Laser Scanner for 3d Printer Application

Since this is my first posting on these forums I would like to start out with a big hello and introduce myself!

I am an all-around tinkerer and electronics enthusiast by night/weekends, and Electrical Engineer by day. I have a pretty strong background in robotics and volunteer my services for a local FIRST Robotics team(if you have a team nearby you should definitely check it out, really cool program for high school and adults alike!).

I have been looking into this 3d printing craze for some time now and I am really excited about the future of this technology. However, I am a bit disappointed with the resolution of the extruder type machines. I feel like the real breakthrough will come with the UV curable resin type machines. These types use a resin that cures with UV light (385nm), and mostly use a DLP type projector with the UV filter removed. I would like to see if it is practical to implement this type of design with a laser setup instead. My uncle worked on some of the first 3d printers back in the day and they all used a laser based system, but until recently the resin they used was hard to find and expensive, not the case today.

I have done a little bit of research and I feel like I am beginning to understand what equipment may be required.

1) A laser that outputs in the UV Spectrum
- I assume (and we all know about assumptions) that a 405nm diode that could possibly work because it will output some energy in the 385nm region, and is considerably less expensive than a 385nm diode.

2) Scanning Galvanometer and power electronics
- I am looking at a 30kpps galvo because the images I want to display will need to be faily detailed. I will need to use a lot of the available angle because the target will be close, a few inches, to the output.

3) Optics to reduce beam width
- I am not sure about this one. From what I have read, in order to get a nice small and clean 'dot' (not sure of the correct word here) I need to expand the beam, then focus it, send it through the galvo, and then through an f-theta lens so that it can focus correctly on a planar surface ( as opposed to the inside of a sphere). This will be important because since the beam will cure the resin, it needs to stay sharp no matter where it is on the target plane.

4) DAC to drive the Galvo
- I will start off with a soundcard DAC but will most likely develop a custom solution for this in the future, after all what fun is a project if you run out of improvements?!

To help you help me I probably need to post a few goals of this project:
1) Beam width at the target: .1mm
2) Beam wavelength: 385nm – 405nm
3) Beam power: Unknown at this point how much power is required to cure the resin.
4) Distance from target plane: 6 inches

I know that some of this information is probably already on this site, but I am hoping that some of you may be able to point me in the right direction. If any of my expectations are unreasonable please let me know so that I can adjust.

The problem I see with such a system is that you are effectively taking a system designed for vector graphics and using it for raster scanning. The image would be redrawn quite slowly to achieve the resolution you want. A 405nm diode running at 600mW or so spread over a large area isn't much power density. An argon or third harmonic generation DPSSL would be orders of magnitude less efficient to get a lower wavelength. Add in the issue of optics to correct the dot size and you will quickly end up with something significantly more expensive than a DLP projector with the UV filter removed.

I see what you are saying, with this type of system though I wouldn't need to necesarily display the entire image at once per layer. I could probably get away with 'printing' out the image one line at a time similar to a 2d printer. The image I have in my head is similar to what I have seen on this laser engraver:



That engraver is a lot more powerful than what I need, but the concept is similar.

As for power, like I said before, I am still investigating how much power is needed to cure the resin, but those that have used the resin say that a little while in the afternoon sun can cure a cup full of the stuff.

I know that I am not going to be able to do this cheaper than the DLP variants, I just want to do it with lasers for fun, surely you guys can understand that?

I know it was brief, but am I off target on the method to reduce the beam size? I'm getting some of my info from here.

Okay, the embeded video did not work, link.

The galvo system might be nice to complement a DLP system (not replace) for explicitly drawing vector graphics with a variable pen size. It could also useful for refining/smoothing the edges of the DLP-exposed mask when the etching/curing is not perfect. Then the beam will only be working at the edges, not doing fill-work. That kind of thing is used for optical waveguides that are fabricated using lithographic methods that don't produce walls that are smooth enough.

As for the hardware, as you will probably not be running this at a very fast rate, you can probably get away with a lower-speed, and lower-cost galvo system. You don't even need to use a sound-card DAC, but could instead rely on a microcontroller's DAC or something of that nature, again, because you're going to run this at a slow rate.

One thing to consider is that most of the laser beams will be quite large for that kind of work, about 3mm in diameter. If you want to focus such beams, it'll still need to be beam-like enough to be aimed by the galvo. So you'll have to experiment with the laser and lens positioning.

Finally, before you do any of this, get a pair of protective goggles. You'll be using a lot of reflective optics and need to do visual calibration. Having a good pair of goggles will protect you as well as ensure that you can see what you're doing. I suggest spending a little more on some quality goggles that have a high VLT, but good UV/violet protection such as the YLW or ARG goggles at OEM Laser Systems. I'd prefer the former for your application, but OEM's site is being updated and I couldn't find a link to that specific product.

Interesting idea.

I bet if you used a laser you wouldn't even need to fill the entire solid. You could do the outside edge and a cross lattice pattern (like #) through the middle of the part to save time, and the UV light in your house or outside should cure the liquid left inside pretty easily. Or you could have focus adjustable by servo and use a wider beam to fill the inside.

@Bionic_Badger:
Thanks for the advice on the goggles, I knew I would need something but wasn't sure where to start looking.

I think I see what you are getting at by using the laser as a supplement to sharpen the image made by the DLP but I would rather do one or the other and lasers seem like it would be more fun (challenging).

You mention that this project would not necessarily need a soundcard DAC and a microcontroller DAC could suffice, which is a plus. I assume then that the galvos operate like an analog voltmeter, where the mirror will move by a magnitude proportional to the input voltage. So that means, and I am thinking out loud a little here, that my X-Y coordinates would translate to my +/- V galvo inputs. That would make for a nice open loop setup, but some of the galvos I have seen claim to be a closed loop system. Do some of these galvos in fact have a feedback signal that I could tap into for a control algorithm? (Sorry if this seems a little elementary, I just want to be sure that I completely understand the process.)

There are many programs out there that can take my solid model file and 'slice' it into thin layers to be built one at a time. Then there is another piece of software that will take the slices and produce what is called 'G-Code' or a machine code to give commands to the NC machine. These commands are traditionally things like turning the Spindle on/off or tool changes but a custom interface board could take that to mean anything, like turn the laser on/off. The really neat thing it does though is to create a route (X-Y-Z coordinates) that the traditional NC machine would take to cut out the slice. This means that I wouldn't have to fool around with converting bitmap or vector images which I believe would be an advantage.

I have a question about 'it needs to be beam-like enough to be aimed by the galvo'. I am a total newb when it comes to lasers, and it’s been a while since I took an optics course, but I take this to mean that the laser needs to have limited diversion. Is this right? I think that the way I would do this would be Laser->Beam Expander (Galilean)->Galvo->Scanning Lens (F-Theta)->Target. I am sure that I will need to experiment with the positioning but I know that the math has to be somewhere and with CAD I can design a fixture to contain everything in its proper place.

@benmwv:
That is what I was thinking for designs that need to be solid; otherwise I would leave a small hole in the part somewhere that could drain the liquid out of the inside.

Hi,
My 2p worth.

Using a laser with focusing optics in order to provide a small spot size, you will need to compensate the focus from the centre of the scan field to the outside (as the distance increases / changes) to ensure the spot size remains constant.
You can do this by a secondary lens moves by a 3rd galvo ( geometric corrector) or using a flat field lens.

ATB
MM

Jimbo2032 said:

@Bionic_Badger:
Thanks for the advice on the goggles, I knew I would need something but wasn't sure where to start looking.

I think I see what you are getting at by using the laser as a supplement to sharpen the image made by the DLP but I would rather do one or the other and lasers seem like it would be more fun (challenging).

Click to expand...

It's worth a shot, and you can always try and make it into a vector-based lithography tool.

You mention that this project would not necessarily need a soundcard DAC and a microcontroller DAC could suffice, which is a plus. I assume then that the galvos operate like an analog voltmeter, where the mirror will move by a magnitude proportional to the input voltage. So that means, and I am thinking out loud a little here, that my X-Y coordinates would translate to my +/- V galvo inputs. That would make for a nice open loop setup, but some of the galvos I have seen claim to be a closed loop system. Do some of these galvos in fact have a feedback signal that I could tap into for a control algorithm? (Sorry if this seems a little elementary, I just want to be sure that I completely understand the process.)

Click to expand...

Well the galvo drivers themselves are PID-controlled. If you really need direct control and feedback you might consider rolling your own driver such as this one. Virtually all the galvos use that exact same circuits (scroll to the bottom), but use a different form of feedback than the ELM-chan galvos (photodiode instead of capacitance). If you go that route you should probably still use a cheap galvo set (talk to hakzaw1) for the galvos themselves, and roll your own driver.

If you're working at small angles but need more precision, you might think about rolling your own galvos too. Something like ELM-chan's galvo system might work, or maybe a larger version of the traditional paddle-photodiode sensor.

I have a question about 'it needs to be beam-like enough to be aimed by the galvo'. I am a total newb when it comes to lasers, and it’s been a while since I took an optics course, but I take this to mean that the laser needs to have limited diversion. Is this right? I think that the way I would do this would be Laser->Beam Expander (Galilean)->Galvo->Scanning Lens (F-Theta)->Target. I am sure that I will need to experiment with the positioning but I know that the math has to be somewhere and with CAD I can design a fixture to contain everything in its proper place.

Click to expand...

I wouldn't worry too much about divergence unless you're planning on aiming at far away targets. At your target 6-inches, you might be able to get away with just focusing the beam at the source lens because the focal length will probably be long enough to maintain a beam-like cone. You should, however, limit the maximum angle of the galvos, as larger angles could cause minor differences in distance, and therefore spot size (though it may not matter in the end for your application).

Bionic-Badger said:

If you really need direct control and feedback you might consider rolling your own driver such as this one.

Click to expand...

Wow, that is pretty intense, and one day I might try to build one myself but for now I don't want to bite off more than I can chew and end up with another half finished project.

I've been looking around for the best place to start and I suppose I need to actually get a laser and start messing around. My first impulse is to get a bare diode and start messing around with designing a drive and power supply, but I am thinking that I may be better off just getting a ready to go module to get started as quickly as possible.

The module I had in mind was this one off eBay from newgazer. I have seen his name pop up in a few posts here and assume he is reputable. I figure that this laser will give me the chance to get in as quickly as possible so that I can begin experimenting with the galvanometer. What do you all think? This laser with a 20kpps galvo should give me a pretty good start (at least in my head...) Additionally, if the laser is as powerful as claimed, I should not have any trouble with curing a UV resin later down the road.

The same seller also has a number of laser goggles in his store, should I just get a pair that he has or go with a specialty shop instead?

That one might be okay to start out with; however, the lens is inside the heat-sink and it may be a pain for focusing to a smaller spot. Those galvos should also be fine. You may also consider just using geared-down servos too if you don't need much speed.

I would buy a better pair of goggles that have been tested by others. Most of those on eBay haven't been tested, and it is not known if they're the same type anyway. Perhaps these ones would be good. I prefer the YLW goggles at OEM laser systems, because it is really easy to see with them on, but OEM's site is broken and those cost about $100 a pair. If you still want those ones, maybe you can call them up and ask.

Is there a datasheet for the PHR-805T Laser Diode? I have looked around but I cannot find one to save my life, even a manufacturer would be helpful. You'd think that with how prevalent they are it wouldn't be this difficult.

Jimbo2032 said:

Is there a datasheet for the PHR-805T Laser Diode? I have looked around but I cannot find one to save my life, even a manufacturer would be helpful. You'd think that with how prevalent they are it wouldn't be this difficult.

Click to expand...

Diodes are often called by the name of the device they were pulled out of. In the case of the PHR diodes as we call them the part number "PHR-803T" came from the sled.

This has info on the PHR diodes: Keep it under 100mA and you should be fine
http://laserpointerforums.com/f42/meta-topic-pinout-performance-actual-ld-26417.html

What information are you looking for exactly?

Also, keep in mind the hotter it gets & the more current you put into it the higher the wavelength will be (up to 415nm~ IIRC)

I was hoping to see if it had a photodiode and if it did what its charactaristics are for use in a driver circuit.

I also just like to have datasheets on file for my projects for reference, habits from work.

There are rarely datasheets for the diodes we use, except maybe those Mitsubishi diodes people have direct access to. Anyway, most are operated well beyond spec so you should take into account that when setting the current, especially if you want a longer lifetime on the diode. Perhaps set the PHR-803T to output 40-60mW or so, rather than push it to 100mW+.

Also note that if you have a 405nm diode outputting in the 385nm range, then you a diode worth a few hundred diodes if not more! The point is, a laser has a small frequency bandwidth, so at most, you may get down to 395nm, and that will probably be on the tail of the gaussian distribution (if you cool the diode).

I can't see a 405nm diode working for a resin that cures strictly at 385nm. However, most optical properties occur in ranges, so I would suspect that it may work, but not because the diode is outputting light in the 385nm range, but range because the resin works up to 405nm.