3x Zodiac

Filled under: Fabrication

Date posted: February 20, 2010



Three acrylic panes illuminated by three sets of RGB LEDs.   Each of the nine LED channels are independently driven by a PIC18F2221 using nine software PWMs. The lighting is controlled by a simple UART wireless connection using a Radiotronix 315MHz serial data interface.

There are several color changing modes setup up such as solid, rotate, and revolve. There are also three redundant tactile control buttons on the lower front of the piece which control the color pattern’s MODE, COLOR PREFERENCE, and SPEED.


Personal Notes:

This is my first full fabrication project using my newly assembled CNC router table and also a prototype design for a future product line. Long ago I made a loose sketch on a piece of paper of a near similar design and said to myself  “I want to make that!”, over the course of time and a substantial investment of self-education and tooling I achieved my goal. And I must say it’s rather gratifying. =) It’s great being both a dreamer and a doer.


The original Sketchup model.


Ok, so currently my table saw does not have power (220v) and my normal router table is not setup so I am unable to make the original design as seen in the above picture. But alas the CNC router is working and ready to go so I reworked the project around my current capabilities, let’s get cutting!

Now that the first step is taken care of, which was identifying current abilities and limitations, I am able to design the project around the known. So instead of making the piece with several strips of wood (as the original design calls for) I will use a single piece of lumber in smaller proportions, 5″ x 13″ sounds good. 1″ thick is good, this will give plenty of interior space for electronics, acrylic, and the power jack.

The next step is drawing up the design in AutoCad.  I need to keep in mind the size of the acrylic panes, the channels for the electronics, the space for the LEDs, and the area for the power jack. Also since this project has a special consideration of it going to be wirelessly controlled there needs to be a channel for the antenna to fit in.

The AutoCad draw up.


OK, now that the part is draw up it’s time to move on to the toolpath generation.


The toolpath simulation.

Now that the toolpaths are laid in and the simulation looks good it’s time to cut out the part. The program I used to generate the toolpaths is ArtCam. It’s a wonderful feature rich program, however, considering the part only requires vector path cutting there is freeware out there which will work well.

Cutting the part.

Yep, don’t laugh, I know I need to make a dust collection fixture. I am waiting for 4″ clear ultra-flexible tubing to come in then I am going to make a MDF dust shoe. But being that I just assembled the machine two weeks ago, it’s OK for now.

The part being cut is actually a test run using a scrap 2×6.

The bit is a 1.4″ straight bit cutting at 16k RPM, 100IPM, and around .125″DOC. For soft pine I could slow down the RPM and increase the depth of cut a bit, but for my first part it worked well.

After some sanding, staining, and clear coating the piece is ready to have the electronics installed.


The three RGB LEDs are driven by nine software PWM channels using a PIC18F2221.

The idea is to have the piece wirelessly controlled. The amount of date being transferred and the speed required are both rather low, so a simple UART connection at 2400 BAUD will work just fine. For this, premade modules made by Radiotronix will work just fine. They operate in the band of 315MHz which is dedicated by the FCC for short range low bandwidth devices. Such devices are garage door openers, vehicle remotes, personal weather stations, etc. One issue that can come up with using these devices in this band is interference from other operating devices. The FCC set strict guidelines in order to lessen the chances of interference by limiting the amount of acceptable on air time down to basically a few bursts per minute.


The left module is the transmitter which works between 5vDC to 12vDC. The right side module is the receiver.

The Insides:zodiac3x_3

The insides.

The lower left you can see the receiver and it’s supporting logic inverter. More towards the bottom is the tactile switch board holding three switches. Then to the far right is the power jack and it’s channel.


Underneath the acrylic pane.

Beneath the acrylic is stiffened felt and neoprene as a void filler. The black felt really helps pop out the illuminated acrylic panes.


The antenna channel.

The antenna needs to be away from metal objects so a decent placement for the antenna is straight down the opposite side of the controlling electronics.


The back. The next version will have dedicated locations for screws instead of nails.


The three front control buttons.


The test setup transmitter.

The above is the testing transmitter setup. Nothing special, just a PIC18F2550 with a character LCD readout for debugging and what not.


The Video:

The flicker caused by the slow camera frame rate.



Well, the prototype project turned out great! However there are a few short comings which require attention. The range of the RF wireless connection is rather dismal at 10 feet. A quick way to increase the range is to lower the BAUD to the lowest point possible which could be around 600bps. This will required a BAUD translator because the PIC’s internal UART hardware can not go below 2400bps and maintain the system clock needed to drive the nine software PWMs without producing noticeable flicker.  But really the best option to gain maximum range is to tune the antenna but being without RF test equipment that’s not an option. So where does this leave the wireless connection? Two possibilities use 2.4GHz RF modules (expensive, added complexity) or us infrared (easy, cheap).

The location for the electronics need to be reworked. For a prototype it’s OK that the electronics are directly soldered without PCB boards but beyond a demo prototype a stable integrated production solution is required. The side channels are .500″ wide and this might be wide enough to make a PCB for the 4 ICs but it still seems a little tight. A possibility is to make a SMD board and place it behind the void of an acrylic pane.

The pine frame is a below passable. Due to the nature of pine’s grain it just does not stain well so using hardwood like Maple or Red Oak should look great. This will require thinning the design down from 1″ to 3/4″ mostly because the availability of 3/4″ thick lumber is vast.

To the drawing board!