Are defined feed rates of any use?

Micromet wrote: It seems the only stand-alone linuxcnc simulator package is for Ubuntu distros..

You don't necessarily need a dedicated "simulator". You can use the new-ish "uspace" version of LinuxCNC.

I actually use an RTAI config in a VM on my Mac for simulations. The realtime performance is awful, but I just ignore the warnings.

I use VNC to remote into the computer on my mill. Then I can run simulations from in my office without going out to the actual machine. I have a sim config that is set up with the same machine limits, max speeds, accelerations as the real config for my mill.

Hi Andy and nitromethane

Just tried to install the uspace version from the git repository using linuxcnc-uspace rather than linuxcnc-dev - seemed to go ok - downloaded the files and exited normally . Went into the debian directory as required and then tried all the stated ./configure modes but only the ./configure uspace option seemed to work partially with the following
unknown distribution: Mageia-4
configuration may be wrong!
successfully configured for 'uspace-Mageia-4'-'uspace'.
but the make failed with "make: *** No targets specified and no makefile found. Stop."

tried ./configure -a but no rt kernels found
so need to install an rt kernel for my Mageia machine before trying again.

But I could follow nitromethane's idea and VNC to my workshop machine - bit of an inconvenience having to go and turn on the workshop machine and then back to the office - but if that is what I have to do - then so-be-it.

Thanks again to both.

Micromet wrote: bit of an inconvenience having to go and turn on the workshop machine and then back to the office.

My lathe is configured for Wake-on-LAN.

(Then I can ssh -Y into it from my Mac and run a remote X-session directly on my desktop)

Good idea Andy - never thought of that.

Hi Andy,

.SVG files are human-readable, so have a look in there.

Had a look in the SVG file - not easy to see what is going on after the start of the "<path" description - the monster series of number pairs went on for ever where I was expecting a series of bezier curve definitions - these appeared to be just a series of coordinate pairs. But when I imported them into a spreadsheet to see if I could graph them in any meaningful way - the number pairs just seemed to be random.

However, I've looked at the potrace FAQ site by Peter Selinger (potrace's author) and it seems I might have been trying to be too clever by making the bitmap image with too high a resolution. He recommends - although at times his advice seems contradictory that in answer to the question

How can I improve the quality of the output of Potrace?

:

The most effective way to do so is to improve the quality of the input! This seems obvious, but not everybody thinks of this. Scan your image at a higher resolution. Scan your image in greyscale, and use a program such as mkbitmap to generate a high-resolution bitmap to trace.
In some cases, it also helps to decrease the resolution of the input image. If the resolution of your input file is incredibly high, the output generated by Potrace might be larger than necessary. This also sometimes happens if the input is speckly or noisy. Rather than trying to simplify the vector image, it often helps to just downscale the input bitmap.

But then again in answer to the question:

How can I reduce the number of nodes in the output of Potrace?

For many applications, the problem of reducing output file size can be solved by reducing the resolution of the input image. This sounds too simple to be true, but can give surprisingly good results. One of the most common "beginner's mistakes" when using Potrace is to start with an image that contains too much detail or noise. Blurring the input image can also help, particularly with grey-scale images.

So I'm going to follow up this route and see if reducing the resolution of my bitmap image makes the complicated gcodetools curves go away