jerky movement plasma table

[kawarider]

Hello,
Recently Neuron Lite installed on plasmatable with Uccnc 300eth, Hypertherm powermax 45. Runs great, super features in the neuron screenset. Top! The only thing is the jerky movement with small circles, at least I presume it is. If I run sheetcam without offset it runs great, with the kerfwith is stops at each segment. I can't find a solution that helps. According to Andrew the problem is in Uccnc. So anyone suggestions?

Here is a bad quality movie that shows the problem

https://www.youtube.com/watch?v=JJaxfJTWYos

Thanks, machine runs great at the moment. Nothing wrong with the cut product, only the jerky movement is not how it supposed to be.
Again, thanks in advance

[kawarider]

Around line 970 there is the line from the youtube.

[kawarider]

Ok, let's get it up again. Nobody? I tried several settings in UCCNC, but nothing satisfying. If I run this programm without offset, no problem. I thing it is in the small radius in the corners and straight lines from sheetcam. On the youtube you see the bridge stopping and causing a lot of stress/noice. It would be nice to have this solved. UCCNC is running great, it is only in this small corners/radius.

[Robertspark]

What is your feedrate for the file?

What is your corner error max setting in uccnc?

Are you running in mm, inches?

[kawarider]

Robert, thx for your reply.

feedrate 3800 mm/min

corner error 0.03

running mm

[Robertspark]

Can I suggest that 0.03mm corner error max is just a little high in my opinion for a plasma table + one that is intending to run at a constant velocity of 3800mm/min (63.33mm per second).

I don't know what kerf you are running in plasma, but say 1mm to 1.5mm. hence in my opinion you should NEVER be running the corner error max at anything less that 1/2 the kerf given that will always be your tolerance MINIMUM.

In reality.... with plasma.... in my opinion and my approach .... I'm happy with a few mm tolerance at the corners.... given I weld most things to something else and I know how to mig and tig just fine and can use an angle and die grinder if I really need anything trimmed to shape.

So this would put your Corner Error Max at say 0.5 to 0.75mm (MIN) at a feedrate of 3800mm/min

You don't HAVE to change this in the settings within UCCNC, you can use "G64 Q" to change this setting on the fly.

I personally change G64 Q settings using the post processor automatically where I consider the feedrate that I'm asking for as well as my machines XY acceleration and also the kerf of the cut intended

(I don't know your machines XY accelerartion or I would have used your figures....)

My plasma cutter is capable of an acceleration of 3000mm/s/s

So at 3800mm/min (63.333 mm/sec), the minimum arc that a machine with 3000mm/s/s would be 66.333^2 / 3000 = 1.337mm (radius).... before the machine would have to slow down.

Generally in plasma with small arcs a 60% feedrate drop is acceptable..... so actually if I allow the feedrate to drop to 60% (3800 * 0.6) = 2280mm/min (38mm/sec), this would mean that the the minimum arc would be 38^2 / 3000 = 0.481mm.... which would be more than enough for me and my tolerance of say 2-3mm at corners on the thin stuff that I do.

https://www.khanacademy.org/science/phy ... celeration

Another way to look at this..... if I set my corner error max at 0.03mm... with the same machine acceleration 3000mm/s/s

SQRT (0.03 x 3000) = 9.49mm/sec (570mm/min)

my machine would slow down to a feedrate of 570mm/min at every junction in order to try to maintain such a close tolerance on the corners as 0.03mm.... which with plasma will be lost because the torch will overburn the corners because it will have to slow down so much that the kerf will get wider.

sometimes it is worthwhile to do some straight cuts to determine your ACTUAL kerf at say 3800mm/min if that is your desired feedrate for a particular material you cut a lot of .... then slow this feedrate down to 60% and do the same thing (if you have THC the voltage will rise (as the kerf will get wider) .... so a THC that can do SAMPLE + HOLD is better.... that way you KNOW you are maintaining the correct cut height which is what you will do in the corners when you disable THC as the feedrate drops.

an acceleration of 3000mm/s/s is not the norm for plasma (as it depends upon your drive mechanism) but it is a target

I've seen quite a lot of machines at about 300mm/s/s acceleration.... for this at the same feedrate (3800mm/min, the minimum arc they could cut WITHOUT slowing down) would be 13.37mm RADIUS..... at 60% of this feedrate the minimum arc would be: 4.8mm RADIUS.
If the machine was set to a corner error max of 0.03mm the machine would HAVE to slow down to a feedrate of 180mm/min to maintain that tolerance due to the machines slower possible acceleration.

[kawarider]

Thanks again for replying. I neede some time to think this over. Just did some testing again. No good. It stays the same, no mather what kind of error I give in. He just stops and than moves on. 90 degree corners are no ploblem, but these small corners? He actual stops and moves. CV mode in on. Acc is here 3000 mm/s2 also. If bridge moves in y direction an has to taken small circle to the angle line, it stops. From angle to angle it is a little better. I think that should be more difficult. I scratch my head on this one. What is going on?

[tõnisson]

In your g-code, there are very small radius moves, between lines.
Try this..

In sheetcam, Options - Machine - Post processor - Edit post

There is a line - "minArcSize = 0.05 --arcs smaller than this are converted to moves"

Change that 0.05 to a bigger number, 0.75 or 1 or something for testing.

Thank me later

[Robertspark]

tonisson that is a GOOD point..... I'll be very impressed if you can cut a 0.05mm arc with a plasma cutter.... (kind of like performing surgery with a hammer!)

[kawarider]

Well, that is new! Thanks for that already, and again if things are fixed. There is some backlash too, but I don't think so. Try this first. At the moment cutting some, after that this is next. Thanks already. It has some to do with the radius. I tried without kerf comp and no problem. So this is it, probably. To be continuid