G code start problem , and g code optimisation.

[kefeur]

When it starts the burning the M3 instruction is omitted in the first line. At the end of each line M5 switches off, then when it starts coming back M3 switches back on. Consecutively, both of them are unnecessary.

[kefeur]

G code Optimisation: In an example of a 50mm*50mm picture that has a 20mm wide figure in the middle and has 15mm of empty space the laser travels that empty space as well but does nothing. It would be better if it would skip these empty spaces and empty lines entirely.

White skipped the purpose / work transcends time and unnecessary abbreviations Movement It would be good if it would skip the white parts during the work process to shorten the time and prevent useless movement. Thank you.

[arkypita]

When it starts the burning the M3 instruction is omitted in the first line.

If so, it's a bug. I will investigate as soon as possible.

At the end of each line M5 switches off, then when it starts coming back M3 switches back on. Consecutively, both of them are unnecessary.

This is necessary and by design.

Take a look at this image, made line by line by horizontal strokes. The horizontal line must have the laser on, the vertical line exchange must ensure that the laser is switched off. Otherwise it would be marked a vertical line to the left and right edges of the image.

It's true that in cases where at the end of the line laser is already off (S0) this M5-M3 seems superfluous, but it is not true in all other cases.

[arkypita]

In an example of a 50mm*50mm picture that has a 20mm wide figure in the middle and has 15mm of empty space the laser travels that empty space as well but does nothing. It would be better if it would skip these empty spaces and empty lines entirely.

White skipped the purpose / work transcends time and unnecessary abbreviations Movement It would be good if it would skip the white parts during the work process to shorten the time and prevent useless movement.

Yes, a lot of improvements can be made in this direction. All leading / trailing white lines may be skipped. Also avoid arriving at the end of line but move directly to the next line. Something like this:

In the past I tried to insert optimizations such these, but the line generation algorithm became too complicated so i abandoned this way.

[kefeur]

It's true that in cases where at the end of the line laser is already off (S0) this M5-M3 seems superfluous, but it is not true in all other cases

Many other g code generator did not use the M3-M5 is anything but start and end. Use S0 instruction .The problem to quickly turn off and then turn the engine driver, and because the micro step is clapping, jumping and noisy. sometimes jumps one step it is a lot . (1 step = 64 micro step) Im Isolating the engine driver enable input from the GRBL.

[kefeur]

M3 error. G code first line:

G90 F400 G0 X0 Y0 M5 S255 G1 F350 G91 G0 F400 X2.6 S0 G1 F350 X.1 S255 G0 F400 X2.1 S0 G1 F350 X.1 S255 G0 F400 X.5 S0 G1 F350 X.1 S255 G0 F400 X2.3 S0

This is correct:

G90 F400 G0 X0 Y0 M5 S255 G1 F350 G91 M3<<<<<<It can not generate it G0 F400 X2.6 S0 G1 F350 X.1 S255 G0 F400 X2.1 S0 G1 F350 X.1 S255 G0 F400 X.5 S0 G1 F350 X.1 S255 G0 F400 X2.3 S0

[arkypita]

If you are able to checkout sourcecode and compile, the last commit has a fix for this bug. N.B. I have not had a chance to test it because I do not have laser engraver available at this time

[arkypita]

GCode optimization was added in development roadmap (when I have time...)

[kefeur]

Very clever man. Thank you.

[arkypita]

Last commit contains GCode optimization (skip consecutive blank lines and move directly to next mark)

[arkypita]

About the problem you tell me about dithering, of sparse small dots

It not depends on a bug in the dithering alghoritm but it depends by the source image. If the source image has not a perfect white background (ie. for jpg compression artefacts) dithering try to represent this non pure white areas/points with small scattered dots.

If the image has a true white background this problem disappear

Maybe I could add some kind of "white thresholds" to clean dirty background in image preprocessing, for now I suggest you to preprocess images with Photoshop or similar to clean background.

[kefeur]

The G-code optimization has been very good. This is a very nicely developing project. Thank you.