Hello Guest it is July 18, 2024, 10:56:05 AM

Author Topic: Radius of arc differs - but a different sort of problem  (Read 1128 times)

0 Members and 1 Guest are viewing this topic.

Radius of arc differs - but a different sort of problem
« on: December 28, 2020, 06:53:07 PM »
I'm getting the dreaded "radius of arc differs" error, but with a twist.
My code begins like this:
Code: [Select]
( Made using CamBam - http://www.cambam.co.uk )
( Bottom3 12/28/2020 12:39:25 PM )
( Post processor: Mach3 )
( T4 : 0.236 )
G20 G90 G64 G40

The program errors out in the simulator on the marked line:

Code: [Select]
(many lines of previous pocketing operation omitted)
G0 Z0.75
G0 X8.2874 Y0.8916
G1 F25.0 Z0.375
G1 F75.0 X8.2708 Y0.8244
G2 X8.2707 Y0.8242 I-0.1213 J0.03
G1 X8.2018 Y0.5483
G2 X8.2016 Y0.5475 I-0.1213 J0.0303
G1 X8.1167 Y0.2176
G2 X7.9737 Y0.1257 I-0.1211 J0.0312
G1 X7.3166 Y0.2429
G2 X7.3149 Y0.2432 I0.022 J0.1231
G1 X6.6596 Y0.3695
G2 X6.6579 Y0.3699 I0.0237 J0.1227
G1 X6.0043 Y0.5053
G2 X6.0026 Y0.5057 I0.0254 J0.1224  <<------------error here
G1 X5.351 Y0.6501
G2 X5.3493 Y0.6505 I0.0271 J0.122
G1 X4.6998 Y0.804
(more code follows ...)

So, it's incremental mode and the website says to use absolute.
OK, so let's change the G91.1 to G90.1
Now things get weird. I still get the radius differs error message. But I also get incorrect tool paths in the preview. Instead of drawing/cutting 10-15° of arc, it draws the compliment - 350° or so.

Since debugging this on the laptop in my office is more comfortable than standing over the machine in the shop, I copy the g-code file and the Mach3 XML configuration file and bring them to my office.

I launch Mach3 using the config file from the shop computer and run the g-code. No error with either G90.1 or G91.1. Same funky toolpath with the G90.1 setting.

One more oddity:
If I launch Mach3 on the shop computer and don't hit reset, then load the g-code and run simulate, I don't get the error. But Mach3 launches in metric units and my initialization string - G80 G20 F1905 S1000 - isn't being executed so the DROs continue to appear in metric units in spite of the G20 at the start of the program. Even if I return to the Program Run tab, reset does not set the DROs to inch.

Both computers are running 3.043.066, and have been for several years with no problems like this.
I've been running CamBam for a while, as well, and have successfully run quite a few programs of comparable complexity (or worse).
The main difference is that one computer is actually connected to a controller.

I'm at a loss where to look next. Suggestions (preferably other than reinstall)?

Offline Tweakie.CNC

  • *
  •  9,214 9,214
  • Super Kitty
Re: Radius of arc differs - but a different sort of problem
« Reply #1 on: December 29, 2020, 07:01:12 AM »
Sorry but my money is on Re-Install - preferably v.062 (download from Artsoft site) rather than v.066.

Re: Radius of arc differs - but a different sort of problem
« Reply #2 on: December 29, 2020, 05:17:16 PM »
the weirdness continues.
I moved the laptop to my office, disabled the plug-in for USB connected controller (Xulifeng). The program runs just fine.
Took the laptop back to the shop, connected the controller (which re-enables the controller plug-in), I again get the error message.
With the controller plugged-in, I deactivate the plug-in (meaning the controller is not visible), no error.
Re-activate the controller plug-in - error.

Why would the presence of the controller cause this error?

Offline Graham Waterworth

  • *
  •  2,706 2,706
  • Yorkshire Dales, England
Re: Radius of arc differs - but a different sort of problem
« Reply #3 on: December 30, 2020, 09:02:37 PM »
I assume you have a genuine licence file and you are not using one that came with the board.  You seem to have been having problems with this since 06-06-2015.

When the controller is connected it is creating the moves from commands sent from mach, when not Mach is calculating the profile.

« Last Edit: December 30, 2020, 09:08:43 PM by Graham Waterworth »
Without engineers the world stops
Re: Radius of arc differs - but a different sort of problem
« Reply #4 on: December 30, 2020, 10:54:19 PM »
yes, a genuine license file once I became aware that the one sent with the machine was bogus. (I wasn't terribly surprised to discover that).
Yes, I had some problems early on, but that was five years ago. It's been working smoothly for quite a while (with a few hiccups). But this is new. I understand the relationship between Mach and the controller board, that's why it's so baffling that Mach can simulate the cut with the board disconnected, but not with it connected. After all, in simulation mode it's not communicating with the controller.