RepRap and lasers

I know we have a few people here that are interested in lasers, and I've put some thought into this (I blame Fernando!). I'm putting down some thoughts.

The two most useful I could see being used with a RepRap is a CO2 laser and a nitrogen laser. TEA lasers actually. Google will provide a more complete answer than I can as to what that is.

The CO2 version would be great for doing cutting, but the problem with CO2 is that for a simple "flowing gas" version, you need an exacting mix (about 10% CO2, 20% N2, and the remainder He), while for a sealed version you'd need some very exotic additives. It's also extremely powerful and hence dangerous. Even though 10.6um is considered eye safe, I don't think any 25W heat beam is really all that safe. Standard polycarbonate safety glasses will act as a block at least.

A Nitrogen TEA laser would actually be far more useful. The main advantage for RepRap is that it is a simple 337.1nm wavelength laser. This is UV-A, so it should work well for spot hardening Fernando's polymers.

While tricky, they work off of welding gas N2 or even plain air. Cooled air N2 (from one design online I've seen) can do 120pps at 0.5J per pulse (2mm beam?) and the useful average power is 180mW. Most normal ones do a few pulses per second. The instantaneous power is something like 50KW due to the pulse width being a few nanoseconds. I think power is reduced by about 75% if you can get the same design to work on air. Many of the parts can be easily acquired or made via RepRap. No glass at all is required, but a rear aligned mirror will boost output by a factor of 4.

Now, for issues with a Nitrogen TEA laser. The common capacitor material is a little hard to come by (extremely thin double sided FR4). The voltages involved are hazardous in the extrme. Normal power for something like this is taken from a wall plug, through an auto transformer, through a HV transformer, then up a voltage multiplier to get a few milliamps at 25KV. This would be difficult to really get working well off a 12V power supply and not poison the area with high frequency noise (but not impossible!). Neon sign transformers are also QUITE popular. The beam shape of this is square, not tubular. That may not be too bad, though. These designs commonly use spark gaps, so there's a lot of UV that gets tossed around.

Darwin itself is well suited for an off-machine laser system. The planar XY system could easily mount redirection mirrors and a focusing lens system.

Other laser systems could be useful, but they're not something I think your average user could assemble.

I think 25KV at a few milliamps is about the spec of the EHT circuit of a large CRT television, and the circuit you describe is pretty much how a TV does it. I expect a lot of these are being replaced by LCD and plasma currently so it might be something fairly easy to get hold of.

You do have to be careful with X-ray emission at these voltages, not sure if lasers are affected but CRTs and valves are.

I have only ever touched a charged CRT once but I lived to tell the tale

---

[www.hydraraptor.blogspot.com]

I just remembered I have 14 20 inch colour monitors in my loft. Probably a bit low on voltage and current at that size though, but I could parallel them up for more current.

---

[www.hydraraptor.blogspot.com]

Wouldn't a optic fiber be the ideal vehicle for the laser beam?
SiO2 fibers and probably some polymer based ones are UV transparent and 1 or 2 meters of it would allow to have the laser away from the cartesian robot and shielded for UV, X-ray and electron beam radiation.

By the way, juts for your info, some polymer resins can be cured with electron beams. This is for high accuracy repraping, out of a vat of liquid resin, as the electron beam only acts on the very surface of the resin volume.

Man, this sounds so much like high energy physics

Presumably curing with electron beams would require the liquid to be in a vacuum?

---

[www.hydraraptor.blogspot.com]

Errr... that is quite a sensible remark... and a pretty big nail in the coffin for any RepRap candidate wanting to implement that tool

Edited 1 time(s). Last edit at 06/26/2007 07:24AM by Fernando.

So, what's next, EDM?

Anyway, my current favorite site to understand the differences in construction is this professor's personal site.

[technology.niagarac.on.ca]

Fiber optics may be the best choice for the laser, but testing will definitely be required. I'm curious about focus and energy loss. One advantage I do see is that you could have multiple lasers focused on a point or in a row, multiplying your fire rate (if you can figure out how to control your discharge point!) for inexpensive units if you're tying to duplicate SLA, or doing a laser dot matrix approach to continue resin curing.

Have you seen those acrylic blocks with a 3-D image burned in them? Multiple beams crossing, with sufficient power at the intersection, to vaporize the plastic, I suspect.

Something similar, using Fernando's UV polymers, might be a way to get vat replication reasonably cheaply. That is, assuming the density of the cured plastic doesn't vary from the density of the liquid, and the index of refraction and transparency remains constant.

I'm not advocating this. I think any of several other approaches are very promising, and don't know if there'd be any advantage to this. A few disadvantages, though. Namely, keeping the vat full for even the smallest, (x,y), object, problems with keeping an open topped container of liquid around, and not expecting to have to mop up the mess occasionally, and keeping the vat free of incidental filler materials. Household dust may in fact make an excellent filler material, but I wouldn't want it clouding up my objects.
Additionally, you wouldn't be able to see the result clearly until it was finished, so no pulling the plug half-way through a build if it was going south, and since you wouldn't be able to melt it down to start over, that would be that much material wasted after a botched build. If you could see the object, you wouldn't be able to continue to build through it, or you'd have to recalculate, with the current stage of the model, the path of the beam to know where it'd end up. A lot more math, and so much for simplifying the built side model.

Now, if you always built from the right to the left, from the front to the back, and up, you could position the lasers at the left, back, and top, and not have to worry about transmission through the completed part.

now thats a really cool and interesting idea. using two lasers that alone would not be strong enough to trigger the hardening reaction, but when they are crossing their combined strength would harden the material. man, i really hope we have some sort of working technology like that in 5-10 years. that is awesome!

That would be an awesome machine! It would be able to build very accurate pieces at a great speed.
But it's a completely different beast from actual reprap designs.
You would need very accurate laser positioning system, lots of mirror... wow, I'd love to see such a machine being built.... you know, laser dude!!

Added: The idea about crossing lasers might be a good one, but i don't see how that could work inside a given volume of UV set resin.
You see, the photoinitiator is precisely tuned to adsorb the UV wavelength so that the intensity of the beam decreases linearly with the path of the beam through the medium. The beam will react with the resin through the whole path and arriving at the crossing point, it will only have a fraction of the initial power.

I think that a better idea would be to make the beam actuate on the surface of the resin only. You could focus the beams very precisely on a point of the surface and make it travel along the layer, then drop the set layer down one level. The surrounding resin would cover the piece again and offer a new surface to act upon.

Edited 1 time(s). Last edit at 06/27/2007 04:20AM by Fernando.

i think the general consensus is that we definitely need FRICKIN lasers!!!

And Fernando has pretty much brought us back to a stereolithography (SLA) system, at least the one I've known in the past.

A high speed laser positioning system does take some expertise and time to make, but it's not impossible. We'd need some small galvanometers and controllers, but it could be done. There are some dedicated people out there making galvos at home for their own laser light show designs. The problem here is not the engineering, but to make it useful with the limits of our lasers. Higher speed needs higher power. If someone did go overboard and shoot for that 120pps 180mW design I outlined before, how fast could we go? If we focus to a 1mm spot and only need one pulse per spot with no overlap, we're looking at 120mm per second.

Also, given the nature of this tool (aka it'll work fine for some, reduced output for others, and might not work at all for yet others), every laser head will be effectively different. Output may even vary with weather conditions if an air laser is used (I'm planning on trying this).

It'll be a complex process of testing, but probably easiest. If a fiber optic head is made as Fernando suggests, a small optical bench for the optics will need to be attached and this can also base the future galvanometers too.

I think for the ease of implementation the fiberoptic head is the easiest one to start with. One could always start building a mirror based SLA (I didn't know these were called that way) later on, when you have a working reprap and can churn out parts on demand for experimentation.

The fact that the fiberoptic diffuses the light may not be that problematic as long as you keep the tip of the fiber close to the surface of the vat. You may even avoid vaporising the resin with to high a power per surface delivery.

How tunable is such an air laser? Maybe we should start documenting this somewhere, just to gather all the pertinent info in one place. it's been a while since the laser idea has been around and i think there a a couple of people or more that have a solid knowledge of what is needed to build one... here's a link to get you introduced:

[www.technology.niagarac.on.ca]

Yes, that's the idea. Oddly enough, the Darwin with some mild modification would also make a good SLA machine, but with reduced capacity. You replace (or modify) the MDF board with a build platform that drops down in the center on four rods to the bottom of the current machine. Remove bottom stiffening rods (will probably have to find a replacement option). Put the vat on a slide on the bottom of the machine if mounted on a table. This should give you about the same build area as the current darwin. UV resins shouldn't change density much between solid and liquid (assumption) so just a single reading of vat height at the start of each build would be needed. A corner "parking" cutout for the head would probaby be needed as the fiberoptics may need to be kept close to the surface. Again, this may not be the case at all, but the cutout would let to get the head very close to the liquid surface and still bring the build platform out of the liquid completely.

Half-height design would allow the entire setup to fit within the current Darwin frame for space constrained people.

As an air laser (or an N2 laser, oops ran out of gas!) can be finicky on power output, a measurement photodiode may be wanted on the head with override capability in case not enough power was fired for the designated pulse. This could also be used as a closed loop delay if the beam pulses were left "closed loop" with the laser controller releasing the next move (or beam stop) in the chain for normal building without firing control being done on the PC. The possible show stopper for the diode would be if it's fast enough to track the pulses.

Research research research...

I think I want to rep-rap some sunglasses...

I'm now wondering if you can get a phenomenon, akin to beat frequency, to allow the light to be UV only at the intersection of two beams, and outside the target wavelength elsewhere. Probably more trouble than it's worth. Especially as I've heard it's hard to keep an X-ray cathode smooth enough to generate a clean beam.

As for cooking with an air laser. If the range between not enough energy and too much energy is wide enough, you could just cook it for an average time, and if it varies by some amount, it would still not be enough to undercook, or overcook, it.

You might, I'm starting to sound like a broken record here, measure the lasers response to the given power to see if it did anything. Does an air laser have a signature power spike as it comes to full power? Can you measure the supply power, and get meaningful feedback? If so, if the energy didn't generate the desired spike, try again in the same spot.

Something else to consider...for maybe a liquid RepRap 2. TI MMA chips, and an ultraviolet source, or a transmissive LCD panel, and again, an ultraviolet source. Hit the surface with an "image" of one layer, long enough to cook the whole layer. Advance the platform, and repeat. This wouldn't be more precise, but you might get faster builds, at a suitable resolution, this way.

> Something else to consider...for maybe a liquid
> RepRap 2. TI MMA chips, and an ultraviolet
> source, or a transmissive LCD panel, and again, an
> ultraviolet source. Hit the surface with an
> "image" of one layer, long enough to cook the
> whole layer. Advance the platform, and repeat.
> This wouldn't be more precise, but you might get
> faster builds, at a suitable resolution, this way.

absolutely.. and you could even print out the holding tank for the solution in HDPE or something similar =)

You know, I think HDPE has such a potential for being the heavy duty printer of the reprap family. I picture big sturdy heads printing out quantities of material for building bulky parts. Here the errors due to the large shrinkage factor of thermoplastics could be kept negligible compared to the size of the whole part: vats, tables, panes, even furniture parts... anything large and sturdy!

Sean:

I don't know if an air laser does have a signature power spike. I don't think it would make it back past the flyback transformer to where it's safe to measure. In the HV side, the electrodes are directly connected to the DC discharge capacitors along an edge in most models. You either get a spark at a hot spot or an even glow of firing. The issue is that at atmospheric the N2 max inversion time is 2.5ns. Most pulses I've seen recorded are sub nanosecond. Talk about timing issues.

Also, I don't think we can get a beat frequency going as you suggested. Frequency doubling is normally the function of a solid crystal lattice (that's how ND:YAG solid state at 1064nm gets turned into a green laser at 532nm, via a doubling crystal).

Fernando:

Add in Adrian getting the balloon system going...

How about HDPE/UHMWPE being the backbone, toss on a layer (with a water sprayer at dispense point?) of RTV-1 silicone. Strength and added chemical stability, although PE is pretty neutral as it is. Of course, you could just use the silicone as a thermal/touch barrier and keep building up from there Silicone mold and HDPE support structure in one.

I have a feeling future repraps will go about as follows.
Get a bag of parts. Mostly structural, some for one extruder head.
Print the parts for a second extruder head...that you need, along with other parts that aren't critical in the short term, but really useful in the long term, (endstop boards, perhaps, parts for a shaft encoder.)
Break the thing down, and install the omitted parts.
Re-assemble.
Print the parts for a third extruder head...that you still need.
Break the thing down, probably not as completely, and install the new parts.
Re-assemble.
.
.
.
Re-assemble your finally completed reprap....
Until someone invents a new modification you just have to have.

SOI-Sentinel,

Darn. Well, you might still be able to turn the thing "on" for a set time that won't overcook and should almost always cook it at least enough. Not the most efficient method, though.

Actually, a spike should be readable past a transformer, unless it has a capacitor across it, or perhaps one of those from the old sola units. I believe they were called resonating transformers.

Again. I'm operating without a net...

Most N2 lasers are DC driven witb the transformers charging a capacitor through a diode.

Anyway, I found a few links that the laser wanters would be interested in. I lost these at my last reformat years ago.

For galvanometer and driver creation...

[elm-chan.org]

The math behind an X-Y scanner (courtesy of Delta Tau)

[www.deltatau.com]

We are going to win.

Would the laser retain enough energy to sinter / cure the target materials if it were set up in a manner similar to a laser printer & prism assembly?

Would be a matter of course of moving the laser+prism assembly very very small increments over the surface of whatever you are trying to print.

If the laser energy would be sufficiently consistent to do whatever job it is meant to do... the laser+prism assembly would definitely give the fastest result?

(Move feasible with UV curing resins than a high energy CO2 laser I presume)

... it's depending of energy vs. scanning speed - with weaker lasers you have to scan slower, so a scanning mirror with microstepping stepper is sometimes better than a synchronous motor ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

I think the first sla used a fibre optic and uv lamp, this is not fast but low tech, with a mercury lamp focused to a point and into the optic with the other end on a plotter head or reprap.

some time ago I bought a cheap digital photoframe with the intention of using it as a mask for uv, the good thing is that the images can be displayed for selectable times and it takes a usb stick, I also cut the electron gun from a couple of 5 inch tv's, the intention was to have them at 90 degrees so that the intersction point produced curing energy, I had visions of building in 3d, both of these project are in the que but I am convinced that time moves faster than it used to.

... i made some tests with 2Watts trough a 0.1mm-fibre focussed in a 0.2mm-spot on coloured ceramic powder - i can melt the surface to solid dropplets with a translation speed around 400mm/min, but you have to reduce the powder-height below 50 microns or the melting dropplets wont connect to the surface benath

Actually i'm busy with some other projects around the laser and constructing/milling different parts for some other reprappers, but one of the next steps is a setup for stacking powder-surfaces for a DIY-SLS with the 5Watt-diodelasers ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

Sounds like the width of the focus spot is more the issue for the level of achievable detail for a DIY SLS machine...

Though I have no idea how strong the resulting object would be, how smooth, how much for the material, or how much energy it would take to do ###mm^3 using the 5W diodelasers.

... it seems to be a matter of some more critical parameters including material properties

I'm testing with my fiberlaser too: it's a 50Watts-laser and i can reduce the spot down to 20 microns or if needed even to 5 microns - but the correct melting/fusing of the powders is a real mystery

If the induced power is a bit over a critical range, the powder evaporates before fusing to fluid droplets or the formed droplets explode and forms craters in the powder-surface.

The surface adhesion of the molten material can result in elevated droplets resting on some single powder-particles beneath the droplet, preventing the fusion of the droplet to the surface ... and other 'magic' parameters more ...

So for every different material and its size/absorption/melting/adhesion-parameters you have to define unique setups for laser-power, focus and translation-speed.

My findings shows that you have to reduce the height of the powder-layers to some ten microns, so the particle-sice have to be something between 2 to 5 microns ... and here the inner adhesion of the powder is a factor - either it agglutinates to bigger lumps, so you can't define a plain surface, or some static processes between the powder and the feeder results in undefined behaviour when pushing the powder over the area ...

It's an interesting but complicated process define and redefine the right parameters for the corresponding materials ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

A 5 micron working area would be astounding, but at the same time I have not seen any commercial SLS devices that claim such impressive detail.

Would melting contiguous spots simultaneously (if a small 5micron*20micron line is feasible) allow for shaped control over how the particles agglutinate?

... comercial SLS-systems use NdYAG-lasers with 30 to 200 microns big spots or CO2-lasers with spots of 60 to 300 microns diamters - aditional there is the diffusion/absorption-rate of the powdered material what's defining the melting area ... so a laser-spot of 30 microns can melt a spot of 20 to 200 microns size only by changing the pulse-time or pulse-shape (energy-profile over pulse-time).

Another essential parameter is the depht of the melted area - in good absorpting materials the melting depth is only a fraction of the spot-size

One of my next steps is implementing pulse-forming for better optimizing of the melting behaviour, then i can finetune the melting parameters and depth by multi-pulses or by changing the energy-density over pulse-time ...

For the diode-laser i have the pulse-controlling simply by ON/OFF-switching the MOSFET and defining the time and PWM-pulse-forming with an Atmega128, but have to implement the controlling in my CNC-controller ...

For the fiber-laser i'm awaiting an AOM which can modulate the CW-beam with some hundred kHz, so this would be more interesting with an optical scanning head than for XY-plotting as with the diodes ... but in the beginning i'm XY-plotting with the fiber-laser too ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

Just seen this zcorp it looks like a mask system, I have no audio on my pc so do not know what he is saying.