Off grid power

NOTE: The contents of this page are not set in stone, and are subject to change! This page is a draft in active flux ... Please leave suggestions on the talk page.

Off grid power

Disclaimer

This page is an overview and discussion of a project in progress. As such, it's virtually guaranteed to contain incomplete, rapidly changing, and possibly inaccurate information that's in the process of being revised. Keep in mind that most of the volunteers on this project are just that - volunteers - with no official connection to OLPC, just a desire to help out and a tendency to ask lots of questions. Don't take anything on this page as fact about what OLPC the organization is doing - this is a grassroots project!

Introduction

Many of the areas that OLPC is interested in deploying into do not have a power grid to use to charge the laptop and run the school server. These sites provide interesting challenges for a successful deployment. The primary option for most of these places is to use some sort of enviroment powered renewable energy source.

Human power options exist but they are "literally" a lot of work. Recharging one of the batteries requires between 15-20 Watthours of power depending on the depth of discharge. 20Wh may not seem like much (its only 17 calories) but in practice its a surprising amount of work when coming from a human powered source.

To provide consistent power things like the school server or VSAT equipment OLPC needs to look at options like solar, wind, water or even really creative stuff like Cow Power.

We're looking for individuals and communities with grassroots groups that want to help with testing and developing non-AC charging solutions for the laptops, particularly off-grid ones. OLPC's Boston office doesn't have space for this kind of testing - there's no roof for solar panels, and they're only exposed to Bostonian weather conditions, for starters, so the grassroots community has to step up.

IMSA has offered to help set up a pilot testbed for the experiment. Once the logistics of such testing labs are worked out, other groups can - and should - follow. Initially, such test sites would gather data and make sure the right information was being collected, then monitor, log, and analyze performance for different designs under a wide variety of conditions.

Read on for more information on this project and how you can get involved.

Get involved

If you are interested in helping, you can help! We need people from all disciplines and experiences. If, after reading this page, you're still unsure of how to get involved, contact User:ScottSwanson or User:Ahw - both IMSA community members and non-IMSAns are welcome!

If you want help with your part of this project, you can post specific tasks below. Please detail what kind of help you are looking for here, a timeframe or expiration date for the posting, along with contact information or how people who are interested should get involved.

Improve this page!

We are looking for volunteers to help improve this project page and keep it up to date. If you see how you can make something more coherent or organized, please jump in and edit; if there's a resource you think would be useful to this project, please add it. If there's something about the project itself you have ideas about, you can add it to this page directly, or start a discussion on the talk page. You do not need to contact anyone about this job - just do it! This posting never expires.

Answer questions

We're generating a list of questions we'd like answered, which will appear below shortly. If you know the answers to any of these questions, or someone who might, please see if you can fill in an answer after the question - or email User:ScottSwanson or User:Ahw if you don't know how to use the wiki.

• Questions coming here
• Questions coming here
• Questions coming here

Participating groups

IMSA

The IMSA] chapter is currently looking like the best candidate solar test lab host, with ample access to supplies, equipment, people, and space (including plenty of sunny roofspace) in Aurora, IL. They are communicating with the IMSA's Green Team (an existing effort to get solar panels on the campus main building's roof, among other "off-grid" projects) about potential collaboration, including a previously-discussed renewable energy museum/lab being planned on campus in partnership with other Chicago-area engineers.

The campus is about an hour from Chicago's downtown, right down the research corridor, close to Fermilab and other locations who would have interest in an alternative energy project. IMSA has no classes on Wednesdays so students can work on a project of their choosing, or do a mentorship at a local company or lab, which means a lot of time and resources can be devoted towards this project.

OLPC Chicago

OLPC Chicago will have a full-time grassroots office running this summer, including full-time mechanical enginer Chris Carrick, who's devoting several months almost solely to power generation and charger design. They'll likely be working very closely with the IMSA team over the summer to support and extend the activities of IMSA's Aurora-based lab.

Potential helpers

Feel free to add to this list if there's someone you feel we should contact.

• IMSA's Green Team
• Appropedia
• Fermilab
• MIT Solar Decathlon team
• Olin chapter - reaching out to Olin's environmental club, professors, alumni working in the sustainable-power field.
• MIT's Amy Smith and her D-Lab and summer IDDsummit] - possible summit projects?

Existing technology

These are existing components that are or will soon be available to work with for testing. It's possible that much of this section should be moved out to the main power page fairly soon.

AC adapters

The default option; normal AC adapters / wall warts / power cords. This is what is currently shipping, and has worked so far because all the initial deployment sites were able to get access to electricity. This will not work in the long run, and is what we need to design and test alternatives to.

Solar panels

The OLPC office in Boston has been working with solar cells intended for use on roofs or flat surfaces. Specs follow.

• Enough to charge a laptop at 12-14V, 5W (so .35 to .4A). There may be a 7 and 10 watt option, but this has not been confirmed with PV Solar.
• Roughly 1.5ft x 3ft in size
• Amorphous Si (Silicon)
• Bound to a flexible plastic backing
• Appear to be fairly rugged, but haven't been solidly abuse-tested
• Manufactured by PV Solar, a part of GoldPeak; GoldPeak is a Hong Kong based company that makes batteries.
• Panels have been measured by OLPC-Thailand in a solar simulator. Here's the IV cuve. File:5W IV Curve.pdf

Questions:

• Are these being mass-produced?
• Are they commercially available? If so, where and in what cost/quantity?

Solar blankets

Other lower voltage panels are used in series to form a solar blanket. This a heavy tarp with 10 panels attached to it in a configuration that yields 60W of output. A blanket would have a 7ft x 4ft layout. Blankets are made by the same manufacturer as panels, from the same substance, and aside from size and power output the same considerations apply.

Questions:

• Are these being mass-produced?
• Are they commercially available? If so, where and in what cost/quantity?

Multi-battery charger (MBC)

The Multi_channel_battery_charger is a charging "box" that has slots for 15 batteries. It is rated at roughly 150W for DC-DC connections, but its power input can vary. The MBC is undergoing active testing and development and is not ready for production yet; only a few prototypes exist.

How to help

This project is a collaboratively organized undertaking, so the best way to get started is to just dive in and introduce yourself. Since the physical lab is likely to be based at IMSA, discussion occasionally takes place on the #olpc-imsa IRC channel.

Scenarios

These are possible field setups using the above technology that could be used (and should be tested) for keeping XOs powered in the field.

Single user single panel

This setup consists of:

• 1 user
• 1 XO with battery
• 1 small solar panel (5W)

Multi user multi panel

This setup consists of:

• N users
• N XOs with batteries
• N small solar panels (5W)

Solar-powered bulk charger

This setup consists of:

• 60W solar blanket powering an MBC
• up to 15 XO batteries at a time

60W would charge 4-5 batteries in about 2 hours. So a 6 hour solar window would charge 12 to 15 batteries. The "small and large school, MBC" scenarios presented below are variants on this one.

Small school, MBC, no extra batteries

This setup consists of:

• 1 MBC, solar-powered
• N users
• N batteries

Small school, MBC, double batteries

This setup consists of:

• 1 MBC, solar-powered
• N users
• 2N batteries

Large school, many MBCs, double batteries

This setup consists of:

• M MBCs, solar-powered
• N users
• 2N batteries

School Server Option

In all but the smallest cases a school will want a school server. The school server will generally need to be powered during the school day and ideally it would be available 24/7 but sleeping in a low power mode when not needed. Unlike charging the laptops the school server will not be very effective if a cloud passing overhead causes the power to droop such that it turns off. Therefore the school server needs a continuous source of renewable power. The current estimate of a fanless low power server with 3 mesh antennas and a hard disk is 24W average power draw. So for an 8 hour day of continuous server use you would need 192Wh of power. If you add in 4 hours of margin then you need 288Wh, and if you wanted to reserve a full 8 hours of use and be able to replenish your reserve in the same time frame then you need 2x or 384Wh. These numbers are a simplification since they assume continuous maximum output and no storage losses. But start to give an idea of the problem.

• need to start creating a grid with all the gory details

Long Haul Communication Equipment

In general an area with no power grid won't have much of a communications infrastructure. In order to connect the sites to the world at large some sort of long haul communications equipment is needed. VSAT is one of the options that can provide this commnuication. So the additional power budget for VSAT equipment must be factored into the overall power budget. A further complication is that in order to keep the VSAT costs as low as possible a lot of the higher-bandwith commnications (loading up e-books, updating library content, etc) may need to be done in off hours. These hours generally don't coincide with daylight hours so unless there is enough power to maintain continuous operation of the communications equipment some sort of scheduler and remote power switch will be needed.

• need info on power draw of the average vsat setup.

Testing considerations

Safety concerns

LiFePO4 batteries are being used in mass-production (not NiMH), so safety issues while testing ("exploding batteries!") aren't as much of a concern. (References to data here would be nice - can someone please add?) However, LiFePO4 batteries are a fairly new technology, so not as much is known about how they will handle in OLPC-specific conditions.

Equipment needed

This is a list of equipment test sites will need to participate in this project. Note that this list may be incorrect or incomplete, and that it is not yet clear where or how all these materials can be obtained.

• Access to XOs
• Access to chargers
• Monitoring equipment
• Computers and software for analyzing log files
• Internet connection

Timeline

Upcoming

• Beginning of March 2008 - A deployment in Peru will potentially need off-grid power access. (More information forthcoming.)
• March 2008 - A pilot in Africa with only solar power begins. (More information forthcoming.)
• June 2008 - OLPC Chicago's Chris Carrick begins full-time work on power solutions, including this project