XO Solar

New

For advanced/recent tips, please see the power-engineering mailing list, and its archives: http://lists.laptop.org/listinfo/power

For an introductory article, surveying key electricity concepts and an example approach suitable for very small schools, see: Power When it's Not Always Available by George Hunt.

See also: Peripherals/Solar_Panel

GP Solar's 10 Watt panels

May 2010 Update: http://iLoveMyXO.com is expected to sell 10 Watt solar panels for XO-1 and XO-1.5 laptops starting in the summer of 2010. For their latest inventory updates, please follow their Twitter newsfeed.

A day by day log of how GP Solar stuff works

The above GP Solar panels may be your best bet if buying in bulk. However GP Solar's minimum sale may be as high as 500 panels or such (someone please confirm?!)

Please help us all find vendors selling similar such solar panels in smaller quantities.

One possibility might be http://EnergyEmp.com in New Hampshire!

Retail Suppliers [XOexplosion.com currently offline]

User:AuntiMame says she's been selling these since June 2008

5W Solar Panel. Yes, these are 5W, but the difference is the ones I'm testing are still very difficult to get, even though they may seem less expensive - you cannot get them... :-) Go with AnutiMame, she's built a reputation for reliability Yamaplos 21:59, 20 March 2009 (UTC)

Foldable polycrystalline solar module

Looking for a solar module small enough to be carried, I was thinking using a solar blanket for its light weight, but it needs a greater surface than a polycrystalline module, which is heavier. For my need I use the Solar Gorilla module which gives 10W under 21 V. Despite the fact that 21 V is within the tolerance voltage of the XO, the voltage needed to be reduced. For that I use a converter and set the output voltage to 12 V. I chose this converter for its efficiency (97 % says the manufacturer).

It works great, it is easily transportable, but it has a main drawback. The price is high: 150€ for the module + 50€ for the converter. --Paquan 08:53, 27 July 2009 (UTC)

Is 10 W the sweet spot under bright sun?

success, 11% to 48%

The current setup is two Harbor Freight 5 W panels, item 41144, rated at 350 mAh each. At $ 50 each, we're far from the $30 solar solution :-(

Sun May 20	finally got the battery to go up!	11% to 48% over about 2 h of sun	clear sky	Yamaplos 15:40, 21 May 2008 (EDT)</sub:>
Wed May 28	first full charge	@ 95 min, 42%, @ 130 min 56%, @ 174 min green light, 97%	partial cloudy to sunny	Yamaplos 12:17, 29 May 2008 (EDT)

Purpose/Scope

Test solar panel setups for the XO laptop

• What is the minimum panel that will work?

expectation is charging time of no more than 3 hours.

Tech Info - I just booted my XO-1 with a meter on the battery and for ~95sec, it draws 1A, after that it'll draw 1A with full backlight(BL). Setting to minimum BL shows ~850mA and with no BL shows ~800mA. Since it is a 6.7V battery, that's 6.7W(V*I=W) to pretty much run the XO with no power management and everything above that is free for battery charging.

The battery says 3.1Ah on it so for a fully discharged battery, you'd need another 1A or 6.7W above what the XO is using to charge the battery in 3 hours. Roughly. Basically, a 15W panel will do the job as specified(3h chrg) but a 10W will probably be acceptable in most cases. Doug LaRue 12:17, 06 June 2008 (PDT)

• Does size/output of panel affect charging time?

Yes. It's all about amperes, apparently. An engineer mentioned a $600+ flexible roll-up solution, currently available over-the-counter, meant for laptops on the go. Bless his heart.

• Controller

A controller downstream from the panel avoids battery overflow, and possibly also frying the XO. More research is needed before I will try a huge panel. Anyway, for now we should be safe, the panel I am using outputs only 120 mA, precisely one tenth of what the XO OEM charger brick gives. The XO battery apparently has such a filter, and there is also something called an Embedded Controller which controls battery charge and some power functions. I need do more work in this department.

Tech Info - yes, there is a built-in charge controller in the XO so there should be no need for an external charge controller. Only if you are connecting the panel directly to a battery pack will you need such an external charge controller.Doug LaRue 12:26, 06 June 2008 (PDT)

• What happens when a battery is slowly charged by a puny panel all day, which then boosts XO battery recharge?

I was given today a 1.2 Ah Pb Yuasa battery by Texas Solar Power Company owner Craig Overmiller to test that Yamaplos 15:00, 24 April 2008 (EDT)

such a small battery should be "safe", but otherwise I would have the same worry of shooting too many amps through. Tried it on, XO works. Now I will try to discharge it and then see what my panel can do. Simple theory is that it would take 10 hours to recharge (120 mAh, feeding a 1.2 Ah batt), that of course not counting any losses... Way too long. Need a larger panel for real life expectations. On the other hand, the XO runs just fine off the panel. Something is peculiar here. We'll see. Yamaplos 17:44, 24 April 2008 (EDT)

Figured out what is "peculiar". The XO is feeding from its battery even when hooked to the panel. Unexpected, and it might mess up things big time. Yamaplos 17:24, 27 April 2008 (EDT)

and so it was... :-( ~~ Yamaplos 19:58, 29 April 2008 (EDT)

Notes on Solar Panels and SLA batteries

• For maximum efficiency Solar panels need a 'peak power converter'.
• Charging SLA batteries above about 80% from solar panels cab be very wasteful (the charging rate drops off meaning you can't use all the energy available).
• Discharging SLA's below 30% (even 'deep cycle' ones) kills them. Quickly.
• SLA's therefore go from about 80% to 30%, roughly half the rated capacity.

Make available for sale XO Solar kits

The whole point of this is that this hardware be available for purchase over-the-counter, by anyone, very soon. debate on the opinion issue and historical data is in discussion page.

Many G1G1 donors would like to play around with alternative power sources like solar.

There seem to be a couple commercial efforts happening, Product News, Solar Panels

• ETA for XO Solar is May 2008

This is for the full-fledged, battery charging system. Still waiting to come across a larger panel, that probably I will have to pay out of pocket... hint, hint :-)

Expected cost of such hopefully will be under $75.

Transfer knowledge & experience

Use a Libre stance as to sharing what we find so that everyone can go solar if they can access the necessary hardware, without needing to purchase directly from us or our partners.

You can call yours the Jerry Solar Solution if your name is Jerry, and you can still call it that even if your name is Martha. I am getting my panels from Harbor Freight Tools, my plugs from Digikey, which we have not proved yet are the right ones, but we hope...

Facilitate XO Solar stuff for needy projects

Yes Nepal, we hear you...  :-)

This will depend, of course, on gathering funds to subsidize kits and shipping. A better accounted G1G1 has been proposed, in which developed world-purchasers would be able to purchase hardware in such a way that their purchase would make stuff available for needy projects.

Getting the Most Out of Your Solar Panels

The following are questions an answers from Richard Smith OLPC Solar specialist, and questions from participants in April 4th Sunday Forum.

Sunday April 4th Presentation

RE: Solar Input and managing charge levels for XO-1 and XO-1.5 Discussion with guests Richard Smith and Robert Pilawa. When using solar panels to charge XOs, the output of the panel varies according to the amount and intensity of the sun. This means that sometimes the voltage spikes up and even exceeds the limit of the XO controller, and can trigger a shutdown of the charge.

The guest specialists discussed the design of the XO input circuit, and a new device under development by Robert Pilawa that can control the input voltage cheaply.

Solar Presentation Agenda from Apr 4:

http://lists.laptop.org/pipermail/olpc_boston/2010-April/000566.html

Robert Pilawa's Slides:

http://www.slideshare.net/metasj/olpc-apr-4-2010

Pilawa-Podgurski, MIT Mathematics department

Robert Pilawa & Brooke Pilawa-Podgurski, MIT grad students, Cambridge, Massachusetts/MA, USA]' to 'Low cost solar charger for the developing world: low cost solar charger with built-in power tracker, adheres to USB charging standards, can charge most cell phones, portable lights, lead-acid batteries. . It can produce up to 35Watts in full sunlight.

Problem: Hot sun issues in Kenya etc creating unexpected conditions. Everything charges fine and then things mysteriously stop.

Richard Smith Richard went back to manufacturer (GoldPeak) to clamp voltage at ~18V to prevent this problem. Using a small circuit added to the solar panel.Samples of this solar panel were built.Testers all said it worked great. But then Richard himself tested it and discovered another small problem with even this new solar panel.This should now be corrected (this $2 added circuit helps) and is now going into production.

XO Power Design

Charging chip is the entire power source of the laptop. 1 input chip creates power rails for all devices on motherboard. Separate power supplies for each of those devices.

RS: XO 1.5 Changes to Support Solar Charging

XO-1.5 overvoltage at 25V instead of 18V for XO-1.XO-1.5 can limit the power it draws.

If you want to set the XO-1.5 to use no more than 10W, you can use battery as buffer.

Allows you to use panel at more optimum range.(Power draw of XO above 10W!) XO-1 sucks up to 17W; XO1.5 sucks up to 20-25W)

Control duty cycle of charging chip "turning it off for x-amount of time"

• EG. Turn it off 50% of time.. you get 12.5W
• EG. Turn it off 0% of time.. you get 25W
• How does the user in the field accomplish the change?
• Power rail 7.5 - 8V

Question? what is oscillation and is there a fix? Answer: Richard Smith: you'll hear a "screaming" noise Output voltage of panel drops / rises / drops.

Q? There has been talk about oscillation issues at low input voltage/power. have these been fixed already? A: 1-2 kHz range is in audible range.Inductors / capacitors.On XO-1.5 it's MUCH louder. Running change Richard will be ordering to production. Richard has attempted to make on power-limit implementation in field automatic. Is this change firmware-only or an ECO? Both.. (must be XO-1.5)

Q? Is oscillationfix HW? A: Richard: yes Easy modif.. resistor changes.Tiny resistors, beyond normal soldering.Microscope required..All that's coming down the pike..

Other ways of Charging

• standalone charger.
• Problem: MBC (multi-battery charger) too expensive.
• MOQ (Min Order Quantity) too high, perhaps 1000 units.
• New concept: single battery charger?

'Q:?' How about a kit for do-it-yourselfers at a much smaller cost? 'A:' "Leave a Little Room" in Ethiopia / Tanzania requesting 200.NS How about always ordering 2 batteries per laptop? Richard: yes!Mulitple Battery Chargers was about removing wires from the classroom..But now a different driving requirement?Workflows already set up so MBC hasn't caught on so far.Several deployments build their own charging racks..

Locked down sometimes for security.But we try to encourage kids to take XOs home. So 2 batteries would make sense. Nancie: Roger in Thailand discussed battery failures "due to improper shutdown" ?

Purchasing More Batteries:

Note: A Battery size changed in XO-1 version, but all batteries work? $25 batteries available at [1]. $28 panels avail now for iLoveMyXO to buy (100 MOQ)

Robert Pilawa’s Low-cost solar controller and why!

1. charge controller basics

2. max power point tracking

3. cost

4. prototype

Note: Batteries commonly damaged in the developing world, because people buy the cheapest etc. Overvoltage protection critical.Similarly at night, if you charge the battery dry (too low voltage) you will trash your battery, sometimes if you do this even twice.

Q? MPPT in charge controller?? A: Is it only for XO-1 because xo-1.5 has it built-in? or has doing it in the charge controller some advantage for xo-1.5 as well?

'A:' 12V panel has an open circuit voltage of 23-27V? but MPPT of 17V? Think of it as gear box, of car of bike. If you don't have gearbox, you can't get up to 55MPH.So you will be running in 1st gear, and wasting your power. You need DC-DC power converter.. staying at Maximum Power Point (MPP).MPP depends on the temperature and sun. You end up wasting up to 30% of avail power of panel. Solar panel manufacturers keep pressing price down and down.Same with lead-acid batteries.You can take advantage in the developing world.Price pressure coming from household uses.Developing countries can take advantage.But now fancy electronics in the middle hasn't yet had these economies of scale yet.Needs to happen.

• Targetting $0.25 / Watt?
• Low-cost Maximum Power Point Tracker.
• Target: 100W tracker $25
• In the US a household version today can be $20.
• Low cost power electronics help..
• High-volume parts..
• New Target Application: XO / off-grid applications..
• Innovative encapsulation techniques (for Robustness)To recharge lead-acid (car) and cellphone batteries.Put it in between your solar panel & XO.Get the real 10W instead of 7W.As Richard was alluding to.Do this on the XO itself would be nice, sure, but we need external circuity in real world here.Many XO installation?Let's say you have 10 of these old Peak Panels.Daytime charge a lead-acid car/marine battery.Then at night charge the XO's / batteries from the lead acid battery.

Q? Cost consideration?

'A: SJ: typically you can capture maybe 30% or more of avail power not used if you have a low power panel... with less than a 50W panel it wouldn't make economic sense, if you can get a 20% increase in power but could just buy another panel for $4/W. (you'd probably pay $6/W for a panel for consumer work) 20% extra power should be used. This makes sense if the additional cost of the fancy controller is less than $0.80/W, otherwise you'd just buy more panels.

'A:' Robert Pilawa : Current Technology and Costs: Todays' designs areabout $1/W commercially. We are targeting something like $0.25/W. this should be available for testing early summer. We'd love to get people with solar installations to have general field testing on what is important feedback.

Q?Why does cost scale with power? 'A:' Robert Pilawa: panels' costs scale with power.Imagine 20% surplus power coming from 10W converter (eg. 12W). From an installation perspective you should look at it as $'s per Watt. Today's designs cost $1 / Watt

'Q?'SJ :should folks use small panels or large panels.Blankets of panels? EG. 500 square feet of panels? 'Q? ' What if big client arrives? "A:'SJ: rule of thumb too simple 10W panel or 1000W panels very different. Metric we should use $/Watt..

'A:' Robert: that's a market analysis, yes

'A:' Nancie: on Vietnam, she had written primer on designing solar/battery setup.

[http://wiki.laptop.org/go/Projects/Vung_Vieng_Fishing_Vietnam#Power]

What if teachers may grab power for fan & other uses

'A:' Robert: extra power..run fans!Richard: there's never extra power!

'A:'Nancie: adults' cellphones will be charged. there will be issues. cellphones are 1/15th the power perhaps, on the bright side.(1/15th the power of XOs?)

'A:'Nancie: power diagrams on the above page [[2]]

Q? what are most people using in the field?

'Answers:'Richard: 10W olpc recmomended panels. Goldpeak will make... other wattages. We have a 10W b/c its a nice number, they will make panels for you in whatever size. The Solomons will use 15W panels because that's a primary vehicle for them... and the extra 5W makes a diff.

Why and how Regulators or Controllers?

'A:' lot of people try to use these other panels, but: the goldpeak panels is suited... the low-peak voltage is lower than most other panels and most others out there trip the voltage limit quickly(which keeps them from charging an XO) Others with a more advanced voltage regulators are (comparatively) expensive.

'A:' If you have a battery in the system, lead acid or otherwise, you absolutely need a controller. You can think of the XO as being a battery in the system; its controller is on-board inside the XO.

Q? How to manage or view the solar input levels?

'A:'Is the application at the panel, or b/t the building and the charge controller? Where is the controller / device? 'A:' the limiter is a stopgap; this would make it(charge controller) unnecessary.so you'd save a small amount with the limiter Richard: this is much better than limiter (charge controller) which will no longer be neccesary.

Q: What is the Cost? A: (maybe $2)

I should say that I hope to be able to get robert's design implemented as an option you can purchase with the panels.then you could order an olpc solar panel and have the circuit built in.

Q: How long does it take to charge an xo from a Solar source?

'A:' 30W hours with a limiter or what have you?

'A:'Empty to full, with a 10W panel, that's 3 hours. That is max uninterrupted output in reality it's somewhat longer, seldom do you get peak output unless you are in a sunny tropical area.the 10W panel also. It's not that 10W is a limit.

Q: How do you evaluate the required input for charging?

'A:' If you're in Kenya, you can get strong sun and more than 10W output. That's why you use the W-hr number to estimate time to charge.

Q:You're saying robert's device will help that?

'A:'no, if the device is doing max power output it won't

help.'A:' robert: right now you'd pull voltage out at 11 or 12V, right? 'A:' richard: yes, but then the XO gets real hot and it averages out...

Q:How can people visualize how efficient their panels are being?

'A:' Suggestion: the charge-controller could have a display.richard: you could put something on the laptop.B.is working on learning activities to do that. Comment: that would be specially good on the 1.5 where you can get realtime data from the ec.

'A:'Richard: you could get a good running average on the 1.0 (in real time)

Q?: in consumer's tech what are people getting for 'that extra $5' if that's what this controller would cost?

'A:'richard: superior, stable performance in low-light conditions; and would be able to charge devices they wouldn't noramlly be able to (at different voltages:cell phones, lead-acid batteries, etc; (there are additional smarts in there)

Q?: Suggestion Learning Activity NS: As you think about an activity to monitor power coming in, it would be useful to monitor / log each week over the course of the year the gradient, so they know what to expect the following year. SJ suggestion - and also let them record the angle at which they support the panels for optimal charging!

'A:' Activity could include info on the course of the sun around the earth at different lat/longs.(related to the optimal angle, over the course of the day, for each day of the year) and then you can include data on the physics of solar panels what their output is related to solar illumination. How power output changes with a sunnier day, a cloudy day with mirrors reflecting sun onto the panel from many places (embedding each panel in a parabolic mirror?)

Q? What about controlling from a interface

'A:' on the xo-1 it is v easy to understand such things; log data is already supported in the file system. OLPC PowerLog records this information. On the 1.5 and new f11 for XO builds

Discussion on Solar intensity and Solar Panel Optimal use

COMMENT: I have heard from people in Peru that elevation (like in the

Andes) interferes with solar panels. Doesn't seem logical. Is it true?

Q? Some people have said that elevation make solar panels work more poorly

'A:that didn't make sense to me: ' no, it works better at elevation earlier.'A:' if you keep adding solar power, you can get more and more power out of a panel until it burns out.

Q?what data sheets can we get about how the panels work? their max/min power output at different input illuminance, standard deviations,etc.

'A:' They are all on the wiki and Serendipity at 1cc-dhcp-204.media.mit.edu

References

http://www.oksolar.com/roof/

Free Charge Controller project: open-source maximum power point tracking (MPPT) for solar panel, wind turbine, etc.

Major Challenges

Amorphous Si panel

A 120 mAh, 12 V nominal panel (advertised as 1.5 W) is enough to "operate" (~6W is normally needed) the XO, but so far appears as not enough to charge the battery,. This is a Harbor Freight 44768 panel. Retails for about $13 plus shipping, similar price on eBay. Yamaplos 15:00, 24 April 2008 (EDT)

notice the exciting orange glow in the first "in the wild" solar XO, April 22 2008

XO

Since I was lent an XO that is solved for now. Applying for a developer XO so I can return this one. Wish me luck, it seems they're not giving them away that easy anymore.

Plug

The weird-beyond-all-understanding plug. Hope I can get some real ones soon for distribution of XO Solar, but it's a simple 20 minute hack for me to fashion one out of plastic straws and hot glue for testing.

got an email from Mike Lee a couple minutes ago with an offer for a sample of a real one. Great!

Yesterday purchased 10 CP-014-ND Digikey plugs, I have been told these are the right ones.

Caution! These are not the right plugs! I had got my information from the person in charge of power at OLPC, so take any information about this carefully, whatever origin it has. $18 bucks down the drain. Yamaplos 16:40, 27 April 2008 (EDT)

I bought some of those too. In the parts box they go. --Mike Lee 12:30, 4 June 2008 (EDT)

There is a current ticket about a better plug so hopefully this will no longer be a problem for Gen 2 XOs. Yamaplos 15:00, 24 April 2008 (EDT)

home made XO plug

Well, it took me more than 20 minutes, but it worked just fine. The central dielectric tube is from a BIC-type pen, the conductor surfaces from some thin copper sheet, the whole assembled with hot-temp glue. Caution do not use thermofusible plastics for anything electric unless you know exactly what you're doing. Same goes for any such XO hack. You're on your own! (don't that feel good, eh?)

Action

• get solar panels

I've been running a 1.5 W, 12 V nominal one since February 2007...

working setup now is two 5 W, 350 mAh panels Yamaplos

• get XO

got one for now. Borrowed from Austin XO

• get solar energy advice

Met today with wonderful people of Texas Solar Power Company. Yamaplos 15:00, 24 April 2008 (EDT)

• wire things together

done. It works!

well, sort of. It appeared to work, but actually kept discharging

Actual charge as of yesterday Yamaplos 15:43, 21 May 2008 (EDT)

tests log

Space reserved to report on what happens.

Uh, need to clean up things. add some schematics. when I'm back from Camp Texas Yamaplos 09:29, 25 April 2008 (EDT)

CIS PVs

New era CIS(Copper Indium(di)Selenide) PVs are arriving,with appealing features & lower prices!

People

please sign in. Being in this list just might give you first dibs as XO Solar stuff starts to get distributed !

If you don't have a User page in the wiki with contact info, please contact me by email so I can touch base with you!

Yama Ploskonka contact for the project, yamaplos@BOLinux.org

buxtor would like to be on the XO Solar list!

Doug LaRue has some tech(hard/soft) backgound

Nicholas Schaefer has free time and an XO-1

Ron Lussier Looking for a power solution for a Fijian island which has no power access.

Stan.SWAN in Wellington,NZ has made significant mid 2008 (yes-NZ winter!) XO charging testing with a 10W CIS (Copper Indium Selenide) PV. Such CIS/CIGS PVs are cheaper than normal Si types,& show superior overcast/haze/low light & off angle output. Automotive 12V SLA based emergency "Portable Power Packs" (~US$35) were found ideal for solar storage. stan.swan@gmail.com

Nick Thompson nickpthompson@gmail.com would love to get any sort of solar XO action going on!

Eric Youngren Is a solar guy who would like to help provide power systems for XOs and school servers and other remote power needs. http://islandenergysystems.com or http://islandenergysystems.wordpress.com

swannjie@gmail.com Click on edit to enter your information here

thomas@biopilze.de got two 6.5 watt-cigs-foldable panels (in fact they deliver only about 3 watts each at weight about 0.3kg each, BUT also cost 100 euro each! -by 2010/03), he experiments with them and LiPo-batteries (LIPO) ... powering an xo (or an Asus asus eeepc 900a) with them and/ or together on his bicycle (equipped with a hub-dynamo -delivers about 4-6 watts during cycling...) all together seems to be enough power to get all running autarc: eg for navigation, steady-real-time-documentation... -is interested in exchange: Skype: biopilze.de (seldomly online, -so email him first to come together...)

special thanks to

• Anne Gentle from XO Austin, for trusting me with her B4 XO, even after I told her I could not guarantee it would not fry.
• Graham Knight for good technical advice.
• Mike Lee, for encouragement and helping me lose some of my noobishness in XO power issues, so that I could sound more like I knew what I was talking about, when corresponding with more expert people.

got the plugs today. Thanks! They're the real thing! Yamaplos 11:48, 1 May 2008 (EDT)

• Craig Overmiller and the people of Texas Solar Power Company, for encouragement, expert advice, and a battery.
• Richard Smith, EC, battery, and power guy for OLPC, for expert advice, encouragement and assorted words of wisdom.
• Wayan Vota of OLPC News and his never satisfied visitors :-), that just couldn't leave solar for XO as something that could wait some more, which eventually got me going.

Location

XO Solar is based in (right now partial cloudy) Austin, Texas