highaltitude.log.20080415

[00:14] Nick change: akawaka -> UbuntuMuntu
[00:16] Nick change: UbuntuMuntu -> akawaka
[00:39] Laurence1 (n=Laurence@host86-141-78-0.range86-141.btcentralplus.com) joined #highaltitude.
[00:44] hallam (n=hallam@18.62.30.22) joined #highaltitude.
[00:46] <fnoble> oi'
[00:47] <natrium42> ahoy
[00:53] flowolf (n=flowolf@unaffiliated/flowolf) left irc: "Leaving"
[00:56] Laurenceb (n=Laurence@host86-160-237-147.range86-160.btcentralplus.com) left irc: Read error: 110 (Connection timed out)
[01:19] <Kornholijo> oi
[01:21] <Kornholijo> sup all :o
[01:24] <edmoore> hi
[01:24] <edmoore> new round these parts?
[01:24] <Kornholijo> 35?
[01:25] <fnoble> 72
[01:25] <akawaka> hmm
[01:25] <akawaka> 46
[01:25] <Kornholijo> 9021?
[01:25] <fnoble> akawaka: no, thats wrong
[01:25] <Kornholijo> 9021
[01:25] <akawaka> damnit
[01:25] <Kornholijo> !
[01:26] <hallam> Mornington Crescent.
[01:26] <Kornholijo> someone help me get started in this "highalt" stuff ;<
[01:26] <Kornholijo> i never did anything in air rofl
[01:27] <edmoore> numberwang]
[01:27] <fnoble> what do you want to do?
[01:27] <Kornholijo> fnoble: just begin with something easy
[01:27] <Kornholijo> i dunno actually
[01:28] <Kornholijo> what about kites xd?
[01:28] <edmoore> there's loads of good stuff on the web about kite photography
[01:28] <Kornholijo> :)
[01:29] <fnoble> most people here are into balloons, although a few are working on parafoils and other such things
[01:29] <Kornholijo> how much the kites cost?
[01:29] <hallam> Not a lot if you make them yourself
[01:29] <Kornholijo> parafoils :D
[01:29] <Kornholijo> 32 inch parafoil for 15$ Oo
[01:29] <Kornholijo> only 32 :(
[01:30] <Kornholijo> what you can do with kites and such? i like baloons better but ill try kites
[01:30] <hallam> Kornholijo: you might be better off with something like this: http://home.earthlink.net/~quade/lawnchair.html
[01:30] <fnoble> 32 is big enough for most people
[01:31] <Kornholijo> loool
[01:31] <Kornholijo> hallam: LOOOOL
[01:32] <fnoble> the easiest thing to start with is probably just a wather balloon with some kind of payload
[01:32] <fnoble> *weather
[01:32] Tigga (n=chatzill@pc-232-235-60.magd.cam.ac.uk) left irc: Read error: 110 (Connection timed out)
[01:32] <Kornholijo> hmm
[01:32] <fnoble> payload would usually include a gps, a radio transmitter and somthing to tie them together
[01:32] <fnoble> like a microcontroller
[01:32] <Kornholijo> like a arm board
[01:32] <Kornholijo> hmm
[01:32] <fnoble> also a camera
[01:32] <Kornholijo> :>
[01:33] <Kornholijo> how much the helium costs? and how much?
[01:33] <edmoore> Kornholijo: do you have some electronics experience?
[01:33] <Kornholijo> yeah
[01:33] <Kornholijo> not much, but I understand stuff
[01:34] <Kornholijo> I just dont do much stuff, i made pic programmer etc
[01:34] <fnoble> helium isnt too expensive, ask in shops that supply for parties and that kind of thing
[01:34] <Kornholijo> well is that party grade good?
[01:34] <fnoble> are you in the US?
[01:34] <Kornholijo> yeah
[01:34] <Kornholijo> chicago
[01:35] <fnoble> any helium is ok really
[01:35] <fnoble> doesnt have to be ultra pure
[01:35] <hallam> party grade is fine, better than what you need but more easily obtainable than the lower-purity industrial stuff
[01:35] <hallam> remember to get FAA permission (not too hard)
[01:36] <Kornholijo> faa permission?
[01:36] <Kornholijo> hmm
[01:36] <edmoore> yeah
[01:36] <edmoore> it's not as ominous as it sounds
[01:36] <Kornholijo> what do you need to get it?
[01:36] <fnoble> also you will need a radio reciever and antenna
[01:37] <Kornholijo> something like packetradio :>?
[01:37] <Kornholijo> i got ham license already
[01:37] <fnoble> a lot of people in the US use packet radio on balloons i think
[01:38] <fnoble> in the UK we arent allowed to transmit with enough power for that to work for us though
[01:38] <Kornholijo> ;/
[01:38] <Kornholijo> i dont think its that much
[01:38] <fnoble> so we have been using low baud rate RTTY
[01:38] <Kornholijo> :)
[01:38] <Kornholijo> it should tranmist too much data tough
[01:38] <Kornholijo> just gps cords and stuff?
[01:39] <edmoore> yep
[01:39] <Kornholijo> log all other info to usb
[01:39] <fnoble> an SD card is easier to interface to
[01:39] <Kornholijo> yeah
[01:39] <Kornholijo> in mmc mode
[01:39] <fnoble> not many microcontrollers have USB host
[01:39] <Kornholijo> im not talking about microcontrollers probably,
[01:40] <Kornholijo> ill just get something versatile as 400mhz arm board with linux
[01:40] <Kornholijo> i can use in many places then, not only balooning
[01:40] <Kornholijo> if it doesnt break :(
[01:40] <fnoble> yeah, some people have had a lot of success with the gumstix boards
[01:40] <Kornholijo> =)
[01:41] <Kornholijo> aother thing..
[01:41] <Kornholijo> what about it falling down on something...
[01:41] <Kornholijo> what about it falling down on something...like a car and damaging it?
[01:41] <hallam> either you buy insurance or you take your chances
[01:42] <edmoore> well you have parachutes
[01:42] <edmoore> which doesn't remove the chance of it falling on something
[01:42] <hallam> nah, parachutes are for wusses
[01:42] <edmoore> but does greatly reduce the chance of it damaging it
[01:44] <fnoble> also, you need to think of a way of deteching the balloon from the payload before it bursts
[01:44] <fnoble> otherwise you can have problems with your parachute getting tangled with the balloon etc.
[01:44] <fnoble> we just use a small pyro with an electric match
[01:45] <fnoble> but can use hotwire through the cord, or a mechanical pin or something
[01:45] <Kornholijo> hmm
[01:46] <Kornholijo> how much the baloons cost? xd, the arm boards are not too cheap too
[01:46] <Kornholijo> i wanna make something cheapy
[01:47] <hallam> don't use an arm then
[01:47] <Kornholijo> i want to
[01:47] <hallam> you could use one of those linux router things
[01:47] <Kornholijo> yeah
[01:47] <Kornholijo> like wrt
[01:47] <hallam> yes
[01:47] <hallam> some people have used those
[01:47] <Kornholijo> but they have little ram etc
[01:47] <Kornholijo> not great expansions
[01:48] <edmoore> ngw100
[01:48] <edmoore> have a look at that
[01:48] <hallam> you don't need very much performance for this sort of thing
[01:48] <fnoble> kaymont are the main supplier of balloons
[01:48] <Kornholijo> no but I dont wanna just use for balooning, maybe for something else too
[01:48] <fnoble> mid sized ones can be had on ebay for like $40
[01:49] <hallam> we've made several 30km flights with video, GPS, software radio etc using a PIC processor with 4kB ram and a 20MHz lock
[01:49] <Kornholijo> k
[01:49] <hallam> clock*
[01:49] <Kornholijo> nice :o
[01:49] <hallam> the balloons are specified by the empty weight of the latex balloon
[01:49] <hallam> you want one that's at least 300g, better 1kg
[01:49] <hallam> if you want to go to a high altitude anyway
[01:50] <hallam> the typical party bottle of helium will fill two or three of those
[01:50] <Kornholijo> how much is that party bottle?
[01:50] <Kornholijo> how big is it
[01:51] <edmoore> Kornholijo: I would say the smaller micros are better for this sort of thing. In as far as you don't need any power to parse a gps stream, make a few decisions, fire off some io, and log it to a card
[01:51] <hallam> small enough that one person can move it
[01:51] <edmoore> an 8mhz 8mit micro could handle it without a problem
[01:51] <hallam> price varies a lot, we've paid around $150 but you might find it cheaper
[01:51] <Kornholijo> bah
[01:53] <Kornholijo> it all adds up to much :(
[01:54] <fnoble> well, you dont need an expensive flight computer like an ram linux board
[01:54] <fnoble> could make one for like $30 parts
[01:54] <Kornholijo> not really
[01:55] <Kornholijo> more like 60$
[01:55] <Kornholijo> + helium/balloon == another $150~
[01:56] <hallam> but you get $210 worth of super sexy photos
[01:57] <Kornholijo> o rly
[01:59] <Kornholijo> i got no income :(
[02:00] <fnoble> you can piece it together bit by bit
[02:00] <fnoble> make the payload first, get it all working on the ground
[02:00] <fnoble> sell some organs, buy helium, youre all set
[02:00] <Kornholijo> not organs :(
[02:01] edmoore (n=edmoore@88.212.167.121) left irc:
[02:03] <Kornholijo> this is hard :(
[02:03] <fnoble> ok... not organs
[02:03] <fnoble> but you can build up bit by bit
[02:03] <fnoble> i guess the main expense is the helium
[02:04] <Kornholijo> yep
[02:04] <fnoble> you might be able to find some cheaper if you shop around
[02:04] <Kornholijo> mybe something surplus from somehwere
[02:05] <fnoble> or get some friends interested and share the cost
[02:05] <Kornholijo> hah
[02:08] <Kornholijo> what about http://chicago.craigslist.org/wcl/bfs/637788362.html
[02:08] <Kornholijo> 1 is full o_o
[02:11] <fnoble> hmm, look around
[02:11] <Kornholijo> i am
[02:11] <fnoble> i think you can get it so you dont pay for the bottle, just the helium
[02:11] <Kornholijo> first just to get the payload working xd
[02:11] <fnoble> and you give the bottle back to be refilled
[02:11] <fnoble> yeah
[02:12] <Kornholijo> the bottle costs half i guess
[02:12] <Kornholijo> its metal..
[02:13] <fnoble> the bottles are expensive yeah
[02:14] <Kornholijo> "I pay approx US$26 for 125 Cu. ft. of helium." <- hmm?
[02:17] <fnoble> also you might find someone who has an account with a gas supplier can get it cheaper for you
[02:17] <fnoble> if your at school or college then that might be worth a try
[02:18] <fnoble> also ive heard welding shops can get it / have it too
[02:19] <Kornholijo> mmmmm
[02:19] <Kornholijo> need to ask some people
[02:21] hallam (n=hallam@18.62.30.22) left irc: Read error: 110 (Connection timed out)
[02:23] <Kornholijo> now I need to decide what will control my stuff
[02:23] <Kornholijo> something from avr maybe..
[02:27] <fnoble> yeah, i like avrs
[02:27] <fnoble> nicer than PICs, but can't get them for free
[02:29] <Kornholijo> you can get pics for free :O?
[02:30] <fnoble> yeah, you can order free samples form their website
[02:30] <fnoble> they come quickly too
[02:30] <fnoble> you get capped at like 1 order a month or something
[02:30] <fnoble> also look at maxim for general ICs as they are also really good at giving free samples
[02:31] <Kornholijo> :>
[02:40] <Kornholijo> what do you consider the "best" pic?
[02:41] <fnoble> there are so many... its hard to say
[02:42] <fnoble> go for an 18 series, they have a bit mire power and can be programmed in C
[02:42] <fnoble> get a selection :)
[02:42] <Kornholijo> ooo
[02:42] <Kornholijo> i hate asm
[02:42] <Kornholijo> C <#
[02:42] <fnoble> yeah the 12 and 16 series pics are asm only
[02:42] <Kornholijo> sucks
[02:43] <fnoble> but the 18s are good, so its no problem
[02:43] <Kornholijo> :)
[02:45] <fnoble> something like an 18f2520 or 18f877
[02:46] <Kornholijo> k
[02:47] <fnoble> get ones with the most memory, and ADC channels come in handy
[02:47] <fnoble> USB is not so useful
[02:48] <Kornholijo> mm
[02:51] <fnoble> actually the 18f4685 or 2685 look about the best spec you can get in PDIP
[02:52] <fnoble> Im assuming you dont have access to PCB facilities, so SMT is hard?
[02:54] <Kornholijo> smd?
[02:54] <fnoble> surface mount
[02:54] <Kornholijo> nope i dont do smd
[02:54] <Kornholijo> i know but you said smt :P
[02:55] <fnoble> smt = surface mount technology
[02:55] <fnoble> smd and smt are pretty much interchangable
[02:55] <Kornholijo> k
[02:55] <fnoble> i guess if your picky smt is an adjective and smd is a noun
[02:56] <Kornholijo> ;)
[02:56] <fnoble> anyway, its getting late here
[02:56] <Kornholijo> ;/
[02:56] <fnoble> so, night, and good luck with your plans
[02:56] <Kornholijo> good luck to you too
[02:56] fnoble (n=fnoble@88-106-206-171.dynamic.dsl.as9105.com) left irc:
[04:33] <Kornholijo> natrium42: sup ;P
[05:19] MetaMorfoziS (n=avr@3e70d12b.adsl.enternet.hu) joined #highaltitude.
[05:39] MetaMorfoziS (n=avr@3e70d12b.adsl.enternet.hu) left irc: "Most akkor mit csinalt aki miert lepett ki?"
[06:32] akawaka (n=akawaka@external.treyarch.com) left irc: Read error: 110 (Connection timed out)
[07:47] borism (n=boris@195-50-211-66-dsl.krw.estpak.ee) joined #highaltitude.
[08:32] Simon-MPFH (n=simon@phantom.mpfh.co.uk) joined #highaltitude.
[09:12] akawaka (n=akawaka@cpe-76-173-152-142.socal.res.rr.com) joined #highaltitude.
[09:53] akawaka (n=akawaka@cpe-76-173-152-142.socal.res.rr.com) left irc: Read error: 110 (Connection timed out)
[10:05] edmoore (n=edmoore@88.212.167.121) joined #highaltitude.
[10:14] flowolf (n=flowolf@unaffiliated/flowolf) joined #highaltitude.
[10:44] Tigga (n=chatzill@pc-232-235-60.magd.cam.ac.uk) joined #highaltitude.
[12:57] Laurence1 (n=Laurence@host86-141-78-0.range86-141.btcentralplus.com) left #highaltitude.
[13:07] Laurence1 (n=Laurence@host86-141-78-0.range86-141.btcentralplus.com) joined #highaltitude.
[13:07] Nick change: Laurence1 -> Laurenceb
[13:11] <Laurenceb> good afternoon
[13:14] <Laurenceb> edmoore: you about?
[13:14] <edmoore> am now
[13:14] <Laurenceb> I was thinking about using an arduino for a parafoil
[13:14] <Laurenceb> reflashing it and modifying the board
[13:15] <Laurenceb> i.e. mutilating an arduino :P
[13:15] <edmoore> well it is an atmega on a pcb with everything on it
[13:15] <edmoore> so not a bad idea
[13:16] <Laurenceb> yes, I'm just looking at the prices
[13:16] <Laurenceb> its nice to have a premade board with usb :D
[13:37] <edmoore> does it have isp?
[13:45] <Laurenceb> yes
[13:46] <Laurenceb> annoying thing is its the 6 pin design
[13:46] <Laurenceb> so I'd have to make an adaptor
[13:46] <Laurenceb> also whoever it was selling "freeduino" boards on ebay has vanished :(
[13:49] <Laurenceb> they were only £14
[13:53] <Laurenceb> probably not worth it as I've got almost everything hanging around, and can pinch the remainders from the electronics lab at uni :D
[13:58] Laurenceb (n=Laurence@host86-141-78-0.range86-141.btcentralplus.com) left irc: "Leaving."
[14:05] Laurenceb (n=laurence@host86-141-78-0.range86-141.btcentralplus.com) joined #highaltitude.
[16:05] edmoore (n=edmoore@88.212.167.121) left irc:
[16:06] borism (n=boris@195-50-211-66-dsl.krw.estpak.ee) left irc: Client Quit
[16:18] flowolf (n=flowolf@unaffiliated/flowolf) left irc: Read error: 148 (No route to host)
[16:23] edmoore (n=edmoore@88.212.167.121) joined #highaltitude.
[16:24] flowolf (n=flowolf@unaffiliated/flowolf) joined #highaltitude.
[16:24] <Laurenceb> hi edmoore
[16:24] <Laurenceb> ed, I'm playing with my kalman filter on my pc
[16:24] <edmoore> Laurenceb: what you get up to in private is nothing to do with me
[16:24] <Laurenceb> it doesnt work, the problem is the way I fudged it to do the GPS measurements
[16:25] <Laurenceb> lol
[16:25] <edmoore> welcome to kalman filtering, dsp in general
[16:25] <edmoore> make it work in matlab first
[16:25] <edmoore> then c on a pc
[16:25] <edmoore> then c in embedded
[16:25] <Laurenceb> if you look at the wikipedia page -> http://en.wikipedia.org/wiki/Kalman_filter
[16:25] <edmoore> if you dive straight to embedded c you'll just upset yuourself
[16:25] <Laurenceb> S has no inverse
[16:25] <Laurenceb> I'm wondering how to cheat...
[16:26] <Laurenceb> could I make up a GPS measurment equal to the current heading?
[16:26] <Laurenceb> with an error equal to the models error?
[16:27] <Laurenceb> at the moment I put gps=0, and H=[[0,0][0,1]], but that means S has no inverse
[16:29] <Laurenceb> I'm planning to put H=identity, then gps=model_heading, with R top left = P top left
[16:29] <Laurenceb> what do you think?
[16:31] <edmoore> what value does model_heading have?
[16:32] <Laurenceb> top of x from the wikipedia article
[16:33] <Laurenceb> x=[model_heading,model_turn_rate]
[16:35] <Laurenceb> just tried that... it doesnt give me divide by zero any more
[16:35] <Laurenceb> but something goes very wrong with P, its all nan
[16:36] <edmoore> first off I'd try this is octave
[16:36] <Laurenceb> yes
[16:36] <edmoore> it's much easier to actually play with the maths that way
[16:37] <Laurenceb> it looks like it explodes due to the error on the heading going to zero
[16:37] <Laurenceb> but this fundging technique looks promising
[16:37] <edmoore> that shouldn't make it explode though I don't think
[16:37] <Laurenceb> hmm
[16:37] <edmoore> i mean you are inputing noise into these measurements yeah?
[16:38] <Laurenceb> yes
[16:38] <Laurenceb> I wrote an octave script to make fake data
[16:38] <Laurenceb> doh
[16:38] <Laurenceb> my errors halve each time
[16:38] <Laurenceb> of course :D
[16:39] <Laurenceb> oh well, thats promising, somethings working correctly :P
[16:42] <edmoore> have a look at tomorrow's js
[16:42] <edmoore> http://www.wunderground.com/modelmaps/maps.asp?model=GFS&domain=EU
[16:42] <edmoore> it would be a *bad* day to launch from cambridge
[16:45] <Laurenceb> yes
[17:08] <Laurenceb> hmm made some progress
[17:09] <Laurenceb> I have a stable P matrix :P
[17:09] <Laurenceb> can you see the error in "predict_and_update" in kalman.c ?
[17:10] <Laurenceb> hint: what state was I updating?
[17:13] borism (n=boris@195-50-211-66-dsl.krw.estpak.ee) joined #highaltitude.
[17:18] mc- (n=mfcastle@cpc4-glfd1-0-0-cust538.glfd.cable.ntl.com) joined #highaltitude.
[17:18] <Laurenceb> hello mc-
[17:18] <mc-> hi lb
[17:18] <mc-> no gps yet
[17:18] <Laurenceb> any good with kalman filters?
[17:18] <Laurenceb> I dont mind
[17:19] <mc-> I don't bother with Kalman, least squares is good enough for me
[17:19] <Laurenceb> lol ok
[17:20] <Laurenceb> I think my filter is getting in a "tangle"
[17:20] <mc-> edmoore, want to meet this evening to get some plastic
[17:20] <Laurenceb> as I'm using pointers, its changing the input with the output
[17:20] <Laurenceb> if you see what I mean...
[17:21] <mc-> lb, why not try LSQ first, they're much easier
[17:21] <Laurenceb> well I'm close to getting it working
[17:21] <mc-> how do you know?
[17:21] hallam (i=hallam@18.214.1.172) joined #highaltitude.
[17:22] <Laurenceb> you can see the statistics are working correctly
[17:22] <edmoore> mc-: I'm seeing a friend this evening might I might be going near godalming
[17:22] <edmoore> but I'm not sure that's particularly convenient for you
[17:22] <edmoore> hi henry
[17:23] <Laurenceb> henry: hi there
[17:23] <Laurenceb> back from the states?
[17:23] <mc-> Godalming is near, but I'm out from 7pm onwards. 6.30pm OK?
[17:25] <edmoore> not really, 7pm is when i leave :) hrm
[17:25] <edmoore> I could make a mad dash to cranleigh now
[17:26] <mc-> or how about 10pm somewhere?
[17:26] <edmoore> I'd like to say yes but thruthfully i've no idea where I'll be - there's a high likelyhood I'll be at the Dog and Pheasant in Brook
[17:26] <edmoore> just near whitly
[17:28] <edmoore> Laurenceb: without being a pain in the bottom, a linear kalman filter may well have problems on a parafoil
[17:28] <edmoore> or anything with quite large angular deviations like that
[17:29] Action: Laurenceb seppaku
[17:30] <edmoore> I thought I'd tell you now rather than later :p
[17:30] <edmoore> don't worry, the extended kalman filter is described below
[17:30] <edmoore> on the wiki article
[17:30] <Laurenceb> rofl
[17:30] <edmoore> and don't try an unscented one on an avr
[17:30] <edmoore> it'll just fall over and bleed out of its ears
[17:30] <mc-> if you left at 6.30pm, I could get to Bramley for 7pm, and then rush off.
[17:30] <Laurenceb> actually extended isnt too bad...
[17:31] <edmoore> nah it's a fairly logical increment
[17:31] <Laurenceb> or at least I can understand how t works
[17:31] <edmoore> mc-: I'm not sure I can make it. She's coming to me here at 7, then we're heading off someone. so logistically it's a bit awkward
[17:31] <Laurenceb> basically differential equations in place of linear approximations
[17:33] <edmoore> fazackerly
[17:33] <Laurenceb> IMO its a matter of getting the design right
[17:33] <edmoore> um... yes?
[17:33] <edmoore> fairly sound engineering advice
[17:33] <Laurenceb> of the parafoil
[17:33] <Laurenceb> as in aerodynamically
[17:33] <Laurenceb> I'm thinking of sticking a fin/tail on the payload
[17:34] <edmoore> for stability?
[17:34] <Laurenceb> yes
[17:34] <edmoore> I am doing a similar thing, except not the fin
[17:34] <Laurenceb> how is it similar then?
[17:34] <mc-> ed, how about tomorrow?
[17:34] <Laurenceb> :P
[17:35] <edmoore> well, I am adding some stability
[17:35] <edmoore> mc-: that would probably work better, yeah
[17:35] <Laurenceb> I see
[17:36] <edmoore> using the wee drogue
[17:36] <Laurenceb> ah yes
[17:36] <Laurenceb> seen that done before
[17:36] <mc-> tomorrow evening?
[17:36] <Laurenceb> ooh the filter is working... but only if I printf the data out from within one of the functions... :D
[17:54] fnoble (n=fnoble@88-106-206-171.dynamic.dsl.as9105.com) joined #highaltitude.
[17:58] <fnoble> hello
[18:00] <Laurenceb> hi
[18:01] <Laurenceb> some_data=function(&some_data);
[18:01] <Laurenceb> ^ is that ok?
[18:02] <Laurenceb> its not going to "tie itself in a knot" ?
[18:02] <fnoble> depends what you do with it
[18:02] <Laurenceb> as some_data isnt changed until the function returns right?
[18:02] <fnoble> some_data is only changed when the function finishes
[18:02] <Laurenceb> yes
[18:02] <fnoble> yeah
[18:03] <fnoble> but if youre passing it by reference, then maybe you can make the assignment in the function itself?
[18:03] <fnoble> what is it doing?
[18:03] <Laurenceb> http://pastebin.com/m6a48f9ed
[18:04] <Laurenceb> line 142 prints out seemingly ok data
[18:04] <Laurenceb> line 147 the gain matrix looks ok
[18:05] <Laurenceb> line 152, my headi ng only changes at a gps update
[18:05] <Laurenceb> ie every 50 iterations
[18:05] <fnoble> ok, thats a valid reason to use that construct
[18:06] <Laurenceb> what do you mean?
[18:06] <fnoble> i mean, if you are doing some_data=function(&some_data)
[18:06] <Laurenceb> k
[18:07] <fnoble> and you never put the output into anything other than the same thing as you pass in by ref
[18:07] <fnoble> then you could make the update in the function for neatness
[18:07] <Laurenceb> my gain matrix looks something like 0.009965 0.006361 0.000013 0.675317
[18:07] <Laurenceb> between updates of the gps
[18:07] <fnoble> but you may want to put the putput into another variable to the input
[18:08] <fnoble> i dont know much about kalman stuff
[18:08] <Laurenceb> due the the error fudge on line529
[18:08] <Laurenceb> ok, doesnt matter
[18:13] <Laurenceb> aha think I fixed it
[18:15] <Laurenceb> hmm, appears to be working
[18:15] <Laurenceb> need a bit more work to tighten up the range limits or it'll give crazy results
[18:16] hallam (i=hallam@18.214.1.172) left irc: Read error: 110 (Connection timed out)
[18:20] <Laurenceb> working with angles is confusing, you have to put in if(angle>180) {angle-=360;} and so on all over the place
[18:21] <edmoore> and on the 6th day, god gave us atan2
[18:23] <Laurenceb> atan2 ?
[18:23] <mc-> atn2 copes with each quadrant
[18:23] <Laurenceb> ooh sweet
[18:24] <fnoble> edmoore: 99?
[18:34] flowolf (n=flowolf@unaffiliated/flowolf) left irc: "Leaving"
[18:45] akawaka (n=akawaka@external.treyarch.com) joined #highaltitude.
[18:45] <Laurenceb> cool thanx for the tip :P
[18:46] <Laurenceb> its looking good now, just need to adjust the setup code
[18:46] <Laurenceb> http://imagebin.org/16479
[18:47] mc- (n=mfcastle@cpc4-glfd1-0-0-cust538.glfd.cable.ntl.com) left irc:
[18:48] MetaMorfoziS (n=avr@3e70d12b.adsl.enternet.hu) joined #highaltitude.
[18:54] <fnoble> does that mean its working?
[18:55] <fnoble> what does the graph show?
[18:55] <Laurenceb> thats heading vs time
[18:55] <Laurenceb> so you can see it jump with each new gps fix (1hz)
[18:56] <Laurenceb> then inbetween it drifts with the gyro
[18:56] <fnoble> should it be trying to keep a heading of 0 degrees?
[18:56] <Laurenceb> it trusts the gps too much, and maybe the gyro too little
[18:56] <Laurenceb> no, i sent it some imaginary data
[18:57] <Laurenceb> its centered on the test data very well, but need the covariance matrices tweaking a bit
[18:58] <Laurenceb> it takes a few seconds to get a good heading
[18:58] <Laurenceb> so after cutdown, I'll turn off the servo and leave it centered for 6 seconds or so
[18:59] <Laurenceb> now I can start doing clever things like sticking HPOD into the GPS covariance matrix :P
[19:03] hallam (i=hallam@GUGGENHEIM-ONE-FORTY-SEVEN.MIT.EDU) joined #highaltitude.
[19:10] hallam (i=hallam@GUGGENHEIM-ONE-FORTY-SEVEN.MIT.EDU) left irc:
[19:13] edmoore (n=edmoore@88.212.167.121) left irc:
[19:13] Simon-MPFH (n=simon@phantom.mpfh.co.uk) left irc: "Leaving"
[19:17] Simon-MPFH (n=simon@phantom.mpfh.co.uk) joined #highaltitude.
[19:20] akawaka (n=akawaka@external.treyarch.com) left irc: "Leaving"
[19:30] <Laurenceb> http://imagebin.org/16483
[19:30] <Laurenceb> theres clearly a bug in the fake data code
[19:31] <Laurenceb> as all the steps are going down not up :P
[19:31] akawaka (n=akawaka@external.treyarch.com) joined #highaltitude.
[19:45] jcoxon (n=jcoxon@host86-144-103-14.range86-144.btcentralplus.com) joined #highaltitude.
[19:47] Hiena (n=Hiena@81.93.195.181.datatrans.hu) joined #highaltitude.
[19:48] <Hiena> Good evening!
[19:50] <Laurenceb> hello
[19:50] <Laurenceb> jcoxon: your good with python, how do I remove the last two characters from a string?
[19:50] <jcoxon> hey Laurenceb
[19:50] <Laurenceb> *you're
[19:50] <jcoxon> chomp?
[19:50] <jcoxon> let me check
[19:50] <Laurenceb> hmm characters yum
[19:51] <jcoxon> there definitely is something
[19:51] <Laurenceb> [:2] or something?
[19:51] <akawaka> [:-2]
[19:51] <akawaka> s = s[:-2]
[19:52] <jcoxon> haha that would do
[20:16] phatmonkey (n=ben@81.2.121.150) joined #highaltitude.
[20:53] Ebola (n=Ebola@unaffiliated/ebola) got netsplit.
[21:03] Ebola (n=Ebola@unaffiliated/ebola) got lost in the net-split.
[21:07] Ebola (n=Ebola@unaffiliated/ebola) joined #highaltitude.
[21:14] phatmonkey (n=ben@81.2.121.150) left irc:
[21:35] Simon-MPFH (n=simon@phantom.mpfh.co.uk) left irc: "Leaving"
[21:45] Hiena (n=Hiena@81.93.195.181.datatrans.hu) left irc: "-=Alkoholmentes-sör és repülõgép-szimulátor, biztos út a guminõ felé=-"
[22:05] <jcoxon> evening all
[22:33] <Laurenceb> hello
[22:47] <Laurenceb> I've got some pretty graphs on the wiki :P
[22:51] <akawaka> don't you dare post a link
[22:52] <Laurenceb> lolz
[23:01] Rocketboy1 (n=grunge@217.47.75.27) joined #highaltitude.
[23:01] <Laurenceb> hi RocketBoy1
[23:01] <Rocketboy1> yo
[23:03] <Laurenceb> I got a kalman filter working :D
[23:03] <Rocketboy1> spiffing
[23:04] <Laurenceb> it introduces a bit of lag... but you can get it down to around a second or less
[23:05] <Laurenceb> I might try to code a lag compensator into the parafoil code... not sure yet, guess it depends how well it works without
[23:17] jcoxon (n=jcoxon@host86-144-103-14.range86-144.btcentralplus.com) left irc: "Leaving"
[23:22] Laurenceb (n=laurence@host86-141-78-0.range86-141.btcentralplus.com) left irc: Remote closed the connection
[23:27] MetaMorfoziS (n=avr@3e70d12b.adsl.enternet.hu) left irc: "joccccc@k"
[23:49] flowolf (n=flowolf@unaffiliated/flowolf) joined #highaltitude.
[23:55] Laurenceb (n=laurence@host86-141-78-0.range86-141.btcentralplus.com) joined #highaltitude.
[23:55] <Laurenceb> cya all
[23:58] <natrium42> i need an edmoore
[23:58] <natrium42> Laurenceb, see ya
[00:00] --- Wed Apr 16 2008