Tuesday, August 24, 2010


selective laser sintering part 8: reciprocating laser cutter

Hi folks,

On Saturday afternoons, like any good grad student, I visit my parents place to do my laundry. While it's spinning around, my dad and I usually get up to some neat projects, and this is what we worked on this week.

I've had an idea for a while that one might be able to build an inexpensive laser cutter using a 1 Watt IR laser diode, as opposed to the larger 20-100 Watt CO2 tubes traditionally in something like an Epilog laser cutter, using a bit of a trick. Traditionally laser cutters have a lot of power, and (to my knowledge) are generally fixed focus, meaning that the focal point -- the hottest portion of the cutting beam -- stays at a fixed depth relative to the material that you're cutting.

Because commercial laser cutters have so much raw power, it doesn't so much matter that they're slightly out of focus at the top and bottom of the material -- even diverged, the beam still has enough power to cut. This gave me the idea that if one could dynamically adjust the focus of the beam, starting at the top of the material and moving it progressively deeper for a given x/y point along the cut path, one might be able to steadily "bore" through the depth of the material by having the most powerful part of the beam moving through the depth of the material. The advantage to this would be that one would be able to potentially use a much less powerful laser diode, and it would be much less expensive than a traditional laser cutter -- but the cuts would take much longer to create. Given that even inexpensive laser cutters are traditionally out of reach of most hobbiests or folks in the reprap community, this seemed like a potentially handy enabling technology. Working on the Open SLS 3D printer project, I also had the idea that since you have the laser there anyway for creating 3D SLS prints, you might be able to use the exact same device without the powder chamber create 2D laser cut objects -- essentially having a hybrid printer capable of 3D prints, 2D cuts, and 2.5D laser cut 3D objects (and any combination thereof). That seemed like an extremely fruitful thing to pursue, so away we went.

The test rig is beautifully simple. Borrowing from an idea my friend Trevyn at the local hackerspace had about using CD/DVD drive head mechanisms as very small CNC axes, my dad and I took apart a bunch of old drives from our junk pile and found the three nicest axes of the bunch. We stripped them down, attached two together perpendicularly to create a 2 axis table, and attached a third axis orthogonal to the table as the Z (depth) axis. Overall, the travel on each axis is very short -- on the order of about 1.5 inches (3-4cm) each, but this is more than enough for experimentation. The entire 3-axis device is very small, and fits in the footprint of a DVD drive -- infact, many of the structural components from the drives were used in the construction of the structural parts of the mini CNC system.

The laser from the SLS project was attached to the vertical axis, and the steppers (which seem to be 5V) were attached to the dsPIC and stepper controller I built for the SLS powder test rig. I wrote some simple code to move the table in a square pattern around 1cm on a side, and reciprocate the laser up and down at each point on the square-shaped path.

I loaded some flat black plastic into the device (I only had the back of a CD case -- about 1mm thick, and I'm not sure what the material is), and gave it a shot. To my surprise, it worked! The process was a little slow (it was probably cutting at around one inch per minute), but it worked -- all without any adjustments or calibration, on the first try.

I need to find some thicker material to test, as well as different material. I'd like to try black acrylic sheets, as well as black ABS sheets, if I can find them. For fun I taped a few sheets of the CD-case together, and it seemed to cut through most of two of them, but I have a feeling that the speed and depth will have to be adjusted for each material type and thickness. Pragmatically, I think my "reciprocating laser cutter" technique will have some limitations, and perform better when the focal length of the laser is large, producing a narrow, sharp focus, and reducing the amount of material around the perimeter of the hole that the beam has to melt to get to the "sweet" spot of the new focus, although this effect may be somewhat mediated at least in part by moving more slowly and in smaller steps down through the material. Through experimentation, I would be happy if I could successfully cut through 3mm material, and very happy with 4.5mm material -- thicknesses that start to become structurally useful in terms of creating 2.5D objects (like my entirely laser-cut linear axis ).

As a side effect of the boring process, one appears to be able to control the depth of the material that's taken off -- such that you will likely be able to mill to a specific depth in addition to cutting or etching, by simply adjusting the depth that the reciprocation process stops at.

Thanks for reading! :)

[part 1] [part 2] [part 3] [part 4] [part 5] [part 6] [part 6 video] [part 7] [part 8] [cogsci.mcmaster.ca/~peter]

Nice progress, I would suggest not just throwing anything under the laser, polycarbonate and pvc are both supposed to produce rather poisonous gas. And gasses that do damage to the machine optics

Acrylic ( plexiglass ) cuts great and requires the least amount of laser power of any material I have seen. It supposedly acts as a wave guide for the laser ( thats with the gas lasers, don't know about the different wavelength with the laser diode you are using )
And so fairly thick parts can be cut on low power machines ( I recall someone with a 50-100 watt machine getting good cuts in 2" thick acrylic )

I only know of cutting that uses an assist gas ( nitrogen for basically everything that is not mild steel ). The higher laser power and high pressure gas allows you to vaporize a narrow strip of material on the order of .010" and blow it through the work piece without transferring a significant amount of heat to the remaining material. I would think the trouble you may run into with such a lower power is too much heat transfer to the work with the result being sloppy cuts, more significant burning, starting fires, etc. It adds a lot of complexity to set up the pressurized assist gas system, but you might be able to keep things from catching on fire by running the laser in a box that you flood with CO2.

Diode lasers are great. I need to replace two small mirrors in our 1500W gas laser, they cost about a grand and I have to open the resonator chamber to replace them. Its a thirty thousand dollar part that would have to be thrown out if it got contaminated with air or anything other than the high purity bottled laser gasses we feed it with. I've seen some 20W diode bars on ebay for around $150, the prices seem to drop rapidly, and 20W is getting into the commercial laser engraver territory. It will be interesting to see what your work turns up. It seems like diode based laser cutters might be the next reprap/makerbot
Nice discovery!

How about of instead of moving the focus through the material at one point and then moving to the next point, you'd just run through the whole outline, then lower the focus, and repeat?

That way there would be less excess heat buildup as the material could cool down while the beam is elsewhere - although that might be a problem for the cutting process in the extent that it relies on melting the material away.
I think a problem will occur when you try to cut thicker material. The cone of the beam will strike the surface but will be too out of focus to melt it, so it will just be obscured, reducing the power at the focal point.

You would have to cut a V shaped channel to match the focus cone in order to maintain the power at the bottom of the cut.
I agree with Nop on the V-shaped-cut/focus-cone thing, though presumably with the right optics you could make the cone's angle pretty small, and hence increase the depth.

But even being able to cut 1mm thick plastic is pretty impressive. Neat!
I suppose you could tilt the laser so one side of the cone was vertical and rotate it so it always had the vertical side towards the object. That way your cut out object would have straight sides and the only downside of the V would be increased kerf.
Hi Peter,

how big is your spot and which width is the resulting cut?

I have a 1Watt-diode with atached 0.05mm fiber and a smallest focus of 0.05mm too in a distance of maybe 30mm, so the resulting cut is depending of the moving speed below 0.1mm but the melting range is bigger.

With a 5Watt-diode with atached 0.1mm fiber i have a cutting focus of around 0.2mm in a distance of 40mm, so the 'clean' cut can be deeper than with the 1Watt-diode.

If you use a lens with much higher focal length, then the focus will be bigger, but the cutting depth will increase accordingly ...

For higher parts you can laminate thinner sheets - so maybe you it will be more interesting to focus on laminating thin sheets, than cutting/engraving thick blocks?

Thanks everyone, these are all very good points. Someone mentioned that if the focus were dynamically adjusted instead of the distance, then one could start with a shallower cone and gradually reshape it into a steeper cone as the focal length increased. That should limit the amount of the beam obscured by the edge of the cut, and allow greater depth.

It will be interesting to experiment with many of these ideas.
Why not use the voice-coil equipped lens already on the DVD sled? It would much faster and while you're at it you could incorporate feedback to keep the spot minimized at all times.
Please go remind your Dad how good of a Dad he is.
www.kickstarter.com/ could help you raise money to invest in buying different materials to test out. I think this work is very interesting and would be willing to donate a few dollars to the cause (if the outcome were formalized results or build instructions it could really spark interest!)
Hey, I like what you did using the cdroms.

I blogged about it
Based on a diagram of a Chinese laser cutter I saw on ebay, it seems the laser is usually located far away from the material being cut, going through a window into the cutting chamber, and then bouncing off a couple of mirrors on the way to the target. So I assumed it was a collimated straight beam rather than being focused to a point, but I could be wrong of course; or do they usually put extra optics on the cutting head to re-focus the beam then? I thought the reason is to avoid having splatters and smoke foul the laser optics. You'd tend to have a hard time with that, using your design, right? But the smaller the focus point, the more it diverges at distances past the focus, so you couldn't get a really fine point if the lens is far away, right? What do you think is the limit, it would be cool if you could have a laser/lens assembly which causes the beam to come to a focus about half a meter or so away, before diverging again.

But this is really cool exciting stuff, makes me want to try it out as well.
CO2 lasers bounce the parallel beam (which is about 6mm diameter) around with mirrors and then focus it with a lens after the last mirror to about 0.2mm.

The longer the focal length the better the depth of focus is, but the spot size is smaller with a shorter focal length.

So it is a compromise between spot size, depth of cut and splash back onto the lens.

With diode lasers the power is less so a smaller spot size is needed to get the power density. That means a shorter focal length so less cutting depth. Being close to the work the lens will be vulnerable to splash back so a jet of air blowing towards the work is a good idea.
An excellent faq about focusing and spot size here.
This is a beautiful piece of hardware hacking art. I enjoy following your work.

I wonder if your method actually requires removal of more material. You remove more of a cylinder rather than an hour glass. Also, as others point out, the edges of the cone need to do more cutting. Even a perfect TEM00 spot has a significant power density fall off at the edges.

Blogged it here.

The calculations on the parallax website have been turned into an interactive calculator here...

I have a very similar machine at home, with the same diode (1000mw 808nm). I don't reciprocate the z axis, rather make multiple passes (like 10!!!).
It can cut through 3mm mdf at the rate of about 1cm / 3 minutes.
It's very slow and I'm tempted to try reciprocating.
Has anyone else had experience cutting anything other then plastic with a laser this low powered?
Interesting idea on moving the focus up and down. Instead of doing a mechanical motion to move the focus, how about a liquid lens controlled type of focusing such as the device from Varioptic. I suppose the only issue might be trying to keep the liquid from getting too hot, but it might be manageable.
Great! But where is the video of it in action?
The "traditional" method of laser focus does not put the focal point in the middle but at the top of the material. This prevents much flareup when cutting. Even when taking multiple passes with a higher power machine the Z is not moved on the recuts.You waste power on the edges of the cut where the beam is not hitting as you move down.

Polycarb does not create poisonous gasses. There is nothing in it that can do that. ABS is probably what you were thinking of, it can create cyanide. The problem with polycarb is it tends to flare up big time.

Realistically you need about 40 watts from a CO2 to cut 3/8" acrylic. I have cut up to 1-1/4" acrylic with a 80 watt laser. 25 watt you will be pushing 1/4" at any useful cutting speed.

But doing anything like this is impossible with a 1 watt laser. Especially in the frequencies that diodes are available in. With the 10.6um wavelength of the CO2 laser just about everything absorbs it. With a diode laser you are limited at high power to a range around 808nm as they are used to pump YAG rods to create 1064nm or 532 and 355nm by doubling or tripling.

Also with CO2 you get a waveguide effect when cutting acrylic. As the material vaporizes from the cut the walls form a wave guide that reflects and keeps the beam collimated allowing it to cut through much thicker acrylic per watt than other materials.
This comment has been removed by the author.
Hello Peter,

please, could you tell us a bit more on how you did the CNC? Do I understand it correctly, that Z axis is fixed (i.e. it moves only in one direction to change its height), then the "table" (CD case) moves? But how does the table move? Does one motor move the whole construction of another motor? In other words, isn't the other one construction too heavy to be moved by the first motor? And also, I assume the accuracy should be pretty exceptional, as the "data spiral" of a CD is quite subtle?
Hi everyone,

Thanks for your comments! Now that the site's back up from its slashdotting, I'd like to take a the time to answer the questions that I can.

incoherent: it's true that a cd/dvd assembly does contain an adjustable lens, usually driven by 4 or 6 wires. there are a few reasons why it would be non-ideal for this situation: (1) the focal length is *very* short, typically only a few mm, which would limit your ultimate depth of cut, (2) aligning an invisible laser diode to use that assembly would be very tricky, and (3) controlling it would be a little tricky. you're definitely right that you could move that little lens pretty quick, but the time that it takes the laser to cut the material at a given depth will be far longer than the maximum travel speed on your axis. (for example, the z-axis on the current system can move much, much faster than the minimum time required to make a successful cut -- or in other words, the axis speed isn't the limiting factor here, it's the laser power. and that's the whole point ;) using that in a power/time tradeoff).

seamus: I will do that today, when I bring a fresh bucket of laundry to him. :) Thank you.

Evan Bowling: I had thought of that, but every moment I spend trying to raise money for the project is another that I can't spend working on it (or, my thesis ;) ). Artificially limiting yourself to having limited resources really tends to spawn innovation by necessity, as in the case of the all laser cut linear axis I designed a few weeks ago: http://builders.reprap.org/2010/07/changing-game-and-using-kapton-tape-as.html. I expect the entire project will only cost on the order of $1000, which really isn't all that much. That being said, if someone would like to donate to the project, you're welcome to e-mail me to talk about the possibilities.

fjr: thanks! I read your post, and it sounds like you have all the parts to build your own mini-3-axis system in an afternoon some day! (you might have to take aparte a few more than 3 cd/dvd drives, and take your pick of the best axes).

ecloud: I think nophead has answered your question below. Basically the way that a traditional laser cutter works is to have a collimated (ie. straight, non-divergent) beam run along the axis, until it hits the 'cutting head', where it is focused downward to a point to produce a high energy-density region that can cut through some materials.

bdring: Thanks! I actually had a lot of fun playing with the design and keeping the coloured spacers on -- it was fun to design, and to photograph.

This reciprocating method definitely removes more of a cone than an hourglass, where a bunch of cones on top of each other end up looking like the cylindrical cut that it's made. By adjusting the focus of the lens instead of the height of the entire laser/lens assembly, one could progressively "sharpen" the shape of the cone as it progresses in depth through the material, and the resulting cut would look something like a \/ with slightly convex edges. This would likely get you more penetrating depth, by removing more of the outter edge at the top, and I think it's a good thing to try next. (It was just easier to move the whole thing up and down for a first test, rather than to make a rig that will slowly rotate the tiny focusing lens screwed into the laser housing).

Definitely good thoughts.
Al: Wow, that's a neat accomplishment! I wouldn't have thought it would do 3mm MDF. At 1cm/3 mins it would take all day to cut anything of size, but it's still neat to hear about :).

Technical Blogger: There are a bunch of different ways you could control the depth of focus, or the entire mechanism. The idea here (and with the Open SLS 3D printer project) is to make things inexpensively, so that they're accessible to a larger number of folks.

Earthshine: I'm actually afraid that the laser scatter might damage the image sensor in my camera, so I didn't take a video of it in action. I would really like to post a video though, and have been thinkinf of using my cell phone as an alternate camera (although the quality is quite poor, so I haven't tried yet in favor of thinking of something better). I had thought of sticking the laser glasses onto the camera, but then what would I use? ;)

macona: Thanks for the information -- I didn't know that acrylic forms a waveguide for 10.6um. The laser cutter I've used for much of the rest of the project is 40 watts, but being an inexpensive Chinese laser it doesn't have an adjustable focus mechanism, and our focus is set (I think) somewhere inside the material itself. It's interesting to note that the focal point is typically at the top of material in a traditional system, I'll have to see if we can adjust ours at the local hackerspace.

ecir: If you have a look at a video of the "Makerbot" in action (there are likely a bunch on youtube ;) ), it might better illustrate the idea of a moving table with a z-axis overtop. Basically the bottom axis of the table (call it X) does have to support the weight of the one ontop of it (call it Y), and you're definitely right that this does require some extra torque. The axes themselves are quite light weight (especially after all the optics are stripped from them to make room for mounting hardware), and the little steppers seem to do pretty okay at it.

The question of precision is an interesting one -- different drives seem to have different pitches on their lead screws, and have varying amounts of precision. In my (quite limited) experience they tend to have anywhere from 150-300 steps of travel across their 3-4cm range, which seems quite poor given the number of tracks that must be on a CD or DVD. I imagine the steppers may be microstepped, and that the coil on the focusing lens may allow it to move side to side, as well. Hope that helps! :)

Thanks for taking the time to write these comments, everyone!
Thanks for the answer!
this company :

has good powerfull lasers and a nice collimating/beam expanding/contracting lens system :


this beam expander can arrange the beam divergence, so, it can easily make the laser able to burn at longer range

Blue laser (compatible with the expander) with 1W

Im planing on builing a cnc laser cutter just like yours, but using a laser from a dvd drive.

Apperently the laser is 400mw and to keep it cheaper im not going to use z axis and try maybe manualy focusing the laser.

Do you have any recomendations on which driver and holding can i use?
also how do i calibrate everything for better cutting?

i dont mind not being able to cut most of the things, the first im only going to use to cut foam, but later i may get myself a 1.5 - 2w laser.
Ok, this is way ro big of a coincidence. "My friend Trevyn" and being from "Toronto", is this Peter of Thinkhaus fame?
glad to see your posting. Wish that we had something open in Hamilton.

Keep up the good work!
Post a Comment

Links to this post:

Create a Link

<< Home

This page is powered by Blogger. Isn't yours?

Subscribe to
Posts [Atom]