[arbuzas]

Polling my hair here bit with this issue. My Chinese KH-5030 laser is having hard time cutting circles (not closing them or sometimes closing off from the start point). Maybe someone had similar issue? I toyed with the settings in RDWorks, which gave me mixed results but non of them perfect.

[Storen]

RDWorks is fine. It is a hardware issue and it is quite common. Unfortunately this is what the cheap Chinese machines do.

Maybe try using overlapping lines with different starting points.

[mysticusa]

it is i believe an alignment issue on mirror/laser path if it is consistent which seems like consistent in one axis... check all mirrors and specifically check your laser's second mirror alignment to third mirror (head unit) and hear unit's mirror down to work table... even half a mm off horizontal on any x/y axis does change the behavior of laser's path (hence the distances it travels to cut... think of arrow/bow shooting... assume you have no gravity... if you aim bulls eye, you hit bullseye, but if your aim is 1degree off, and further it is the target, further away from the bullseye arrow will hit... if you understand this in layman terms, basically it is simple trigonometric... you gotta be perfectly horizontal in x/y axis for your lines to finish in the same spot... in vertical axis (head unit to table) assuming your table is consistent and level, if laser is not 90 degree vertical, your edges will be angular even in perfect horizontal alignment... at will get worst with thicker materials... but your current issue is mirror alignment...if it was locational problem, it would be most likely rails where head unit moves in x/y axis...

see this pic


if you have problem in Y axis or X axis, and further the head moves away, further away you sway from your target! also more the misalignment the chances of laser clipping from edges of mirrors or head unit, which will reduce power at impact point on your target due to less laser beams hitting the target...

[Labmaster]

Beside the fakt, that a missalignment causes various problems and should be avoided, in this case it could not be the case for such problem

In this case, where two points that should be next to each other are not (point where the circle starts, point where the circle ends) its a dynamical problem where in the machine something has changed during the time of the process between first and second point (during the circle).
If it would be a statical problem, like a missalignment or geometry error, the two points next to each other would still be next to each other even if the circle is not a circle or the cut angles are wrong.

So it is a dynamical problem.
This could be loose belts (Backlash problem), other loose or flexible part, linear mechanics not moving freely or it could be loosing steps. (loosing steps is not very likely in this case as it normaly would show this on other parts of the circle)

As here the Y-axis is driven in opposit directions when aproching the start and the end point, i would say it could be the belt tension or in worst case the linear rails are bad or wrong mounted (not leveled to each other) and so they are not moving free enought.
So i have to agree with "Storen" this is a typical problem of cheap asia machines, as the build quality of the mechanicas is sometimes very bad.
The good thing is, with some investigation and worktime it could be solved.

On my 500x300 Chinese machine i had simmilar problems.
When i removed all the belts and tried to move the y axis manually i found out, that this took a hugh amount of force to overcome the first moment. It was like the slides where sticking on the rails. The whole linear movement was horrible and needed a lot of force. After diggin deeper and removing the y axis, i found out, that the sliders itself will move freely but as soon as i mounted the y axis back, they stuck.
A angle measurement showed me then, that both y-axis-rails where not mounted leveled to each other. The left rails was tilted to the left, the right rails was tilted to the right. When the y-axis itself was mounted back then on top of the sliders, both sliders got forced to stay leveled (when screwes to the y axis was tighend up) where they could not as the rails are not leveled and so they started to stuck.

best regards,
Walter

[arbuzas]

I spent good couple days aligning beam, that was first thing I suspected for this issue. Beam is aligned now but problem still exists.

Y axis both belts where bit on the loose side so I tightened them but not too mad. Though I can't see any difference really.

I did couple more tests with different RDWorks settings as I read somewhere that option "Backlash reapy optim" in "Output" tab helps with this issue. It does something, see attached pictures for the results. BTW my machine is also 500x300.

Attachment 310318Attachment 310320Attachment 310322

[Storen]

Interesting results. It looks like "Backlash reapy optim" helps a lot. Did you experiment with the "Pen Drawing" X and Y values?

[mysticusa]

I spent good couple days aligning beam, that was first thing I suspected for this issue. Beam is aligned now but problem still exists.

Y axis both belts where bit on the loose side so I tightened them but not too mad. Though I can't see any difference really.

I did couple more tests with different RDWorks settings as I read somewhere that option "Backlash reapy optim" in "Output" tab helps with this issue. It does something, see attached pictures for the results. BTW my machine is also 500x300.

right it seems like labmaster said, belt/gear/motor is driven over or under... it is possible that either a motor issue or dirty/worn belt issue? maybe some dirt causing skipping gears/teeth on the belt? it seems clearer on these set of pics... if you can measure what labmaster say, to eliminate the platform/rail alignment/level...

in my first cheap laser unit, i had the similar issue but slightly different cause... base of the platform was extremely flimsy, was causing the work table misaligned non linearly in either axis and any location.. my cuts were getting similar effect to yours...

[arbuzas]

@Storen I tried toying with "Pen drawing" options but I don't think so it's part of this "Backlash reapy optim". It did nothing anyway.

I did couple more tests but this time with rectangle (attached results). Fine lets say I'm having belt issue, but my Y axis have two belts on each side and its driven by one motor (both sides through shaft). Correct me if I'm wrong but if one belt would slip surely other side would compensate? If its motor issue then I should get more random results, I run same tests over and over and result is always same.

From attached pictures you should see that for rectangles this "Backlash reapy optim" fixes issue. So if its motor/belt slipping then how this option would magically fix this. Maybe its firmware issue?

Attachment 310512Attachment 310514

[fixtureman1]

You say your Y has 2 belts if one slips a little then the gantry can rack just a little and not be able to come to the exact spot it started at.

[Nolan Yang]

the issue should be caused by the belts or gears, please check it carefully.
Laser cutter?s cut starting point and ending point mismatch - Thunderlaser
you can visit our website to see how to do it.

[arbuzas]

the issue should be caused by the belts or gears, please check it carefully.
Laser cutter?s cut starting point and ending point mismatch - Thunderlaser
you can visit our website to see how to do it.

Belts and gears looks OK, though I keep looking if I can spot anything unusual. But what "Backlash reapy optim" option does in RDWroks that it manages to fix the issue in my rectangle example?

[Nolan Yang]

try these 2 ways to fix the issue.
1. add "edit cut in propery" .


2. set "open delay" and "through power".
let me explain what they mean.
sometimes when you need to cut wood/acrylic, the begining power is not enough to cut through the material,
we need to increase the power and keep it stay at the begining position this moment.
it is why we need to set "open delay" and "through power".

if you still can't fix the issue, I just say that it's difficult to fix it. maybe the table is not flat, maybe the power supply is not stable,
this issue is caused by hardware.

[arbuzas]

Actually I just noticed that my rectangles and squares are not perfect when measured diagonally. I.e 15mm square is off by +1.1mm diagonally on one side (other side is more or less OK). So that's probably frame issue? Or is there something I can tweak in Vendor settings to correct this?

[Storen]

Actually I just noticed that my rectangles and squares are not perfect when measured diagonally. I.e 15mm square is off by +1.1mm diagonally on one side (other side is more or less OK). So that's probably frame issue? Or is there something I can tweak in Vendor settings to correct this?

I have exactly the same problem with my machine - the X and Y rails are not completely perpendicular. For large errors you can loosen the gear belts and try to reposition the "X" actuator. That way I managed to limit the tilt to only 0.029 degree.
What I did further to deal with this smaller error is - I tilted the graphics (with .029 in my case) in RDWorks. It works perfectly but you have to do it manually for every file.
As far I know, there is no general setting to compensate for this error.

[mysticusa]

if it is cut fine, it is mirror issue.... mis aligned mirrors will cause those offsets so your cuts will look incorrect... see the pic on #3 post... if your beam is not aligned, your lines skew from straight line based on misalignment angle... instead of drawing a straight line, you will be drawing triangle kind of line... depending on how bad is misalignment, your beam can hit corners/edges in the head/mirror assembly and can cause uncut or partially cut parts due to reduced beam power.

[Nolan Yang]

Actually I just noticed that my rectangles and squares are not perfect when measured diagonally. I.e 15mm square is off by +1.1mm diagonally on one side (other side is more or less OK). So that's probably frame issue? Or is there something I can tweak in Vendor settings to correct this?

please look at our troubleshooting on our website, maybe you can solve the issue soon.

[Nolan Yang]

http://www.thunderlaser.com/troubles...ving-size.html
forget the link, sorry.

[arbuzas]

One more interesting finding (well more sad then interesting). I marked all four corners using "Pulse" and measured, it came out 502x302mm instead of 500x300mm. Not sure if its related issue or its new issue. Wondering if I should correct this with "Vendor settings -> Home Offset (or Breadth)", anyone?

@Nolan Yang - I did that. My step length was bit off, but now its bang on. I tested with 150mm square, though that does not resolve the issue when its not perfect diagonally.
@Storen - Thanks! That's next thing I'm going to try after I figure out why my bed size is off by 2mm
@mysticusa - Beam is aligned, I tested with masking tape, it's shooting at the same spot. Also beam is shooting on centre from the nozzle.

[Storen]

One more interesting finding (well more sad then interesting). I marked all four corners using "Pulse" and measured, it came out 502x302mm instead of 500x300mm. Not sure if its related issue or its new issue. Wondering if I should correct this with "Vendor settings -> Home Offset (or Breadth)", anyone?

@Nolan Yang - I did that. My step length was bit off, but now its bang on. I tested with 150mm square, though that does not resolve the issue when its not perfect diagonally.
@Storen - Thanks! That's next thing I'm going to try after I figure out why my bed size is off by 2mm
@mysticusa - Beam is aligned, I tested with masking tape, it's shooting at the same spot. Also beam is shooting on centre from the nozzle.

2mm is huge error for a CNC machine.

“Breadth” is the setting that limits the travel of the head (the work size). If the “Breadth” value is 500 but you are getting 502 clearly the machine has wrong “Step Length” in combination with some tilt (in your case).

My advice is to fix the tilt first because you will have to update the “Step Length” afterwords.

Also, for better accuracy when you calculate the “Step Length” and the tilt angle, measure the largest rectangular you could cut. Something like 495x295.

[fixtureman1]

What size table is showing in your software. I expanded mine to 525 X 325 without hitting any thing