Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: John Watkins on April 13, 2009, 11:37:29 AM
-
Hello everyone. This is my first post so thanks in advance for helping out the new guy.
I have a problem which I believe I have narrowed down to Mach III mis-interpreting the g-code I give it. I get a grooved pattern on the preview screen that is exactly duplicated when the part is cut. No change in feedrate or depth of cut has any affect. The gouges are always in the exact same place.
First instinct says that the g-code is based on bad geometry, but this whole curve is only three lines of code, two G02s and one G03. Any ideas? I get perfect cuts if I go diagonally or straight. It only does this on curves and it does the exact same thing if I use all G01s. Any help is greatly appeciated.
Here's how the curve looks in my CAM software's simulation screen.....
(http://cncguitarparts.com/images/visualturn_screen.jpg)
and here's what it looks like in Mach III. I would normally just attribute this to imperfect graphics, but the parts are very similar to what's on the screen.
(http://cncguitarparts.com/images/mach_screen.jpg)
-
Can you attach the code and also the xml for the turn profile you use.
Hood
-
Hood, thanks for replying. Here's the relevant code;
G18
G2X0.3378Z1.0790K-0.1328I0.4212
G3X0.4566Z1.2939K0.2208I-0.1455
Not sure what you mean by xml for the turn profile. I created the profile in Rhino, created the tool path in VisualTurn, and posted to Mach III
-
The xml has all of your settings, if you are using the standard turn profile then it will be found in your Mach3 folder and it will be called Mach3Turn.xml. You will need to copy it to a different location on your drive and then rename it as the forum will only accept a file being posted once. Suggest you rename it JohnWatkins.xml
To attach it to the forum use the additional options button on the reply page and browse to the location of the renamed xml.
Hood
-
a picture is worth a thousand words.....
(http://cncguitarparts.com/images/rough_surface.jpg)
-
going to get the xml file now. thanks again...
-
Also could you attach the actual G Code file as well please.
Hood
-
trying this again. not sure what happened.
-
Dont see any real issues with your profile and code looks good. I am not sure what type of motors and drives you have but you may want to try increasing the pulse width in motot tuning, try it at 3 or even 5.
Hood
-
Going to do that right now. I'll let you know what happens. Thanks again for taking the time. I greatly appreciate it.
-
Same result. Maybe a little less pronounced. Any other ideas? I've gone as far as to move the job around within the travel area to see if it was a blem in one of the screws. Nothing seems to change it. One IPM or 40 IPM; same thing.
Here it is with the step and dir pulse set to 5 on both axis.
(http://www.cncguitarparts.com/images/2nd.jpg)
-
Try setting the active states to high on each axis, it probably wont help but is worth a shot.
How tight/rigid is the machine/tooling?
What kind of motors and drives?
Hood
-
How do you get it to travel in the right direction after doing that?
The machine is a harbor freight retrofit using some 90v NEMA 34s and a camtronics control box. The Z screw is a known-good four-start screw with a very tight polymer anti-backlash nut. The X-screw is the factory screw held tight with the coupler to the motor. I suspected the X screw right away, but anything it's doing that would cause this would also show up in a diagonal/taper cut, right?
-
Homing and Limits, reverse axis.
Your pics show an interrupted cut, is that the shape of the stock you are turning? Was the diagonal on a round or similar to that (if it is as shown in pic :) )
Not familiar with Camtronics stuff, will have a search around in a while.
Hood
-
Here are a couple of pics of the setup.
-
If you can try a piece that will have the full 360 degrees getting turned.
Hood
-
I have done that in the past with the same result. I don't mind trying again, but I'm not really able to do it now because all the material for this batch has already been through the milling operations that give you what you see in the pics. I may have to wait until the weekend to troubleshoot this further because we have a deadline for those parts. We'll just run them this way and sand them for now, but if anyone has another idea that I could try when I'm able, I'm all ears.
Can't thank you enough for your help.
John
-
Hello John
Your setup is very prone to vibration, part is far away from the chuck,tailstock could be shortened.
is the tool at the right height
are the tailstock and the shaft that holds the part center .
Just a tought
Marcel
-
Check that the encoder is not loose on the shaft, you could also try swapping the x & z motor encoders over.
Graham
-
Hi John,
May be all wet but here are a few things you may want to try.
I echo what Marcel posted.
Try leaving a little more stock to be left in the rough cuts and do more than one finishing cut.
You may also try a different tool tip, one that shears / peals off the wood instead of scraping. Similar to doing the turning manualy on a wood lathe. High clearance and rake, razor sharp, depends on the wood. Also can try doing two at a time by glueing two halfs together with a piece of paper between and splitting them thus you get continuity of the cut around the profile.
Like i said, may be all wet, just some thoughts, been a long time since i did wood turning.
RICH
-
I don't think the setup is weak. The machine is of steel and he is turning wood.
Back to the first post. These imperfections is clearly visible in the Mach3 toolpath. This fact is a strong indication it is nothing wrong with the workpiece setup but that the fault is within Mach3 or its setup or the code or anyting other influencing how Mach3 interprets the G-code.
-
Could it be a VERY coarse "steps per" resolution ?
JAT,
RC
-
One thing you can try ,is taking a free cut or a cut of .005 if the profile is still not smooth then he as other problems . As for the machine is made of steel it does not matter
deflection is one of the biggest problem of the machinist world .
-
I don't think the setup is weak. The machine is of steel and he is turning wood.
Back to the first post. These imperfections is clearly visible in the Mach3 toolpath. This fact is a strong indication it is nothing wrong with the workpiece setup but that the fault is within Mach3 or its setup or the code or anyting other influencing how Mach3 interprets the G-code.
He is turning with interupted cuts which can induce deflection and resonance, whether that be in the machine, the tool or more likely in this case the stock itself.
The imperfections in the toolpath in Mach are due to the graphics and nothing to do with what will actually cut.
Hood
-
He is turning with interupted cuts
Hood,
Exact same results with un-interrupted cuts.
Reply #15-16
RC
-
RC
I saw that reply but the OP is going to test again so we will see.
It would be interesting to see a line drawn on a board with a pen mounted vertically in the toolpost.
Hood
-
Yes..it would.
Also, to do the end of a piece of alu or hard plastic directly in the ghuck. As ridgid as possible.
Thanks,
RC
-
John,
Could you please post a dxf file of the profle.
Do you realize that that your cutting air with "part" of the finish cut so that your left with marks in the
profile. I BACKPLOTTED your code and it is not MACH but your code as created.
See attached.
RICH
-
Hi All,
HM......... I just tried that code again and got a different display........ ??? ???
So maybe just ignore reply #27 for now! :-\
RICH
-
Rich,
I tried it too and got what you did initially.
What I noticed though, was that even in the area that WAS cut with the finish pass, there are still many grooves/humps.
RC
-
Hi All,
Can someone else try John's code and see what it looks like in MACHTurn?
SEE ATTACHED as i never came across this one. ??? So maybe i am not going bonkers! ;)
RICH
-
RC,
Maybe " WHAT TAUGHT THOUGHT HE SAW ONLY THOUGHT HE SAW" ???
Hate to shed doubt!
RICH
-
Rich,
I can vouch for you.....you're NUTS ! :D
No really... I get exactly as you with RAD. and DIA. settings.
RC :)
-
HM....
You should try it with the SS Plug in!
I can deal with the "humps and bumps" but if the display is screwy your not sure what
your dealing with without turning it. I don't even trust my backplotting on this one.
RICH
-
Something else....None of our displays show the zig-zag tool path that John is showing with his rough AND finish passes.
-
Maybe Graham can comment on the code as there are G18 G17 ?? then also about the G70.
I never used the G70 and my book say's it a Fanuc code??
RICH
-
RC,
The roughs are straight lines even though they look rather smooth.
When i backplotted the file i got small segments instead of a smooth curve for the finish cut.
Still would like a DXF from John.
What Mach Version are you using? I using 3.042.024 and the standard screen set.
RICH
-
R.020 and std lset
RC