Old School Swarf and New Swarf?

[crazy^millman]

Okay been programming 5 Axis machines for a day or two in Mastercam. I just ran into an issue where I cannot use the new swarf. The old school swarf proved more robust and capable on the part I am programming. I got filtering a huge plus, I got sync by entities another huge plus, I didn't have to lie to the toolpath and tell it minus the radius of the tool to get it to cut on center line. The new swarf was dancing all over the place and in Vericut it looks like a hammered surface. The old School toolpath processed faster and produced a much better toolpath. I like the controls of the new toolpath with regards to lead in/outs, but if you know how to work with the old school you can get good results. I cannot share the file, but a copy of this will be going to the right person to look over.

 

What are others seeing with the new swarf verse the old swarf?

[Guest]

I have actually had geometry that the old swarf just didn't like.....New swarf chewed right through it.....

 

I generally use the older one first.......

[BenK]

I like the old one overall but the new one has better controls on a lot of things. It's nice having the two options so if one doesn't work the other generally will. I normally use the old one first also.

[gcode]

I havn't used old swarf in at least 5 years

The difference in surface quality between old an new is night and day, at least with the parts I'm cutting

If you can.. send me a file, I'll check it out

[gcode]

The new swarf has

really powerful linking options..

feed control

robust stock awareness

the ability to shorten or lengthen the toolpath without modifying geometry

start at point (combined with length can give you an overlap on a closed swarf)

powerful retract/unwind/rewind control

built in transform/rotate

 

it does not have filtering ablitiy, but the big Okuma I run likes points

though I have run into occasions where filtering would be useful

[crazy^millman]

The new swarf has

really powerful linking options..

feed control

robust stock awareness

the ability to shorten or lengthen the toolpath without modifying geometry

start at point (combined with length can give you an overlap on a closed swarf)

powerful retract/unwind/rewind control

built in transform/rotate

 

it does not have filtering ablitiy, but the big Okuma I run likes points

though I have run into occasions where filtering would be useful

 

Agreed why I fought and fought with it till finally giving up and going back to the old school toolpath. I have about 20 hours in just these areas of the part. Critical areas that have to be right. The Vericut checked them out. The new toolpath process was over 8000K and the old school was was only 251K to cut the same area. Trunnions do like a lot of code, but you take that same program to the JOBS and you will see the lesser code for that head to run through the smoother it will run. That is what I am seeing so far on what I have been running through the 2 JOBS these programs are running on.

 

Edit/ I will throw this into the conversation. The new toolpath will hyper thread and the old toolpath will not. You don't realize how much you like the hyper threading process for toolpaths till you get into the toolpaths that don't use them. Yes I think it is still processing faster, but I am locked out of Mastercam while waiting for the old Swarf for process where I don't have to with the new one.

 

Let me get back at it I am behind schedule on this one. Thanks for the replies and I did check the newer version has both. :clap:

Edited July 13, 2016 by 5th Axis Consulting Group

[crazy^millman]

The New Toolpath will not support KeySeat cutters and the Old School Toolpath will support them.

[Jim at Gentex]

 

start at point (combined with length can give you an overlap on a closed swarf)

 

 

I just upgraded to X9 and started looking at the new swarf interface.

Yes, I can see there is much more flexibility here.

 

I'm having trouble with the start point feature.

The old swarf interface had the 'select first surface' option which would set the start point there.

I see the new dropdown start point menu, but when I select one of those options I am expecting to be prompted to select the geometry.

That doesn't happen, and the start point stays where the 'automatic' setting placed it.

What am I missing here?

[gcode]

I just upgraded to X9 and started looking at the new swarf interface.

Yes, I can see there is much more flexibility here.

 

I'm having trouble with the start point feature.

The old swarf interface had the 'select first surface' option which would set the start point there.

I see the new dropdown start point menu, but when I select one of those options I am expecting to be prompted to select the geometry.

That doesn't happen, and the start point stays where the 'automatic' setting placed it.

What am I missing here?

 

Its a little klutzy

Once you select Start Point, you'll notice a "+" sign appear next to "Cut Parameter"

Expand that and you'll get t a point selection box

[BenK]

Its a little klutzy

Once you select Start Point, you'll notice a "+" sign appear next to "Cut Parameter"

Expand that and you'll get t a point selection box

 

Is there any way to get the tree items to default to open? Clicking that plus all the time really adds up. 

[Jim at Gentex]

Its a little klutzy

Once you select Start Point, you'll notice a "+" sign appear next to "Cut Parameter"

Expand that and you'll get t a point selection box

 

OK got it!

 

Yes, klutzy is a good description.

It does work so that's cool. 

 

Thanks.

[JoshC]

Many of you have really good points on this topic. I always try the "classic" swarf first as its easierr in my opinion and has less settings to worry about. If however, the classic doesnt give me enough control i move to the Surface/solid swarf as this has 4 serpeate collision strategies that can be used amolg many other advanced options like roughing, etc. that are far more powerful than classic offers.

 

so in summary, simple stuff run classic, advanced stuff run Surface/solid, but if you have a complete understanding of Surface/solid swarf i dont see why we couldnt use that for everything, just preferance mostly i believe.

[crazy^millman]

Many of you have really good points on this topic. I always try the "classic" swarf first as its easierr in my opinion and has less settings to worry about. If however, the classic doesnt give me enough control i move to the Surface/solid swarf as this has 4 serpeate collision strategies that can be used amolg many other advanced options like roughing, etc. that are far more powerful than classic offers.

 

so in summary, simple stuff run classic, advanced stuff run Surface/solid, but if you have a complete understanding of Surface/solid swarf i dont see why we couldnt use that for everything, just preferance mostly i believe.

 

Josh I documented with CNC instances where the New Swarf failed and the traditional Swarf shined.

 

Part of what we do as a developer partner for CNC Software is document issues and provide real world feedback to help improve the product.

[crazy^millman]

Had a situation where I wanted to machine the walls of some ribs. New Swarf was not respecting the overhanging feature the rib was coming right up next to. It was also not paying attention to the Spindle Head it keep trying to crash into the part. (Yes using all 4 of the Collision Controls)(No Mastercam toolpaths respects the kinematics of the machine) I was going to old school Curve it with Tool Axis control lines and ten gave it some more thought. I created my upper and lower chains and made an old school swarf toolpath using sync by entity. Normally with curve I would have to keep tweaking my end tool Axis vector line to get the movement I wanted. With using the way I just keep extending or shorting the upper chains of my swarf till the tool cleared the overhanging feature and was not slapping the head into the part. New Swarf will not use only chains to drive the toolpaths where the old school Swarf will and in this case I figured it saved me a good 10 to 12 hours worth of modeling and making tool axis control lines using the a way that has worked perfectly for years with curve 5 Axis and tool axis control lines.

[Aaron Eberhard]

Interesting discussion, guys, thanks   I'm going to mostly stay out of it, but I will be mining it for info...

I just wanted to pop in on this one:

New Swarf will not use only chains to drive the toolpaths where the old school Swarf will

 

That's actually incorrect, but not obvious.

 

The reason it looks like that is because the strategy is set to Automatic, which relies on using the surface/mesh information to calculate the correct (or what it guesses is the correct) orientation and position.  If you choose a different strategy, like say, sync with tilt lines or sync with curves:

 

 

You'll find that you can now uncheck your drive surfaces:

 

 

Hope this helps.

 

Cheers,

[crazy^millman]

Interesting discussion, guys, thanks   I'm going to mostly stay out of it, but I will be mining it for info...

I just wanted to pop in on this one:

 

That's actually incorrect, but not obvious.

 

The reason it looks like that is because the strategy is set to Automatic, which relies on using the surface/mesh information to calculate the correct (or what it guesses is the correct) orientation and position.  If you choose a different strategy, like say, sync with tilt lines or sync with curves:

 

 

 

You'll find that you can now uncheck your drive surfaces:

 

 

 

Hope this helps.

 

Cheers,

 

See learn something new everyday. Thanks for chiming in.

 

Question does the sync with curves work like entity? I assume the other sync method like to points is not supported though I have not used that process in years.

[Aaron Eberhard]

No, the syncing is by length.  50% of the top spline is synced to 50% of the bottom one, etc..

 

It's not even affected by node-point density, which is nice.  The only way to sync otherwise is to use tilt lines.

[crazy^millman]

No, the syncing is by length.  50% of the top spline is synced to 50% of the bottom one, etc..

 

It's not even affected by node-point density, which is nice.  The only way to sync otherwise is to use tilt lines.

 

Aaron will have to play with that when I get some time. You have the file and should be able to see where I did some of the toolpaths using old school swarf. Would you mind seeing if they work the same using the new school toolpaths.

 

I will see if I can explain what I was doing to help. Have some ribs that as .125 wide that have bosses stuck right down in the center of them. They have a .25R fillet in the corner. They change in height and go from a feather edge up to 1.25 tall. I used swarf to follow the many different surfaces the floor of the rib cam in contact with but one surface was flat and the boss was normal to that plate surface. At one of the ends I needed to sneak the tool into a tight area when a over hanging feature only gave .01 clearance to the shank of the tool, but bringing the tool normal to that surface would have slammed the head of the machine into the part so I have to keep cheating it short to be withing the .04 profile tolerance and do my very best to machine it, but not crash the machine. I think if you look at the results and run it the through the machine sim model I sent you you can see that we have clearance on the head and the .005 clearance of the over hanging feature with a longer than standard gauge length tool was accomplished. There was only 4 of them on the ID. We have over 40 of these to do on the OD, but what is cool is not over hanging feature so I think we should be able to know these 40 out in about the same amount of time it took me to do the 4. :crazy:

[crazy^millman]

Soon as I said I had not used points for SYNC I ran into a swarf toolpath that points proved to be the only thing that would work. Swarf shape that has to be not normal the floor for a boss then come up and sweep a radius and then normal to the top and then back down another wall hitting the other radius then to another boss not normal to the floor.

 

Here is a download to show you want I did.
https://www.dropbox.com/s/erqpnu0yonpm6nr/SWARF%20POINTS%20SYNC.mcx-9?dl=0

 

 

Edit/// I will just keep being me.

Edited January 16, 2016 by 5th Axis Consulting

[gcode]

the new 5X advanced swarf does a pretty good job of this too

 

https://www.dropbox.com/s/l2b0go8yv8d4s1g/SWARF%20POINTS%20SYNC_gcode.zip?dl=0

[Mick]

This is a great thread. I have to revisit 5x Swarf both old and new. I played around with both a while back, but haven't had a chance lately.

[crazy^millman]

the new 5X advanced swarf does a pretty good job of this too

 

https://www.dropbox.com/s/l2b0go8yv8d4s1g/SWARF%20POINTS%20SYNC_gcode.zip?dl=0

 

Gcode thanks for the file. I thought about this and the new swarf should fall under a new category. wireframe/surface/solid I noticed my NCI was 1.4k and your was 5k with 40 or 50 operations that is going to add up fast for file size and NC code. Old School Swarf like I mentioned earlier in the thread allows for filtering and on a 2000 operation part that is over a gig in size every little reduction in file size is important. 

[gcode]

Ron

I looked at this again this morning

Regarding filesize, we were comparing apples to oranges

In your old school toolpath you had the tolerance set to .001 and distance set to .05 yielding an NCI  of 1410K

I had my new school set to .0005 and with max distance unchecked  yielding an NCI of 4880K

 

If I set my new school path to your settings the NCI is reduced to 1917K

an added benefit is regen time

old school 9266 ms  = 10 seconds

new school  63ms = .06 seconds

[crazy^millman]

Another little trick. Say you have an upper and lower chains and are covering a bunch of different features. You can break it up into separate operations or you can combine. Now if you combing and have sharp corners you will need to use points to help you link I don't know any other way around it. Round corners not a problem, but sharp corners will need points from my experience. (yes always open to suggestions)

 

Another trick is to make the upper and lower chains into one spline for each. Use create/spine/curves and I like to keep my original curves and copy what I want to make splines with to another level. I then have my upper and lower chains as single entites. I use the create/point/segment I normally use 10 points and then grab all the chains that were the original ones, not the spline. (yes Mastercam could allow you to window all the entities you want to put points on verse the archaic method of having to pick them one by one). Now I had 14 operations in my file down to one operation. Very nice clean looking motion. 3/8 Ball endmill on a 5 Axis Head-Head running 24000 rpms and 360ipm. I wanted to run it a 27000 rpms and 405 ipm, but people were afraid.

[justchips]

Ron how did YOU create those points??? and would you mind explaining the theory for a young bull?

 

with new solid swarf the top corners normal to the floor? looked kinda wonky.