scruffy days

The shows get better and better every year. This one featured a slew of eco-minded projects: BioBronc, Solar Jewelry, Solar Time (pictured here with my parents ), BikeJus, Device Power Monitoring, Blue Phoenix, and the debut of Sustainable ITP - a showcase of student and faculty work in the program’s sustainable practices initiative.

Solar Time, a project with Gilad Lotan, is a meter displaying the total amount of energy in watts available from ITP’s 80-watt solar panel located on the Tisch building’s roof. Readings are retrieved via a wireless radio network and logged in a database.

Also in the show was Under The Level, with Catherine Colman, a mapping project of New Orleans’ post-Katrina destruction on the streets of New York. More details at underthelevel.org.

A simple metering interface to increase awareness of the solar energy available in relation to daily electricity consumption. [some documentation to date]

Concept: Solar powered Bluetooth modem for recording environmental data (temperature, air quality, etc.) for broadcast to your mobile phone. Microcontroller would still require separate power source, but blips of power could supply enough & trigger the Bluetooth function in this type of application.

———————————————————————————–

I attended the Eyebeam workshop last summer on DIY energy and built a basic solar engine. For this project, I wanted to learn more about how these BEAM bot components store/release energy and apply that to something beyond the bot design.
SOLAR ENGINE
I tried another take on the solar engine, with an LED as load. I also tested this with a mini vibrating motor I had around, but required too much current so it would never power (1.5-3VDC, 62mA). I thought I could try out some small fan motors, but these required even higher current - one at 0.06A and the other 0.22A (both at 5VDC).

Measurements:
Voltage - max of ~2.6V
Current - max of 5.8mA
Capacity in joules - 5.8mA * 1 hour * 2.6V = 54.28800 Joules
Power consumption in watts - 5.8mA * 2.6V = 0.01508W

Energy storage method:
1 capacitor, 4700uf 6.3V
8212 ENGINE / PM3 POWER MODULE
My next attempt was the 8212 solar engine. This uses hysteresis for power discharge, which is triggered at certain voltage levels, dependent upon the combination of resistors used with the Maxim 8212 chip.

Highlighted below are values I tried for R2 (trigger) & R3 (hysteresis). Didn’t work out though.

Currently, for another class, I’m working with a BlueSMiRF to receive and broadcast data. This can be powered with 3-6VDC, 25mA average. You can reduce stand-by consumption to 2mA with an AT command, but so far the modem’s only responding with error messages. I wanted to try and see if it was possible to periodically provide the modem with bursts of power so it could possibly broadcast a bit of data to a local computer or mobile phone. Measurements of the solar cell under a desk lamp (26W CFL): 6.4V, up to 10.4mA. Datasheet also rates it capable of 20mA in a short circuit setup.

Below is my first attempt at the circuit, using 100kOhm resistors for R1, R2, R3 & a wire jumper for R4. This didn’t work out as the configuration of resistors was incorrect, at these values according the chart above would trigger at values too low.

Next attempts below, changing the resistor values. R1 was still wrong, and Jeff F. recommended setting trigger value higher. Second image below uses these values:

R1 - 100k, R2 - 400k (trigger at 6V), R3 - 244k (hysteresis at ~3V), delta-V of 3V

Still didn’t work though, as the voltage remained stagnant at around 5 volts.

Referred to Jeff LeBlanc’s layout of the same circuit, and made these changes to the resistors, now seeing a change in voltage. Toggles between 3.6 - 4.5V. Problem is that the TR3904 isn’t discharging that voltage though, and unsure as to why.

Measurements - need to calculate when load issue is resolved:
Voltage - triggers around 4.5, down to 3.6V and back up again
Current - load not releasing
Capacity in joules - (Amps*Time*Volts = Joules)
Power consumption in watts - (Watts=Volts*Amps)
Energy storage method:
2x 4700uf 6.3V capacitors
(Also tested with a 1F 5V cap, but this significantly changes the circuit’s behavior. The climb in voltage is much, much slower…)
Where to go from here
Figure out problem of load issue in PM3. Research other BEAM circuits that maybe better suited to supply power for Bluetooth broadcast application. As I continue with the Bluetooth project, I’ll record consumption rates for specific functions.

Megan & I decided to use the playfulness and motion of the salad spinner as a children’s toy. One Kid Band is a portable, modular toy that encourages individual expression and musical exploration. Kids can customize their own LED light displays, attach speakers and other interfaces and power them through their own physical activity.

Features:
- Creative play
- Learning by making
- Enables imaginative solutions

Before OKB, we attempted battery charging:
100 tugs ranging from 10-20V charged up AA battery 0.02V
We tested this twice, and voltage on battery rose 0.02V each time. Scenarios this might be implemented: gym, possibly physical therapy. But the salad spinner as a charger alone wouldn’t be very practical - it would be best to integrate with systems that are already utilizing a similar action and allow the charging to take place over time.

So we’ve superglued our salad spinner to our servo & running the 4 leads through 2 sets of rectifying diodes:

Voltage & Current Readings:
Slow, full tug - up to 50mA, 5V
Increase speed - about 100mA, 10V
More force & speed - up to 180mA, 12-18V

These are all maximum values, so they only reach these when the spinner cord is drawn to its fullest length. See how the voltage values spike in this video

Added 1F 5.5V capacitor to circuit (had about .5V left from experimenting with it yesterday). After a couple tugs, reading at 1.02V. After 10 medium tugs, voltage up to 1.96V. Tried carefully not exceed the 5.5V reading so as not to damage the capacitor, but even as I increase force, it’s more difficult to get past 3.5 or 4V. Cap now at 2.71V. Plugged in LED, charge started immediately dropping. Within about 20 seconds, reading of cap went down to 1.93V, and continuing to trickle out with the LED acting as a load to drain the charge. In the case of the shake lights, a switch is used to allow the flow of electricity to the lights. Is that how we should control the flow in this case? Can we prevent/control the trickle with just a switch?

Next step: Read datasheet on Fast Charge IC Management chip, bq2000 (available as free sample from Texas Instruments). “A programmable IC for fast charge management of NiCD, NiMH, or Li-Ion in single applications. The bq detects the battery chemistry and proceeds with the optimal charging and termination algorithms.” Whoa! what algorithms? Ahem, let me continue: “This process eliminates undesirable undercharged or overcharged conditions and allows accurate and safe termination of fastcharge.” Sounds promising…setup below still needs to be tested:

Bought a charger, 2 AA Ni-MH rechargeable batteries and a battery holder from Radioshack. Specs on packaging note the charging current:
140mA +-10% for AAs
80mA +-10% for AAAs

Using a hacked stepper motor from a junked printer and a salad spinner to convert kinetic energy to small amounts of electricity. This initial test produced a max of 6V and 40mA, but the setup is just quickly hacked together. A more stable setup should yield better results. The motion of pulling the drawstring is fun & playful - Megan and I are still brainstorming on possible uses for the converted energy. Check her notes for more documentation.

I recently attended this workshop led by former ITPer, Jeff Feddersen at Eyebeam. He discussed the basics of energy, how humans have manipulated/stored it through history, electricity, solar energy and its current status. Here’s what I learned:

So how far are we from seeing wider usage of solar power in cities, products, etc? At this point in time, solar power is about 8-35% effective, with the higher grade stuff being used by NASA. Solar panels react to certain frequencies of light, which is another reason they’re not so efficient. The cost is also still pretty high, as it’s still labor intensive to produce and requires growing crystals within a very controlled environment; the size needed for the supporting infrastructure is also another factor keeping the cost up.
There are a few different types of panels you can purchase:

• monocrystalline - darker in appearance, more efficient and costly. Example shown in workshop was the SINONAR SC-9225
• multicrystalline or polycrystalline - least efficient, cheapest to make. However, these panels have a more interesting, reflective pattern and Paul commented on its possible appeal for usage in fashion & wearable technology
• thin film - can almost be printed. Instead of grown crystals, these are sprayed on. Currently, these are not very efficient at all, but also have application toward wearable tech

Batteries end up doubling the cost of trying to implement solar energy use. In addition, our limitation in manipulating materials make batteries a bad storage of energy (esp. in comparison to how our bodies store energy).

The last hour or so of the workshop was spent building a “solar insect” (kits came from solarbotics.com & the instructions are featured in Vol. 6 of MAKE magazine). This was a lot of fun, and extremely satisfying to see our bots jitter slightly under the desklamps.

All in all, this was a great workshop and a good intro to the basics of energy and solar power. Since my undergrad studies in product design, I’ve been interested in sustainability, and renewable/recyclable materials. The building of the BEAMbots was like an appetizer and I think most of us were eager to try out more and experiment with other projects. It would be great to have a part 2 and up, or a more fully developed class in this topic.

Additional related links & resources:

• Jeff’s recommended reading: The Bottomless Well
• Schematic for basic solar engine