I'm a student, very new to this, and need help. Want to build my first scanner.

[swamidog]

i use an etherdream DAC with LSX. it works very well, is well documented, and all the source code code is available for download.

it's also being used for some industrial marking: http://www.beamconstruct.com/etherdream.php

I'd be leary of the Etherdream DAC. I'm sure it works well but it's basically a few people on this forum who decided to design a DAC. From what I gather it costs as much or more than some commercially available DACs that also have APIs so ask the right questions before you buy.

[james]

If you are looking for a more DIY approach, you might want to start with a modified sound card. For some uses, there is no special driver required. It's a sound card! It plays waves! On the other hand, you can get a special driver for it that make it look like a proprietary laser show DAC and it is supported by quite a few applications, both ways.

One of the nice things about it is that it clearly demonstrates what laser signals really are and where they really come from in a very generic way.

Another nice feature is that it works in just about every OS in common use.

Plus, there is a very well established application called LaserBoy that is also platform independent, free and open source.

Even if you get a specific laser DAC, you might want to build yourself a LaserBoy DAC. It is not very expensive at all and anything you write for it will be very useful to a lot of other people who also use them.

[mixedgas]

Respectfully disagree, unless you have no budget....
Modified sound cards are a PITA, there is better open source hardware these days.

Steve

[james]

Well, it is the only one that is truly generic in every way. You can choose from a wide variety of sound devices, PCI, USB. For not much more than about $20, you get the most important guts of the whole thing; the DAC with an interface and drivers. Add to that a tiny handful of totally common electronic parts and you have a working system that has the potential of producing a set of analog signals that are every bit as good or better than any of the other DACs. If it comes down to it, you can record the analog outs of any other DAC and play them as a wave file through the modified sound card. The simplest form might not have some of the features found in other DACs made specifically for laser display, but these things can be added with more outboard electronics. Some of the PITA factors can be alleviated by your choice of sound card and the quality of its provided driver.

Before anyone chimes in on the "every bit as good or better" quote, any sound card can do 48KHz. Some can do 96KHz and even some do 192KHz. They are all 16-bits. Galvos don't benifit much from high sample rates and laser modulation (for DPSS lasers especially) doesn't either.

In any event, choosing a DAC / software solution is a lot like choosing a religion. You'll figure it out.

[colouredmirrorball]

I'm going to scratch the "clouds" idea completely... it sounds way too
complicated to be safe and compliant, and as a beginner I know I can't do this.
(Thanks Jon! You certainly convinced me that this is extremely unwise.)

Hey, no need to! Cloud scanning is my goal too, but you need some big laser systems for them (multiple watt laser systems) and I don't think I'm experienced enough to handle them. It's just not a good idea to start right off with cloud scanning

What three lasers (perhaps R, G, B?) would be best to combine to produce
a wide range of output colors in a cost-effective manner? When I do eventually
work outdoors, how powerful will they need to be to shine on a building and
still produce vibrant color? I hesitate to use wattages that are too powerful
for a beginner though. I don't trust myself to be 100% safe until after I've had
experience; I do not want any accidents. I can always upgrade the lasers later.

RGB is how it's mostly done. As for the powers, that depends on the circumstances. 1 W of each colour will be more than enough for a white wall a few tens of metres away in a dark environment, but you might need more when there is ambient lighting involved. Others are more experienced so I'll let them give better suggestions.

colouredmirrorball:

I get the sense that heat dissipation is going to be a big problem with some of
these lasers. If I stick with low-power lasers, will this be a non-issue?
When I do scale things up, can I develop a large housing to keep the optics as
far apart as possible to minimize their contributions to total heat? Aside from
a heat sink, do you typically employ fans?

That depends on the lasers theirselves too. Some modules have a cooling system built in and will automatically shut off if the laser becomes too hot. Others require a fan constantly blowing air over them. Some lasers are more than happy if they are just passively cooled when mounted on a big base plate. Most projectors do have fans in them, and that brings the issue of filtering the dust out of the air before it reaches the optics.

Optics generally have no problem with heat, unless you really go crazy with laser power (say, 5-10 W?). If you design a large case you might get away with no fans at all because the heat can be sufficiently dissipated through the case, and you will have room for improvements such as bigger lasers in the future (that might spoil the passively cooled case). But this is of course at the cost of your back, and a big case isn't always better when you don't have a lot of space.

You just have to look at your system, set it up, let it run for some time and touch the laser housings from time to time. If you notice they become too hot, better add some cooling.

My knowledge of electrical and mechanical engineering is pretty poor, but do you
think a beginner such as myself could build a laser from the diode, driver,
etc. directly without too much trouble? I've only done extremely basic
soldering work, but as I mentioned before I should be able to get some
assistance from the campus EE department. How hard is it realistically? Would I
be better off getting something already built? (Granted, I _do_ want to learn how
to do this. I just don't want to destroy the equipment on my first go.)

Depends on you I guess. I could do it so there is no reason you wouldn't, but expect to blow a few diodes in the process. It's pretty unevitable but part of the learning curve.
Something you want to look up is ESD. It's the main reason of death of diodes, and you can avoid it with a few basic steps. A Lasorb is a little device that protects the diode from ESD, but it doesn't protect against stupidity. Just a few hints.

Also when you really screw up, don't feel bad because someone has already screwed up more badly (like me... hehe).

Misc. questions:

Can the 30K galvos that JohnYayas mentioned reach 1000 points max with minimal
flicker, or do I need higher KPPS? I really want to minimize flicker, maximize
points, and maximize drawing distance, so I'll invest more in the galvos if it helps
and still falls within my budget. I don't want to have to scale down the
image--if anything, I want it to scale up. I may be forced to reduce draw points if
it's absolutely necessary.

1000 points can be achieved with not much flicker, if you reduce your scan angle. You need to experiment (or attend a LEM, it's really recommended!), but you'll find 30k can handle 1000 point frames just fine. There will always be flicker but it's up to your taste if you find it too disturbing or not.

Could I get a second laser/xy-galvo/DAC to work synchronously with another set
in order to effectively double the number of points or the scene complexity?
With a clever algorithm, it may do more than double scene complexity since the
vector draw points could be clustered to each galvo, giving them both smaller,
optimal drawing paths. Would aligning and rectifying two separate galvos be
hard? (This setup might be way too expensive to even try. I need to get my feet
wet with the basics anyway.)

PL user masterpj tried that at the most recent DutchLEM. I'd say the technique works, but it's hard to set it up. You need to make sure the images overlap each other perfectly and you need to apply geometric correction.

How inefficient is it do draw "solid" figures instead of outlines? I've seen
several videos where lasers are used to "paint" solid surfaces. Perhaps I can
cheat a little bit and draw alternating scanlines that give this illusion.

I don't know what you mean exactly. There are two ways I can think of what you mean: using a raster, or using lenses to quickly expand the beam.
Raster graphics are commonly used and they work similar as a television, only you can tell the different lines apart and you can't have too many detail in them. It's a neat trick for every now and then. You can't make the image too big because a raster is usually rather complex (about 1000 - 1500 points) so the flicker is noticeable. Also you need to optimise blanking. For other graphics you can get away with a poorly blanked line now and then but with rasters each line needs to be blanked accurately, or the effect isn't visible.

The other option is a lens that can quickly alter the divergence of the beam. I think Pangolin developed such a system, but as always with Pangolin, it's not cheap. Also the opinions over the result of the effect are divided. I haven't seen it in person so I can't recommend.

And good luck with your exams! I better start preparing myself for mine as well...

[james]

quote:
How inefficient is it do draw "solid" figures instead of outlines?

There is a rather esoteric technique called beam brush that places an electrically controlled lens assembly in the beam to create a controllable divergence. So you can control the beam from being a tight spot to a large flashlight like wash. It's a very impressive look, but it's also rather difficult to achieve and there is also the need for an extra control signal and a way to intelligently create it.

It's pretty "far out" stuff.

The only other way to fill in an area is with a lot of scanning and that will tend to flicker a lot.

[colouredmirrorball]

Ah, beam brush, that's the word I was looking for. Thank you James!
Also nice pictures... now I want full colour too

[james]

These pictures came from my first rig that was Argon and HeNe with a PCAOM. They look just as good and much brighter on my solid state rig.

Perhaps I should mention that they are from waves I made in LaserBoy back in 2005.

[JohnYayas]

Respectfully disagree, unless you have no budget....
Modified sound cards are a PITA, there is better open source hardware these days.

Steve

Yes and no. The pain you speak of is mostly associated with the cheap CM106/108 USB sound cards due to driver issues and voltage drift. There are some soundcards that would be a dream to work with and in fact the VERY VERY expensive DigiSynth software does exactly that. Streaming abstracts to a sound card is much more natural than to a frame based DAC, especially when streaming at 96khz. But otherwise, I agree that sound card DACs are more trouble than they are worth.

[james]

Very very expensive compared to what?

You can find all sorts of sound devices made for pro audio applications that are designed to be compatible with ADAT technology. They have been around for so many years that there are tons of them available used. A lot of them are designed to have both unbalanced, single-ended -10dBv outs AND balanced +4dBv outs that require no correction amp at all. I have an Echo Layla 24/96 that has the first 6 channels' decoupling caps shorted out and that's it! It works just like a modified ADAT. It IS a hard drive based ADAT!

Using an ADAT to record a show made on a proprietary laser controller back in 2003 is exactly what gave me the idea to do it all in software.