the text? Again, given the OP's aims with this machine, he is not going to be happy chasing down motion errors. Engraving is just the beginning of what he wants to do (hint: most folks aren't talking about writing their own routines to generate G-code).

Never had gouging issues engraving part numbers on round stuff. Just as clean results as if engraving flat stuff. This is all on "throw away" Haas machines & rotaries or C-axis lathes. But I guess YMMV.
"Shit show" results on high end machines? Sounds like operator error.
 
An issue with the Brother machines is that they have 3rd party tables and require a bit more parameter setup to work really well, depending on the rotary table used. A direct drive Yukiwa table is less finicky than my Sankyo table, for example (less gubbins between the encoders and the workpiece with direct drive; no gear train). Yamazen and I went down this road and we have it figured out, but very few applications in production machining require simultaneous 4th axis motion, so a learning curve for all is involved. Andy and Yamazen were (as always) fantastic about this and will help the OP get set up, but it isn't as simple as plugging one of these things in and getting it running.

Brother is also committed to it, hence their introduction of inverse time feed, but this is all still very new on the control and I understand some enhancements to it are inbound (it currently runs out of digits required to make it go at programmed feed rates, this is being corrected, and Mode B path smoothing doesn't work in inverse time yet).

As far as NX; wrapping text is easy. Contouring cuts with a 4th are easy (if the underlying CAD data is clean). Detailed engraving is where this gets difficult, NX really doesn't have great engraving tools being the core issue. I mentioned Fusion because it has the opposite problem - great (V Carve-like) engraving tools, no good way to wrap them around a 4th axis reliably.

If I wanted a 4th for my S1000 that would be good for cams, tooling, and items requiring simultaneous 4th axis motion, which would you recommend?
 








 
Back
Top