5-axis Swarf

[Justin Beebe at Folsom Tool]

I am making two similar parts that both require 5-axis swarf toolpaths. On one part the swarf toolpath runs very smoothly with a constant feed rate. On the other part the the swarf paths stalls almost coming to a stop at times then speeds up. All of the parameter settings are the same for both parts. Also the upper and lower rails for both parts are all splines and both are being posted with the same post and machine def. The code for the part that runs smooth has a single G93 before the cut. The other part that stalls and speeds up switches back and forth between G93 and G94. It slows down when it reads the G94 then speeds up when it reads G93. I believe it is switching between IMP and inverse time. I can't post the code on here until I get back to work on Monday but I am going to add some screenshots of the parts. Any Ideas why one toolpath would use a combination of G93 and G94 and how to prevent it from happening.

[Justin Beebe at Folsom Tool]

Here is the part that runs good. The rails are red along the top edge

[/img]

[Justin Beebe at Folsom Tool]

This is the part that stalls. It takes about four minutes to take two passes which should take less than a minute. Rails are also in red.

 

[/img]

[Bruce Caulley]

A couple of questions:

Are the rails lines and arcs, or splines? Did you try curve 5X? Have you tried using the surfaces with the new advanced 5X?

 

Personally I prefer curve 5X when things get messy, but the geo on your parts should be straight forward.

 

Bruce

[Justin Beebe at Folsom Tool]

Bruce,

 

The rails are all splines. I tried edit/simplify but I had to increase the tolerance to .002 in order to convert the geometry to arcs then the converted geometry was not connected so I could not chain it. I have not tried curve 5X or any of the advanced 5X toolpaths because I do not know how to use them. What surfaces should I select for advanced 5X. I would think the surfaces between the rails. Any advice would be appreciated.

[Justin Beebe at Folsom Tool]

Update,

 

I had to go into work this morning on a non related issue. While there I picked up my machine def and post for these jobs. I'll post the code later. I also tinkered around with curve-5X last night and came up with a toolpath that looks promising. I will post that code also.

[gcode]

Justin..

if you'd like to email me the file I'll take a look at it

[Justin Beebe at Folsom Tool]

gcode,

 

I can probably email you the splines that I used for the rails but I can't share the entire file. Let me know if this will be of any use to you for the purpose of helping me. I'm back at work on another non related issue. This has been some day off. When I get back home to my SIM I will post the code. I am posting these with a custom Fadal post from in house that has been bullet proof.

[gcode]

send me the splines, the surfaces betwen them

and your swarf toolpath I see what I can do

 

[ 11-27-2009, 07:32 PM: Message edited by: gcode ]

[Justin Beebe at Folsom Tool]

gcode,

 

I will send you the file when I take out the extra geometry. In the mean time here is a sample of the code that stalls out.

 

 

(T5 - 1/2 FLAT ENDMILL VIPER - H5 - D5 - D0.5000" )

N100 G0 G17 G20 G40 G80 G90

N110 (SWARF BREAKTHRU RIGHT PLANE)

N120 (TOOLPLANE NAME - RIGHT SIDE)

N130 T5 M06 ( 1/2 FLAT ENDMILL VIPER 2.0 LOC)

N140 G0 G17 G90 E1 B81.094 A-86.131

N150 G0 X3.0125 Y-9.9206 S3500 M3

N160 H5 Z6.6873 M8

N170 G94 Z1.0623 ----Runs slow after reading G94

N180 G1 Z.6873 F10.

N190 X2.9495 Y-9.9951 Z.7065 B-81.093 A-86.13 F40.

N200 X2.8865 Y-10.0696 Z.7256 B-81.092 A-86.129

N210 X2.8236 Y-10.1441 Z.7446

N220 X2.7606 Y-10.2187 Z.7634

N230 X2.6976 Y-10.2932 Z.7822 A-86.13

N240 X2.6346 Y-10.3677 Z.8009 B81.093

N250 X2.5716 Y-10.4423 Z.8195

N260 X2.5085 Y-10.5168 Z.8382

N270 X2.4455 Y-10.5914 Z.8568

N280 X2.3825 Y-10.666 Z.8754 B-81.092

N290 X2.3194 Y-10.7405 Z.8939 A-86.129

N300 X2.2564 Y-10.8151 Z.9123

N310 X2.1933 Y-10.8897 Z.9307 B-81.091

N320 X2.1303 Y-10.9643 Z.949

N330 X2.0672 Y-11.039 Z.9672

N340 X2.0041 Y-11.1136 Z.9853 B81.092

N350 X1.941 Y-11.1882 Z1.0034

N360 X1.878 Y-11.2629 Z1.0214

N370 X1.8149 Y-11.3375 Z1.0393 A-86.13

N380 X1.7518 Y-11.4122 Z1.0572 B81.093

N390 X1.6886 Y-11.4869 Z1.075

N400 X1.6255 Y-11.5615 Z1.0928

N410 X1.5624 Y-11.6362 Z1.1105

N420 X1.4993 Y-11.7109 Z1.1281 A-86.131

N430 X1.4361 Y-11.7856 Z1.1457 B81.094

N440 X1.373 Y-11.8603 Z1.1632

N450 X1.3098 Y-11.935 Z1.1807

N460 X1.2467 Y-12.0098 Z1.1981

N470 X1.1835 Y-12.0845 Z1.2155

N480 X1.1203 Y-12.1593 Z1.2328

N490 X1.1159 Y-12.1644 Z1.2331 B-81.093 A-86.13

N500 G93 X1.0345 Y-12.2376 Z1.2438 B-81.082 A-86.114 F363.41-----Runs fast after reading G93

N510 X.9314 Y-12.2926 Z1.2498 B81.092 A-86.124 F341.78

N520 X.8175 Y-12.3198 Z1.2557 B81.095 A-86.143 F341.21

N530 X.7068 Y-12.318 Z1.2604 B-81.094 A-86.133 F360.92

N540 X.6999 Y-12.3171 Z1.2603 A-86.131 F1000.

N550 G94 X.607 Y-12.303 Z1.2667 A-86.132 F40. -------Runs slow again

N560 X.514 Y-12.2888 Z1.2729

N570 X.4211 Y-12.2747 Z1.2788

N580 X.3281 Y-12.2606 Z1.2846 A-86.131

N590 X.2351 Y-12.2465 Z1.2901

N600 X.1421 Y-12.2323 Z1.2955

N610 X.0491 Y-12.2182 Z1.3006

N620 X-.0439 Y-12.2041 Z1.3056

N630 X-.137 Y-12.1899 Z1.3103

N640 X-.23 Y-12.1758 Z1.3148

N650 X-.3231 Y-12.1616 Z1.3192

N660 G93 X-.4162 Y-12.1475 Z1.3233 B81.095 A-86.128 F424.62 -----runs fast again

N670 X-.5092 Y-12.1334 Z1.3273 B-81.094 A-86.131 F424.91

N680 X-.6132 Y-12.1056 Z1.3234 B81.122 A-86.039 F370.69

N690 X-.7069 Y-12.0559 Z1.3094 B81.186 A-86.021 F370.8

N700 X-.7816 Y-11.9845 Z1.3382 B-81.026 A-86.194 F355.49

N710 X-.8498 Y-11.9093 Z1.3173 B81.094 A-86.131 F382.38

N720 G94 X-.9142 Y-11.8361 Z1.3098 F40.

N730 X-.9787 Y-11.7629 Z1.3022 A-86.132

N740 X-1.0431 Y-11.6897 Z1.2944

N750 X-1.1075 Y-11.6165 Z1.2865

N760 X-1.172 Y-11.5433 Z1.2785

N770 X-1.2364 Y-11.4702 Z1.2703 B-81.093

N780 X-1.3008 Y-11.397 Z1.2621

N790 X-1.3652 Y-11.3239 Z1.2537

N800 X-1.4296 Y-11.2507 Z1.2452 A-86.133

N810 X-1.494 Y-11.1776 Z1.2365

 

 

Notice how it switches between G94 and G93 during the swarf cut. The code on the other part that run faster and smoother has a single G93 before the swarf cut and no G94's.

[Justin Beebe at Folsom Tool]

gcode,

 

You got mail!

[gcode]

back at you..

[Justin Beebe at Folsom Tool]

Tom,

 

Thanks for all the info. You path looks really good in backplot. I also verified it using the solid of the part as the stock and it looks very accurate. Does that path exist in X2. We have X4 and we even had In House update this Fadal post for X4 but this was such a rush job I decided to go with X2 because I had similar parts already done in X2. As for the path I am using it is making good parts. They are being checked on a CMM. The path does not look as accurate as yours but the parts look good and the only close tolerance features on these parts are the bores. I am going to call In House monday and send them a copy of the code. I see what you mean, my post is posting G94 on certain 5-axis moves. When I get back to work I am going to post your path with my post see how it runs on the machine. I will check all the retracts.

 

Thanks again for all you help. I will keep you posted as to the outcome.

[gcode]

Justin.. advanced 5axis is available in

X2. If you recreate that path in X2 pay close attention to surface normals..

Advanced 5X uses surface normal to decide which side to comp to.. a backwards surface could get real ugly

[Justin Beebe at Folsom Tool]

Thanks for all good advice. I will make sure all the surface normals are correct. This is one I will actually proof out myself on the machine just to be on the safe side.

 

Thanks again

[Justin Beebe at Folsom Tool]

I just wanted to update this thread. My post was sent out to In House and the developer made some changes. It now outputs the correct code. Thanks In House you guys always do such a good job with customer service.

 

gcode,

 

Thanks again for all your help. I never got to try out your toolpath on this job but I look forward to trying it in the future. I am also planning on attending some advanced multi-axis classes at the start of the new year. I just need to check with my reseller to see what is available.

 

I appreciate all your help

[Jim at Gentex]

Justin,

 

One more point to make concerning this issue.

 

I have had the same problem you described here. Sometimes it is as simple as the number of node points on the spline. When I see a swarf toolpath that is 'stalling' or making weird movements, I usually end up deleting the original surface, recreating the upper and lower edge splines with a similar number of node points on each, then re-create the surface. It takes a while sometimes if the surface is big, but I have yet to see this method fail. I always end up with a smoother toolpath than I started with.

[MotorCityMinion]

If you need to break up splines that simplify can't handle, try drawing a line perpendicular, or close to perpendicular, through it in 2 places, break at intersection then delete the center of the spline. Now draw a line between the 2 remaining entities. You now have something that can be extented, trimmed, filleted, whatever. I use this method all the time when projecting lines for use as boundrys or blend lines.

[shazam/TPP]

+1 Jim

[Justin Beebe at Folsom Tool]

Thanks for the all the good advice.

[Colin Gilchrist]

You can also use Create - Spline - Curves Spline to rebuild an existing spline. If that still fails to give you a good spline, here is another trick:

 

Use Create - Point - Endpoints to create a point at the exact start and end of the spline. Then use Create - Point - Segment, and set the distance to .01. This will give you a ton of points that lie on your spline (but only if the spline is parametric, if not, you will need to convert to a parametric spline).

 

Now use delete duplicates to remove the duplicate point at the start of the spline.

 

Select all points, then change the point style to the little dot (makes it easier to see).

 

Now use Create - Spline - Automatic to create a new spline from these new points (I delete the old spline or move it to a new level first).

 

The Automatic Spline method will prompt you to select the first, next, and last point in the spline.

 

This will also give you a spline with a ton of node points. I then use the "RMNodes" Chook to remove the extra nodes in my new spline. I use a tolerance of .0002-.001, depending on the level of detail in the spline.

 

This is a surefire way to get a clean spline in Mastercam.

 

The other thing I recommend to everybody is to use the "WholeSpline" chook whenever you break or trim a spline. It helps, especially if you are getting weird tangency conditions at the end of your spline...

[Justin Beebe at Folsom Tool]

Colin,

 

Thanks for all the info. I'll have to print this out and play around with it.

[Colin Gilchrist]

Any time Justin. Hopefully the ramblings I spew forth will help someone.

[danielm]

You da man, Colin!

 

Good to see you here....wouldn't be the same w/o you.

[gcode]

quote:

---

Hopefully the ramblings I spew forth will help someone.

---

rambling??? hardly... gems mined in the

school of hard knocks is more like it...

good stuff