Just ordered in 4 more flexible outdoor solar panels from our friends at Modern Outpost and will have a total of about 100Watts available on the Frond.

This is the new layout for the solar panels as “branches” to the Frond.

We had a successful run with the Frond at Recompression — the interactive parts worked well! Next step is to make the Frond twice as big and integrate all the gear from the podium into the Frond stock itself.

Toward this redesign, we’ve identified a single full range speaker that we can position in a custom horn shaped enclosure (which doubles as the stock).

3″ Full range speaker: w3-319sf Spec Sheet

We also designed a custom exponential horn enclosure to get the most efficiency from this single driver.

The sound frond podium under construction. This is where we put all the interactive components for a trial run at Recompression 2010.

Inside the podium case: top is the theremin with control board face, under it is another black box containing the Arduino Mega, solar charge controller, Rougue Robotics MP3 board, and miscellaneous fuses and power distribution blocks. Also mounted in the exterior are 3 infrared motion sensors and 3 ultrasonic range finder sensors.

Next shelf down are two weatherproof speakers, and on the bottom is a deep cycle battery to hold the charge generated by the solar panels and used to power the system and lasers at night.

Over the weekend on Lasquiti, we were having a great deal of trouble trying to get the uMP3 audio shield to work with the Arduino. In fact, we thought we were going to be working with the latest rogue robotics product, the rMP3. Oddly enough, they are both priced the same but what’s the difference? Well, here’s the info from their website: The rMP3 includes headers for the Arduino board connection mating (header mounting optional). Unfortunately, we have an Arduino Mega and I’m assuming that the connections where made to line up with the Duemilanove board. In order to make it work with our board, we would probably require the hardware fix described in the earlier post.

http://forums.adafruit.com/viewtopic.php?f=31&t=10939

Another issue is wondering whether the shield can play multichannel audio? I’m going to email to confirm later today.

If not, in terms of audio, we may have to rethink our approach. Whether or not the uMP3 can play multiple audio tracks, the cost to buy one is rather high ($99 before shipping). If creating multitrack audio means buying additional shields for each sensor, Adafruit Industries makes a wave shield for $22 USD but each one requires an Arduino Duemilanove board to function (at $30 USD).  In terms of power consumption, I have no idea what the costs are, but I’m guessing that having multiple audio shields is not the most efficient use of power.

So I started to look for alternatives that were under $300 and fortunately, there are many options available. The issue of course, is whether to use a computer and the power consumption that this entails, but these days, you can find a single board computer that actively consumes anywhere between 2-7 watts.

For me, the most impressive single board computer I’ve found is made by the Israeli company Compulab.

It  is capable of running open source software, Windows XP or Windows 7 and has built in stereo in and outs! Also, it’s active power consumption ranges from 5-7 watts, depending on the CPU load. Here are the links for more info:

http://www.compulab.co.il/fitpc2/html/fitpc2-sb-datasheet.htm

http://www.compulab.co.il/fitpc2/html/fitpc2-sb-price.htm

http://www.compulab.co.il/fitpc2/html/fitpc2-price.htm

http://en.wikipedia.org/wiki/Fit-PC

To buy a single board, with the added option of audio in and outs ($3),  the company asks you to multiply the listed price by 1.2, so that ends up being $237 USD They also make their own enclosures with a base price of $245 USD.

It definitely works well with an arduino controller and it seems to have found a small niche in use with robotics:

Looking at it in terms of cost, it’s double the price of an Arduino Mega board + uMP3, but the fit-PC2 board offers functionality that is impossible at this time with an Arduino board (can’t record decent quality audio, multitrack audio playback with one shield?) . The idea of being able to run Max/MSP patches and not having to write out too much code attracts me a great deal. Using Max/MSP, having the ability to have multiple buffers of prerecorded audio, create a mixer, and having the option to record audio (record the ambient environment, create random loops in realtime) is pretty awesome.

Maybe this is something we should consider after recompression and for burning man?

Another low power 6 watt pc option: http://wwwd.amd.com/catalog/salescat.nsf/doclookupweb/25918BEBCC7071CF862572970038C606?OpenDocument&id=WinMate+Communication+INC.+~AMD+Geode+LX-800+500MHz

I was looking to find a solution to our problems with the uMP3 shield over the weekend. Luckily I have it! Ends up that the Mega board is not compatible with the shield and requires a hardware fix. Below is a copy of some of the text found on the adafruit forum, just to give an idea of the problem. It deals with a different shield but I’m sure it applies to the uMP3 and I’m going to inquire with rogue robotics to confirm that the mega board is not compatible without this fix.

From gregsadetsky on the adafruit forum:

After getting the Wave shield, I’ve discovered that although the shield’s hardware is “generally” compatible with the Mega board (e.g., the shield sits well on top of the Mega, ground and power pins are aligned, etc.), the pins used to communicate between the Mega and the shield must be remapped in software and unfortunately (for the time being) in hardware.

I’ve found most of the answers I was looking for by comparing the Duemilanove’s and the Mega’s schematics. There were two problems to solve:

a) the Duemilanove’s SPI pins (used for communicating with the SD card) are located on digital pins 10, 11, 12 and 13. The Mega’s SPI pins are 50, 51, 52 and 53, on the right side of the board, out of the shield’s “reach”.

b) The digital pins 2, 3, 4 and 5 used by the Mega board to communicate with the DAC (digital to audio converter) have a different software port mapping from the Duemilanove’s.

For more info, pictures, and the solution,  here’s the link to the forum page:

http://forums.adafruit.com/viewtopic.php?f=31&t=10939

Also, From Rougue Robotics site, the uMP3 mated to a Mega with a jumper:

http://www.roguerobotics.com/wikidocs/code/rmp3_on_an_arduino_mega

Micheal has developed an organic horn-shell design for the speaker enclosure. The idea is to maximize the volume output and frequency response of a single driver using passive filtering and amplification methods inherent in the shape and structure of the enclosure characteristics itself.

Also, he identified this software that will aid in the design parameters of the high efficiency horn design. See the ISHTEK speaker applet page.

Horn loaded speaker cabinets have been around since the early days of audio recording. Very early devices used horns to amplify very small amounts of sound. The invention of the transistor rapidly increased the available amplification allowing manufacturers to make smaller, lighter, more compact speaker enclosures with lower efficiencies.

Front  horns are very common on smaller high frequency drivers. The cross sectional area of the mouth of the horn limits the lowest frequency which is effectively amplified. Larger cross sections can amplify lower frequencies, and smaller horns only amplify higher frequencies. Because of this relationship and the availability of high power amplifiers, large horns to amplify bass frequencies have fallen out of favour. The higher efficiency of horn loading has a following in the audio community where low power tube amplifiers are coveted.

Other designs use very small drivers with folded horns to amplify the lower frequencies giving a wide response out of a very small package. Many of these designs also employ complex active filtering to balance tonal irregularities produced by the extremely compact design.

Innovative designers have also rethought the mouth of the horn. Some use speaker box placement within the room to increase the horn length and mouth area by using the physical properties of the room as part of the speaker. These may be as simple as placement a certain distance from a wall or a reflecting into a corner. Others build listening rooms which have been conceived as the mouth of the speaker.

The frond was conceived as a low power interactive sculpture. The capability to extract the energy necessary for operation from the sun as plants do along with design elements based on organic models lead the initial build. The addition of amplified sound could very easily be accomplished with conventional techniques. Off the shelf products could even be disguised, painted, covered, masked, or hidden so as not to call attention to their obviously inorganic design. Power requirements would still be a concern, as the finished product might require additional power sources. Or if the power requirement was deemed too important to compromise, sound pressure levels within the interactive area may be inaudible to the participants whose actions are causing change in the audio program.

Outdoors there are no walls, or corners with which one might extend the frequency response of a high efficiency horn speaker. Building a listening room to enclose a solar powered flower is not an option either. But there is one structure that will be available at all the intended installation sites which can substitute: the ground.

The concept is to use build a horn which uses the ground as a one side of the horn. Looking at the bottom of the structure, one sees what appears to be the mouth of the horn emerging. Careful measurements of this apparent mouth when compared to the gap from mouth to ground would show that the gap is a coninuation of the horn. As the ground becomes one side of the horn and extends outward, the bottom of the structure curves gently upward away from the ground and maintains the expansion rate of the horn. The outer curved surface is the other wall of the the horn. The curve of the horn turned inside out creates a bulb like shape, or maybe onion, but definitely organic looking.

In keeping with the organic design concept, the bulb is held up by a stem which seemingly grows thicker to root the structure. The stem and root structure supports the bulb. It also has a curved sloping shape mimicking shapes found in stems and roots. The easily identified mimicry belies the complex shape necessary to bend the horn back on itself, spreading what was a single direction port into a multi directional radiator of enormous area.

As this is still a conceptual design, actual effectiveness is yet to be determined. Early software simulation suggest that the final enclosure may be capable of perceptually flat frequency response from 50 hz to 17,000 hz and sound pressure levels in the range of 103 to 106 dB at one meter listening distance when driven by a 1 watt amplifier.

Micheal found this full range speaker with very high efficiency ratings of 90 dB 1W/1m. Could be a good way to generate loud enough sound volume, with the frequency range we need, given our limited amplifier power constraints.

Pioneer B20FU20-51FW 8″ Full Range Driver

Initial audio sample we’ll use for testing purposes.

SoundFrond_Test_Excerpt1

The Frond at Burning Man 2009 shown at night during a full moon. The frond provided the camp with power it collected from its solar panels during the day. No interactivity at this stage — just lighting from 36 lasers.