Fibre laser with moveable bed- queries

[Ian Stewart-Koster]

I've had a few biggish stainless steel engraving jobs that we've done on the ULS or on the bigger CO2 laser, with Cermark.
The last one was 680 x 380 mm approx, and took 8 hours on the CO2, at 0.05mm gap (500 ppi).
It turned out great, but for the whole 'day' it took.
I've done a few smaller ones, that just squeezed into the ULS cabinet, and poked out past the bed limits, but the text area fitted inside the limits- though they still took 3-4 hours or so to laser. No probems really.

A potential job has just come I have to quote on: wanting 100 of stainless plaques engraved, at 100 x 230mm or 4" x 9".
I ran a test on the ULS-M-300, and it took 35 minutes.
I ran another test on the flatbed CO2, doing 2 at once- 54 mins for the pair, or 27 mins each.
(The RDWorks 'preview' window said 36 mins, and I find it is generally wrong by close to 1/3, or a 150% increment is needed- sometimes more like double, but 150% here in this case) I was surprised that the ULS wasn't quicker than CO2, but maybe ganging two plates beside each other gave me the gain.

So I started thinking Fibre, 50 watt, Raycus...,and have some questions:

1. with a bed of 300x300 or 12" square, will the extremities laser paler than the middle, due to the further distance from the source (inverse square law of diminishing power here) ? Or does the software compensate ?
2. does the focus need changing for bigger plates like that, for the same reason, or does the lens accommodate that?
3. for the likes of 600x300, the company in China we got our flat bed laser from, said they make a moveavble bed fibre laser with Raycus parts.
They have two options: one that's effectively a 50W with a 300x300 bed, and the bed has a 600 long table, and a stepper-motor driven Y axis.
The other has a 200 x 200 bed area, and a moveable X & Y axis, to give you 300 x 600 bed area.
4. EZcad software- is there any other that works?

The company in China we got our flat bed laser from (threecnc.com) suggested the 200 x 200 bed area, and bigger movement on the 2 axes would be the better unit, but that it would still take half an hour to darken the stainless sufficiently to match the cermark blackness we have, on the 100 x 230 mm job we're doing.

How feasible is it to use a smaller fixed bed fibre laser, and have a fixed/anchored origin, or X axis at least, to butt the plate against, then do the left half, move it and do the right half, or does warping/heat affect the focus a bit and make the join too obvious?
(I ask this as I've done enough cnc routing jobs where I've routed half, rotated the substrate, and routed the other half, and both pieces aligned OK, eg on a 15-18 ft item. I'm familiar with the headaches and routines of alignment, and the beauty of known fixed origins!)

I'd appreciate any thoughts or comments on this kind of job, or this kind of mcahine- does anyone here have a moveable bed fibre laser?
Thanks a lot!

[Kev Williams]

First off, the problem with marking stainless TO BLACK with a fiber is slower than Cermark, substantially. If you can talk a customer into 'bright' SS etching, which is blazing fast, then-- you're better than me ...

Extremities- Yes, there can be power loss, but not too bad. Depends on the lens. What else that can happen at the extremes is detail loss, text letters can get a bit funny lookng. Focus is handled by lens grind, and there's a big difference between focus distance from lens center straight down and distance to the far corner of a 12" work area. Using long lenses, good focus becomes an issue due to parallelism of the work table to the scanhead. A table that's out of parallel to the scanhead by only 1/2° on a 12x12 worktable will put the opposite ends out of focus by over 2.5mm, one end low, one end high, more than enough to affect the engraving...

As to moving tables, I'm seriously considering one myself, to fiber engrave the black anodized operator panels I build. Problem is, some of them are 550 x 805mm with outlines running about 3/4" inch inward from the perimeter... just making room under the scanhead would mean building a new tower mount of some sort, and the moving table would have to be dam accurate to align the outlines in 3 or 4 moves!

All that said, I'm curious as to what you're engraving on those big plates that takes 8 hours?
These plates are 610 x 708mm, and while there is quite a bit of un-etched space, there's still a lot of black going on...
These plates take about 3 hours 10 minutes to do in pure laser time...
STlarge.jpg
I know this much, to anneal just the big S on that plate to black with one of my fibers, would take around 1/2 hour. Fiber's are NOT fast at doing black! At least, a 'nice' black, you can do a 'burn' black that's quicker, but it's not all that pretty...

[Ian Stewart-Koster]

Thank you very much, Kev.
Much to think about.
I'll attach photos of three jobs in the last year, that have each been over 12 x 24 inches of cermarked stainless steel.

The biggest, I timed, on our flatbed laser, and it was 8 hours...despite it seeming like it should have been less.
That was with an 0.05 gap, or 500 ppi, ie 20 passes of X per millimetre advance in Y.
Slow and hot... though it was fairly cool by the time each next pass got there.

I have the big machine tuned for accuracy- and I found that accuracy by making with slowish accelerations, or low speed start-offs.

Facebook is full of newbies with new lasers, who ask in despair "why is my etched photo blurry?", or "why do the start and end lines not line up", or "why are my straight lines wobbly?" etc. I diagnosed and tuned the reverse interval, backlash, wobbles and other issues out of the big machine within 6 weeks of getting it. The loss, might be rapid acceleration, but I am 'fairly' content... for now!








20190430stainlessno2low res.jpg 2019043stainless_lowres.jpg Yarraman-acknowledgements-plaque.jpg

[Kev Williams]

I have to ask, did you raster engrave everything on the big plates with the snake? Did you do any color mapping, or all black? I'm thinking you may have fully rastered about 85% of the entire plate from top to bottom, which would definitely explain the 8 hours--

Depending on your layout -as fed to the machine- and engraving speed, I'd be willing to bet that by vectoring the snake's perimeter, and applying some color mapping to avoid engraving a ton of white space, you could whack at least 5 hours off the engraving time...

[Ian Stewart-Koster]

Thanks for the reply, Kev.

I'll add photos of the moveable Y, and the X & Y versions below, for interest's sake...

Yes, I rastered the whole plate/s.
I found with the chinese machines, and Cermark, that a slowish rastering speed gave me blackest blacks- but 80mm/s or so is sloooow!
I also have a chinese -made CO2lasermark solution that I tested, and it went a good black at 380 mm/s, but it cannot take being diluted at all, and I don't have access to more, at the moment. It goes on in a dark bluish grey, and is marketd as sticking to stainless, aluminium, glass & ceramics.
(I've applied it to a ceramic white tile, but not yet tested it under the laser. It does stick to bare aluminium, but looks fairly 'frosty' or coarse-ish)


By colour mapping, do you mean sorting areas into different colours, to be lasered as one block?

I've tried the "independent output" option in RDW layer parameters a few times, and although the preview window shows a HUGE time saving, reality has proven the preview window to be like Pinocchio with a broomstick or a 10 ft bargepole for a nose.
In one case, a preview time of 35 mins, lowered to 12 mins with Independent Output checked. Great! When I went to laser it, at the 20 minute mark it was barely past 1/3 of the way through. It took 55 mins to complete, not 12!

After being messed around, I tended to stay with full fastering, as the plaques have few areas where there is much dead travel space.
Maybe I should try more vectorising of text, with centrelines and inlines to fatten things up- an outline is insufficient- and on the big laser, the advantage is I can gang 25+ of the 4x9 inch plates, and just leave it alone, whereas with the inertia of the gantry, rapid vector cutting is not on the checklist beside accuracy, unless we're slow.

Still. I may test it out...on both ULS and chinese... Thanks for the suggestion.

[Ian Stewart-Koster]

Chinese fibres with moveable beds- Y at 300x 300 and X,Y at 200 x 200 bed
CatchF847.jpg CatchBD87.jpgCatchBD87.jpg CatchF847.jpgCatchF847.jpg

[Ian Stewart-Koster]

Kev, I ran that file I have to quote on 100 of, as a test, with only the name in the middle rastered/engraved, and the rest vector marked, by doing a centreline and then 2 inlines and a perimeter line of all text.

On the ULS, it brought the total time down by 1/3. I probably could have omitted the final text outline-it looked nice before we got there!

I'm trying it now on the CO2, to see how it goes for accuracty in tiny vector letters... Preview says 35 minutes...
Took 42 minutes, but it was on a slow vector speed. Soo slow & hot- the stuff was warped underneath.

(I covered half the plaque under the ULS-I've covered the other half on the big CO2, so we can check the relative differences easily.)


Conclusion- rastering 2 at a time on the CO2 was quicker.
Vector marking multiple inlines on the ULS was quicker than rastering- but they need lots-as many as 16 inlines on bigger text, and mildly out of focus helped.
___________________

Ran a separate file on cardboard, to just check quality, and ALL text as multi-inlined vectors at 0.15mm gaps between vectors.
I found I could raise the speed on the CO2 to 50mm/s, and it was no worse- but I suspect I have my small circle limits set to maintain quality, so they might limit it at 22mm/s, or something.

I'll check back with the analysis later...

[Ian Stewart-Koster]

Client has decided to shink the plaque size to maybe 1/4, (thank goodness!)

[Scott Shepherd]

I agree with Kev, fiber on stainless is the slowest method.

For what it's worth, I do plenty of stainless plates that are 16" x 28" and they run about 35 minutes on a Speedy 300.

[Tony Lenkic]

Steve,

What is your speed setting marking stainless with cermark?
35 minutes to mark 16" x 28" is quick.

[Gary Hair]

Steve,

What is your speed setting marking stainless with cermark?
35 minutes to mark 16" x 28" is quick.

You could even say it's "speedy"...

[Ian Stewart-Koster]

Thanks, Scott, Tony & Gary.
Yes, 35 mins is fast- especially when Cermark needs a sbit of slowness- still, much time is wasted in the turn-around stage of the lens' path on most systems.
And we'd don't have a Speedy! (I wish!)

He came back and accepted my quote for the 1/4 size version, and is getting 50 of them, for starters.
I can gang them up on the ULS and the big chinese machine, routing a template to hold whatever number, for repeat placement.
50 will almost use up all the Cermark we have left, so just ordered another 250g jar.
_______
ON a different subject, I bought a dead VLS 450 2 years ago. The tube was no good, and I had the motherboard reconditioned, but the machine still won't go.
But for $130,it was worthwhile for parts, and I figured I'd convert it to Chinese for the exercise, when I can get some time.
I also have 2 closed-loop steppers/servos (Leadshine) that I got sent over from China with our flatbed laser, and a spare W1 Reci tube, mounts, and PSU.
I also got 2 spare Ruida controllers, and PSUs, and other needs, when they were on special at Cloudray a while back, so I have the parts to try and make a Universal 'Speedy' with a CO2 tube! (I have a spare chiller of a sport- the original CW3000 that I never used, that came with the Golden Laser.)

The only slowup might be that the leadshine closed loop servos/steppers and drives are big- unnecessarily heavy duty for a lightweight ULS gantry & head.
One day I'll make a start...

[Scott Shepherd]

Steve,

What is your speed setting marking stainless with cermark?
35 minutes to mark 16" x 28" is quick.

Email sent Tony.

[John Lifer]

Ian, I can run my chinese at 160mm/s with cermark LMM14 or 6000 and it comes out well. I'm surprised you have to run at only 80....
And by color mapping ,yes, Kev is saying do each section of say your text separate color in RDworks. And snake in different also. (set powers the same correct power and speed for cermark, but each will run separately) . One trick to estimate your time is to send the job to the machine. It should be really close when you use the machine time estimate. I've found it can vary a few seconds on a 30 minute run the first time, but not software 25 to 50, to 150% error..... (seems the bewer the version of the software seems to make the error go up. Early versions were only 10-20% off)

[Kev Williams]

I just ran this batch of SS plates, the total rack is 24" x 11",
each row is 27 lines, .185" high Arial Bold. Left row black, center row green, right row blue,
so each line engraved individually...
ssp.jpg

my settings, and run time, 40w Synrad / Gravograph LS900-
pwsp.jpg

So what exactly is 17% speed on this machine? As it turns out, about 100% speed on a Chinese machine--
I ran 20 total sweeps- 10 right, 10 left- of one line of text, I measured the line at 7.08", and the time took 12.32 second per my watch...
--7.08" x 20 = 241" (6136.64mm), / 12.32 seconds = 19.6" (497.84mm) per second, including turnaround time, which is basically zero on this machine at 17 speed

Just barely over 1 hour, and that's a substantial amount of engraving in an 11 x 24 area vs 8 hours for your roughly 12 x 24 area....

Could be you need to try testing at higher speeds and less power? If there's one truth about laser engraving, very often less is more

edit-- I said 'rows', I should've said 'columns'--