[DLDude]

Hey all. I have a Rabbit Laser and I cut large, complex parts with really small wall thicknesses. The piece has over 1000 cuts. As the part progresses, the laser beam seems to fall out of the program as if it is drifting. It can seem to drift at least 1/32" or more off of the expected cut path. Is it unusual for the laser to drift over a 1-1.5hr cut program? I've tightened the belts and it hasn't helped. I slowed the program down to 30 speed and it does the exact same thing it did at 50 speed. Any ideas?

[Exsecratio]

lose cap screws on the belt cogs? failing that crap on the linear bearings can cause the same.

If possible, get those two out of the way before looking at missed steps on the motors

best wishes

Dave

[DLDude]

Is there any advantage to using the standard cutting mode in LaserCut rather than 'Immediate". I have a feeling once the laser drifts on 'immediate' mode, it can never go back, whereas having that option off may keep it in the right place?

[pontiacg5]

There's quite a bit in a laser that would make it do that. Some pictures might help tell what's going on.

Is it only in one direction or is it in both axes? Does it also drift in engravings?

You more than likely have something loose or dirty like Exsecratio said. You'd see something remotely similar if the belts are loose too. I'd clean and regrease the rails and check the belt tension before doing anything else. You might try moving the laser head through the full range of motion and listen for odd noises too.

Immediate has nothing to do with this and has more to do with setting origins. The reason your laser can never recover from drifting is because it uses stepper motors. Steppers are open loop, meaning that the computer asks the stepper to move, but it has no way of telling if it physically moved so there is no way to correct drift or lost steps. Servos have encoders or resolvers, so they are closed loop. If a servo doesn't move after getting a command to move the computer knows about it and corrects it.

Ah, I also have a Gweike (FSL) laser. If I stop it in the middle of a cut program it will NEVER repeat when I restart it. I'm not sure why, but the head seems to move towards the origin more violently when I stop the cycle and I'm pretty sure it moves a bit after it thinks it's stopped. If I let the program finish running it will repeat all day long, but it approaches the origin slower after it finishes running a program.

[DLDude]

Alright, thanks you guys. I will try to check the rails and such.

[Exsecratio]

Hey Ponti as a quick sidenote,

Try reducing datum speed in the machine options / feeding see if that helps? I had a similar problem so reduced and played with the machine settings and now don't get it. Must admit though that the stop / datum function does seem more violent on my 1290 (basically a Rabbit 1290 but UK version)

cheers

Dave

[Exsecratio]

I've tightened the belts and it hasn't helped.

Hiya Dude,

That could actually make it worse, careful you don't over tighten the belts.

best wishes

Dave

[DLDude]

We did notice that a lot of our machine settings had chanced after we had our controller replaced. What are good standard numbers?

Right now we have:

Start Speed: 5.0
Quick Speed: 10.0
Acceleration: 15.0
Datum: 8.0
Test Speed Fast: 10.0
Test Speed Slow: 8.0
Original Offset Length: 10.0

We cleaned the rails and made sure all belts are tight (and not too tight, they seem good). The problem is still the same.

[Exsecratio]

Is your machine Imperial Dude?

Mines metric so can compare

cheers

Dave

[DLDude]

I think it is metric

[Exsecratio]

Thats set quite low then..

Try,

Start: 10

Quick: 100

Accel: 700

Datum: 30

Test Speed: 300

Test Speed (slow) 30

Org-Offset: 2

[DLDude]

I will try this! Do you think it has anything to do with the drifting?

Also, what is the difference between "Worktabe" and "Feeding" settings?

[Exsecratio]

To be honest? probably not

best wishes

Dave

[DLDude]

Alright. My aplogies. The numbers I were looking at were for the Z axis and not the X and Y.

Here are the XY numbers

Start Speed: 5.0
Quick Speed: 150
work-Acc: 400
space-Acc: 550
work-AccAcc: 6000
space-AccAcc: 6000

Datum Speed: 40
Const speed: 5.0
Test Speed Fast: 150
Test Speed Slow: 80
Lasers Space length: 0.0
XMaxChangeSpeed:20
YMaxChangeSPeed:20

Here's a picture of a grid of squares being offset. It seems each column is offset, but I couldn't find any spots inside the grid where the laser jumped.

[avetek]

Looking at the photo, it suggestes that the machineX and Y rails are not square to each other?, If you get a piece of mdf board and get the machine to mark out as bigger rectangle as you can on the table, you can check it for size and also check the diagonals for squareness. I had to adjust the count of each stepper to achieve the correct sized 800 x 500 rectangle I tested.

[DLDude]

I think doing a big rectangle is a good idea. I'm not sure about the XY being not perpendicular. On that program it cut like that, but on other programs it does horizontal lines just fine. I'm thinking it's an issue with the stepper count or the controller. Thanks for the tip!

I'm guessing I should do a big rectangle with maybe every cm marked to see how much it travels per cm?

[Exsecratio]

I'm thinking stepper count as well, crab gantry tends to hop and jump rather than progressively get worse.

best wishes

Dave

[DLDude]

If it is the stepper count, what is the BEST way to calibrate that?

[Exsecratio]

Cut a square and increase / decrease the step count in the axis that's at fault.

Repeat until the square matches the size (allow for kerf naturally) you should end up with a square that can be flipped and rotated till it fits in all direction (permutations).

Long time since I've had to do it so forgive the rather simplistic answer

best wishes

Dave

[DLDude]

A small update:
Ray from Rabbit Laser came down and replaced the controller and it's not fixed. The model that was on it just didn't play nice with the rest of the system.