2500 x 1300 flat bed laser RD parameters question

[Ian Stewartkoster]

Warning, long post - hopefully clearly explained!
HI, we've had our big 2-head, 2-tube laser a few years now, and after a lot of tweaking, I'm mostly very happy with it.
I saved the vendor settings right at the beginning, although they're pretty useless - the chinese mob who set it up were very good, but things like reverse interval settings etc, and small circle speed limits were not a part of their expertise.

For a long while I had slight start and end cut alignment issues - fixed by minor tweaks to backlash.

Engraving is great. Very happy - as long as the X speed is max 400 mm/s, and Y acceleration and speed is low- very low.

Cutting is another matter. For anything say 35mm high or bigger, cutting is great. It's just very small stuff I'm finding tricky:

I've had 2 jobs recently - diecutting card for a printer - 7000 discs, 3/4" diameter, and 5000 at 1" diameter, set up on A4 sized sheets.
The problem with our big laser has been at that small diameter, the circles are not perfectly circular.
They all have a slight flat-spot at 5-past 12, and 35-past, or between 12 and 1, and between 6 and 7 o'clock.
The start and ends line up perfectly. If I set a seal of half to 2 mm, (or path overlap) it is imperceptible, which is good.

I did have other issues, mainly being the 'bounce' at the beginning of every cut - a dampened sinewave shape that ceased after about 3 or 4 cycles, which I attributed to the higher acceleration in idle, then slowing down and overshooting the mark as it started each new cut. It did not matter where on the clock the start point was. I had to crank the cut start speed right down low, and the acceleration right down low in order to eliminate the bounce or wobble - read, edit and write to the controller, in User settings.

That made everything pretty slow though.

With the big laser, both tubes - W1 and W8 Reci are carried in the extra-big gantry, so there is a fair bit of inertia to have to shift and absorb - much moreso than smaller bed lasers where the gantry only carries the head and the belt for the head.

I tried turning on and turning off the Backlash Reappy checkmark. All that did was alter the starting place of each new file, and change the total time taken for the jobs due to extra travel time to begin each new cut at the opposite side of the clock. It did not change the flat spot which is like a failure of Y to move microscopically at the direction change time after the head passes 12 o'clock or 6.

When I tried to go faster, we got the bounce at each cut beginning point.

I tested this by lasering lots of concentric circles, not by wrecking the client's job!

IN the end I guessed that although in Vendor settings, Y has a much lower acceleration factor and lower overall speed because it is a bigger Leadshine stepper motor moving a bigger weight, compared with X's stepper which only has to drag the head - that maybe the RD6442 might have problems sending instructions ot 2 different drives with different parameters, and perhaps that's the reason for the flatspot - so I set both X & Y in Vendor settings to the same lower speeds and acceleration as Y used to have.

It sufficed to run this job acceptably - and the resulting discs look round to anyone else, but I still see the minute flats. And minute overcorrection bounce in the curves.
No, the belts are not loose, nor is the lens loose. if I laser a 1200mm square on cardboard, the four sides are accurate, as are the diagonals.


So, I was wondering if anyone else with a machine with a heavyish gantry like ours, might not mind sharing their vendor settings for overall X & Y speeds & accelerations, and User settings for the same, and small circle speed limits please?

It is great at bigger sizes, but I'd like to be able to have it better at the tiny vectors, at a greater speed than snail's pace!

Thanks for any thoughts!

[Gene Uselman]

Hi Ian- I have seen your posts on several forums over the years and you seem to be always on the frontiers of whatever you do. We do not have a lot of flatbed lasers among the members but I hope someone can address your problem. I have been involved with larger and larger 'cabinet' machines and they definitely present new and different problems requiring more inventive solutions. We have a machine [1800x1300]which came with a W2 and W6 tube [both with bad 'mode burn' patterns] and I have finally gotten a SPT 130w tube in the main cutting side [and nothing in the other] and have found the 130w does everything I have asked of it, down to 2mm holes in .25 plywood for instance. Engraving is also very impressive- long story short, I am eliminating the low power side completely. You might consider removing your smaller tube to reduce what must be a lot of mass on your gantry. Good luck, Gene

[Ian Stewartkoster]

Thanks for the reply, Gene, and the compliments... It's a struggle to figure out, with so many variables, what tiny modifications really have what effects, when it really matters...

Also our Chinese Goldenlaser, which cost a squillion dollars more when new in 2012, than this big one,(and is still on its original W6 tube), is just 'always right' - it works beautifully. The proprietary software I dislike, but like you've found with your unit - it works well, so I accept it. (I have bought a spare RDC244x controller to one day put in, but for now I won't)

The dampened wobbles have to be a speed/acceleration thing. If it was servo-controlled, I'd say the PID settings were not tuned right. It is hurrying to an x,y position, overshooting, correcting itself, slowing and overshooting the correction, till it lands in the right place, then finishes the circle. This last job involved cutting 7000 discs in heavy card - they had to be close to perfectly round.

The very slight flat spots at just past 12 and 6 o'clock, I'm putting down to some sort of 'give' in the belts or Y gear couplings. Not backlash as such, but a similar effect. It is such that for a very brief moment, the gantry is not moving while the Y stepper changes direction. but it is not at 12 & 6 o'clock like you'd expect.

I did try setting the backlash compensation in User settings, and increasing it from zero, to 0.01mm, 0.02mm, 0.04, 0.08, 0.16, 0.25, o.50, 0.75, 1.0, 1.5 and 2.0 mm to see what the resulting paths looked like, and from that I deduced that our problem isn't quite backlash - the flat spot is in a different part of the clock in a circle, than where the backlash correction appears, when you look at them all - besides, there was no setting at which the flatspots vanished. I tried negative numbers too - but they did not work.

I'm going later to check the couplings to see if there is a possible bit of slop , but my current gut feeling is to recheck the tension on the 2 long Y belts. It is possible that if the teeth are not perfectly in synch, that a miniscule variation could be the cause of this.
I watched a video on Youtube on harmonic drives that I found fascinating -it shows how a minor variation in alignment between two belts, can have a tremendous torque effect. They apply this to robotics, but I'm reverse-thinking it back to being a possible constributing factor to our problem? Who knows...
https://www.youtube.com/watch?v=7QidXf9pFYo

I appreciate your reply.
The weight of the W2 tube is not significant really compared with all the rest of everything in the gantry, incluing both steppers, all gearing, and the W8 tube, and the windows & roof etc.
Besides, I find I use the W2 tube 95% of the time!

[sebastien laforet]

I'm going later to check the couplings to see if there is a possible bit of slop , but my current gut feeling is to recheck the tension on the 2 long Y belts. It is possible that if the teeth are not perfectly in synch, that a miniscule variation could be the cause of this.

i'm pretty sure that such a big machine should not use belts. the load and the distance is too big for this system. in serious CNC machine, this is always rack&pinions, or threaded spindles. belts are cheap and works for quick movement / low charge, which is totally not your case (especially since the gantry supports not only the head, but the whole laser).

maybe your backlash is ok, but you could step into the elastic response of the belt with so much weight.. a domain that is obviously out of reach of our smaller machines.

[Ian Stewartkoster]

That 'elastic response' or microscopic 'give' in the two long Y belts is a possibility also... although I don't remember the same problem being quite so bad when it was newer.
Thanks for the reply.
I've not yet checked it out - I had a bigger job to laser yesterday, and plus or minus 1mm was plenty of accuracy in that!

[Gene Uselman]

being quite so bad when it was newer

Ian- I have developed a few new problems over the last 75 years too .