Re: foot power
On 01/26/2011 10:05 PM, Carlos Nazareno wrote: Hey guys. I haven't taken that much look into foot pedal chargers and do know what their internals look like -- I just remembered coming across the link I posted so I shared The efficient way to do a foot pedal powered generator would probably be to have a small weighted wheel spin as a dynamo when you step on it? You just keep pumping on the pedal to keep the wheel spinning and it shouldn't take much effort, the wheel's momentum should keep power flowing. A lot of these ideas have been analyzed in detail. Hardly any of them except the bicycle idea would be feasible at our power levels: http://tinyurl.com/6zozbcn I would note that in your sample above, the reverse EMF from the generator will be significant; the wheel will slow down rapidly. There's no such thing as a free lunch. It seems that the only thing that's feasible to make the XO kid-powered is to significantly lower its power usage. -- Alan Eliasen elia...@mindspring.com http://futureboy.us/ ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: [OLPC-AU] hwclock accuracy
Very interesting. Strange. Unexpected. On Wed, Jan 26, 2011 at 11:30 PM, James Cameron qu...@laptop.org wrote: Until these problems are fixed, please change your build; Could this be added to http://wiki.laptop.org/go/OS_Builder#Recipes ? cheers, m -- martin.langh...@gmail.com mar...@laptop.org -- Software Architect - OLPC - ask interesting questions - don't get distracted with shiny stuff - working code first - http://wiki.laptop.org/go/User:Martinlanghoff ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: [OLPC-AU] hwclock accuracy
On Thu, 2011-01-27 at 09:01 -0500, Martin Langhoff wrote: Very interesting. Strange. Unexpected. On Wed, Jan 26, 2011 at 11:30 PM, James Cameron qu...@laptop.org wrote: Until these problems are fixed, please change your build; Could this be added to http://wiki.laptop.org/go/OS_Builder#Recipes ? I've split this into two parts, part of the build-system needs patching, modules/base/kspost.10.core.inc. The second part adds the rpm in the ini file, and uses the custom_scripts module to write changes to the image. http://dev.laptop.org.au/projects/xo-au/repository The two main files are examples/olpc-os-10.1.3-xo1.5AU.ini and setAU.sh Jerry diff --git a/modules/base/kspost.10.core.inc b/modules/base/kspost.10.core.inc index 4daa967..aecb68d 100644 --- a/modules/base/kspost.10.core.inc +++ b/modules/base/kspost.10.core.inc @@ -64,6 +64,14 @@ sed -i -e 's/TEMPORARY_STATE=no/TEMPORARY_STATE=yes/' \ # Remove resolv.conf from rwtab so that it can be updated atomically (#2748) sed -i -e /resolv.conf/d /etc/rwtab +# Remove adjtime from rwtab so that it can be updated atomically (#10605) +sed -i -e /adjtime/d /etc/rwtab + +# sigh.. must append UTC to /etc/adjtime or ntpdate writes localtime +cat EOF /etc/adjtime +UTC +EOF + # ensure temporary state directory doesn't get too fat (#9636) sed -i -e 's/RW_OPTIONS=/RW_OPTIONS=-o size=1M -o nr_inodes=1024/' /etc/sysconfig/readonly-root ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: foot power
Carlos Nazareno wrote: Hey guys. I haven't taken that much look into foot pedal chargers and do know what their internals look like -- I just remembered coming across the link I posted so I shared The efficient way to do a foot pedal powered generator would probably be to have a small weighted wheel spin as a dynamo when you step on it? You just keep pumping on the pedal to keep the wheel spinning and it shouldn't take much effort, the wheel's momentum should keep power flowing. Basically, never do mechanically what you can do electrically. There are cells now which are reasonably light, cheap, and can take 30W charge. http://cgi.ebay.co.uk/Lifepo4-A123-Cell-26650-2300mah-3-3v-battery-3-Pcs-NEW-/260726135323?pt=US_Batterieshash=item3cb47d1a1b for example. A flywheel is going to be _lots_ heavier. My design that I posted about earlier in this thread had basically a 'gearbox' electronically that optimised charging of the battery at the optimum biodynamic speed, and then used the generator as a motor to pop up the pedal back to the operating position. It was designed to get enough charge to charge a mobile phone battery fully by an adult over 3-5 mins. This is only at best 50-30W or so. If you want something to generate power efficiently, you need to use both legs - either a cycle, or a step-machine. Cycles are rather more efficient than step machines though the mechanism is pretty much inherently bulky. ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: [OLPC-AU] hwclock accuracy
On Thu, 2011-01-27 at 15:30 +1100, James Cameron wrote: On Wed, Jan 26, 2011 at 06:09:02PM -0600, Jerry Vonau wrote: That actually makes things worse, try installing ntpdate and run it with the stock /etc/adjtime file, now the real time clock is set to localtime as that is the first time hwclock is run. http://osdir.com/ml/fedora-olpc-list/2009-08/msg00102.html http://dev.laptop.org/ticket/9705 http://dev.laptop.org/ticket/10605 in particular. Until these problems are fixed, please change your build; remove /etc/adjtime from /etc/rwtab, change the third line of /etc/adjtime to be UTC, copy the zoneinfo file for to /etc/localtime, install ntpdate, and write a script in /etc/NetworkManager/dispatcher.d/99-ntpdate Just thinking ahead, does any of this have any negative impact for olpc-update? Thinking /etc/adjtime should be treated the same way /etc/sysconfig/i18n is. Testing hasn't got that far, just curious. Jerry ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: [Server-devel] Adding puppet to the school server.
On Wed, Jan 26, 2011 at 5:03 PM, David Farning dfarn...@activitycentral.com wrote: Our initial goal will be create puppet modules to setup and maintain these key services. Very interesting - thanks and welcome! As I've mentioned to David Van Assche recently, you can think of initial setup as something that the rpms themselves do, or scripts in xs-config. The calling of puppet is to manage thousands of servers -- following a bit the tradition set forth by infrastructures.org Initially, I would propose - good puppet recipes - integration with dashboard You may find there's more benefit in creating dashboard modules (I assume dashboard is modular) -- maybe those write puppet snippets that are just a recipe, or maybe they drive something that is a puppet module. But the real ease of use (for the admins) comes from having a nice UI in dashboard. Example dashboard modules that control... - enable and control OpenDNS filtering - deploy a new XO OS img, and trigger the upgrade - deploy/manage Activities that the XS serves - deploy/manage static content (html, css, ogg) that the XS serves I suspect there is enormous value in developing this area. cheers, m -- martin.langh...@gmail.com mar...@laptop.org -- Software Architect - OLPC - ask interesting questions - don't get distracted with shiny stuff - working code first - http://wiki.laptop.org/go/User:Martinlanghoff ___ Server-devel mailing list server-de...@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [support-gang] DSD: Feature requests for 11.2.0 - seeking deployment and community input
Adam and Daniel, I remember some time ago seeing (I cannot remember where just now) about possible functionality to include the ability for users to Chat from any activity as a way of collaborating. I would like to add that to the list. I am happy to put together requirements, if need be. Thanks. Gerald On Wed, Jan 26, 2011 at 3:00 PM, Adam Holt h...@laptop.org wrote: Plz CC de...@laptop.org if you have ideas with kids/mentors/teachers' votes behind them!! Subject: Feature requests for 11.2.0 - seeking deployment and community input Date: Wed, 26 Jan 2011 19:26:38 + From: Daniel Drake d...@laptop.org d...@laptop.org To: OLPC Devel devel@lists.laptop.orgdevel@lists.laptop.org Hi, OLPC recently announced the development of a major new OLPC OS release for XO-1 and XO-1.5:http://wiki.laptop.org/go/11.2.0 We are seeking feedback and ideas from the community for feature requests for the upcoming feature development stage. Feedback from other parts of the community is welcome too. We're particularly interested in things that would make a big difference for deployments, solve big in-the-field challenges, etc. Please post any ideas here, and they'll be considered when we come to plan which features we'll aim to include. Features includes anything on the software level, really. From activities and UI to low-level system stuff. To give some examples, here's a selection of items from the last time that a feature plan was constructed:http://wiki.laptop.org/go/Feature_roadmap/Page_of_all_features_that_target_9.1.0 Thanks, Daniel ___ Devel mailing listDevel@lists.laptop.orghttp://lists.laptop.org/listinfo/devel ___ support-gang mailing list support-g...@lists.laptop.org http://lists.laptop.org/listinfo/support-gang ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: foot power
On Thu, Jan 27, 2011 at 6:41 AM, Ian Stirling o...@mauve.plus.com wrote: Carlos Nazareno wrote: Hey guys. I haven't taken that much look into foot pedal chargers and do know what their internals look like -- I just remembered coming across the link I posted so I shared The efficient way to do a foot pedal powered generator would probably be to have a small weighted wheel spin as a dynamo when you step on it? You just keep pumping on the pedal to keep the wheel spinning and it shouldn't take much effort, the wheel's momentum should keep power flowing. Basically, never do mechanically what you can do electrically. There are cells now which are reasonably light, cheap, and can take 30W charge. http://cgi.ebay.co.uk/Lifepo4-A123-Cell-26650-2300mah-3-3v-battery-3-Pcs-NEW-/260726135323?pt=US_Batterieshash=item3cb47d1a1b for example. A flywheel is going to be _lots_ heavier. My design that I posted about earlier in this thread had basically a 'gearbox' electronically that optimised charging of the battery at the optimum biodynamic speed, and then used the generator as a motor to pop up the pedal back to the operating position. It was designed to get enough charge to charge a mobile phone battery fully by an adult over 3-5 mins. This is only at best 50-30W or so. If you want something to generate power efficiently, you need to use both legs - either a cycle, or a step-machine. Cycles are rather more efficient than step machines though the mechanism is pretty much inherently bulky. ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel Looking at this thread reminded me of clocks with weight pulleys. I saw one recently outside UCSF hospital. http://www.flickr.com/photos/revger/4381874187/ The weight unwinds the pulley slowly. This is also seen on roasting jack (For those who attended the OLPC SF dinner/bbq, the Kleiders have this at their house) http://en.wikipedia.org/wiki/Clock_jack Will such a wind-up mechanism work? Sameer ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: foot power
On 01/27/2011 04:08 PM, Sameer Verma wrote: Looking at this thread reminded me of clocks with weight pulleys. I saw one recently outside UCSF hospital. http://www.flickr.com/photos/revger/4381874187/ The weight unwinds the pulley slowly. This is also seen on roasting jack (For those who attended the OLPC SF dinner/bbq, the Kleiders have this at their house) All of these questions are easily answered with some basic physics. E = m g h will calculate how high you need to lift a mass to get a certain amount of energy. I wrote some analysis of a gravitational system in the OLPC wiki: http://wiki.laptop.org/go/Battery_and_power#Pulley_Power I'll quote it here: 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. Again, I will note that the only way to make the XO reasonably powerable by a child (or an adult) is to reduce its power consumption. -- Alan Eliasen elia...@mindspring.com http://futureboy.us/ ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: foot power
On Thu, Jan 27, 2011 at 07:54:16PM -0700, Alan Eliasen wrote: Again, I will note that the only way to make the XO reasonably powerable by a child (or an adult) is to reduce its power consumption. Well, this isn't entirely true. For example, something like a potter's wheel pedal could potentially generate sufficient (25+ W) amounts of power, given a reasonably efficient dynamo (and a reasonably efficient child!) However, the endurance problem is still *huge*. The XO-1 takes about 2-3 hours to charge from 10% battery level. If you have a foot-pedal system that generates 25 watts, you still need the child to do this for _several hours_. Alright, you may say, maybe you make it ultra-efficient and very easy to push the pedal down (repetitively, for hours on end!) But that's just the tip of the iceberg. Noise is another big problem. The crank charger is quite loud (for only putting out 15~ W or so). Imagine a classroom with even 5 foot-chargers going at a time... it's an issue. The only reasonable solution I can think of is playground equipment turned to power generation (a la PlayPump). Swings, seesaws, and carousels all can generate _significant_ amounts of power (100watts as a conservative estimate for a converted carousel!), and _store_ that energy in a battery or flywheel for later laptop charging. Thus, we don't just need a 15w charger for one XO... the endurance problem makes this near impossible. We need a 150watt charger for one XO (full charge), so the child(ren) can put out significant amounts of power for a short time, and charge the XO's batteries over extended timescales. Higher-wattage chargers also work well with solar panels (they can both charge the same set of batteries.) Just some thoughts, Nate ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: foot power
On 01/27/2011 08:18 PM, Nathaniel Theis wrote: On Thu, Jan 27, 2011 at 07:54:16PM -0700, Alan Eliasen wrote: Again, I will note that the only way to make the XO reasonably powerable by a child (or an adult) is to reduce its power consumption. Well, this isn't entirely true. For example, something like a potter's wheel pedal could potentially generate sufficient (25+ W) amounts of power, given a reasonably efficient dynamo (and a reasonably efficient child!) However, the endurance problem is still *huge*. The XO-1 takes about 2-3 hours to charge from 10% battery level. If you have a foot-pedal system that generates 25 watts, you still need the child to do this for _several hours_. I've seen a huge deal of confusion in the human-powered discussions here and on the Wiki; there seems to be some sort of strange idea that having a flywheel lets us get out more energy than we put in. The potter's wheel example above is one such example. It seems to neglect the idea that drawing off 25 watts of power from the flywheel's rotation means that we have to put in (more than) 25 watts of power to keep it from stopping. Anyone who has ever used a generator on a bicycle knows that you can *feel* the drag from the reverse electromotive force as soon as you turn on the light. It will stop you. And those generate maybe 3-6 watts. The braking force from pulling off 25 watts of power is huge. Alright, you may say, maybe you make it ultra-efficient and very easy to push the pedal down (repetitively, for hours on end!) But that's just the tip of the iceberg. I'd never say such a thing, because I realize that if I'm going to be getting 25 watts out, I need to be putting in more than 25 watts, constantly, to keep the flywheel from stopping! There's no such thing as a free lunch, or a magical linkage that puts out more power than it takes in. It doesn't matter if it's a treadle, a kickwheel, a playground carousel, or whatever... that kid has to be sustaining more than 25 watts or the wheel will stop very rapidly due to the significant drag from the generator. No amount of magical efficiency will make it easy to push the pedal down. The kid has to be constantly putting out an average of 25 watts (or significantly more, due to friction and generator efficiency.) Anything less and the wheel stops. Rapidly. Just how rapidly? It's calculated below. Let's analyze the above situation mathematically. We're not doing any justice to the cause of education if we can't analyze these situations with basic physics and real calculations. Let's take a pretty hefty potter's wheel. I looked up some on the web, and we'll use the figures of a 130 lb concrete kickwheel with diameter 30 inches (radius 15 inches.) I'm going to put the following equations in Frink notation, ( http://futureboy.us/frinkdocs/ ) Frink is a programming language/calculating tool I've developed for just this sort of purpose. It tracks units of measure through all calculations and ensures that answers come out right, even when mixing units. The moment of inertia, I, of the disc is given by: I = 1/2 * 130 lb * (15 in)^2 Let's say that the kid is able to spin it to a rather high rotation rate of 120 rpm. (This is quite high for a foot-powered potter's wheel.) omega = 120 rpm The kinetic energy of the spinning wheel is given by: E = 1/2 I omega^2 Which gives 338 joules of energy. We're now going to draw this energy off with a generator. Let's say the generator produces 25 W of power and is, very generously, 60% efficient. (A bicycle hub generator that efficient is very expensive.) The wheel will then come to a complete stop after: E / (25 W / 0.6) Or a dead stop after only 8.1 seconds if we stop adding power. This is significant braking. Note also that the kid has to be putting in 41.6 watts of power for the generator to produce 25 watts, due to its efficiency. If the kid lets it slow down at all, they'll have to put in more than 41.6 watts to accelerate it back up! We're also neglecting the significant amount of energy that was needed to spin the wheel up to get it to this point, or any other frictional or electrical losses. A flywheel does *nothing* to reduce the amount of sustained power that someone has to produce. Especially if we can't posit some sort of magic source of energy to spin it up in the first place. If you disagree with any of this analysis, please respond with physical equations. Engineering is done with numbers. Analysis without numbers is only an opinion. --Akin's Laws, #1 http://spacecraft.ssl.umd.edu/old_site/academics/akins_laws.html -- Alan Eliasen elia...@mindspring.com http://futureboy.us/ ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
trac email verification temporarily enabled
We've seen a large increase in spam tickets in the past couple of days [1]. I've reacted to those by deleting the tickets, and adding patterns to wiki/BadContent. I'm not confident that the latter is working, because I've seen further tickets. I'm fairly certain accounts could be created without e-mail verification. If anyone knows different, let me know. I've enabled EmailVerificationModule [2], which means you may be asked by Trac to verify your e-mail address. It seemed to work for me, but I've got TRAC_ADMIN. Unfortunately this raises the bar slightly for bug reports, in that an e-mail address is required. (Rafael Ortiz, thanks for closing tickets as invalid. Deleting works well too though, if I can get to them before you do. ;-) ) References: [1] http://lists.laptop.org/pipermail/bugs/2011-January/thread.html [2] http://trac-hacks.org/wiki/AccountManagerPlugin/Modules#EmailVerificationModule -- James Cameron http://quozl.linux.org.au/ ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: [Server-devel] XS-0.7 plans -- your thoughts please...
On Tue, Jan 18, 2011 at 11:29 AM, Martin Langhoff martin.langh...@gmail.com wrote: Ay ay ay! This last 12 months have been frantic, as I've ended up biting off a ton more than common sense would recommend to chew. And then chewing, slowly, awkwardly. Good things have come out of the hard work of this year, but the XS has been delayed. One good piece of news is that I've helped deploy more XSs in the field, both in person and through this list. So I think we have good feedback on what to streamline and make easier. And I have a window of time to hack on it. So... the plan is roughly the same as it was in http://lists.laptop.org/pipermail/server-devel/2009-October/004139.html - except that I'll start drafting the build on top of F14. My actual plan is to have the packages for RHEL6 / CentOS6, which will give us a more stable platform. With less churn, I get to spend more of my limited time on interesting work :-) So -- going back on the traffic we've see in the last 24 months, what would you highlight? What have people asked for (that wasn't easy/trivial/possible)? What problems have we heard that were hard to diagnose...? cheers, m -- martin.langh...@gmail.com mar...@laptop.org -- School Server Architect - ask interesting questions - don't get distracted with shiny stuff - working code first - http://wiki.laptop.org/go/User:Martinlanghoff ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel Sorry about the delayed response (school is back in session), but can I assume that XS 0.7 will run on a XO 1.5? cheers, Sameer ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] XS-0.7 plans -- your thoughts please...
On Thu, Jan 27, 2011 at 5:50 PM, Sameer Verma sve...@sfsu.edu wrote: Sorry about the delayed response (school is back in session), but can I assume that XS 0.7 will run on a XO 1.5? Definitely on the wishlist, but not a given. Even if it does, one of the desired features is very unlikely to happen: use the wlan as AP. m -- martin.langh...@gmail.com mar...@laptop.org -- Software Architect - OLPC - ask interesting questions - don't get distracted with shiny stuff - working code first - http://wiki.laptop.org/go/User:Martinlanghoff ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] XS-0.7 plans -- your thoughts please...
On Thu, Jan 27, 2011 at 3:08 PM, Martin Langhoff martin.langh...@gmail.com wrote: On Thu, Jan 27, 2011 at 5:50 PM, Sameer Verma sve...@sfsu.edu wrote: Sorry about the delayed response (school is back in session), but can I assume that XS 0.7 will run on a XO 1.5? Definitely on the wishlist, but not a given. Even if it does, one of the desired features is very unlikely to happen: use the wlan as AP. Is that because the wlan card does not support hostap (I haven't verified this)? What if the card simply ran in ad-hoc mode? It might support a small pool... cheers, Sameer ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
