General

Why am I doing this project?
It seemed like a neat idea. The project actually contains several areas of interest. Sustainable technology, computing, wireless technology, and design all interest me and this was a way of squishing them all together. Providing services in quasi-remote areas seemed like a good way for me to learn about these areas in a hands-on manner.

What will the project look like?
Here is a simplified schematic for a visual understanding of how the solitude would work. The solar array produces electricity which is stored in the batteries. The batteries then provide the energy to the wireless computer. With the weather station project, the computer would be gathering data from the datalogger and then sending it to a wireless access point on campus where the data would be stored and served to the Internet.

Computer Questions

Web Server Questions

How much power does a computer consume?
A number of experiments have been made to determine the approximate power consumption of Laptop computers. For my prototype, at least, I intend to use a laptop as they are already designed for low power consumption.

Procedure: I conducted my test by connecting the different laptops to a modified KWH meter. (Like what is outside of your house.) The laptop connects to the meter, and the meter then connects to an outlet.

How data was obtained: The meter is unique from others that I've seen. It is not obvious what one revolution of the disc indicates in terms of KWH. Because of this, I decided to use a known power consumer, a 75 Watt incandesant light bulb. It took the light bulb 170 seconds to rotate the disc once. I used some special laws of proportion to then determine the unknown watt consumption of the computer. I have yet to varify if my method is correct with someone who really knows the laws of proportion. I connected the laptop to the meter and got all the settings configured. I then simply timed how long it took the current configuration to turn the disc one rotation. I read that normal (home sized) meters are fine tuned enough to measure such small wattages so it should be accurate.

Below are descriptions of the computers I have tested so far. The Toshiba is what I plan to use for my protoype because it was what is available. (Thanks to Josiah and his brother for the donation.) Unfortunately, it only has 4 MB of RAM which may make things pretty difficult.The iBook was available for testing and I was curious. I knew that PowerPC used less power than Intel chips, but it would appear from my tests that the iBook actually uses less power than the old 486. I was surprised by this and did the test on the laptop twice. I've done some testing with a PowerMac G4 just to see how laptops compare to desktop computers.

A couple strange things happened that probably make the data slightly off.

• The meter isn't easy to read. My times could have been off by a few seconds in either direction.
• The iBook's hard drive (even though clearly instructed not to) spun down in the last minute of the test. I had to turn the screen on and run an application to get it spun back up.

I don't think those two things change the data too much. In addition, I'm not that interested in exactly how much power is consumed, I'm just trying to get an general number. There are however some odd results. If you compare test 3 to test 4, the results indicate that keeping the hard drive spun up actually reduces the power consumption. Also if you compare test 1 to test 2, it would appear that keeping processor cycling off decreases power consumption.

Next I need to do some tests where the drive and processor are actually doing something. In the below tests, there weren't any pages being accessed.

Tested Computer Descriptions
Manufacturer	Model	Processor	Proc. Speed	HD	HD Speed	RAM	Screen	OS
Apple	iBook	PPC G3	300 Mhz	6 GB IDE	???	96 MB	12.1" SVGA TFT color	OS 9
Toshiba	Satellite T2115CS	Intel 486	75 Mhz	???	???	4 MB	broken	win 3.1
Apple	Power Mac G4	PPC G3	450 Mhz	10 GB IDE	7200 RPM	128 MB	17" Apple CRT	OS 9
Apple	PowerBook G3	PPC G3	400 Mhz	???	???	???	14.1" SVGA TFT color	OS 9

Energy Consumption
Test #	Computer	Watts	Screen	Hard Drive	Proc. Cycling	Approx. KWH/month (computer running for 730 hours/month ave.)
1	iBook	7.8	off	spinning	on	5.69
2	iBook	6.59	off	spinning	off	4.81
3	iBook	13.14	on	spinning	off	9.59
4	iBook	13.42	on	spun down	off	9.80
5	Satellite	9.20	off	spinning	off	6.72
6	Satellite	9.18	off	spinning	off	6.70
7	Power Mac G4	20.02	off	spinning	off	14.61
8	Power Mac G4	103.66	on	spinning	off	75.67
9	PowerBook G3	8.88	off	spinning	off	6.48

How do I decrease the amount of power required by a computer?
From what I've read, heard, and found in my tests, the screen is the biggest energy consumer in a laptop or desktop computer. Luckily, I don't need a display for a server. The next largest consumer is the hard drive. It is in theory possible to get rid of the hard drive. Having content that changes becomes an issue but there are ways around it. The next largest consumer is the processor. From the tests it appears that a PowerPC chip does indeed use a significant less amount of power than an x86. I'll work with what I got. :-)

How much energy will the server consume per hour, day, month, year?
According to the numbers above (using the satellite tests 4/5) it needs about 6.75 KWH per month. Keep in mind this is a rough estimate and the above tests weren't actually serving any content. With actual disk access and processor tasks running (which will be tested later) energy requirements will be higher.

Are there computers devices that do similar tasks (wireless web server)?
I haven't come accross anything yet. Looks like if you were a big company and had lots of money, you could easily have a low powered embedded solution, but I am neither a big company nor loaded.

Wireless Questions

What kind of antenna should I use?
It would be nice to use a omni-directional antenna which would provide other areas with wireless connectivity. However, the signal may not be strong enough to reach a remote area. A directional antenna would need to be used for remote solitudes. Using directional antennas can provide connectivity to distances of 10+ miles. The key is to have a line-of-sight which isn't always possible.

Can I make my own antenna?
Yes. People in Australia, Seattle and other areas are setting up wireless networks using low cost equipment. Many of them are building their own antennas. The Australia group has made a antenna how-to page that describes how to build a directional antenna. It looks a little sketchy. If funds can be found, it would be better to go with a manufactured antenna. Pacific Wireless makes cheap directional antennas and HyperLink Technologies makes antenna kits that work with wireless cards.

I've managed to get an antenna which was mounted within a PowerMac G4. Here are a few photos: The whole antenna. the infamous connector which is called (unconfirmed) a, "SMZ" or, "BT type 43" connector. I have yet to find a source for these in the states. Here is the one of the two end plates. And here is Apple's little circuit board that connects the wires.

Can I use an antenna that others around campus can connect to?
Yes. If I can use an omni-directional antenna as mentioned above, anyone with range could connect wirelessly. In fact, I've already started to make this happen by starting Hampshire Wireless.

Should I use products from Cisco (Aironet), Orinico (Wavelan), other?
Any product that supports 802.11b should in theory work. And from what I've read, they do work. However, the Orinico products seem to be the most popular and well suppported.

What sort of range can I expect?

How can I increase range?

How much do trees and other obstacles reduce the singal?

Energy Questions

Wind Questions

What repairs or modifications does the windmill need?

How much power does the windmill produce?
From "More Other Homes and Garbage":
The Savonius, or S-rotor, rotates much slower than a modern two- or three bladed propeller and is less than half as efficient in capturing the wind's energy. It must , therefore, be geared and its cross-sectional area (to intercept the wind) must be quite large to make up for its inefficency. The Machines requrie a large surface area of material and so are heavy and hard to balance. Small do-it-yourself units can be made out of 55-gallon drums (see Hackleman's nice booklet in the Bibliography), but they deliver so little power that they hardly seem worthwhile. For example, a machine made from three 55-gallon drums stacked one on top of the other, having a cross-sectional area of about 27 square feet and an efficiency of about 20 percent of the theoretcial maximum, produces about the same amount of power as a simple high-speed prop with a diameter of less than 4 feet. Hackleman's figures indicate that one such stack will probably yield less than 10 kWh per month in 10-mph average winds. Figure 3.10, in case you hadn't noticed, includes an approximate power-output curve for such a unit. S-rotors are not particularly appropriate for the generation of electricity, although their good-torque, slow-speed characteristics make them useful for such applications as pumping water.

What are average windspeeds in this particular area?

How do I convert the rotational motion of the windmill into electricty?

Photovoltaics Questions

How much power does a PV panel produce?

What are the average sun conditions in this area?
Below are energy values compiled by Solar Pathways Corp. from horizontal data given in University of Wisconsin EES report no. 44-2 (F-Chart Data). This will soon be translated to kWh.

How can I get my hands on a PV panel?

Batteries

What type of batteries should I use?

How many batteries do I need?