Click here to check out the latest version of the Hoverlay or continue reading to see my humble beginnings 🙂
I got some feedback that this actually is not 3D and/or holographic. Well guys, it depends on the projector you use 🙂 a regular off-the-shelf light field projector will do. If you do not have a light field projector at hand, a few drops of LSD in the vaporizer vat do somewhat the same effect*.
(*Srsly, never do that!)
Seeing a so called “fog screen” or “vapor-display” for the first time can be very irritating if you still believe, floating-in-midair screens were reserved to sci-fi-content, but never reached reality. The following video shows the Heliodisplay, a commercial vapor-display invented by Chad Dyner, who also holds the corresponding patent for this technology:
The very first time in my life I saw such a screen on a trade show I had to stick my hand through the thin airflow again and again to grasp the thing. It’s a good example for how astonishing effects can be achieved by actually simple means. Here you can see an original Fogscreen® (another commercial product), it works likewise. My guess was, that there probably would not be another commercial project too soon, because of the patent-barrier and also the narrow market for this kind of geeky stuff, so I decided to start this project named “Hoverlay”, to gather and develop information, new ideas and designs for this particular kind of canvases.
The core principle behind the fog screen is probably the fact that it’s made from fog. Water of fog-fluid is vaporized and the mist is then somehow formed into a flat (or bent) airflow on which a projector throws imagery from behind. At the moment, there is a company out there named “Fogscreen” that produces, sells and rents out massive large fog screens for events, as well as other commercial and even home-brew builds, each one approaching the idea a bit different. The following is true for all of them:
- fog is created
- fog is formed into a laminar flow
- images are rear-projected through the fog by one or more regular video projectors
For the first two tasks there are different solutions already available: Fog-generation:
- cooling the fog down (e.g. by letting it run through a bucket of crushed ice) then letting it fall from above through a slit
- trapping the fog between two air-buffers which are formed by blowing air (e.g. using fans) through an honey-comb-like array of straws
- blasting fog out of a large tube stuffed with straws (actually the straws are only at the outer border or the tube
- Most builds that combine diffrent ideas
So some use regular smoke machines, others ultrasonic atomizers to generate the fog. Some throw the airflow from above, others from below. Some cool down the fog and make use of gravity, and almost everybody uses straws. Additional thoughts were:
- adding an electrostatic charge / ionizing the airflow and fog to prevent them from mixing to early (a student friend of mine came up with this very interesting idea)
- instead of using ultrasonic misters or smoke-machines, boiling water, e.g. from a steam cleaner, could be a pragmatic alternative
- the design should be modular and expandable, adding modules increases the size of the screen
- it would be nice to add sort of a dmx/artnet controlled mechanical flap to switch the fog on and off quickly for achieving additional effects by “chopping” patterns into the fog stream
From that thoughts, I set up my first prototype design. For this step I am likely to use Adobe Illustrator. There are better tools for CAD, but as my roots are in the design/media branch, I am just way faster with Illustrator for now. So this is how “Hoverlay” could look like:
Basically, this is a simple box containing fans, the vat with water and atomizers and some stream forming elements. The crossection schematic shows how the Hoverlay works on the inside:
This prototype is not supposed to be perfect, it’s meant to be fast to make with a high initial learning effect. I am going to use it as a proof of concept and for setting up a checklist of considerations that are supposed to be made for the V1. I expect to use for this:
- 4 sheets 1500 x 500 x 3 mm PVC, (at my local hardware store for 15 €/pcs)
- 8 pieces 80 mm fan with ~40 CFM (at pollin.de for 1,10 €/piece)
- 4 pieces ultrasonic foggers (at ebay, 3-5 €/piece)
- 24 V AC transformer (min. 70 W) for the foggers, operating at 24 V alternating current
- 12 V DC power supply (min. 15 W) for the fans, operating at 12 V decay current
So I made the layout in Adobe Illustrator, exported the parts as dxf from which I made g-code using Estlcam. The process is pretty straightforward. If you want to do a similar project, but do not have access to a mill or lasercutter, maybe you want to checkout my PVC hard foam plate tutorial on how to build this kind of parts without heavy machinery. This is the layout for a holder that is gonna hold all the atomizers in position which I exported as DXF with the settings on the right (sorry english users, I hope at least the buttons and checkboxes are somewhat in the same position):
Mind that Estlcam does not support fancy stuff, just basic shapes. At least it’s available in English, if I have more time in the future I will add english screenshots to the english version of this post. The layout has exactly the dimensions I want the part to have, without any tool considerations such as cutter diameter or laser kerf. After importing the dxf into Extlcam, you will want to add your tools to the tool list, and this is exactly the moment where you have to worry about cutter diameter or laser kerf, which you will have to enter into the ø-field. Cut materials at the right feedrate (F) and spindle speed (S) if you know them.
I am quite new to CNC milling, but from what I can tell, the PVC hard foam that I used for this (and almost any kind of other prototype I ever made) is best to machine with a feed of at least 3000 mm/min at low spindle speeds and a two blade cutter. With this, I get great edges instead of molten plastic. The next step is defining the parts (left below), and then the cutouts (right below). Make sure you define them with the right tool selected:
Once all parts and cutouts are defined with the right cutting tool, we will arrange the order in which the are going to be cut. General rule: Cutouts first, then the part. You’ll find it in the menue as “Fräsreihenfolge festlegen”. Select “Vorwärts” and click the parts and cutouts in the order you want to cut them.
After that we’ll send it off to the gcode generator. Estlcam will automatically show you a simple simulator where you can test your gcode or at least play it back.
With the gcode on a thumb drive in hand, I’ll approach the next step, which is actually milling the parts. If you’re new to CNC-milling, you really ask a professional to walk you through the entire process from layouting the parts over g-code generation to operating the machine. In general, CNC-machines are:
- quite expensive, you don’t want to destroy them
- extremely strong, you don’t want to get hurt by them
- well maintained and in 100 % working condition, otherwise you don’t want to use them
- extremely dumb, you do the thinking. If 1 or 2 happen, it’s probably your own fault and not the machines, even if 3
So, know what you do 🙂
The vacuum table on this mill was to weak to hold the plate in position (it’s powered by an old vacuum cleaner), so I grabbed 10 car dashboard silicone pads from the cheapo store for 10 $. I really can say that those pads work great for CNC milling and outperform every low-budget vacuum table for a fraction of the cost.
Although it takes some dedication to clean the pads, since the loose their stickyness once the get covered with dust and milling chips.
I ended up with some nice parts. The connection bridges in the right image above were totally unneccessary. The dashboard pads were strong enough. The other parts were milled without connection bridges.
Then it was time to glue everything together with plastics glue. As I said, this is a prototype, and glueing is fast and cheap for this purpose. Only the outer frame will be attached with screws, since I need to install the electronics on the inside later.
In addition to the PVC housing, I had to take care of creating a laminar flow, that is supposed to stream out of the machine. I chose to go mainstream and use straws. There was a cheap offer about round thousand black straws on ebay I purchased, then cut them 4 cm pieces using a laser cutter. This particular laser cutter has an 1W laser diode, so this step took almost a whole day. Nevertheless, use protective eyewear whenever you use those kind of tools, the camera I took the photo with is replaceable – your eyesight is not (yet).
I then had to put the straws together somehow. I found some of the straws already were melted together during the laser cutting process, so welding them together seemed to be a good idea. I put together a homebrew-straw-welder out of a old soldering iron and a paper clip (left picture below). With this tool it was quite easy to weld about 500 straws together to solid blocks within less than an hour. I just hold the straws together side by side and melt a little hole into the touching walls of both straws with the little hook of the welding tool. Works:
After everything was finally assembled, I could of course not wait to see the whole thing in action, check it out in the video:
As said, the prototype got many issues. The main thing I will focus on in the V0.2 will be:
- depth of the fog layer, currently the fog layer is too thick (2 cm), which results in blurry images
- stability of the fog layer, the fog layer flutters a lot, bit like a flag in the wind
- stop leaking condensed water out of the case, unfortunately I totally underestimated how fog behaves, ending up with condensed water everywhere in this first build
So whish you all a happy new year and stay tuned for Hoverlay updates 2014!
Updates are there, check out the Hoverlay II here.