Battery and power

From OLPC
Jump to: navigation, search


Removing the battery from a B4 model OLPC laptop.

OLPC has put a lot of effort into building a device with very low power consumption as it will be used in areas with poor or non-existent power infrastructure, with electricity often very expensive for the user.

Early concept devices were shown with a hand crank on the side to demonstrate that they would work in areas where the only electricity available comes from devices like the Freecharge portable charger. This was removed due to concerns about stresses on the casing, and also ease of use.

Replacement Batteries and Chargers

Replacement XO batteries and chargers (and other parts) were available for sale from this independent Michigan (US) store:

http://iLoveMyXO.com (as of 2016 now a blog talking about WANs)
http://twitter.com/iLoveMyXO (inactive since 2013)

XO-1 Power Input Specifications

Read more details in Hardware specification. See also XO_power_draw and plug_packs.

Also see Battery replacements for why opening up your XO's batteries is a bad idea.

Electrical Specs

XOs need a DC input ranging from 11 volts to 18 volts to charge the battery inside. This is far more flexible than most portable devices. Polarity is positive (+) center pin and negative (-) outer. Reversing the connection will not harm the XO, but it won't charge either.

Switching regulators as employed in OLPC can be a very demanding load for a power supply. The input current typically exhibits a 1/Voltage characteristic (this e.g. means that an OLPC needs 3 times as much current at 5 Volts than at 15 Volt).

  • Rsmith 15:18, 28 February 2009 (UTC) Only in theory. The OLPC switcher will not turn on at less than ~11V.
    • --Frederik.Questier 14:40, 23 June 2009 (UTC) 9V 600mA proves to be enough for charging (slowly) the battery during off-mode. During on-mode it's just enough to keep the battery charged during light use (reflective screen mode).

The impact differs depending on the type of power source and can range from small over poor efficiency to malfunction.

Following contributed by Sreeram Dhurjaty:

The switching power supply of the XO will tend to be more efficient at the higher voltage. A higher voltage, for a given power rating, will draw less current. Switching FETs (used in the XO) have an on-resistance and therefore the losses will be lower at lower currents for a given FET. Furthermore, diodes that are used in rectification will be less lossy for lower currents as the forward voltage of the diode is multiplied by the average current in order to determine the loss.

  • Rsmith 15:43, 28 February 2009 (UTC) Again in theory yes. But for the XO the overvoltage protection in the front end of the switcher has an input impedance that _increases_ as the voltage increases. The optimum input voltage varies from XO to XO depending on component tolerances but its about ~11.5VDC.

Be aware that when "wall warts" are used to generate the requisite voltage,there may be a greater loss, as classic bulky "linear" types (still popular in developing countries)have efficiencies ranging from 30-50%. For example, consumption of 2 Watts by the XO (assuming 90% efficiency) will necessitate a power consumption of 6-7 watts from the mains. Wall warts also tend to be more efficient at higher voltages due to lower copper losses that are entailed by lower current-draw.

Mechanical

DC socket specs

  • 1.65mm center pin diameter,
  • 6mm outer barrel diameter,
  • 10mm contact length (measured from innermost surface of green casing countersink).

DC plug specs

The DC plug is actually standard. However, these particular dimensions are not frequently used.

  • 1.7mm centre pin diameter,
  • 5.5mm outer barrel diameter,
  • 9.5mm contact length.

The large spring in the socket pushes the plug sideways so that the contact surfaces mate.

Mechanical Drawing File:DCJack.pdf

Note that a mating plug is shown on that drawing.

This connector is somewhat non-standard. The "glass is half-full" people say that a completely weird connector is a good idea to avoid kids being mugged to have their charger bricks stolen. The half-empty guys say it's a clever conspiracy by some unmentionable dark force to keep control over OLPC by throttling the delivery of plugs to hardware hackers. I need plugs to deliver XO Solar kits. Yamaplos 17:18, 19 April 2008 (EDT)

With luck, this connector will be revised for Gen 2. See this ticket.

Mating plugs that fit

  • Digikey part #CP-2199-ND (right angle) Current recommended part, tested at the OLPC office in Boston
  • Digikey part #CP-2195-ND.
    The barrel for these parts is 4.75 mm and the length is 9.5mm but they still fit nicely. The right angle part won't swivel around 360 degrees because it interferes with the high side of the plastic casing. 9.5 mm looks like its about the smallest you can go with the length. These parts have a 6ft wire and a molded plug housing, and terminate in 2 bare wires. The outer cutout in the plastic casing of the XO is 10mm so make sure that the plastic molding of the plug is < 10mm. These plugs conform to the inner dimensions of the barrel better, which is optimal because the spring contact that touches the outer barrel has more flexibility to size variations.
    (courtesy Ed Leslie)
    The Digikey CP-2199 and CP-2195 look like standard EIAJ RC-5320A-03 jacks which are really meant for 6.3...10.5 V. "In theory", these should be easy to get.
  • Farnell offer an existing adapter suiting Aust/NZ & the Pacific Islands mains plugs. Australian firms Jaycar & Dick Smith both also sell (~Aust$20) a compact & versatile mains plug pak, delivering 12V @ 700mA, that'll nicely charge an XO.
    OLPCplug.jpg
  • 5.5x1.75mm connector as part of a laptop plug set made by Velleman. A distributor in New Jersey, Tequipment, sells the plug set (Model# PLUGSPSET3 adapters for USD $3.80. You'll have five extra plugs in other sizes to add to your parts box for other projects. Ordering from Tequipment was easy and delivery was prompt. I have photos posted on Flickr showing the plug and the slight modification necessary to use it with the Radio Shack 2-pin adaptaplug socket or you can solder directly to the pins on the back of the plug and seal appropriately. --Mike Lee 21:37, 4 June 2008 (EDT)
    • Velleman also sells a PLUGC5 "lighter plug" for USD $3.00. One would expect that they would mate, but the PLUGC5 end is round and the PLUGSPSET3 adapters are rectangular so they don't mate out of the bag (if you look at the pictures closely, you can see the round/rectangular shapes). I used an X-Acto to cut the rectangular shield off the adapter which allowed them to mate. A little 1/2" heat-shrink tubing (big enough to slide over the plug, small enough to shrink tight) and it looks great. --gvb 29 June 2008.
  • Kensington tip N2 and N2B work with User:AaronPeterson's XO. the package of his universal Ultra Portable Notebook Power Adapter K33197US says that plug is for certain IBM/Lenovo laptops. It's not a perfect fit.
  • Someone at a San Francisco OLPC meetup had actually cut his original XO power cord and put RCA connectors where he cut it. This way he was able to use the original connector with various power sources (solar, car battery, etc) -- perezda
  • A small number of free raw connector samples are available from an OLPC experimenter here.
  • In Europe, a number of computer appliances (Modems, Routers, Switches, Hubs) come with power supplies that have the right power range and connectors that fit with the outside diameter but have an inner diameter that is too big. These can be adapted by putting a bit of aluminium foil into the inside of the connector. Be sure not to short-circuit!

Mating plugs that DO NOT fit

  • Digikey CP-014-ND is not the right thing. I ordered a few on 24 April 2008. $18 bucks down the drain. Yamaplos 16:42, 27 April 2008 (EDT). Richard Smith adds: The problem is that it's too short, by about a .5 to 1mm. The interesting note is that the datasheet specifications are exactly the same as the plugs that _do_ work. So I dug a bit deeper. The end result is that a barrel length of 9.5mm is on the hairy edge of too short. The reason one set of plugs work and the others do not is that the the CP-2199 and 2195 plugs have better holding force inside the connector. They grip the inner pin much better, so they still work. They feel like they have some sort of internal taper that holds the center pin tight. So if you are trying to find connectors:
    • Even if the specifications on the datasheet are the same they need to be tested.
    • I think the minimum length needs to be >=10mm and perhaps 10.5 mm.
  • Radio Shack part 274-1569 is a pair of 5.5 mm OD, 2.1 mm ID plugs, one of which can be soldered to a source of power (in my case, Radio Shack part 270-1559 -- 12VDC cigarette lighter plug with a 2.5 m cord with an on/off rocker switch and green LED power indicator) to make an inexpensive 12VDC power cord. However, when trying to use these Radio Shack parts, the connection was too loose to provide a solid connection. I wouldn't recommend that people purchase them.

Multi channel battery charger

OLPC is designing a charger for schools that can charge many OLPC batteries at the same time, and can also power running laptops so that their batteries can be removed and charged. This charger can be powered by a high power AC or DC source.

Mains Power

Mains wall power- IF AVAILABLE -(with a transformer based device)
Little to no problems.
Mains wall power (with a switching regulator based device)
Most of these devices use a current limit or current foldback technique, & may be problematic on startup. A 60 Watt notebook (19V, 3.16A) supply could handle 6 OLPC but it may have difficulties providing their startup current.

Alternate Power Sources

Diagram illustrating problems to be considered in delivering power to the laptops

The Peripherals page tracks some commercially available alternative power products.

Diagrams like "Current versus Voltage, charging/not charging" and "Current versus Time, @6V, 12V, 24V" would be needed to give good advice on dimensioning supplies. A means to limit or control the input current of the OLPC by software is needed to make the most of some of the above given energy supplies.

For most of the below cases, a simple regulator, that only connects input to the OLPC when the voltage rises above a programmed threshold (determined by the power source - 11V for a lead-acid battery to avoid deep discharge) would avoid the problems at switchon.

Home Power Magazine regularly explores alternative power generation solutions targeted toward off-grid locations.

Animal Power Alternatives

Cow Power

See Cow Power for documentation of the cattle powered generator at the OLPC pilot project site in Khairat Village.

Hydroelectric Power Alternatives

In areas with sufficient stream flow and elevation drop, small hydroelectric installations may be possible.

I'd like to expand on this idea. What about some sort of waterwheel hydro generation to put on the downspout of a school's roof during the monsoon season? Of course it's seasonal, and erratic, but it seems like a decent amount of energy to harvest. Hbonwit 11:21, 1 May 2008 (EDT)

Salvage Old Motors

People could salvage old electric motors and turn them into a portable generators which could be run by bullock power or running water.

Wind Power Alternatives

Wind mill (turbine)
Poor efficiency if the voltage is "stuck" at a level at which the switching regulators start. Substantial support towers & stays will be needed,& all up small wind turbines are not really cost effective in the 10W range. They're additionally noisy & act as "visual magnets" for stone throwing vandals.

Hand held Wind Power

A wind power source on a individual level could be more efficient when it comes to wind power instead of the conventional 20 foot tall 6 feet wide wind mill trying to power 20+ XO laptops. Individual wind chargers will have the children find the best spot for wind gusts or become the wind creator themselves through kinetic energy.

The portable, power storing HYmini comes to mind. It is 5.4 inches high, 3.4 inches wide, and 1,3 inches thick. Its 1200mA/h lithium ion polymer rechargeable battery stores power from three sources.

  • Wind power from a small palm sized fan where a 9 mph gust will offer 5 volts
  • Solar power from their cardboard solar panels
  • Wall socket power from a AC/DC power adapter.

The stored charge can power most devices from the supplied USB cable or one of the five adapters. their website gives the example of 20 minutes of uninterrupted 19 mph wind will provide an iPod with 30 minutes of power. While the price is high at $50, a bulk order or a chat with the company could bring the price to reasonable limits.

It even comes in a white and green color scheme.

Solar Power Alternatives

Please also see: XO Solar

Basic Electricity -- Batteries

For an introductory overview, see: Power when it's not always available

Server with Solar Power

This server version could be built with a project laptop with the lid sealed shut. This machine should be a low power device -- a server should operate by solar panel in the lid or wind. It can be left unattended at high points in the area. This could be mounted on a pole or large tree at a hill/mountain top.

Furthermore, solar panels might be built into the laptop lid to charge the battery. I don't know if this is possible, but I am sure that the sun is the most readily available source of energy in the third world. Chemists have found a way to make cheap plastic solar cells flexible enough to paint onto any surface or mold into the case and able to provide electricity for wearable electronics or other low-power devices. Several available technologies such as Quantom dots have recently, made 10 x leaps in efficiency and are very cheap to produce.

We should separate the server idea (which belongs on some other page) from the solar power idea. Since the laptops will accept any DC power source, external solar panels could very well be used to provide that power. They don't need to be built into the lid or anything.
Solar panel-directly connected
Photo Voltaic (PV) panels deliver an output current roughly proportional to illumination,& need BRIGHT sunshine to achieve their maker's ratings. Outputs lower under load anyway- a typical 2W PV rated 16V OC (open circuit) & 125mA SC (short circuit) may deliver only 12V @ 100mA when charging- XOs need a minimum charging supply of 10V as well. See [1] - since there's no regulation with a direct connection, efficiency depends on diverse factors. Four types of PVs are now usually available- Amorphous,Poly Crystalline, Mono Crystalline & CIS/CIG(Copper Indium Gallium Selenide),with choice influenced by ruggedness & overcast performance as much as efficiency & cost.

Cartrickle.jpg

Even in bright sun, a 2W PV is really much too small for realistic XO charging however, & "out of the box" XO's have been found to need a bare minimum of a 10W solar panel to charge their 6.7V 3.1Ah battery in any sort of user friendly time frame! Even in midday sun a 10W PV will take 3-5 hours,& under overcast/hazy conditions this may frustratingly take all day. A 20W PV (costing ~US$100+) is hence more realistic for regions with marginal solar resources. Aside from "mad dogs & Englishmen",strong clear sky tropical sunshine is of course usually AVOIDED by most humans,& cabled solar setups (with the PV in the sun & human XO users in the shade) are MUCH preferred. Alternatively a rechargeable battery pack can be conveniently solar recharged & THEN brought indoors to the XO. Jumppv.jpg
N.B. XO solar charging is perhaps THE most viable,reliable,unobtrusive,versatile & cost effective technique in the OLPC target regions, & further testing -including promising CIS/CIGS types that suit hazy skies-can be found on the XO Solar page.

Some solar cells in the case can be useful, they will be charging batteries slowly every time of the day, while walking or just using the OLPC outdoors. And they can provide some extra time when the computer is off, in mesh mode.

Even the cheapest Chinese calculators (worth of USD 1) are equipped with SOLAR BATTERY - so it must be very cheap. And it is very light. And it is very wear-resistant. So why isn´t it mentioned in OLPC design? Not as the only source. Not as the main source. But why not allow it to contribute in energy balance of OLPC? Katerina Tlusta, hajtl@volny.cz

Most cheap 'Chinese' calculator's solar cells need only supply power in the order of a few microamps, while no less than a miliamper (or more, depending on temperature) would be needed in order to counter the self-discharge of a typical NiMH (this is not a capacitor as in BEAM robotics ;->). IMO, at least half a Watt of power should be considered worthwhile for most uses, providing roughly an hour of uptime after a day of sunbathing. Don't forget, that the user already has the option of connecting a proper small panel to the DC connector of the current design! -- bkil 20:08, 14 April 2007 (EDT)

The military frequently uses a folding solar panel to charge laptops and devices in remote places. Though the panels can be a bit expensive they are very rugged and a 10 watt or larger rollable panel which is cheaper if not as compact could easily charge a few XO laptops and possibly even operate them while charging. It might be feasible to apply one of the flexible solar modules to the back of an XO screen. --Brandonj 00:41, 16 November 2007 (EST)

Solar or Fire heated Stirling Engine

The Stirling Engine [2] is an old but none-the-less smart thermomechanical device that transforms heat energy into mechanical power (rotating shaft) that could drive a small bicycle generator. The stirling engine has two heat exchangers. One that is heated up by e.g. a fire or solar heat (maybe concentrated by reflectors). The second one has cooling ribs to stay as cool as possible. The larger the temperature difference of the two heat exchangers the more power it can produce given a fixed engine size. The automotive industry has improved the stirling engine for years now and has achieved an astonishing efficiency. As a small, simple and rugged battery charger produced in large quantities it could be low price (a few dollars) and it could be heated by a whole range of heat sources. Solar heating should be preferred since kids and fire is often a hazardous combination.

Solar or Fire heated Thermo-electric Generator

'No moving parts' makes these attractive and reliable. They are direct heat-to-electricity, instead of heat-to-motion-to-electricity. Direct fire might take them to solder-melting temperatures - maybe use a sealed heat-pipe, boiling water in the fire and condensing steam on the TEG. Solar concentrator (lens and/or reflector) can be designed to limit power input. Cold water from a stream or water evaporation could provide a 'heat-sink'. Tellurex make devices to generate 1.5 to 5.7 Watts of electricity from a 150 degC heat source and 50 degC heat sink, costing $22 to $37. Might need 2 in series to increase the voltage, or if the temperature difference is lower. Quantity discount applies to 25+. Sounds worth investigating ! Bismuth Telluride isn't the nicest material, but not generally high risk unless you grind it up and breathe or eat it.

Chemical Power Alternatives

Wood Gas Generator

Electricity can be easily produced by generator powered by an internal combustion engine converted to run on wood gas. Gasifiers create combustible gas through the burning of wood or charcoal, that can then be used to power a gasoline generator with a slightly modified carburettor. Gasifier can be made from common materials such as 50 gallon oil drums and metal cooking pots. I am enclosing a link to a research document written by FEMA on the use of gasifiers as a fuel source during a petroleum crisis or fuel shortage. http://www.webpal.org/webpal/b_recovery/3_alternate_energy/woodgas/fema_wood_gas_generator.pdf

Direct Methanol Fuel Cell

MTI Micro, a subsidiary of Mechanical Technology, Inc. has developed direct methanol fuel cell (DMFC) technology that promises 2-10 times more up-time in normal portable devices, as compared to lithium and nickel-based batteries, and can re-charge without wires. I don't know specifics (costs, voltage, how recharge works exactly), but there could be a very interesting partnership with OLPC here... MTI Micro has been supplying power solutions for U.S. government laptops and packs, but are now moving into more consumer-base markets, and have just entered a deal for Samsung cell phones, to be shipped in 2009.

(Added comment: Methanol is quite poisonous. -- Nick Bodley)

(Added comment: So is nickel, cadmium, lead, sulfuric acid etc. Let's not lose focus here -- Matt Burnham (whatwouldmattdo )

Fuel Cells

When you look at military PDAs,they are powered by fuel cells that aren't so available in these countries and are not in the price range being considered. We have to do something with the options available.

Buying 5 to 10 million fuel cells, and taking into account that the rest of the world would need or produce something between 100 and 1000 millions fuel cells, they could be in the price range. BUT fuel cells are going to be disposable or refillable, with all the distribution problems it may have in many of the areas where OLPC are supposed to be used.--S112 02:07 16 August 2006 (EST)
The US military does not use fuel cells in PDAs or in any field equipment that is general issue. The main source of power is batteries. The problem of powering all of the electronics soldiers must carry is very similar to getting an XO charged in a remote location. The military's research into powering those devices can cross over to powering XO laptops. --Brandonj 00:21, 16 November 2007 (EST)

Biological Fuel Cells (aka. Microbial Fuel Cells.[3]) are fuel cells that generate their electricity not from hydrogen, methane or methanole as known from conventional fuel cells but they actually digest organic matter (leafs, fruit, meat) and generate electricity. No kidding! There are already prototype robots driving around the countryside feeding themselves on what they find on the way! The structure is rather simple. You have a bioreactor where you put the organic material. Special bacteria living on electrode surfaces digest the "food" and deliver electriciy. Therefore such a power supply lives from the land as people and animals do. [4] [5]


Human Power Alternatives

Bicycle Adapter

Practical trials

I've been trying to get the XO to charge from a bicycle alternator (bottle dynamo) this week, but no results so far. I am using the standard "Miller 12V 6W" alternator here in India that puts out between 0 and ~24 Volts AC. The set up is currently: AC voltage in goes through diode rectifier with three caps on the output (4700 uf, 0.1 uf, 0.01 uf) and an 7812 to limit the voltage going into the laptop. The four diodes produce a fully rectified voltage, as shown here. The capacitors help smooth out the bumps.

The XO charging light and icon come on, but there is insufficient current draw available to actually charge. To provide more power, I hooked up a second, identical alternator to the input of the electronics above. Again, the charge light came on. I wasn't able to measure current draw, but the 8712 was heating up quite a bit. My thought is that the AC was not in phase, and so there might have been destructive interference. I did not pedal for more than 20 seconds for fear of letting smoke out somewhere along the line.

These bottle alternators can only provide 0.5 Amps, and so I'm not even sure that two would charge the laptop. The stock wall adapter is rated for 1.42 Amps, so we still may be short.

From here there are two options:

Here is a possible schematic.
  1. Try and see if there is a simple circuit to get the two (three?) AC signals in phase.
  2. Try and track down a larger (1 Amp?) dynamo that seems to exist for bicycles.

(Two days later)

Can't get the second option (larger alternator), no bicycle shops seem to know about this.

I hope someone else can try the first option. The "x3" by the capacitors mean to use what is mentioned above (4700 uf, 0.1 uf, 0.01 uf)

Hbonwit 10:27, 1 May 2008 (EDT)


Giving each dynamo its own bridge rectifier then connecting the outputs together at the capacitor will solve the problems with them being out of phase 86.16.36.41 10:50, 27 August 2008 (UTC)

Some ideas for future tests:

  • (optional) the OLPC can use up to 18 V for charging. Perhaps a 18 V regulator (7818 or equivalent) would work slightly better than a 12 V regulator (7812). (You can add 2 resistors to a 7812 or 7805 or LM317 to turn it into a 18 V regulator).
  • It's possible this 7812 is getting so hot that its thermal overload protection has tripped and it no longer lets current through (until it cools off again). Perhaps a heat sink would help?
  • The laptop already has a regulator built in that is likely more advanced and efficient than anything cheaply available, so do away with the one in the bicycle charger circuit to significantly increase power efficiency. Seeing as a lack of power from the 6W generators is the problem, over voltage is not a problem, as the laptop will keep drawing more power up to 17W or 25W (depending on XO version) while charging, keeping the voltage low. As Frederik.Questier observed a powered off laptop will start to charge at 9V at 600mA for 5.4W power usage. When finished charging the laptop will be exposed to the full open circuit voltage of the charger, if this is from a 24V AC source this will be less than 34V DC after rectification, still lower than the 40V the laptop can withstand. If the open circuit voltage is expected to be too high a 20~36V zener diode across the capacitors may be used to for over voltage protection. Don't forget to protect the capacitors from over voltage either, if they have a 25V maximum.
  • Switching regulators often have a latchup problem on startup. If that is the problem,
    • One simple but crude way to avoid this latchup is to add lots of capacitance on the input and output of the the voltage regulator, and wait until those capacitors are fully charged before starting the switching regulator. Start pedaling to charge up those capacitors (this should take much less than 1 second), *then* plug in the OLPC and continue pedaling. The capacitors need to be large enough to supply all the energy during the brief start-up transient.
    • "soft start" or, even better, "maximum power point tracking" avoids the surge current that causes latchup. Does this require changes to the charge controller inside the OLPC? Is this something we could include in the next version of the OLPC?
  • Two inductors, one between the high output of each rectifier and the capacitor, would help share the load between the dynamos. It may improve efficiency with even one dynamo.

Debate on Bicycle effectiveness as a method of recharging

Bikes are very good pedal units and are often used by many students and workers. A simple add-on could allow to charge the computer while, for example, going to school or to work by bicycle as usual, this kind of add on could be really welcomed in many countries.

The simple add on it's just 2 rims (one to be added to the user's bike, one to to the hand-crank on the charger ),and a chain. Something like 2 U-bolts applied to the bike rack used to transport the laptop could avoid any stress to the computer case when charging This tool is not a replacement for the hand crank but a complement to it. It would cost less then any pedal units for users that already have a bike.

Another approach to using a bicycle would be to build or adapt something like an indoor bicycle trainer. a trainer is a kind of simple stand, the back wheel goes on a roller which is attached to some kind of resistance device (in this case a generator). sometimes a simple flywheel can be included to smooth the output. this is good if for instance there is only one bicycle (or only one generator/addon) in town: you can set up the "charging station" at a central place like a school. I think you will see from the picture that it should be possible to make these cheaply: just some tubing and a roller. Here's a ~100W (@ unreg ~230V) ex NZ "Smart Drive" washer stepper approach. Ex N.Z. F&P "smartdrive" washer!

http://stonek.com/montevideo/afilador/afilador4707bx.jpg
Bicycles have been adapted to many things... if the third world is good at something is to grab something and reuse it for their own needs. To wit: here's a typical picture of an 'afilador' (somebody that sharpens your knives, scissors, etc.) whose bicycle has a stand that swings under the traction wheel so that when you pedal the sharpening stones rotate—much 'simpler' and portable (it's the bike itself). --Xavi 08:48, 2 May 2007 (EDT)
The roller that presses onto the tyre in the trainer is not efficient - neither is a 'bottle cycle dynamo' - they are hard work and wear out the precious tyres. Much better is a hub dynamo even one costing UK £15. Building a wheel or adding it to a bike will also be an educational experience and the 'flywheel effect' makes it much easier than 'foot-pump' systems. Maybe 'One bike per family/village' would make a good sister project ! --195.137.93.171 15:16, 25 November 2007 (EST)
Maybe this should go into a discussion page, but I disagree with the above post. I think bottle dynamo is the way to go, based on what I've seen here in rural India. Bottle dynamos are widely available and understood. As well, you can use the kick stand and be pedaling in place to charge the laptop (with 0% of your pedaling going towards forward motion) I'm working on getting our XOs powered this way. (Challenge is that some are AC and some are DC.) More info to tie in here: a Wiki.laptop page and a wikipedia page
--Hbonwit 09:35, 26 April 2008 (EDT)

RFI

While cranking my radio at the same time that it was playing, I noticed that there was static. But I didn't think much of it because the whirring sound of the generator was even louder. But then my wife complained that the cranking was disrupting our television reception when I was too close to the wall-mounted TV aerial. Any power generator for the OLPC should be tested for RF interference generation because it will not go down well if the child's cranking interupts vital economic information such as weather broadcasts or market price reports. In remote regions radio is a vital communications link and a device which disrupts radio reception will be frowned upon. Perhaps proper shielding will prevent this or attention to the generator design.

Child Powered Play Charger

Battery charging playground equipment could be connected to a Multi channel battery charger. Unlike other existing human powered options, this solution does not require users to think that their energy is being used to charge the laptop. This is similar to the Playpump concept where water is pumped by children at the school's merry-go-round.

Kinetic energy recovery

As laptops, those computers are going to move around. A Kinetic energy system to power OLPCs might not be little enough to be integrated into the computer. But their users will probably run, dance and jump around even more than they will walk. As the project is already suggesting external power systems, what about recycling or reusing human energy that's already wasted?

It might need a second OLPC battery, instead of another energy storage device.

Just for keeping the kids untied from their laptops, while recharging... I mean, while playing.--S112 02:31 16 ago 2006 (EST)

Darpa has done lots of research on this, and several companies are producing small cheap energy harvesting devices that could be employed and would be superior to the hand crank or pedal. The charging device used in Faraday flashlight that produces power by shaking is very simple and cheap, and could be tuned to charge by almost any movement especially walking.

Another possibility in this category is a backpack that captures wasted vertical motion when walking or running. There have been various news stories about such devices. The top of a backpack is also a decent place to put small solar panels.

IBM has published an exhaustive paper with viability calculations of solutions related to human-powered computing a decade ago, it's available here. They conclude that legs are the best. -- bkil 21:20, 14 April 2007 (EDT)

Among TIME's Best Inventions of 2008 is number 33, the Biomechanical Energy Harvester invented by Max Donelan a kinesiologist at Simon Fraser University that attaches around the knee(s) and recovers up to 5W of "braking" energy while walking. Originally reported in Donelan, J. M., Q. Li, et al. "Biomechanical energy harvesting: generating electricity during walking with minimal user effort". Science 319(5864): 807-810, 2008. DOI: 10.1126/science.1149860.

In addition to human carried energy harvesters, it is possible to install gear in a floor to harvest energy from the people (or livestock) walking above. ex a busy town square or playgound or buillding or animal pen. See an interesting Wired Magazine Article. See also a company with some very cool ideas about energy harvesting: The Facility.

  • nPower® PEG (Personal Energy Generator - 1st (?) actual commercial product that does human powered-energy recovery for handheld devices. Captures & stores kinetic energy generated by everyday activities.

Website: http://www.npowerpeg.com

  • Orange Dance Charge - device is strapped to arm, harvests kinetic energy of people dancing and stores it for charging of cellular phones later on.

Articles:

Keyboard Power

It may be possible to harness the energy expended by typing. Maybe not as primary source, but possibly useful to extend battery life, in combination with other measures. article study

Pulley Power

Instead of cranking something small by hand, use a rope wrapped around a flywheel to increase momentum and output. The operators then pull the rope occasionally keeping the flywheel at a high rate of speed. Has the advantage of circular engineering instead of linear solutions that all lead towards a dead end.

Having a flywheel in no way decreases the amount of power input you'd require. It is, however, much more mass that the kid would have to carry. The child would still have to put in at least 6 watts of power to produce 6 watts of output, and that's just to run a standard OLPC (without charging.) Assume something like 25% efficiency at best, and the solution is untenable. The only way to make the XO human-powerable is to work very hard to decrease its power consumption. The easiest gains would be in changing its software and implementing true CPU idle shutoffs, turning off the radio when unneeded, and reducing power to the screen.

Trevor Baylis, the British inventor whose work led to the original Freeplay radio, has also discussed using a bucket full of water or sand which would raised via a pulley and human power, and would drive a dynamo as gravity pulled it back to the ground.

  • The physics of such a gravitational potential energy system are easily analyzed. Assuming the figures on this page of ~6 W draw are reasonable, one would need to haul a 5 kg (11 pound) bucket to a height of 440 meters (1445 feet) feet every hour, and that's assuming perfect conversion of potential energy to electrical energy. The reality would be probably more like 25% efficient, or less, so assume a 2 km-tall tower. Not very realistic. Does the kid carry around and erect an immense derrick of that height every time they need to charge? Alternately, you wouldn't have to lift it as high if you lifted a larger mass. If you only wanted to lift it, say, 3 meters into the nearest strong tree, you could instead carry around an immense, sturdy bucket capable of holding 734 kg (1620 lb) of sand or water, fill that full of sand or water every time you stopped, rig a complex set of pulleys to give sufficient mechanical advantage so that the bucket's weight doesn't fling the child into the low stratosphere, and then haul it up those 3 meters into the tree every hour. 70.59.34.204 01:16, 6 September 2008 (UTC)

Hand cranked power

First of all, remember that the hand crank shown on the side of the original prototypes will not be used in the shipping units. The mechanical forces of cranking turned out to be incompatible with the laptop case. In addition, the hand crank relies on some of the smallest and most delicate muscles in the human body, namely hand and wrist muscles. As any anatomist will tell you, the strongest human muscles are in the leg. If you want to prove this to yourself then try the "Freecharge portable charger" (see below). You can buy both a hand crank unit for mobile phones and a foot pedal unit.

  • The FreeCharge Weza ($269 by ebay) is a foot-pump device similar to the one that will be used with the OLPC.

Freeplay Hand-Cranked Generator in development

Freeplay , the 'clockwork radio' people, are developing a hand-cranked generator specifically for OLPC.--195.137.93.171 15:37, 25 November 2007 (EST)

That page claims a "firm order" for Peru. How did they work there? How many did they sell? Are they still selling them? Are they available to the general public? How much power do they generate?

Potenco’s Pull-Cord Generator

Potenco’s Pull-Cord Generator (PCG) looks like a large yo-yo : you pull a cord out and a dynamo generates electricity. Claims 20W peak average power.

Which is it--peak or average? Also, their FAQ claims "Neither version of the PCG is currently available for purchase. We’re hoping to get the funding to go into production in the next few months, and that product would then be available in early-mid 2009." (Retrieved Jan. 2011).

Quads are among the body's strongest muscles. Why not put leg/ankle straps on the Pull-Cord Generator? That way, the child could strap the generator to one leg for stability and essentially do leg lifts with the other. It would be easy to switch legs, and this set-up is more portable than a pedal-powered one.

Mechanical Shaking

This weekend I bought an ingenious LED torch (flashlight).

It operates by using a 15mm magnet that slides inside a 9 mm tube that in turn is surrounded by a 25 mm bobbin at the centre of the tube. There is a rubber buffer at each end of the tube to absorb excess impacts and to encourage the magnet to re-enter the coil in the reverse direction. The user just gives the torch a shake, and this shaking charges up the capacitor and hence the battery. The minimalistic pc board contains an on-off switch, full wave rectifier connected to the coil, a current limiting resistor in series with the LED and switch, and is connected to what looks like two disc batteries in series, connected in parallel across a capacitor, and in parallel with the direct-current side of the full wave rectifier.

Similar devices are shown e.g. at http://www.vidcam.com.au/sales/faraday.htm

and at http://www.hc-gifts.com/3_Eternity_Flashlight_Hand_shaking_Forever_Flashlightsmall_623.htm

I was wondering if such an idea could be incorporated horizontally into the base of the OLPC so that when the child walks to school carrying the OLPC, the natural movement charges the laptop.

I have been running this torch most of a weekend, without any deliberate shaking of it, and it has not yet extinguished itself!

Certainly, the OLPC load is more that a single bright LED, but I think this approach will have promise, especially in developing countries where there is no power at home or the school, without the problems incurred by the mechanics of an external crank. --Olpcme 08:49, 31 May 2007 (EDT)

I think the power and efficiency are quite low. I am sceptical about 'most of a weekend' : this site says 5 minutes light for 30 seconds shaking, and the light only needs ~0.01 Watts of power, rather than 2 or 3 we want for OLPC! The cranked models are slightly better - 30 mins light for 1 minute cranking.--195.137.93.171 16:43, 25 November 2007 (EST)
Correction: The OLPC requires approximately 6 watts for normal operation.
Some of those shake-lights are sham. The disc batteries aren't really rechargeable, they're just ordinary long life lithiums that will run out eventually, although they will accept some charge. Try removing the batteries and running the light on the capacitor alone.
A few notes about shake flashlights/torches -- They use a supercapacitor, probably one farad, which can be damaged by overcharging; the few that I've seen don't have overcharge protection. The magnet is extremely powerful, as well. I've owned a couple of these, and imho they are not practical (they work, but the supercap discharges fairly quickly). In mine the magnet sticks to the bumpers and does not slide freely at all, but vigorous shaking does work. Far better is a well-designed crank flashlight. -- Nick Bodley

Balancing Musculoskeletal System

The human body, at its best, is a nicely balanced symetrical machine. My experience of hand-cranking is that it is too one-sided. You always crank with the right hand which means that it tires while the left hand does little work. Over time this means that the muscles of the right hand and arm will become a lot stronger than those of the left. This musculoskeletal imbalance can lead to various problems including back pain caused by unbalanced muscular tension. This is likely to be a greater problem with children than adults, because a child's musculoskeletal system is still developing.

Two-sided cranking

I tried to crank with the left hand but there were two problems. First, I couldn't see the charging LED which was disconcerting. I don't know if it is necessary to see it, but since it is there, using the crank system backwards and left-handed seems unnatural. But the more fundamental issue is that I have to crank in the opposite rotational direction when I use my left hand.

There are two possible ways to solve this for the OLPC. One way is to have a removeable crank that attaches to either the left or right end of an axle that runs through the generator. This way, I don't have to flip the device around so any indicator is still visible. And, when I crank left handed, it rotates in the same direction. Fatigue sets in after ~10 minutes anyway

Pedaling

The other possibility is to get rid of hand cranking entirely and use a pedal on the charger. The FreeCharge Weza (site indicates no longer for sale and none in stock) was a foot-pump device with a pedal that works like those for inflating airbeds except the pedal runs a charger. In addition to being easy to pump right-footed or left footed with no modifications, it uses the much stronger foot and leg muscles. If the generator still maintained an exposed axle, then it would be easy for users to rig up animal powered devices to turn the generator instead of the foot pedal. For an example of the pumping mechanism, see if you can find a working model of a Singer treadle sewing machine, a treadle potter's wheel or a pump organ. The treadle mechanism is a larger version of what needs to be built inside a pedal charger.

The chief advantages of a treadle pump system, or the reasons why such a system was successful historically are:

  1. Legs are stronger than hands or arms
  2. Legs have essentially infinite continuous stamina at a rate in the magnitude of 2 to 20 Watts for two legs.
  3. Leg power generation doesn't interrupt work (this is why treadle sewing machines and pump organs are successful)

This is the better option because the kids can use the XO and generate the electricity at the same time.

YoGen Max

Foot pedal that can be placed on the floor then pumped while sitting down, similar to the kick pedal of a drumset.

Website: http://www.easy-energy.biz/products/yogen-max/

Storage of Generated Power

Solar is an excellent hybrid option - but it cannot be a sole solution. There are plenty of places on the planet that coud use the olpc system but have very short days in the winter. It may be much better to separate the power generation issues from the power storage issues - and find a series of collaborative generation options including both solar and turbine with water/wind/human extensions for charging batteries.

I came to this website trying to find a solution for a school in the western Himalayas - where there is very little sun in the winter - and not much water or wind either - and of course there is no mains power! I guess we will look at finding a longlife laptop battery with good human-powered generators available. Maybe some flywheel-based pedal generator that can be built with local materials is best. If we can find a reliable solution for pedal generation, then it may be possible for batteries to be made redundant - which would be great, as they are expensive, consumable, and don't work well in low temperatures. (86.142.27.202 14:54, 7 December 2006 (EST))

Above poster, please look in this article at the section on Bicycle Dynamo practical trials! Hbonwit 11:21, 1 May 2008 (EDT)

Car and Motorcycle Batteries

Car battery, single OLPC
Sensitive to deep discharge (due to thin plates compared to lead acid batteries specialized for solar cells). Premature end of life if discharged down to under 10 V.This battery should be the lightest/cheapest,& easy to find/replace in the launch countries, and should be of good quality. NB spill proof SLA (Sealed Lead Acid) types,especially the global standard 12V 7Ah, may be available "slightly used" almost free from mission critical security firms etc. Such a 12V 7Ah SLA, which a child can readily carry by hand/shoulder bag, will charge an XO several times before it's own recharge is needed. These SLA batteries are often the "engine" in the increasingly popular automotive "Jump Starts",& simple recharging from a car 12V accessory socket may be possible. XoSLA.jpg
Car battery, class room situation
Acid fumes, and high current electrical shorts may be a Health & Safety issue if the batteries are stored indoors. Long cables to an outdoor battery locker may waste supply energy unless they're thick, in which case the valuable copper may need securing against theft for scrap metal!
Car style electrical generator
Older ones need a torque roughly proportional to the current. High internal losses, large torque and poor efficiency if the voltage is "stuck" at the voltage at which the switching regulators start.

People who have the use of vehicles can charge up extra batteries and rent them out to OLPC users.

Reuse of dead car batteries

Most car batteries fail due to decreasing ability to supply starting current. Such batteries can still be able to supply the much smaller amount of power to the laptop and could be charged by solar cells or a manual generator. They could even be placed in a car or truck and charged from a cigar lighter adaptor during work trips. A discarded vehicle battery should be able to run an OLPC machine for a few days after a full charge.

Now that is a prime topic for an e-book to be distributed with the OLPC. Some guy will take his kid's generator, adapt it to water buffalo power and use the beast to charge up car batteries for a fee.
Failure due to not supplying starting current is not that common. Most lead acid storage batteries fail due to a cell sulfating up. At that point the battery is useless. In regard to toxic materials; lead-acid storage batteries contain metallic lead, lead salts, sulfuric acid. Benefits may outweigh hazards, and proper handling minimize them, but they are there. Lead is particularly toxic to young children.
Damaged yes, useless no. Desulfators can help repair sulfated cells. The battery won't ever be as good as new, but it might recover a significant percentage of its capacity, depending on how deeply it sulfated.

Boulder Wheel

For instance: Just as a fairly constant supply of electricity can be produced by a water wheel from the gravitational energy of water descending from a higher to a lower level, so it should be possible, in dry conditions, to use the potential energy of -- for instance -- a large boulder which descends from a height of (say) a meter or so to "ground level".

The way this might be done is to arrange a rack-and-pinion drive with quite a lot of gearing-up; the boulder's platform sits on top of the rack, whose descent is slowed by the pinion and its associated gear train, at the far end of which a small generator is turned by the now speedily-rotating shaft and produces a very small, but fairly constant, current.

The torque-resistance of the small generator would then act, as a flywheel or an air propellor acts, as a "brake" on the pinion's shaft via its gear-train; and the end effect would be that the (very slow) descent of the boulder on its platform produces a tiny but constant electrical current at the terminals of the generator, which is turning at a similarly constant speed.

All that we now need to arrange is that, at intervals and when it has almost reached ground level, the boulder is removed from the platform on top of the pinion, which is allowed to rise (or is pulled upward again) to its 'meter-high' position. The boulder is then lifted to that height -- perhaps via a pulley-system if it is sufficiently heavy -- and replaced upon the platform, whereupon the cycle of descent-and-power-production starts again.

The advantage of this suggested system is that, like a water-wheel, it's simple and robust. Its output isn't quite as constant as that of the water-wheel, since it's interrupted by the need to lift the boulder at intervals. This disadvantage can be smoothed out in various ways, though, from the use of several such boulder-wheels in tandem to the storage of the electrical watt-hours produced in fairly large (perhaps car-type) batteries, these being then 'tapped' by whatever actually uses the avaailable current -- usually a small computer (or several computers) of the type already discussed here.

--Later comment: This isn't going to be very effective. A 10-kilogram weight descending one meter releases 100 joules (100 watt-seconds) of energy. That's probably less than a minute of operation. The human body is not well designed for lifting weights, and children certainly shouldn't be doing it. Anyway, this is not the One Laptop per Sisyphus project. Generally, people, take the time to run the numbers for your proposals; if the numbers don't work out, maybe you should think a while longer before offering the proposal.

The physical analysis of the energy needed and produced by this system is exactly the same as the "Pulley Power" section above, but even more ponderous and unusable. I agree with the previous commenter that you should apply some high-school physics to your solution before posting.

See also