First Laser Project, raises more questions

[Walt Langhans]

Hi Guys!

Well I found a place here in Atlanta that I could rent a laser, and I got my initial designs done.

241524241525241526

In case you wondering what happened to the windows on the brick building, I was a bit heavy handed while cleaning the residue off it, and took a bit of the material off in the process. If your wondering what the heck the middle building is all about, it's a style called SteamPunk.

The material is 1/8 MDF, the laser is an Epilog 30 watt Zing 24 (07 I think).

While doing this I had a long chat with the Daniel (the guy who owns the laser) about the pros and cons of different lasers, and laser systems. The biggest thing that stuck out for me what how long it took the laser to get these jobs done. It was around 2 hours to get all that cut and engraved. I realize that a big part of that was the settings that we used on the laser, but it got me wondering how much faster a more powerful laser would actually be? Obviously the 30 watts handled the job, but after the initial "this is cool" watching the laser work, it soon turned into watching paint dry. Also, Daniel mentioned that the different types of laser tubes can effect how quick you can get a job done as well and that all laser tubes are not created equally so to say.

So... in a nut shell, now that you can see what I'll be doing, what kind of laser will get it done quickly, or at least quicker:)

Thanks!

[Steve Busey]

So... in a nut shell, now that you can see what I'll be doing, what kind of laser will get it done quickly, or at least quicker:)


There are only a zillion variables to that question, but the basic answer is: get as much power as you can afford!

[Joe Hillmann]

If your goal is speed it may be worthwhile to change the detail of the parts. The laser slows way down on curves compared to straight lines so the curves on the bottom of all those shingles and the circles on all the rivets add a lot of time. Also I would assume that of that 2 hours it took to make the parts only 20 minutes of it was actually cutting and the rest of the time was engraving the detail (I am assuming that you vector engraved the detail rather than rastored it).

With that in mind even if you step up in power you won't save a whole lot of time, the only place the extra power will make much of a difference is where you are actually cutting. With all those short lines and curves the laser never has enough time to get up to speed to make having more power make a difference.

As a side note I had played with making some castle toy kits and decided that If I wan't to actually make them to sell I will screen print the detail onto the wood then use the laser to cut the parts out. Of course using screen printing would mean you would have to do quite a few sets at one time to make it worthwhile.

[Bruce Dorworth]

Walt, Very cool!! I have an 80 watt laser and I think when doing the bricks and roof you can run faster with a more powerful machine. If you want someone to do a test send me a file and I will test cut it.

Bruce

[David Fairfield]

Ditto what Joe said. Everything you add to your graphic will add to machine time, its cumulative. You might save time by using commercial brick and shingle material. Architectural and hobby suppliers carry it in various scale. Also in some cases you can orient parts and use color mapping to take advantage of the side to side engrave speed.

Dave

[Walt Langhans]

Yes, the majority of the time was in the engraving. I was kind of afraid that the additional power wouldn't result in a big savings of time.

One of the questions I'm now debating is would it be better to get 2 lasers from some one like hurricane at $7-9K a pop or spend $20k on a single Epilog, Universal, or Trotec???

I read in one of the treads around here (it was about doing architectural models) that printing out the details was an option, but I don't remember if it was screen printing or not.

[Joe Hillmann]

Another way to speed up the engraving time may be to rastor rather than vector. On my machines rastoring thin lines doesn't look very good but on an epilog it may look okay. When rastoring I figure it takes about 10 minutes per square foot. Compared to vectoring where it can take hours to do a square foot if there is a lot to engrave. And since Bruce is offering I would suggest having him run the file and see how much faster an 80 watt machine is.

[Joe Hillmann]

Walt.

What are you planing to do with the models? I assume since you are looking at buying a laser to make them you plan to sell them? Also how big are they? I am guessing about 3 inches wide and 6inches long? Knowing the size and there final use may help people to give you ideas.

One idea, if they are as small as I think they are, is to use iron on transfers to put the detail onto the parts after you have ran the transfers through a regular printer.

Another idea, if they are larger than I think, would be to make stencils and paint the detail on that way.

[Walt Langhans]

Walt, Very cool!! I have an 80 watt laser and I think when doing the bricks and roof you can run faster with a more powerful machine. If you want someone to do a test send me a file and I will test cut it. Bruce

I think I'll take you up on that, thanks!

https://docs.google.com/open?id=0B_f74JMee5cKZFFDbGhpSmN4Y2s


Walt.

What are you planing to do with the models? I assume since you are looking at buying a laser to make them you plan to sell them? Also how big are they? I am guessing about 3 inches wide and 6inches long? Knowing the size and there final use may help people to give you ideas.

The models here are examples of the styles that I'll be doing, the finished product (which I do intend to sell) with actually have a second layer of MDF that will be puzzle pieces, so you will be able to take the buildings apart for easy storage and transportation as well as make them bigger and mix the styles. I'm going back to do some more work next week and I'll post some more pictures so you can see exactly what I mean. The foot print of this building is 4 * 6, but the finished one will be in more pieces. I need the detail cut into the MDF because the people that buy them are going to want to paint them. They are people who like to build models from kits.

[Joe Hillmann]

I make quite a few wooden toys and before I start on them I have a price in mind that I want to sell them for. Lets say $5. When I design it, it might take 10 minutes to make on the laser. In order to sell it for $5 I have to get it down to 5 minutes to make, so then I start taking detail out of it and lining up the parts so they nest together in a way that one cut can be the ends of two pieces rather than one, and getting rid of curves in order to be able to make the toy in 5 minutes.

When I first started to modify my designs like that I felt guilty about producing something that wasn't the best I could make but my sales went way way up because I was able to price my stuff at a point that my customers were willing to pay. You may want to look into redesigning your items to see if you can make them quicker that way rather then by using more power.

[Bruce Dorworth]

Walt, do you have any idea how long it will take to do the wall that you attached? For a comparrison..

Bruce

[Walt Langhans]

Yeah... you and I think along the same lines. I have a pretty good idea what my people want, and it's what I made with the good detail. I too feel guilty not doing the best I can. And I know I can sell these around the $15 - $20 mark, so it's how quick can I get them done, or am I willing to let it take the time that it needs and set up a TV and lazy boy in the garage so I can kick back while I keep an eye on it to make sure the house doesn't burn down, lol.

[Walt Langhans]

For the individual wall, not a clue. I figured I'd see what yours did, then when I go back, run just that wall and compare.

[Rich Harman]

I ran the file on my machine. A couple of observations.

1. There are overlapping lines. The wall is made up of four sections, the ends of the bricks overlap.

2. The lines do not match up exactly. There are several places that that the horizontal lines do not meet evenly. So instead of the laser treating the two lines where they join as a single contiguous line, they are engraved separately making for a visible line where the laser stops/starts.

Total time, including cutting out the windows and the outline of the wall was 2:50. This was using a 2.5" lens. If I were doing more of this type of thing I would switch to a shorter lens (or use the beam expander) for finer detail.

[Walt Langhans]

Thank you very much Rich!

Before I did these I didn't understand that the laser will cut / engrave if there is more than one line. I'm going back and redoing a lot of stuff because it was copied and pasted and lead to things like that and the misalignment. But I wanted you to have the same file so I was comparing apples to apples.

And now for my next noob question. 2.5" lens???? and switch to a shorter lens or beam expander???

[Rich Harman]

I use a 2.5" (65mm) lens vs a shorter lens such as a 2" or 1.5" because the longer lens has a longer depth of focus. That is useful for large pieces where it tends not to lay perfectly flat, in other words, the distance between the workpiece and the lens can vary by as much as a couple millimeters when using a longer lens without a loss of cutting performance. If a shorter lens were used the same variations in distance would cause the cutting point to be out of focus and therefore not cut through in those places.

The tradeoff is that longer focal lengths have larger spot sizes.

A beam expander increases the diameter of the laser beam and collimates the beam to reduce divergence which makes for a more consistent power density everywhere on the table - most useful on larger tables. The larger the diameter of the beam going into the lens, the smaller the spot size. However, using a beam expander decreases the depth of field. So using a beam expander along with a short focal length lens will result in an extremely shallow depth of field.

Most of what I do is cut acrylic and plywood from 1/8" to 3/8". I have found the 65mm lens, without using the beam expander, to work best.

[Walt Langhans]

Thanks for the explanation. One of the things that I wasn't completely happy with was the thickness of the engraving lines, I would like to have them a little wider and I wasn't sure how to go about fixing that.

[Rich Harman]

You can make the lines wider by defocusing slightly.

I ran the file again but this time I rastered it. I think it looks a lot better. My machine is quite slow for rastering, mainly because it has a lot of mass to speed up and slow down. The total time was a little over three minutes. The Zing should be much quicker.

[Bruce Dorworth]

By raster do mean the engrave function in LaserCut? I ran it as cut @ a speed of 350 and it took a little over 6 minutes.

It would probably help if all the small lines were joined so it would cut continuosly instead of all these little segments. Also this is made up from four wall sections, so where they overlap you have double lines that cut twice.

Bruce

[Walt Langhans]

Rich, if it's not too much trouble could you post a pic with the 2 runs? No hurry. Thanks

[Paul Phillips]

Walt, it seems like you are firm on rastoring/engraving these so if you are simply looking for max engraving speed there are two "best options" that I know of, first would most likely be a Trotec speedy 300 (Trotec owners, correct me if I'm wrong, I'm going by the website which lists the 300 as being a bit faster than the 100) or a Universal laser with the super speed/dual beam feature, (requires two laser tubes though so it's expensive), not sure which is cheaper or faster but maybe someone on here who has one can run the test for you and compare the times.
Paul

[Rich Harman]

By raster do mean the engrave function in LaserCut? I ran it as cut @ a speed of 350 and it took a little over 6 minutes.

It would probably help if all the small lines were joined so it would cut continuosly instead of all these little segments. Also this is made up from four wall sections, so where they overlap you have double lines that cut twice.

Bruce

The first run (2:52), vector only, was 100mm/sec 30 power - except for the parts that were cut through, they were 30mm/sec, 90 power. It would go a bit faster if the horizontal lines were continuous.

The second run was rastered (Scan), bi-directional (X swing) .1mm interval at 600mm/sec, 75 power. I first changed the line width to .75 points in Corel then converted to bitmap (254 dpi).

Overall I like the second one better but I would vector the bricks that surround the windows so that they stand out more.

241548

Here's a video - you might want to turn the volume down.

[Glen Monaghan]

You didn't say how long the second, rastered version took.

All those short vertical lines for the sides of bricks would require rastering the entire surface.
Here's a quick noodling that ignores the windows and assumes the grid of bricks is a regular array rather than a bunch of separate rectangles trying to more accurately represent individual hand-set bricks at slightly different spacings and orientations, and assumes all the lines are a single vector's or raster's thickness:

The bricks appear to be about 4 times as tall as the lines are thick so it would take four full side-to-side passes to raster each row of bricks, and there are 32 rows of bricks so that would be 128 full width passes in raster mode. In vector mode, you'd need to make 31 full width passes for the horizontal brick lines and then there are 63 alternating vertical lines of brick sides there would be 63 vertical passes for those. If you know the time for a full width raster pass, full width vector pass, and full height vector pass, then you can predict the results. Assuming that raster and vector times are identical for horizontal scans, then the key factor is how 63 vertical vectors compare to 128-31=97 horizontal scans. Vertical speeds are generally slower than horizontal speeds and in this case you have 1.5 times as many extra horizontal passes rastering as you have vertical passes vectoring. So, if horizontal passes are less than 1.5 times faster than vertical passes, vectoring should be faster. If horizontal passes are more than 1.5 times faster than vertical passes, rastering should be faster.

If you double the size of the bricks so it takes 8 horizontal passes per row, there are still 128 raster passes. However, with half as many bricks that are twice as large as before, there would only be 15 horizontal vector passes and 31 vertical vector passes, so you'd be looking at the difference between 31 vertical passes and 128-15=113 horizontal passes. In this case, there are 3.65 times as many extra horizontal passes when rastering as there are vertical passes when vectoring, so horizontal speeds would have to be much faster than before in order to favor rastering over vectoring.

Small bricks which don't need many horizontal rasters per brick but involve lots of vertical vectors will tend to tip the scale in favor of rastering when horizontal speeds exceed vertical speeds. Making bigger bricks that have fewer horizontal and vertical edges would quickly shift the balance toward vectoring.

Additional details such as the decorative bricks around the windows are "free" when rastering but add to vectoring time, while it seems to me that the tall window cutouts might help vector times slightly more than raster times because rastering skips over the gaps many times more than vectoring does.

-Glen

[Scott Shepherd]

I rastered the brick and came up with about 1:30 on it. I vectored the brick and it took closer to 2 minutes, maybe even a little more, I don't recall the exact number because as soon as I saw it take longer than rastering, I stopped paying attention.

[Rich Harman]

You didn't say how long the second, rastered version took.
In post #18 I said it was just over three minutes.

[Bruce Dorworth]

I don't know how you did it that quick. First time I ran this file it was 6 minutes. I connected all the horizontal lines and it was stil almost 6 minutes. I tried to raster this piece but the software said I had open polygons. I use lasercut 5.3

Bruce

[Rich Harman]

You need to convert to bitmap before rastering. In Corel, change the line widths to .75 points then convert to bitmap.

[Rich Harman]

I don't know how you did it that quick. First time I ran this file it was 6 minutes.

It could be that your speed is limited by software. I know that if I set mine to cut at 200mm/sec that it won't cut any faster than if set to 100mm/sec.

[Walt Langhans]

You need to convert to bitmap before rastering. In Corel, change the line widths to .75 points then convert to bitmap.

I understand how to do all that, but why does it make a difference, and is it a good rule of thumb to do that for all of my stuff that I'm doing?

[Rich Harman]

The difference is that LaserWorks and I assume LaserCut, does not recognize line widths. So a grid of lines, regardless of how thick they are are seen as just a series of unclosed polygons (lines) - they can't be rastered because they have no "area". Converting to bitmap is an easy way to be able to raster them.

If you draw a circle in Corel Draw then try to raster it in LaserWorks the result is that it will raster a filled in disk. It recognized the circle as an enclosed shape. It is irrelevant what the line width of the circle was set to because that is ignored by LaserWorks.

If you want to raster that circle as a ring instead of a filled in object then you need to either convert it to a bitmap in Corel first, or use the "Convert Outline to Object" tool or simple draw another slightly smaller circle inside the first. There may well be other ways too but those are the ones I know.

Using the brick wall as an example - I think the best way to raster it would be to first separate out the windows that need to be cut so that the work is done on only the parts of it to be rastered.

Select all of the wall and Weld. Then Convert Outline to Object. If you do not weld first then you will end up with many objects overlapping when you convert the outline. In Corel hide the fill and you should see that the entire wall (except the semicircular windows) are one contiguous shape. In other words one click inside any of the lines with the Smart Fill tool will fill the entire "grid" of the wall.

Hopefully you have saved the original because if you want to vector engrave the bricks now, it will take twice as long because there is a great deal more line for the laser to follow.

[Walt Langhans]

Ahhhhh, that makes sense. Thanks :)

[Don Mayhew]

Walt,
You are welcome to come to our office in Atlanta and do a comparison. We have the RayJet, Speedy 100 and Speedy 300 to show you different time studies. Don Mayhew, Trotec Laser.



Hi Guys!

Well I found a place here in Atlanta that I could rent a laser, and I got my initial designs done.

241524241525241526

In case you wondering what happened to the windows on the brick building, I was a bit heavy handed while cleaning the residue off it, and took a bit of the material off in the process. If your wondering what the heck the middle building is all about, it's a style called SteamPunk.

The material is 1/8 MDF, the laser is an Epilog 30 watt Zing 24 (07 I think).

While doing this I had a long chat with the Daniel (the guy who owns the laser) about the pros and cons of different lasers, and laser systems. The biggest thing that stuck out for me what how long it took the laser to get these jobs done. It was around 2 hours to get all that cut and engraved. I realize that a big part of that was the settings that we used on the laser, but it got me wondering how much faster a more powerful laser would actually be? Obviously the 30 watts handled the job, but after the initial "this is cool" watching the laser work, it soon turned into watching paint dry. Also, Daniel mentioned that the different types of laser tubes can effect how quick you can get a job done as well and that all laser tubes are not created equally so to say.

So... in a nut shell, now that you can see what I'll be doing, what kind of laser will get it done quickly, or at least quicker:)

Thanks!

[Walt Langhans]

I finished up cutting / engraving the rest of my buildings. For first run / prototypes from someone that has never used Corel or a laser before I'm pretty pleased how they turned out. There where a few issues but nothing that can't be corrected.

This is a Gothic style ruin

241848241849241850241851

What makes my stuff different from what's already out there is the puzzle piece design. With it, the terrain can be easily broken down for storage and transport, as well as mixing styles and build different structures.

This is a 2 story stone design:

241855241856

It also breaks down into two of these:

241857

Now I just need my own laser :D

[Larry Folk]

Walt,

Here are some photos of what I do with my 75 watt Epilog.
I am an architect and I make models for builder clients. These are at 1/4"=1'-0" scale

Workflow

Draw everything in SketchUp.
Export parts as .svg files.
Open .svg files in Corel and nest parts onto a 12"x24" page size.
Laser cut and vector engrave.
Siding and shingles are 3d engraved on "Polyback" material and painted.
Windows are 1/8" white acrylic.


Larry

[Glen Monaghan]

Larry, do you route or otherwise work the edges on ridge lines to make them fit together, or just leave gaps and cover with shingles?

-Glen

[Walt Langhans]

Wow! That's very impressive Larry! Thanks for sharing :)

[Larry Folk]

Glen,

I cut these out of 3mm Baltic Birch for the whole structure.
For the roof pieces I take the piece to my belt sander and bevel the edges to get a pretty tight fit.
At the valleys I use a small strip of metal tape to represent the valley metal trim to cover the seam between the wood pieces.

The shingle caps are just strips of the shingle material to cover the seam. Just as in the real world.

In this photo you can see the valley metal piece and the shingle caps better.

[Walt Langhans]

Larry, why Baltic birch and not MDF?

[Larry Folk]

Walt,

I just started with Baltic Birch but I like the strength verses weight and I get it locally from a woodworking store for $0.75 sq. ft.

I actually have not used MDF ever for anything.

I also use it for the large scale R/C aircraft that I design and kit. These are from 1/4 to 1/2 scale of the full size.

Again I use SketchUp to create this work also.

[Walt Langhans]

Very Very Cool!

[Glen Monaghan]

larry,

How do you deal with getting proper shapes/fit and alignment among all the myriad parts that aren't simple parallel or orthogonal joins? I've played around a bit with doing a relatively simple doll house in Sketchup and found it difficult to get things to fit together (but then, I haven't had much experience with the software...)

-Glen

[Bruce Dorworth]

There are good shetchup tutorials all over the net..

Bruce

[Larry Folk]

Glen,

I don't know how much experience you have with SketchUp but after you start using it a lot lights will go off in your head and you will see a work flow develop.
There is a forum called SketchUcation (just Google it) where you will find hours of information, just like Sawmill Creek.

I have some of LaserBuzz's designs and I took the Humvee and redrew it with SketchUp and was able to use a plug-in that gave me the exploded view in the attached photo. This always helps to see things.
Also make sure as you make each part you make it a component. Otherwise everything sticks together as you draw.

I have used many cad software packages the last 30 years and SketchUp is probably the easiest to use and you can make anything with it. Great for making things with the laser.

Larry

[Glen Monaghan]

I've only spent a few hours stumbling around with Sketchup (as well as three other 3D packages that I've been looking at to help with some 3D designs that are difficult to complete and fine tune in 2D coreldraw). I've read a few tutorials to get started but hadn't come across anything that really helped with doing something like these 3D laser projects with tabs and slots, odd angles, etc. I'll definitely check Sketchucation.

I have one laserbuzz design as well, which needs cleaning up for my tastes, and coreldraw just isn't the right tool for that job because it's too hard to verify inter-part 3D dependencies. Later, I saw the design in their booth at NBM Indy and you could easily see some of the more obvious fit/alignment problems I'd encountered. I emailed them about some of the problems I'd found when assembling their pattern and got something of a brush-off with the suggestion that I'd done something wrong since no one else had complained about them.

-Glen

[Larry Folk]

Glen,

I got the same response when I showed them the Humvee reworked. I cut and built the large doll house and had a much harder time assembling it than I should have.
I couldn't imagine in a million years trying to do this with Corel. Just the wrong tool for the job.

Larry