Advanced Tech | Light & Shade

Doria Fan
doria[at]nyu[dot]edu

Advanced Tech
Fall 05

Info on solar components,
more links and resources

Window Shade Project (Light & Shade):
Adding solar cells and a rechargeable battery to make it semi-sustainable

This is a continuation of a project from Design for the 5 Senses that I worked on with Michael Jefferson and Jae In Lee. It's a window shade that doubles as a light. One set of strings would be to control the level of sunlight by adjusting the height of the shade. The other set of strings controls the light level. I didn't have time to add a solar component to it for that class, so I decided to work on this for Advanced Tech. There is some general "design" info here and a list of resources, but this page is mainly tech documentation of the solar component. Please see links on the left for more background & project info and images.

I spec'ed out some flexible solar panels that could be inserted into the back of the window shade (which would be made out of a transparent material) which would be connected to a rechargeable battery that would partially power the light at night. For proof of concept, I connected some more powerful solar cells (same surface area as the window shade) - which I place in my window - to a rechargeable battery. I then added a PIC and an LCD and LEDs to the circuit so I could monitor and measure the rate of charge. I finally connected these components to the circuit for the window shade.

Overall, I was able to get the battery charged, and partially power the window shade with it. Bear in mind, I was doing a solar energy project a few weeks before winter solstice, so the results may seem discouraging. Two overcast days in a north-facing window in Brooklyn in the winter was the time it took to fully charge the battery, which could run the circuit for 1 1/2 hours at night. If you have this thing on for 4-5 hours a night, it will at least cut your energy bill by approx. 20% which is better than nothing.

Prototype

Link to quicktime movie - footage of it in action

Circuit diagrams & pic code

window shade only:
• window shade circuit digram (PDF)
• PIC code

This is the basic circuit to connect solar panels to a rechargeable battery and a load that Gary Shrober drew for me. The circuits I built (diagrams below) are based on this.

solar cells & rechargeable battery (without load):
• solar cells & rechargeable battery circuit diagram (PDF)
• PIC code

connecting window shade to solar cells and rechargeable battery:
• window shade with solar cells & rechargeable battery (PDF)
See above pic code

(Circuit diagram files are Illustrator CS2 pdfs. Feel free to download and use components for your own circuit diagrams.)

Photos of setup and circuit coming soon...

Some basic notes on the circuit, PIC code, and setup:

(Most of the useful info below came from Gary Shrober, Rolf Levenbach, and Tom Igoe.)

Spec'ing out battery and solar cells:
• The rechargeable battery won't take current unless the voltage applied is a few volts higher than the nominal battery voltage.
• Solar panels need to have a higher operating voltage than the battery to charge it. Take sunlight conditions into consideration. Solar panel specs are for peak sunlight conditions.
• General rule: do not charge a rechargeable battery with more current than 10% rated capacity.

Connecting solar panels:
• Connect solar panels in series to get a higher voltage, or in parallel to get a higher current (or a combo of the two).

The 1N5817 (Shockley) diode is used to prevent the solar panels from draining the battery under dim conditions. There is a .2V drop across this diode.

The resistor (in series with the 1N5817 diode):
• There is a limiting resistor (100 ohms) in series with the solar module, and diode, so you don't overcharge the battery.
• I also use this resistor to measure the "voltage drop" across it to figure out the charge rate using Ohms law. V=IR. R=100 ohms, in this case. I can measure the voltage before and after the resistor, using the PIC and/or a multimeter to get V, and calculate I. I/delta T will give me the charge rate.
• You can vary (increase) the resistor value to get a greater difference (greater differential to measure) between voltage coming from the solar cells and the "load side" (battery). However, this will lessen the current going into the battery (Ohm's law). Much higher than 470 ohm may siginificantly limit the amount of current going to the battery.

Power shunt/simple voltage divider to "scale down" the measured voltage to ADC in at PIC:
• I took voltage readings (where the blue dots in the circuit are) before and after the 100 ohm resistor.
• The max value of the battery is 7.2V and for the solar panels (under unrealistic optimum conditions) would be 20V, both significantly higher than the max value of 5.5V that the PIC can safely read.
* For each point where I am taking a voltage reading, I have a 20K ohm resistor in series with the measured voltage, and then a 10K ohm shunt to ground. This way only 1/3 of the measured voltage will be applied to the chip (within its safe range).
• High resistance values (20K, 10K) were selected to drain as little current as possible froom the rest of the circuit (i.e. deplete the battery).

Voltage readings/ADC values:
There's a bit of noise and fluctuation with ADC input values.
• Tip from Gary: "Put a .01 or .001 uFd capacitor (ceramic would be fine) across each ADC input pin, it will make the measurement more stable by averaging any noise that appears. Just be sure to delay a few milliseconds after you select a ADC pin before you start the conversion."
• Tip from Todd Holoubek: Take multiple ADC in values and take an average.

In the circuit with solar charger and window shade, the 1N4004 diode and limiting resistor in series are there to keep the current from the battery flowing to the load (window shade), instead of back to PIC that is measuring voltage. I have the switch (power supply, on/off switch, and circuit of this PIC and LCD in series), so I don't need to power the the circuit while the battery is charging. I usually just turn it on to check battery level.

LCD display
• I used the Powertip PC-1202-A because it was cheap and the specs were available.(See pic code & circuit - links are above- and specs for pin call outs, etc.)

Here's what I did and the results from experiment:

I calculated the power consumption of our window shade to be 7.2V (you need at least much going into the voltage regulator) / 1.1 amps (.1 amps to PIC, and .1 amp for each of the el strips). I spec'ed out a rechargeable battery of 7.2V /1800mAH ($20.69) -- which should be able to power the circuit for maybe 1 1/2 hours. In reality, I should go with something with more amp hours, and a higher voltage (so the el strips glow brighter), such as the 9.6V/5000 mAh ($88.65). Due to costs though, I went with the cheaper battery, because I just wanted to get this thing up and running as proof of concept, for now.

I found some flexilbe solar panels that actually fit into the sleeves of the window shade. When wired in series and parallel, they could give me high enough voltage and current to charge battery. However, due to its to higher price (yes, I'm being cheap here) and lower operating current, I first decided to run the tests with more powerful solar cells (10 V, 150 mAmps) with approximately the same surface area as the window shade, since I'm running this in the middle of winter with low light levels. I wired 2 in series to get 20V, 150 amps.

I connected a PIC to read in voltage values, an LCD to display the values, and 7 LEDs to indicate the voltage level (for 0-7.2V -- how charged the battery is). Initially, I had the PIC do the math to give me the scaled voltage values. I first tested the setup by using to pots as my ADC ins to make sure the PIC code was up and running. However, after connecting the solar cells and batteries, my voltage readings were a little strange, so I stripped that, and other functionality, out to isolate the problem. You could take multiple ADC in readings and average it with simple PIC code. I ended up taking down these values as a number between 0-1024. I manually taking down multipe values from the PIC/LCD and processing the values in a spreadsheat, which I compared to values I took with a multimeter. (Spreadsheet with results).

What kept throwing me off most was how low the current and charge rate was. Gary had initially guessed that I would be in the range of 50mA (for solar cells with max operating curent of 150 mA), considering the sunlight conditions. The results I got were an order of magnitude smaller - somewhere under 10mA. The voltage differential between what was coming in from the solar cells and the battery level was so nominal, I thought it might be neglible due to fluctuations with ADC input values. I compared these values I got off the PIC with multimeter readings, and they were in the same order of magnitude.

I was able to charge the battery, but very slowly. If I had more time, I'd try this again with different resistor values, or a different set of solar cells, or the battery to try to isolate the problem. I was also doing this with natural sunlight (not a controlled testing environment), and would try this with some artifical light. If I were to use the thin flexilbe solar cells, I'd expect an even lower number. I would be curious to see how this charge rate compares with the supposedly more powerful cells I tested.

Some notes on materials/supplies, etc.:

SOLAR CELLS:
• Solar panels: 10V .150Amp 6" x 6" Solar Panel ($9.95/each from Electronic Goldmine)
These things are relatively fragile. I put them in a clear plexiglass box with foam on the backside to protect it, so I could easily move them in and out of the window. (Note: The plastic is an extra layer that will reflect/refract some of the sunlight and slightly reduce charge rate.)

flexible thin film solar modules:
• Solar powerfilm, TX 3-25, 3V, 25mA, 114mm X 25mm X 0.2mm / 4.5X1.5X .01 inches ($2.89/each at Jameco)
This size fits perfectly into the sleeves for the window shade. I'd have to wire 10 of them (the number of sleeves are protoype has) in series and parallel (2 sets of 5 (in series) in parallel to get a total of 15V at 50 mA.
• SOLAR,POWERFILM,MPT15-150, 15.4V,100MA, 253mm X 150mm/ 10 X 5.9 X .02 inches ($44.89 at Jameco)
About the same size as our window shade.

BATTERY:
• BAT,6-PK,AA,NIMH,1800MAH, 7.2V RECHRG, SOLDER TAB ($20.29 at Jameco)

POWER SUPPLY:
* These AC to DC regulated table top power supplies are the way to go. It's easier and neater than hooking up multiple wires/power supplies when you don't have enough each amperage. If you know what the required voltage and current consumption of your circuit is, you can spec this out. In our case, we knew the el strips are optimized for 8-10 V, and needed at least 1.1 amps. The el strips are connected directly to the regulated power supply, and does not go through the 5V voltage regulator (as does the PIC). The output plug (specs below) will mate with the power supply connector we use at ITP.
• AC to DC Table Top Power Supply
Specs:
Input voltage: 90-264VAC @ 47-63Hz
Output voltage: 9.0VDC
Output current: 2.78A
Wattage: 25
Output plug: 2.1mm ID x 5.5mm OD female center positive (regulated)

ELECTROLUMINESCENT STRIPS/PAPER
• ELECTROLUMINESCENT STRIP, 5" X 1", ($3.50 each from All Electronics)
• INVERTER FOR EL PANELS AND STRIPS ($6.95 from All Electronics)
• 2 CONDUCTOR CONNECTOR, MALE ($1.75 from All Electronics):I highly recommend getting the connector. Trying to solder wires to this inverter is really tricky because the leads are surrounded by plastic casing. This will save you the headache, and give you an easy connector to work with.
Altogether, it ends up being almost $13 to get a 1"X5" strip to light up. This stuff is not cheap. But, it's an easy set of components to deal with to get up and running. And, this ended up being the cheapest option for el paper from the different vendors we looked at.

• www.elwirecheap.com - they have some kits for el-wire and (1/2" wide) el strips. Too thin for what we wanted, and not the right colors
• hwww.e-lite.com/: a variety of el paper, el tape, and starter kits. Aside from the starter kits, they have a min order of materials of $100 or so dollars.They sell a dimmer, too, but it's $100+ dollars.

• Another cheap el tape source ($5.95 from Jab Tech):
includes dual inverter and 5 feet of el tape, approximately .25" wide. Right now, I think they only carry pink. It's ugly, but very cheap. I used it for another project, where I needed a lot of surface area, as a cheaper option to el paper.

Dimming with el stuff:
Previously mentioned --Some vendors sell dimmers, but they are very expensive (over $100++.) Tom helped us hook up a dimmer to an el strip. We only got about 15% variation -- not very noticeable. Little payoff for big effort. This is what I've also heard from other people (Ram & Ty on the solar wallpaper project, and Todd) who've tried to dim el panel. We ended up ditching this functionality.

• Velleman Kit VK-8064 DC controlled dimmer (269 Canal is known to carry this, too.)

Some thoughts on electroluminescent stuff:
• This stuff is really expensive and it's a total power hog. The latter was a great motivation for adding the solar element to this project -- to help offset the energy consumption.
• There is a lot of voltage/current running through these things.Each strip of 5 sq. inches draws about .1 A. The voltage at the input of the inverter is 5V (powered by 5V), at the output, going into the el strip, is 85V. That can result in a significant shock. The strips I used are optimized for 8-10 VDC. Power requirements depend on the surface area of the electroluminescent material. For the final prototype, I ran this off of a 9VDC -- which would probably get us 150V. Make sure the connections with the el stuff are well insulated, or at least, out of reach.
• If this shade were to really sit in a window, we would really have to make sure the wires were well insulated. (People often accidently leave their windows open when it rains.) The energy consumption of this material (if you're using a lot of it) can make it hard to spec out a battery for portable, mobile, or wearable projects. Not sure if I'd want to be packing that much heat -- might be deemed unsafe.
• However, it's thin profile makes it ideal for many applications. It can require less wiring than LEDs using edge lighting, etc.
• It comes in a whole suite of incredibly tacky neon colors. It's typically used to pimp out trucks and boats, and I'm sure it's quite popular at Burning Man. I have seen some projects that make less tacky, if not, good use of it (Check out loop.ph.)
• Arguably, it does not give off the nicest quality of light. The color of the light in its off stage and the color of the material in its off state do not match. For example, we went with a blue light, because the material is white when off, and we wanted a neutral white shade. The el strips that give off a white light is pink in its off state
• I used long (5 foot) strips of el tape for another project. I noticed the tape got dimmer the longer it got.
• El stuff will dim over time.

LASERCUTTING:
• general info on getting stuff lasercut at NYU (it's FREE)
• This laser cutter cannont cut metal. This includes tin foil. I tried getting tin foil cut. It didn't work.
• file for this pattern

I used a thicker foil/paper (white on one side/reflective surface on the other) from Adorama, (surly service, but conveniently close to school). I mounted the cut paper on a transparency and cut it into strips. This material was thin enough to fit into the sleeves of the window shade to give a "white" shade. I like the white material better than a color transparency images. With more time, I would've like to have worked on developing some more interesting patterns. [Unfortunately] Getting the tech part up and running always seems to eat into the time I'd rather be using for the design elements.

Some parting thoughts on the project:

The sleeves for the window shade were initially designed so it would be modular, and you could change the image or pattern. The el strip would be a solid material with overlays on top.The original thought there was that people could customize it, or if get they tired of their furnishings and they could change it. I'm actually not sure how big of an effort people would make. Also, with the el strip being solid, you wouldn't have too much light diffusing through the window shade when it's down. If the shade was fully drawn, exposing the solar cells, the room would be fairly dark. I'm not sure how many people would want their rooms so dark in the daytime. (If the shades are never down in the day, the battery will never get charged.) One alternative would require giving up the modularity of different patterns. Instead of using overlays, the el paper and the solar cells could be laser cut to a pattern and enclosed in the sleeve, which is transluscent on the indoor facing side. This would allow a little more light to filter in, like a lace or sheer curtain, or blinds.