power & speed settings for 1/8" acrylic

[Linda Smith Alabama]

I have a 2-year-old 60 watt Chinese laser engraver, and I'm having to run it very slowly to cut 1/8" acrylic. It takes 80% power and speed of 3 to get through it. I'm a newbie but that seems weak to me. Does this sound like a bad laser tube? We have verified alignment and focus are correct.

[Dan Hintz]

You'll have to tell us what a speed of 3 equates to in inches per second. For US-based systems (Epilog, ULS), 100S is roughly 75ips. At a true 60W, you should be in the area of 1ips for 1/8".

[Linda Smith Alabama]

According to manufacturer's specs, max speed is 24,000 mm/min. That's 400mm/sec or 15.74 inches. Max speed setting in the software is 400. So, if my math is right that's 0.12 inches/sec at speed setting of 3 (15.74/400 *3)
Reeeealllly sloooow.

[Kim Vellore]

That is assuming it is linear. I am not sure it is on my epilog.

Kim

[Viktor Voroncov]

Two years for glass tube 99% you need replacement. Also check lens and mirrors - they also may be need replacement after such long time.

But may be it's OK for your tube? Did you cut same acrylic before with faster speed?

[Linda Smith Alabama]

It's 2 years old but I've only had it a month, so I don't have any history to compare it to. It cuts through 1/8" hardwood easily and will cut 3/8" acrylic at the lowest speed. I just thought it would cut 1/8" acrylic faster than that and was wondering what I should expect. Since I bought it used, I went in knowing that the tube might need to be replaced. Of course, I want to get as much use out of this one as I can.

[Mike Null]

It sounds to me like it's working ok.

I believe Dan's reference is to raster speeds rather than vector.

[Dan Hintz]

My suggestion of 1ips for vectoring 1/8" with a 60W should be the main point of reference. For your machine, it looks like a setting closer to 25 for speed is appropriate (again, assuming your cranking out a true 60W). If you're going 8-10 times slower, your tube is losing power...

[Linda Smith Alabama]

Thanks for the advice. Yes Dan, about 25 should be right for speed according to your estimate of 1 ips. That doesn't even come close to cutting through it.

But there is something strange going on... the power does not seem to be consistent on different files, and it cut hotter after reinstalling the software, until I loaded a different file. When I first got this laser engraver I had a problem with "fading in", and that was solved by upgrading from LaserCut 5.1 to 5.3. I wonder if this is another software glitch? Wouldn't surprise me!

[Dan Hintz]

Make sure the colors you're using in your file are pure RGB, not shades of RGB.

[Frank Corker]

Linda if you are new to the machine, before you go leaping into buying another tube, have you cleaned the mirrors and lens' on your laser before? If not, I would suggest you give that a try, also check your alignments.

[Linda Smith Alabama]

Thanks Frank. Believe me, I will get as much use out of this old laser tube as possible!

I just cleaned the 3rd mirror and lens yesterday. We cleaned them all and aligned it about a month ago.

It's so frustrating I didn't even want to deal with it today, but I decided to try something... uninstalled the MPC03LV driver thru device manager, rebooted into SAFE mode, reinstalled the driver (first time I've installed it in safe mode), reinstalled LaserCut 5.3, rebooted, opened up an old file in CorelDraw, and lo and behold... it's cutting hotter! Not as much as it should be, but better. I can cut 1/8" acrylic at a speed of 6 now. This thing has the buggiest software I've ever seen!

[Richard Rumancik]

You'll have to tell us what a speed of 3 equates to in inches per second. For US-based systems (Epilog, ULS), 100S is roughly 75ips. . . .

Dan, does ULS spec the maximum VECTOR speed for your machine? I had the impression that some manufacturers (GCC is one) spec the "speed" of a laser system as the raster speed. Which may not be anywhere near the speed that can be achieved when cutting.

If you have time someday, perhaps you could verify the 75 ips number experimentally on your machine. You could draw the largest rectangle that your machine can make, and time it at S100%. (Might need to do a few passes for more accuracy.) Then distance/time = avg speed. I think this would represent the maximum average speed you could get in production (it is not the peak max speed due to accel/decel on the corners. But peak speed would be meaningless anyway.)

Realistically, you probably would not be able to achieve this speed in production as most parts are much smaller than full bed size. It's not very often that we cut full-size rectangles. I would not be surprised if on a small part you could only get 25% or less of maximum avg vector speed.

But I'd be curious to know whether the newer machines can really vector cut that fast.

[Dan Hintz]

Richard,

I'll have to test, but I believe you can still achieve that speed with straight horizontal/vertical lines (how often do we get those, right?). Since the driver adjusts power based upon the actual carriage speed (well, it's supposed to, but quite often it ends up sucking pretty badly due to driver bugs), moving from a horizontal line to a 45 degree line probably drops the speed by 2/3 or more, but it also drops the power by the same amount to keep the power density the same.

For example, if I command it to use 100S/100P while moving horizontally, it will do just that at 75ips... if I tell it to use the same parameters but move diagonally, it will drop the parameters to 30S/30P because (frankly) the drivers suck. In case anyone is wondering, the linear speed should remain at 75ips and power at 100%, but for whatever reason I have yet to see a machine that can handle that appropriately. This is not an issue of physics/dynamics of the head as the problem also resides in long straight lines that are not along the horizontal or vertical. Tight designs require a lot of direction shifting and therefore the speed ramps must be taken into account, but there's no excuse for a straight line to not carry the same speed.

Want proof of this cockup? Draw 20 horizontal vector lines a few inches shorter than your bed length so you don't spend a lot of time ramping the speed up/down, spaced a 1/16" apart from each other, and run them at 100S. Time it. Now take the same file and rotate it by 1 degree. Run it again. If the driver is written correctly, the times between the two runs should be identical. They most likely won't be (though I admit to never having tried this on my machine, having never really thought about it until now).

EDIT: If I remember, I think I'll try this when I get home and post the results here...

[Dan Hintz]

Anyway, the point I think you were getting at is I shouldn't be using the raster spec to determine the vector spec, and you'd be correct. That said, we have no other values to go on, so YMMV. An eighth of an inch per second for a 60W machine on 1/8" acrylic sounds painfully slow...