Vector cutting paper on HPC ls 3020 (like k40)

[jamesd256]

Hi,

Recently been trying to cut out some intricate designs via importing DXF into Newlydraw.

What I've found is that the internal shape cut out very often hangs on by a mini 'tab', because for some reason, the start and end of the cut didn't end up joining.

Is this a common problem? I have exactly the same issue when cutting vinyl sheet on my CNC router (using one of the Chinese bullet shaped drag knives)

Is there a way to 'overcut' paths?

Cheers

Edit: I forgot to say, the focus and power levels are close to correct at least, because the cut comes out clean and consistent. It has crossed my mind to increase power to see if it fixes, but I would rather not cut the whole piece hotter than I have to to get around this.

Similar Threads:

• Paper Cutting pages of a book
• Newbie- Paper cutting
• Cutting Paper

[Tweakie]

The vinyl dragknife issue is because the cutting tip lags behind the centre of rotation (this is necessary to get the knife to 'steer').

If your laser is slow to turn ON (even a small delay) then there will be a gap (if the start and end positions are intended to meet). Perhaps try a test of drawing two identical parallel lines each starting from different ends - do the ends line up ?

Tweakie.

[jamesd256]

Thanks Tweakie, you've correctly identified the issue with my laser cutting. I performed the test as you mentioned, and there is about 0.15mm difference between the start and end of a 100mm line at 50mm/s.

As for the drag knife, what you say also makes sense there.

The question remains, is there some way to over cut paths? Is there any other solution apart from slowing down the cut, or slightly de-focusing the beam (might char a bit)?

I'm looking at some pieces where there are hundreds of cuts in an a4 sheet, and I've found it hard to 'weed' the cut outs without creasing. Not to mention the boredom factor!

Appreciate your help.

[Blackened]

If this is only a recent problem, you might want to hunt down the cause of the slop. Maybe adjust belts, tighten things up a bit. It's only going to get worse if you don't address it IMHO. Otherwise if it's a delay in the power turning on before motion begins, and if your machine uses G-code, a short 50-100 ms dwell after the power on command might help.

[jamesd256]

Thanks Blackened.

No, this is a new acquisition, so I don't know if it's slop, or if it's getting worse.

These machines use a crappy proprietary board, so I can't do gcode.

A planned upgrade to a proper controller board is around the corner, and a gcode solution seems like a good suggestion that can be tweaked to fit different problems.

[Blackened]

I'm building a laser from scratch ATM. One of my must-haves was to run on G-code, or at least the option. I use industrial machines daily and editing G-code is something I do to greater or lesser extent with most jobs. It's invaluable to know. If your post-processor is spitting out some poor code you can fix it. Good luck with your machine! Who knows, maybe it's just a loose screw and you'll be up and running without tabs on the scrap.

[Tweakie]

Certainly the first and easiest thing to do is check all the mechanicals. As Blackened has suggested it may just be something as simple as a loose grubscrew on the motor pulley / coupling etc..

Tweakie.

[jamesd256]

I detect wise words Tweakie, from both of you, thanks again.

I had tested for slop a bit - engage steppers and try to move head wrt to x and y on gantry. Not much slop detected - proper mechanical deep dive next I thought.

Then I remembered from before slop resulted in skewed circles on another machine, so I cut out 3 lines of circles; 3mm, 2mm and 1mm. Sure enough the 3mm ones were rough, 2mm ones a funny shape and the 1mm ones were eye shaped.

I noticed the movement sounded very clunky on the 1mm circles, so I tried halving it from 100mm to 50mm and reducing the laser power accordingly. Result - beautiful circles. I am learning the limits of my new machine!

Oh, and no more tabs, so with your help problem solved. Thanks again.