[az_356]

Having some problems going from Dolphin to Mach. Doing a simple pocket with a 0.575" radius and get the "Radius to end of arc differs from radius to start on line 12" error in Mach. I did the exact same part earlier today but with a 1.15" radius and no problems. Checked Mach to be sure it's set to absolute, I have re-drawn the part 10 different ways in Autocad and Dolphin. Everything seems OK but I can't get this very simple part to run. Any thoughts? Here's the first few lines of code Dolphin is creating.

( Produced :- 17:09:44 Monday, September 26, 2011 )
( CNC File :- nojig )
( Post Processor :- M_MACH3 )
( Part Number ID :- )
N5G00G20G17G90G40G49G80
N6G49
N7T1M06 ( End Mill )
N8G00G43Z1.9685H1
N9S3500M03
N10G94
N11Z0.1181
N12G01Z-0.1333F3.937
N13G03X-0.0752Y0.0752I-0.0752J0.0F3.25

[ftech]

I think I see what the problem is, I ran into this when I first started to use Dolphin.
If you look at the code you attached, you will see that there is no move to any X or Y position before it tries to do a CCW arc. Unless you start the program run at exactly the right position, it will not be able to start the arc properly.
I found that Dolphin does this when you re-open a file and make a change or 2 and immediately post process. It omits the initial X,Y move. However, if you run a simulation first and then do the post, it will include the initial move and you will be OK.

Hope that helps,
Rick

[AiR_GuNNeR]

I've seen this with other CAM systems as well. If you are doing a helical cut, you have to set your CAM to NOT use arcs. Use line segments instead. I typically follow up with a profile operation to clean up.

[az_356]

Thanks guys. Dolphinusa sent me the new post processor, and while introducing a new minor problem, it has cured the first. I will try out these sollutions moving forward.

[judleroy]

I'm not sure about dolphin so don't read to much into what I'm saying. I have had this problem in several programs. What I've found helped with this type of error is to change the accuracy in your system settings to .0001. It eliminates issues with the way machines round numbers in there systems.
Judleroy

[andre-dolphin]

On a general note - most post-processors should be configured to output different number of decimal places for metric and inch - normally 3 metric and 4 inch.

There are some posts that don't have this, if you want to check, in CAM select Edit > Post processor, when you see the post scroll down to the section WORDS.

You should see something like this.

:XAXIS = {" X"DDDD.ddd} {" X"DDD.dddd}

The first set of letters refer to metric - 4 place holders for whole numbers DDDD and three places holders for decimals ddd. The second set are for inch.

If you post doesn't have these entries you can edit the post and add them.

The syntax is important so please be careful.

Hope this helps

ATB
Andre

[harley4ever]

What new problems occurred? Is it fixed?

[ftech]

Does anyone still read G-code?
Although it is surprising how quickly these threads get distracted from the original topic, it is even more surprising that no one else seems to have paid any attention to the g-code that was originally posted. It is pretty obvious to anyone who reads basic g-code that the original poster’s code does not contain any X or Y moves before the initial G03 circular interpolation. Obviously, the specified curve will only work if the program is started at EXACTLY the right place to begin with. ANY other position will generate the error sited. I find it hard to believe that no one else picked up on this. My first post explained why this happens in Part Master.