highaltitude.log.20080527

[00:46] akawaka (n=akawaka@external.treyarch.com) left irc: Read error: 110 (Connection timed out)
[00:54] Tigga (n=chatzill@pc-232-235-60.magd.cam.ac.uk) left irc: Read error: 110 (Connection timed out)
[01:03] MetaMorfoziS (n=avr@3e44b94c.adsl.enternet.hu) left irc: "/quit"
[02:12] borism (n=boris@195-50-204-179-dsl.krw.estpak.ee) left irc: Read error: 145 (Connection timed out)
[02:13] yansa_ (n=yans@host-89-230-209-7.lublin.mm.pl) left irc: Read error: 110 (Connection timed out)
[02:28] akawaka (n=akawaka@cpe-76-173-152-142.socal.res.rr.com) joined #highaltitude.
[02:30] Akadecca (n=akawaka@cpe-76-173-152-142.socal.res.rr.com) joined #highaltitude.
[02:32] akawaka (n=akawaka@cpe-76-173-152-142.socal.res.rr.com) left irc: Nick collision from services.
[02:32] Nick change: Akadecca -> akawaka
[02:41] Laurenceb (n=laurence@dhcp37-133.sthughs.ox.ac.uk) left irc: Remote closed the connection
[04:01] akawaka (n=akawaka@cpe-76-173-152-142.socal.res.rr.com) left irc: Connection timed out
[04:26] mogwie (i=_johan_@cpe-24-24-211-18.socal.res.rr.com) left irc:
[04:57] akawaka (n=akawaka@cpe-76-173-152-142.socal.res.rr.com) joined #highaltitude.
[07:48] yansa_ (n=yans@host-89-230-209-7.lublin.mm.pl) joined #highaltitude.
[07:54] edmoore (n=edmoore@pomegranate.chu.cam.ac.uk) joined #highaltitude.
[08:07] akawaka (n=akawaka@cpe-76-173-152-142.socal.res.rr.com) left irc: Read error: 110 (Connection timed out)
[08:21] borism (n=boris@195-50-200-221-dsl.krw.estpak.ee) joined #highaltitude.
[08:23] yansa_ (n=yans@host-89-230-209-7.lublin.mm.pl) left irc: Read error: 110 (Connection timed out)
[10:14] flowolf (n=flowolf@unaffiliated/flowolf) joined #highaltitude.
[10:28] Tigga (n=chatzill@pc-232-235-60.magd.cam.ac.uk) joined #highaltitude.
[11:10] Simon-MPFH (n=simon@phantom.mpfh.co.uk) joined #highaltitude.
[12:09] MetaMorfoziS (n=avr@3e44b94c.adsl.enternet.hu) joined #highaltitude.
[12:34] RocketBoy (n=grunge@217.47.75.27) joined #highaltitude.
[12:35] <RocketBoy> does anyone know what the zero pressure balloon piece on sky news was this morning?
[12:37] RocketBoy (n=grunge@217.47.75.27) left irc: Client Quit
[13:50] Ebola (n=Ebola@unaffiliated/ebola) joined #highaltitude.
[13:59] Ebola (n=Ebola@unaffiliated/ebola) left irc: Connection reset by peer
[14:18] natrium42 (n=alexei@CPE000625d867e2-CM0014045885be.cpe.net.cable.rogers.com) left irc: "pwned!"
[14:53] yansa_ (n=yans@host-89-230-208-118.lublin.mm.pl) joined #highaltitude.
[14:58] Ebola (n=Ebola@unaffiliated/ebola) joined #highaltitude.
[17:17] yansa_ (n=yans@host-89-230-208-118.lublin.mm.pl) left irc: "Ex-Chat"
[17:27] jcoxon (n=jcoxon@80.88.222.59) joined #highaltitude.
[17:28] <edmoore> hi jcoxon
[17:29] <jcoxon> hey edmoore
[17:29] <edmoore> how goes?
[17:29] <jcoxon> not bad
[17:29] <jcoxon> long day
[17:30] <edmoore> yup
[17:30] <edmoore> they always are
[17:30] <edmoore> especially leading up to bloodybexams
[17:31] <jcoxon> indeed
[17:32] <jcoxon> done anything exciting
[17:32] <edmoore> no
[17:32] <edmoore> revision
[17:32] <jcoxon> oooo seen the cool phoenix picture with the parachute
[17:32] <edmoore> yeah
[17:32] <edmoore> amazing
[17:34] <edmoore> clever engineers
[17:35] <jcoxon> aren't they
[17:36] <edmoore> the video of the orbiter pointing to take the photo shows some pretty complex kinematics
[17:36] <jcoxon> its so cool that they now just turn bits of it on
[17:36] <jcoxon> i like the idea that now that its in place they are taking their time to do everything
[17:37] <edmoore> well you really don't wanna screw up!
[17:37] <jcoxon> http://news.bbc.co.uk/1/hi/world/americas/7421641.stm
[17:37] <edmoore> minimise all your movements and that
[17:37] <jcoxon> oh dear
[17:38] <edmoore> there was a wonderful look on the rovers project managers face when someone suggested getting rid of the dust by sort of folding the solar panels back up to vertical and wiggling them back and forth
[17:39] <edmoore> might get the altitude record
[17:40] edmoore (n=edmoore@pomegranate.chu.cam.ac.uk) left #highaltitude.
[17:40] edmoore (n=edmoore@pomegranate.chu.cam.ac.uk) joined #highaltitude.
[17:40] <edmoore> poor bugger
[17:40] <jcoxon> yeah
[17:43] <jcoxon> haha, i've got to be back in cambridge next monday
[17:44] <jcoxon> for a hospital appointment
[17:44] <jcoxon> only just remembered
[17:44] <edmoore> petrol costs must be scary
[17:46] <jcoxon> probably get the train
[17:46] <jcoxon> well i'll be in london anyway so will get the train to cam and back
[17:46] <jcoxon> and then drive down to canterbury
[17:46] <edmoore> best mentos vid ever
[17:46] <edmoore> http://break.com/index/mentos-and-diet-coke-at-1200-fps.html
[17:47] <jcoxon> nets too slow
[17:48] <jcoxon> bbiab
[17:56] akawaka (n=akawaka@cpe-76-173-152-142.socal.res.rr.com) joined #highaltitude.
[17:57] Akadecca (n=akawaka@cpe-76-173-152-142.socal.res.rr.com) joined #highaltitude.
[18:14] <jcoxon> back
[18:15] akawaka (n=akawaka@cpe-76-173-152-142.socal.res.rr.com) left irc: Read error: 110 (Connection timed out)
[18:41] Akadecca (n=akawaka@cpe-76-173-152-142.socal.res.rr.com) left irc: Read error: 110 (Connection timed out)
[18:47] MetaMorfoziS (n=avr@3e44b94c.adsl.enternet.hu) left irc: Read error: 104 (Connection reset by peer)
[18:55] MetaMorfoziS (n=avr@3e44b94c.adsl.enternet.hu) joined #highaltitude.
[19:10] <edmoore> jcoxon: am back too
[19:10] <jcoxon> yay
[19:13] <edmoore> lol
[19:13] akawaka (n=akawaka@external.treyarch.com) joined #highaltitude.
[19:38] jcoxon_ (n=jcoxon@80.88.222.59) joined #highaltitude.
[19:40] Rocketboy1 (n=grunge@217.47.75.27) joined #highaltitude.
[19:44] <jcoxon_> evening Rocketboy1
[19:44] jcoxon (n=jcoxon@80.88.222.59) left irc: Read error: 110 (Connection timed out)
[19:45] Nick change: jcoxon_ -> jcoxon
[19:48] <Rocketboy1> hi jcoxon_
[19:48] <Rocketboy1> hows it going - I havn't been round for a week or so
[19:49] <Rocketboy1> I saw a zero pressure balloon on Sky News today (they have it on monitors at work) but I don't know what it was about any ideas?
[19:51] <jcoxon> oh
[19:51] <jcoxon> it was a guy who was going to do a stratosphere jump
[19:51] <jcoxon> but they released the balloon too early by a mistake
[19:52] <Rocketboy1> it looked like it was in the UK and they were about to launch it
[19:52] <jcoxon> Canada
[19:52] <Rocketboy1> ah Ok
[19:52] <jcoxon> http://news.bbc.co.uk/1/hi/world/americas/7421641.stm
[19:52] <jcoxon> Rocketboy1, got a question for you about packet radio
[19:53] <jcoxon> the gumstix has audio out and has packet radio code built in to the kernel
[19:53] <jcoxon> but what would be required to interface this audio out to an actual radio module
[19:53] <akawaka> jcoxon: just an audio connection
[19:54] <jcoxon> but the radio modules i'm thinking is something like a Radiometrix NTX2
[19:54] <akawaka> and a circuit to trigger the ptt
[19:54] <jcoxon> rather than an actually full radio
[19:54] <akawaka> i see
[19:54] <jcoxon> cause its going to need to be 10mWs
[19:55] <jcoxon> unless i can find a nice low power module that accepts audio
[19:55] <akawaka> so that radio just takes serial data?
[19:55] <jcoxon> yeah
[19:55] <Rocketboy1> for the NTX 2 you need a 100K resistor to bias it to a central frequency + a capacitor to block DC
[19:55] <Rocketboy1> thats what I did for NOVA 1
[19:56] <jcoxon> so if i outputted audio thats all thats needed?
[19:56] <Rocketboy1> yep
[19:56] <jcoxon> wow
[19:56] <Rocketboy1> (and the NTX 2)
[19:56] <jcoxon> a ntx2 on a jack, just plug it in :-)
[19:57] <jcoxon> (with the bits in between
[19:57] <jcoxon> )
[19:59] <Rocketboy1> actually it depends on the output levels from the gumstix audio out
[19:59] <Rocketboy1> you need about 3V peak to peak drive to get the most out of it
[20:00] <jcoxon> the reason why i'm asking is that i've been playing with its usb host function
[20:00] <Rocketboy1> for NOVA-1 there wasn't enough audio out from the TNC I used - so I had to use an amplifier
[20:00] <jcoxon> so i can quite easily get images off a usb webcam
[20:01] <jcoxon> which could then be transmitted down (after a bit of processing)
[20:01] <Rocketboy1> thats why I used this circuit http://wiki.ukhas.org.uk/communication:nova1_radio_system
[20:02] <jcoxon> i haven't seen that page!
[20:02] <jcoxon> weird
[20:03] <jcoxon> so i'd need to measure the output peak
[20:03] <jcoxon> but potentially its pretty easy
[20:03] <Rocketboy1> yeah
[20:04] <akawaka> http://planetary.s3.amazonaws.com/misc/PSP_008579_9020_descent.jpg
[20:04] <Rocketboy1> actually if the output is DC coupled you can probably go streight into the NTX-2
[20:06] Ebola (n=Ebola@unaffiliated/ebola) left irc: Remote closed the connection
[20:06] <jcoxon> Rocketboy1, how would i work out if its DC coupled? the schematic?
[20:07] <Rocketboy1> yeah from the schem and the specs.
[20:07] <Rocketboy1> is that beagle?
[20:07] <jcoxon> no phoenix
[20:07] <Rocketboy1> wow
[20:08] <jcoxon> Rocketboy1, have you seen the simulation video of its entry and landing on mars?
[20:09] <Rocketboy1> nope
[20:11] <jcoxon> just finding it
[20:12] <jcoxon> http://www.youtube.com/watch?v=UgnWr3j7TCw&feature=related
[20:12] <akawaka> http://saturn.jpl.nasa.gov/multimedia/CASSIE/
[20:12] <akawaka> that one is fun
[20:12] <akawaka> the orbit is insane, constantly using titan for adjustments
[20:12] <jcoxon> Rocketboy1, so if can get this packet radio working, we should be able to get some mid flight images this summer :-p
[20:13] <akawaka> we're looking to do that with slowscan tv
[20:16] <Rocketboy1> uber cool
[20:18] <Rocketboy1> if you want me to take a look at the gunstix audio out let me know
[20:19] <jcoxon> oh yeah
[20:19] <jcoxon> especially with the goliath board on its way
[20:19] <Rocketboy1> actually it would be worth doing a packet radio test again
[20:19] <jcoxon> as if it could be just a matter of plugging in the radio
[20:19] <jcoxon> you'd then have gps, gsm, radio in one package
[20:19] <jcoxon> i've got a few spare audiostix daughterboards
[20:20] <jcoxon> which could be played with to
[20:20] <Rocketboy1> the test we did fro NOVA-1 packed up just as we seemed to be getting to the limit of communication range
[20:20] <jcoxon> well if we fly with say badger board then that could do the boring stuff
[20:20] <jcoxon> and we could push the new stuff
[20:20] <Rocketboy1> - but we havn't really tested the range of 10mW of packet fully
[20:21] <Rocketboy1> it may be better than I think
[20:22] <jcoxon> i'm quite happy to work on a system
[20:22] <jcoxon> the packet radio software should be pretty easy
[20:23] <edmoore> akawaka: is that real?
[20:24] <edmoore> it's not actually landing in the crater is it?
[20:24] <akawaka> edmoore: what?
[20:24] <akawaka> no, it landed about 20km away
[20:24] <edmoore> I didn't even know there were any craters near where it was landing
[20:24] <edmoore> the s3 jpg
[20:24] <edmoore> oh wow
[20:24] <edmoore> why did they not release that pic then rather than the crop?
[20:25] <akawaka> http://www.nasa.gov/mission_pages/phoenix/images/index.html
[20:25] <akawaka> happy mr skeptic pants?:)
[20:26] <edmoore> oh that's new since earlier
[20:26] <edmoore> well
[20:26] <edmoore> OMFG
[20:26] <edmoore> is all I can say
[20:27] <akawaka> better
[20:28] <Rocketboy1> ah - its only a bit or rocket science
[20:28] <Rocketboy1> piece of cake to do
[20:29] <akawaka> i'd do it but my week is already pretty busy
[20:29] <edmoore> There was this amusing graph we were shown at a talk on the Huygens lander.
[20:30] <edmoore> It was a plot of acceleration against time, time axis being a few minutes, acceleration being up to several G. There were two lines on it, both of approximatelly the same shape, except one was about 4 seconds later than the other
[20:31] <edmoore> It was predicted accelerations against time, and actual acclerations against time, for the lander descending onto Titan
[20:32] <edmoore> And someone asked him why there was 4 seconds difference - and he said that after 3.5 billions kilometers are 6 years, jpl got them there 4 seconds late
[20:32] <Rocketboy1> ha ha
[20:32] <edmoore> f'kin amateurs.
[20:33] <akawaka> haha
[20:33] <Rocketboy1> bbl - I'm going to play harve
[20:33] <edmoore> cool, cya later
[20:36] <edmoore> that picture is just so spectacular
[20:36] <edmoore> i can't get over it
[20:43] borism (n=boris@195-50-200-221-dsl.krw.estpak.ee) left irc: Read error: 145 (Connection timed out)
[20:44] hallam (i=hallam@GREEN-THREE-NINETY-EIGHT.MIT.EDU) joined #highaltitude.
[20:44] <hallam> hallo
[20:44] <edmoore> henry
[20:44] <edmoore> http://planetary.s3.amazonaws.com/misc/PSP_008579_9020_descent.jpg
[20:45] <hallam> got a reference for that? I saw the earlier one, but that looks too good to be true....
[20:45] <edmoore> until something cooler happens in the whole of engineering, this is my wallpaper
[20:46] <edmoore> my first reaction, then akawaka linked me: http://www.nasa.gov/mission_pages/phoenix/images/index.html
[20:48] <edmoore> I mean it seems it's a composite
[20:48] <edmoore> but still
[20:48] <hallam> bad ass
[20:49] <edmoore> I'm just astonished
[20:49] <edmoore> it's the best photograph ever
[20:49] flowolf_ (n=flowolf@host147-208-dynamic.3-79-r.retail.telecomitalia.it) joined #highaltitude.
[20:49] flowolf (n=flowolf@unaffiliated/flowolf) left irc: Read error: 113 (No route to host)
[20:50] <hallam> pretty damn amazing
[20:50] <hallam> I guess some people were a little worried about the descent systems after last time
[20:50] <edmoore> which time?
[20:54] <hallam> MPL 99
[20:54] <edmoore> that didn't even get as far as edl
[20:54] <edmoore> or rather, it never had a hope of edl'ing
[20:55] <edmoore> well.... I guess it actually did all three, just depending on how you look at things
[20:55] <hallam> yes it did, you're probably thinking of MCO
[20:55] <edmoore> you are correct, I am
[20:55] <edmoore> MPL just didn't communicate, i remember now
[20:56] <hallam> and they think it was a misidentification of vibration from the landing gear being deployed as surface contact
[20:56] <edmoore> that's a bit of a crap way to bugger up
[20:56] <hallam> quite
[20:57] <edmoore> I wonder if there are less deterministic but perhaps more robust ways of knowing when you've landed
[20:57] <edmoore> I guess they've got it working now anyhow
[20:58] <akawaka> radar
[20:58] <edmoore> It's an interesting topic. We'll have to think of a nice way for the flight computer to know what to do if it has a reset on the bench/on the way up/descending
[20:59] <hallam> well yeah, that should be designed in, it was in the old Nova software
[21:00] <edmoore> I spotted it. but copying straight across would be no fun!
[21:00] <hallam> save the most important state to nonvolatile memory regularly, and restore on reset
[21:00] <hallam> haha, I think the software could use a rewrite
[21:01] <edmoore> we'll have to start tarting up all the badgerboard software soon
[21:01] <hallam> and on reset, transmit a frame of telemetry before turning on any other hardware
[21:01] <jcoxon> evening hallam
[21:01] <hallam> hi James
[21:01] <edmoore> it would be good to have it determine where it is though
[21:01] <jcoxon> life good?
[21:02] <hallam> better to first transmit where it was before reset
[21:02] <edmoore> let's imagine hypothetical situation a) the lipo expands on the way up and we loose power. or a connector contracts. or something. it just dies on us
[21:02] <hallam> yes thanks
[21:02] <hallam> yourself?
[21:02] <edmoore> balloon carries on regardless, bursts, and on the way down we get pressure back/temperature up and get power back
[21:03] <hallam> sure
[21:03] <edmoore> it'd be nice for it to be robust enough to have no problem dealing with that
[21:03] <jcoxon> hallam, yup
[21:03] <hallam> right, I'd say it should be a requirement
[21:03] <edmoore> nice = requirement in ed-speak
[21:03] <edmoore> cool = nice to hae
[21:03] <hallam> hehe
[21:03] <edmoore> nice = must have
[21:04] <edmoore> cool is the high speed camera
[21:04] <edmoore> oh james, you'll love this. forward it to doug
[21:04] <hallam> so you send the (previously stored) frame of telemetry, then turn on the GPS and continue as before
[21:04] <hallam> if the GPS doesn't come back, at least you know where it was, even if not where it is
[21:04] <jcoxon> edmoore, forward what?
[21:04] <edmoore> hang on, getting there
[21:05] <edmoore> http://gizmodo.com/341634/casios-exilim-pro-ex+f1-hands+on-1200-fps-demo-video-and-sample-60+shots+per+second-gallery
[21:05] <edmoore> oh hang on, you have a poo connection
[21:05] <edmoore> basically it's an amazing camera
[21:05] <jcoxon> it might work
[21:05] <jcoxon> just give it time
[21:05] <edmoore> hallam: why forward the last telem?
[21:05] <edmoore> oh I see, sorry didn't read
[21:06] <edmoore> hallam: so we just go on external data if we have any
[21:06] <edmoore> if after a series on resets of gps and trying the backup on the telit we still got nuthin, we failsafe
[21:07] <hallam> right, but there's some chance that there's a power problem that could be exarcerbated by turning on other parts e.g. GPS
[21:07] <edmoore> fire the chute mortar and be done with it
[21:07] <hallam> yes but include a failsafesafe to only fire the chute if you're sure you're still airborne
[21:07] <edmoore> we will have current and voltage monitoring built into badger v2
[21:07] <hallam> would suck to have it land on someone's face, jolt the batteries back into life and then explode on them
[21:07] <edmoore> assuming it's not that bit that's gone down
[21:08] <edmoore> yes quite, but equally it'd suck to have it land on someone's face
[21:08] <edmoore> without chute
[21:09] <edmoore> so the next level of failsafe is an inherent machanical one - make it fall non dangerously a-la nova 4
[21:09] <edmoore> well the payload fergus and i are building should be that anyway, which was why we have gone deployable chutes
[21:09] <edmoore> basically nova 2.ahem = never again
[21:11] <edmoore> I was thinking we could use the accelerometers to see if we're descending or have landed if the gps goes down. equally that wouldn't work if someone has picked it up
[21:11] <edmoore> how about, hallam, in the style of a securicor van : "ATTENTION. PARACHUTE ABOUT TO DEPLOY. PLEASE STAND BACK!"
[21:11] <edmoore> on a loud-speaker
[21:12] <edmoore> my monologue co-efficient is increasing - somebody say something!
[21:13] <jcoxon>
[21:13] <edmoore> clever
[21:13] <jcoxon> :-p
[21:14] <hallam> I think barometric is the way to go
[21:14] <jcoxon> edmoore, was talking to Rocketboy1 about packet radio on the gumstix uadio
[21:14] <hallam> that SCP1000 is pretty solid
[21:14] <hallam> no deploy if < 500ft
[21:14] <edmoore> i guess we could put the footprint on
[21:15] <hallam> has issues with mountains, but not many of those in the UK
[21:15] borism (n=boris@195-50-200-221-dsl.krw.estpak.ee) joined #highaltitude.
[21:15] <edmoore> actually, we have a pressurised warm box on the new payload, so it'll have to be external
[21:15] <jcoxon> castle hill!
[21:16] <edmoore> I say 'warm', it's not heated, just insulated and has a pcb with linear regs inside it.
[21:16] <hallam> pressurised??
[21:16] <edmoore> I say 'pressurised', it's just in a box with no obvious air gaps
[21:17] <hallam> that'll leak like a sieve then
[21:17] <edmoore> and an o-ring
[21:17] <hallam> like a smaller sieve
[21:17] <edmoore> it'll be sealed on 5 faces, if you see what I mean, and the in-out face will have an o-ring on it
[21:17] <edmoore> it'll probably take a 1atm...
[21:17] <hallam> ask Ferg about vacuum leaks
[21:18] <edmoore> it's not critical, just better than nothing
[21:18] <hallam> guess so
[21:19] <hallam> I'm going to have the rocket computer heat itself up to +20C before rocket launch, then hope the components stay within operating limits when it hits vacuum
[21:19] <hallam> and more importantly within calibration limits
[21:19] <edmoore> I'm not entirely convinced by the telits and cold temperatures (sure, we'll test it thoroughly), and if it's to be subject to lots of test flights, may aswell try and minimise thermal stress cycling
[21:20] <edmoore> given it costs us little and doesn't make any new failure points. I hate them.
[21:20] <edmoore> right, i need to be off. will be back later. 11ish
[21:20] <edmoore> byeeeeee
[21:21] Action: jcoxon reminds himself that running kermit through kermit makes it hard to quit from the first kermit
[21:21] <jcoxon> or was it the 2nd kermit
[21:21] <jcoxon> oh well
[21:23] <hallam> later Ed
[21:30] <jcoxon> hallam, been up to anything interesting then?
[21:30] Simon-MPFH (n=simon@phantom.mpfh.co.uk) left irc: "Leaving"
[21:34] Rocketboy1 (n=grunge@217.47.75.27) left irc: Read error: 110 (Connection timed out)
[21:36] <hallam> designing a cute little rocket flight computer
[21:36] <hallam> doing a bit of flying
[21:36] <hallam> relaxing after finals
[21:36] <hallam> how is the hospital treating you?
[21:47] jcoxon_ (n=jcoxon@80.88.222.59) joined #highaltitude.
[21:47] <jcoxon_> stupid hospital internet
[21:55] flowolf_ (n=flowolf@host147-208-dynamic.3-79-r.retail.telecomitalia.it) left irc: "Leaving"
[21:55] RocketBoy (n=grunge@217.47.75.27) joined #highaltitude.
[21:55] flowolf (n=flowolf@unaffiliated/flowolf) joined #highaltitude.
[21:56] jcoxon (n=jcoxon@80.88.222.59) left irc: Read error: 110 (Connection timed out)
[21:57] natrium42 (n=alexei@CPE000625d867e2-CM0014045885be.cpe.net.cable.rogers.com) joined #highaltitude.
[22:15] Laurenceb (i=Laurence@dhcp38-010.sthughs.ox.ac.uk) joined #highaltitude.
[22:15] <Laurenceb> hi all
[22:15] <Laurenceb> RocketBoy: got your email
[22:15] <Laurenceb> edmoore: you about?
[22:19] <edmoore> Laurenceb: yes
[22:19] <Laurenceb> how good are you with statistics?
[22:20] <Laurenceb> I have two measurements, errors on each are 0.5, and the correlation coefficient is 0.5
[22:20] <Laurenceb> how do I find a covariance matrix?
[22:22] flowolf (n=flowolf@unaffiliated/flowolf) left #highaltitude ("Leaving").
[22:23] <edmoore> do you know the standard devations?
[22:23] <edmoore> actually you don't need it, it's just 1
[22:24] <edmoore> oh no wait I misread you - can you define error a bit more?
[22:24] <Laurenceb> hmm if the matrix is a,b,c,d
[22:24] <Laurenceb> standard error
[22:24] <edmoore> ie. is that normalised?
[22:24] <Laurenceb> hmm
[22:25] <Laurenceb> I put in a vector of normalised noise at one side
[22:25] <edmoore> ok cool, not to panic
[22:25] <Laurenceb> the matrix turns it into measurement errors
[22:26] <Laurenceb> so if my matrix was the identity, standard errors would be 1, and there would be 0 correlation
[22:26] <edmoore> yep correct
[22:27] <Laurenceb> but working it out from the info I have is a bit confusing...
[22:28] <Laurenceb> hmm does 0.5^2=a^2+b^2 = c^2+d^2 ?
[22:28] <edmoore> well the question is really about your standard errors and how they relate to the standard devaition
[22:28] <edmoore> and I'm not sure you can calc one from t'other with the info you've got
[22:28] <Laurenceb> I already said
[22:28] <Laurenceb> with the identity example
[22:29] <Laurenceb> I think 0.5^2=a^2+b^2 = c^2+d^2 , and also ac+bd=0.5^2 or 0.5^3 ... not quite sure
[22:30] <Laurenceb> ie either correlation factor * error or error^2
[22:31] <Laurenceb> I think its ^3, as basically its <error1 * error2> = <error1*error1* correlation factor> by definition
[22:33] <Laurenceb> its horrible isnt it :P
[22:33] <edmoore> is that the definition?
[22:34] <Laurenceb> I think so.... but not 100%
[22:36] <Laurenceb> ok, I think this might work, but then you average 10 measurements... arggg
[22:36] <edmoore> oh so there are 10 measurements?
[22:37] <Laurenceb> I guess you can propogate the matrix into the state frame
[22:37] <edmoore> oh I am confused
[22:37] <edmoore> you'll have to ask me again in about a year
[22:37] <Laurenceb> then use some more standard techniques on the 10 results
[22:37] <Laurenceb> nvm
[22:37] <Laurenceb> :P
[22:38] <Laurenceb> it involves kalman filters
[22:38] <Laurenceb> but not in the normal engineering sense
[22:39] <edmoore> my course in proper man statistics in next year
[22:39] <edmoore> for now I should probably not upset my brain with stuff it won't be examined on
[22:40] <akawaka> so based on the data we get during ascent, we plan on doing a prediction for our descent and correct it based on data received during the descent, anyone done anything like this?
[22:41] <edmoore> yes... sort of
[22:41] <Laurenceb> I've never tried on the way down
[22:42] <edmoore> is as far as it uses the discrete cumulative method on the way up
[22:42] <Laurenceb> theres code for that on the wiki
[22:42] <edmoore> i.e. adding up the landing spots in discrete altitude chunks, a la the wiki
[22:42] mc- (n=mfcastle@cpc4-glfd1-0-0-cust538.glfd.cable.ntl.com) joined #highaltitude.
[22:42] <Laurenceb> hi mc-
[22:42] <akawaka> url?
[22:42] <akawaka> descent rate is a big unknown really
[22:43] <mc-> hi lb
[22:43] <edmoore> but then on the way down, predict where you should be in 500m (altitude) time, then see where you actually end up
[22:43] <edmoore> and add the difference to the final landing spot
[22:43] <edmoore> there is probably a much better way of doing it
[22:44] <mc-> hi edmoore, did you manage to check the bits delivered by R&Things?
[22:44] <mc-> akawaka, did you build your class E tx?
[22:44] <akawaka> the onboard computer is probably the best place to do the prediciton since it has second-by-second data
[22:45] <akawaka> mc-: yes i did, seems to be working
[22:45] <edmoore> hi mc- my car is at home atm - it will be back up in 2 days. I will check then. It had to go home about a week and a half ago for mending
[22:45] <mc-> ok, np
[22:45] <akawaka> writing some software for it right now
[22:45] <mc-> did you see the scope waveforms?
[22:45] <edmoore> akawaka: you can see it just keeps a running total of the position
[22:45] <edmoore> it has no memory of previous states
[22:45] <akawaka> not sure how effective it is, hopefully i can get some contact reports from it
[22:46] <mc-> on 7MHz band?
[22:46] <akawaka> mc-: yeah, the output looks good, and it seems to be switching right
[22:46] <akawaka> mc-: no, i got my general license so its on 10.143
[22:47] <edmoore> a better algorithm would keep track of all the data it accrues and dynamically adjust the predicted landing spot by updating it's descent rate model based on the actual data it's getting in, and running the landing model throughout the entire data you stored on the way up
[22:47] <mc-> driving into a dummy load? what I and C values did you use?
[22:47] <akawaka> mc-: yeah, i used the values in bens schematic
[22:47] <mc-> and wound your own inductor?
[22:48] <akawaka> no, bought 1.2uH inductors on digikey
[22:48] <akawaka> connected two in series, perpendicular to each other for the 2.4uh
[22:48] <mc-> good thinking
[22:49] <akawaka> http://picasaweb.google.com/martin.donlon/Random/photo#5203321294655086386
[22:50] <akawaka> hot blurry perfboard action
[22:51] <akawaka> ben recommended adding a fet to slow down the rise and fall times, gonna add that tonight
[22:51] <akawaka> after paying $7 for a pfet from frys (fuckers)
[22:52] <mc-> looks like you bought the digikey oscillator for 10.143MHz
[22:53] <mc-> what RC values will you have for the pfet?
[22:54] <akawaka> 10kohm and 1uh, .22uf or .1uf
[22:54] <akawaka> haven't decided yet:)
[22:54] <akawaka> would like a rise time of between 1 and 5ms
[22:55] <akawaka> not sure exactly how the rc time constant relates to that
[22:55] <mc-> the fet will switch over a small voltage range though
[22:55] <mc-> I think you'll need larger RC values
[22:56] <akawaka> yeah?
[22:57] <mc-> it's a few 100 mv to switch between off and on, AFAIK
[22:58] <akawaka> at what point does it reach saturation though?
[22:58] <akawaka> fets are all new to me to be honest
[22:58] <mc-> and won't be well balanced as the switching point is near Vdd
[22:58] Action: edmoore got a digikey catalogue in the post today. My, it's a beast.
[22:58] <mc-> ie it will ramp up slowly, and ramp down quickly
[22:59] <mc-> so there will be spikes on switch off
[22:59] <akawaka> so two different resistor values?
[22:59] <mc-> yes, that could work
[23:00] <mc-> or if you want to be fancy, a PWM waveform
[23:00] <akawaka> i don't want to be fancy:)
[23:02] <akawaka> i should be able to take a look at it in my scope, but it is difficult to catch one-shot events like that
[23:03] <mc-> drive the pfet with a square wave so the scope triggers
[23:05] <mc-> it maybe easier to have a small square wave around the turn on/off point?
[23:06] <akawaka> yeah
[23:08] <mc-> the Gate-Source threshold on the pfet I have is -1V
[23:09] <mc-> you'll need to drive it from the micro with a NPN
[23:09] <akawaka> yeah
[23:11] <akawaka> so the fet is complete on at -1?
[23:11] <mc-> so you could have 2 Rs as a voltage divider from the NPN collector to the pfet
[23:11] <mc-> -1V is the threshold
[23:11] <mc-> but can vary from part to part
[23:12] <mc-> sorry -1V is the min from the dsheet
[23:14] <mc-> another way might be to build an op-amp integrator, so you can ramp up and down the voltage slowly
[23:15] <mc-> but not too slow, or the pfet will get too hot, and burn out...
[23:16] <mc-> cya
[23:16] <akawaka> ythank you!
[23:16] <mc-> have fun
[23:17] mc- (n=mfcastle@cpc4-glfd1-0-0-cust538.glfd.cable.ntl.com) left irc:
[23:17] <Laurenceb> http://wiki.ukhas.org.uk/ideas:landing_spot_prediction <- akawaka
[23:18] Laurenceb_ (i=Laurence@dhcp38-010.sthughs.ox.ac.uk) joined #highaltitude.
[23:19] <Laurenceb_> edmoore: I have the S matrix , its (0.129,0.483,0.483,0.129)
[23:19] <Laurenceb_> huhahaha
[23:20] <Laurenceb_> well I think that works...
[23:21] <edmoore> isn't the S matrix quantum mechanics?
[23:21] <Laurenceb_> erm
[23:21] <Laurenceb_> I was talking in the kalman matrix sense
[23:22] <Laurenceb_> as thats what its about - in our lectures covariance is usually called S
[23:22] <akawaka> just calculate the eigenstate
[23:22] <Laurenceb_> yes yes
[23:22] <Laurenceb_> :P
[23:22] <akawaka> anyway, the answer is two photons
[23:23] <edmoore> fair do's
[23:23] <akawaka> sorry, forgot to carry the 1
[23:23] <akawaka> three photons
[23:24] RocketBoy (n=grunge@217.47.75.27) left irc: "Leaving"
[23:24] hallam (i=hallam@GREEN-THREE-NINETY-EIGHT.MIT.EDU) left irc: Read error: 110 (Connection timed out)
[23:26] <Laurenceb_> http://www2.b3ta.com/host/creative/65052/1211808870/CreditCrunch.jpg
[23:28] <edmoore> yum
[23:32] <akawaka> you can really taste the recession
[23:36] jcoxon_ (n=jcoxon@80.88.222.59) left irc: "Leaving"
[23:36] Laurenceb (i=Laurence@dhcp38-010.sthughs.ox.ac.uk) left irc: Read error: 110 (Connection timed out)
[23:38] hallam (i=hallam@PIKA-ONE-TWENTY-ONE.MIT.EDU) joined #highaltitude.
[23:49] edmoore (n=edmoore@pomegranate.chu.cam.ac.uk) left irc:
[23:52] Laurenceb_ (i=Laurence@dhcp38-010.sthughs.ox.ac.uk) left irc:
[00:00] --- Wed May 28 2008