Where Did My Raster Detail Go?

Suppose I had a nice crisp bitmap file that I wanted to raster engrave. Say, this one:

original

But when I print it to RetinaEngrave3D, the fine details are gone and the lines are too skinny! What the heck happened?

The original swirls look nice and black and smooth to the eye, but at the pixel level that’s not exactly the case. Gray pixels have been used to make the curves appear smoother. This is called anti-aliasing.

The solution is to adjust the B/W Threshold slider in the Raster Properties box to treat more of those gray pixels as black, until the effect is what you want.

Also, if you raster engrave a design like this at low power and high speed, you’ll find that the vertical and near-vertical lines are stronger than the horizontal and near-horizontal lines. This is because the laser unavoidably puts out an extra burst of power whenever it turns on, which it does at the left and right edges of any vertical feature as it rasters back and forth horizontally. This extra burst emphasizes the edges of vertical features. To minimize the visual impact of this effect, increase the power or decrease the speed, at least enough so that horizontal features are strongly engraved. The vertical features will be engraved a bit deeper, but that will be much less noticeable.

Combining Raster and Vector

It’s often useful to have both raster and vector data in the same job. For instance, if you want to create an engraved medallion with your logo on it, you’ll probably want to raster engrave the logo (or parts of the logo) and also vector cut out the shape of the medallion (i.e., the outside circle). RetinaEngrave3D requires you to be a little careful with colors when you want to do something like this.

Whenever you print an image from, say, CorelDraw to RetinaEngrave3D, all the features in your drawing will try to show up on both the raster screen and the vector screen (unless you’ve told RE3D to ignore raster data). Your task is to arrange for only the features you want rastered to show up on the raster screen, and only the features you want vectored to be vectored. The two screens behave differently.

The Raster Tab

On the raster screen, everything you see in the on-screen preview will be rastered. There’s only one control that can change what you see: the slider marked “B/W Threshold” in the “Raster Properties” box on the right side of the RE3D screen. When you’re working with a grayscale photograph, this slider is used to set the level of gray that will be treated as black. Everything lighter gray will be white, and everything darker gray will be black. If you set it to 255, the whole screen will be black; if you set it to 0, the whole screen will be white. If you set it somewhere in the middle, light colored features will be hidden and darker features will be visible. The secret here is that colors are treated as shades of gray, too. Yellow will be treated as a light gray, while darker colors like blue will be treated as darker shades of gray. Black, though, is always the darkest color. So what you need to do is make sure your raster data is true black, and your vector data is some other primary color(s). Then in RE3D, on the raster screen, simply adjust the B/W Threshold slider until all the vector data disappears, leaving just the raster data visible.

The Vector Tab

On the vector screen, you don’t get an actual preview. Instead, the color of each feature in your design is important. Each primary color you have used in your design is automatically assigned a line in the “Vector Layers” table. That includes black, which is what you used for your raster data. If you designed your raster data in a vector program like CorelDraw, chances are good that some parts of it will show up on the vector screen as well. Most likely, you’ll see an outline of your raster features, which you probably do not want. Since you drew them in black, they will all be controlled by the settings you put on black’s row of the Vector Layers table. If you set the “Passes” column for that row to 0, all black vectors will be ignored. They will still be visible on the screen, but they won’t be used when the job is run on the laser. Perfect!

Of course, for this to work, you have to make sure all the features you do want on the vector screen are a color other than black. Stick with the primary colors (red, green, blue, cyan, magenta, yellow) since that’s what RE3D is looking for. (Also, remember that you probably want these features drawn with zero stroke width, which CorelDraw calls “hairline”.) Again, each color automatically gets a row in the Vector Layers table. Set the order (if it matters), speed, power, and passes (usually 1) for each of these colors, according to what effect you want.

Raster then Vector

Once you have all this stuff set correctly, you can use the “Raster then Vector” mode (selected by a pull-down near the top of the window) to run both parts of your job. Note that switching to the Vector Cut tab in the main window resets this pull-down to Vector Mode, and likewise switching to the Raster Engrave tab resets it to Raster Mode. So finish your checking of both screens before selecting Raster then Vector mode. If you forget, you’ll get raster or vector but not both. As long as the material doesn’t move, you can recover by simply running the other mode manually.

Separate Raster and Vector Jobs

Sometimes it just isn’t convenient to use colors this way. For example, if your raster data isn’t pure black and white but also contains some grays or colors, you’ll want to adjust the B/W Threshold slider to get the best-looking treatment of those non-black colors. That may not be the same threshold setting you’d need to hide all the vector data. You could always go into a bitmap editing program like Photoshop (or Corel Photo-Paint, which we do have installed) and alter the raster data to be pure black and white, but that’s an extra hassle. Instead, you might find it easier to just handle the raster and vector parts of your job separately. You can put them into separate files, or from a single file you can temporarily delete the parts you don’t want, then print to RE3D, and then undelete them. Or, at least in CorelDraw, you can select the parts you do want, and in the Print dialog check the choice marked “Selection” in the “Print Range” box to print just those parts.

There’s a catch, though. If you print the raster and vector separately, they might not be lined up the way you want. Recall that when you print anything, RE3D finds the smallest rectangle that encloses everything you printed (the “bounding box”), and shoves that rectangle up into the upper left corner. If you print raster and vector data separately, each will have its own bounding box. If you’re not careful, they won’t be the same, and your raster and vector data won’t line up properly. There’s a standard trick to solve this problem: before you separate the raster and vector data, draw a yellow rectangle around the outside of your design (and don’t use yellow for anything else). Then include the yellow rectangle with both the raster data and the vector data. Yellow is a very light color, so your B/W Threshold setting will probably let it disappear. In the Vector Layers table, you can set Passes to 0 for yellow to eliminate it from the vector job. But in both cases, the yellow rectangle will be included in the bounding box, so everything will be lined up just the way you want.

(See also this Raster and Vector Rendering post, for a more visual explanation.)

How to Schedule Time on the Laser

[Updated. Change of plans as requested by Colab managers.]

There is now a short online form to fill out to schedule time on the laser, which you can find by clicking on Schedule at the top of the colaser.org web page. The form goes directly to a Colab volunteer, who will try to respond promptly (usually within a few hours). The volunteer will try to arrange for a Colab key holder to be available to open the building for you. If that’s not possible at your requested time, they’ll work with you to find a mutually agreeable time when you can use the laser.

The Google Calendar is now read-only, and gives you a view of what times have already been reserved or blocked out. You can view the calendar by clicking on Schedule at the top of the colaser.org web page, or directly by following this calendar link. On that view you’ll find a button to add the calendar to your own calendar, if you wish.

The Google Group I previously set up to provide access to the calendar is no longer needed for that purpose, since the scheduling volunteer will update the calendar.

Please schedule the time you need, as far in advance as you can. The Colab key holders are happy to arrange a convenient time for you, but they can’t be expected to show up right away without advance notice. If you’re already at Colab and the laser isn’t already scheduled for someone else, you can ask any key holder who happens to be available if they can let you use the laser.

If you can’t make it at the scheduled time, please do your best to let us know. The scheduling volunteer will let you know who to contact if you need to cancel.

In the future, we hope to have regular open hours when you can be sure Colab is available, and a fully automated way to schedule time on the laser during those hours. In the meantime, and for hours when Colab wouldn’t otherwise be open, going through the scheduling volunteer lets you get time on the laser and access to the building in a single step.

Crud on Autofocus Probe

If the autofocus probe is covered with black tarry goo (see photo) at the end of your job, you’re cutting something that makes too much smoke and soot. If you can’t adjust your settings to avoid this kind of smoke and soot, please avoid using that material again. That same crud will land on the optics and cause another expensive repair.

Contaminated autofocus probe.

Contaminated autofocus probe.

Third Class Scheduled

A third class for this month has been scheduled for Thursday, November 20, at 7pm at Colab.

Please email me paul@mustbeart.com to reserve your place in the class. Because the laser room is small, the class size is limited to 6.

Expect about 90 minutes in the classroom, followed by practical hands-on exercises in the laser room. Once you complete the class, you’ll be able to schedule time on the laser to do your own projects.

Update: this class is full. Waiting list open. When would be a good time for more classes?

Classes scheduled

Basic laser operation and safety classes are scheduled for this coming Sunday, November 9 at noon, and the following Thursday, November 13, at 7pm.

Please email me paul@mustbeart.com to reserve your place in the class. Because the laser room is small, the class size is limited to 6.

Expect about 90 minutes in the classroom, followed by practical hands-on exercises in the laser room. Once you complete the class, you’ll be able to schedule time on the laser to do your own projects.

Update: Thursday’s class is full already. I will accept a few wait-list entries. More classes will be scheduled soon.

Laser is Fixed!

The replacement mirrors and lens arrived this morning, and I got them installed and working this afternoon. The laser is functional again, maybe even a little better than before. Let the artmaking resume!

I replaced mirror 2 (which might have been OK), mirror 3 (which was trashed), and the focus lens (which was destroyed). Did a full alignment on the optics, and a rough spot-check on the table leveling.

People have noticed that the laser seems weaker on the lower right and works best in the upper left region of the work area. It’s normal for there to be some small differential, because of the longer beam path to reach the lower right, but it was much worse than normal. I think it’s improved now. A cut that worked at speed 17 in the upper left required a speed of 13 in the lower right.

The new lens has a slightly different focus point than the old lens, for whatever reason. I have adjusted the autofocus probe to what seems to be the best focus point for the new lens. If you prefer to measure, you’ll find the new focus point is several millimeters longer than before. That is, the table is a bit lower with respect to the cutting head.

People have had trouble cutting half-inch plywood. I was able to cut a small square out of half-inch plywood today, in a single pass at speed 17 (or 13 in the lower right corner) at 100% power. There was some charring on the edge of the wood, but nothing that wouldn’t sand off. The only trick I used was to focus a bit down into the material, instead of at the surface. The Z axis setting I used was about where the autofocus probe touches the material, but that was just an educated guess. You’ll still need to experiment with settings and focus heights to find a configuration that works with your specific plywood.

I still think we can get better cutting of thick plywood (half-inch and maybe thicker) by getting a stronger air-assist compressor. I’ll look into that soon.

I hope to see you and your laser-cut art at YOUtopia!

Laser is broken

The laser is offline for repairs. Replacement parts are expected to arrive on Wednesday, October 8, and if all goes well I’ll get them installed on Thursday. Apologies to those of you planning to use the laser in the next few days for last-minute YOUtopia projects.

In the photo gallery below, you can see the damage. The zinc selenide lens has a hole blasted in it, and most of the surface of the final mirror (a few inches above the lens) has been ruined, with the worst damage in one localized area. These parts need to be replaced (about $300). Besides that, there’s a notch out of the aluminum mirror holder at the top of the cutting head. That won’t cause any problems, as far as I know. There was also a coating of reddish powder on the inside of the tube between the mirror and lens, which I presume to be oxides of the material blasted out of the lens.

I am not sure exactly what happened. Clearly, the laser beam itself did all this damage, but what went wrong first, and are all three types of damage the result of the same sequence of events? I don’t know. There are basically two things that can go wrong: gross misalignment, causing the laser to hit the holder instead of the mirror, or crud on the mirror and/or lens. I didn’t find any gross misalignment when I checked, so I suspect there was crud on the mirror and lens.

If there was crud on the optics, it was probably a build-up of smoke deposited by burning plywood. If there’s crud anywhere on the optics when the laser passes through, it gets hot, possibly burns, and damages or destroys the optical surface it’s on. Once the optics are damaged, the laser beam can bounce around unpredictably and cause additional damage. Once the surfaces are no longer pristine, the heat left behind by the laser beam causes the damage to accelerate quickly.

Put this down as a lesson learned about keeping the laser’s optics clean. I will try to do more preventative maintenance, but every laser user needs to be aware of the issue. At the beginning of your laser session, shine a flashlight at the mirror in the cutting head (look sideways at the left side of the top of the head) and see if there’s any visible crud on the mirror. If you see any, you can clean it (if you know how to clean it without causing damage!) or stop and report the issue.

If you end up making a lot of smoke during your laser session, please inspect the mirror again at the end. Make a note in the log. Post a note to warn the next users. Let me know if the mirror needs cleaning.

Unfortunately, the lens can’t be inspected without taking things apart, so I don’t recommend you do that routinely.

If you notice the laser isn’t behaving as expected, please let me know directly. Email is best (my address is on the wall). Make a note in the log book, too. Sometimes we can catch a problem and fix it before it causes permanent damage to the optics.

Maximum Material Width < 48"

I thought I could fit a 48-inch wide piece of material onto the laser’s knife table (removing the honeycomb table that usually sits on top of it). After all, it’s supposed to be a 48×36-inch laser, right?

Actually, no. If you lower the Z axis a bit, the 48″ material can be fit in between the walls of the chamber. But when you raise the Z axis to focus on the material, it hangs up on the lip surrounding the table. This makes a horrifying noise, and very likely screws up the leveling alignment of the table, and could damage the Z axis mechanism.

Try to avoid doing that!

The laser is more accurately described as a 1200x900mm machine. That works out to 47.24″ x 35.43″. And in reality, the X/Y positioning system can’t even quite cover that entire area. It can really do 47.24″ wide, but the Y positioner only spans 35.19″.

And if you have the honeycomb mesh on the table, which you usually will, your material has to stay inside the frame that holds the honeycomb. Those dimensions are 45.875″x34″.

All numbers above are approximate. Leave plenty of margin if you can.