highaltitude.log.20090129

[00:09] <hallam> Hi Laurenceb
[00:10] <hallam> I'm getting the C to do nav
[00:11] <Laurenceb> nice
[00:11] <Laurenceb> going with least squares?
[00:11] <hallam> I guess
[00:11] <hallam> just going to do the nav bit and ephemeris extraction for now
[00:12] <hallam> trying to code it in a way that will work well on the blackfin
[00:12] <Laurenceb> yeah
[00:12] <Laurenceb> will this run with blackfin linux?
[00:13] <hallam> hadn't intended to
[00:14] <Laurenceb> but will it work?
[00:14] <Laurenceb> I mean it uses dma, so shouldnt mind should it?
[00:15] <hallam> yeah in principle it should be ok
[00:15] <hallam> as long as it can run every ms
[00:15] <Laurenceb> it would be kind of cool
[00:15] <Laurenceb> to have it rn under blackfin linux
[00:16] <hallam> sure
[00:17] <hallam> maybe i'll look at it later
[00:17] <hallam> but I usually prefer bare-metal
[00:28] <hallam> fergusnoble:
[00:28] <hallam> APPLY FOR CME
[00:28] <fergusnoble> hallam: having second thoughts
[00:29] <fergusnoble> hallam: spoke to my supervisor and he said that a lot of physics students didnt rate the teaching there
[00:29] <fergusnoble> or course, not teaching quality i guess
[00:30] <fergusnoble> is sondy on skype? it might be good to have a quick chat with her
[00:30] <hallam> yes she is
[00:30] <hallam> I didn't take the classes, but got the impression the physics teaching is top notch, especially junior lab
[00:31] <hallam> id springmann
[00:31] <Laurenceb> CME?
[00:32] <fergusnoble> yeah, im sure the quality is good but apparently its not at the same level, i guess owing to the fact people start eariler and specialise less
[00:32] <fergusnoble> hallam: im looking at the course catalogue, how do yo tell which ones are for which year?
[00:33] <fergusnoble> or do you just do the ones you have the prereqs for
[00:33] <hallam> well, you can take any of them
[00:33] <hallam> yeah
[00:33] <fergusnoble> ok
[00:33] <hallam> come on skype
[00:33] <fergusnoble> and coming from cambridge you dont have to worry about prereqs i guess
[00:33] <fergusnoble> as you cant have done any of them
[00:33] <hallam> right, basically look at the prereqs and see if you think you can handle it
[00:34] <hallam> and you can ask the profs about it
[00:34] <hallam> it's very flexible
[00:34] <fergusnoble> ok
[00:34] <hallam> it's your decision and all but I think you'd be absolutely crazy not to apply
[00:44] <Laurenceb> whatas this?
[00:44] <hallam> cambridge-mit exchange
[00:44] <Laurenceb> ah yeah
[00:44] <Laurenceb> cool
[00:44] <Laurenceb> yeah I'd apply
[00:44] <Laurenceb> guess the only thing to be concerned about would be air fare
[00:45] <Laurenceb> but its hopefully not to bad
[00:49] <SpeedEvil> win size 40
[00:49] <SpeedEvil> meh
[00:50] <SpeedEvil> Could you stop talking so much, you fell off my scroll buffer again :)
[01:01] <hallam> http://www.pegasushabproject.org.uk/zeusbot/highaltitude.log.20090128
[01:02] <Laurenceb> anyone know what auto CR+LF does in gtkterm?
[01:02] <Laurenceb> my radio sets DTR low when it gets /n
[01:03] <Laurenceb> but with auto CR+LF off, when I enter return, its not reconised
[01:03] <hallam> i would guess that it adds an automatic LF after every CR
[01:03] <hallam> or something along those lines
[01:03] <Laurenceb> but a \n seems to come through the radio link
[01:03] <hallam> you know, unix vs windows file format
[01:03] <Laurenceb> its really odd
[01:03] <Laurenceb> yeah it works in windows fine
[01:03] <hallam> what \n spits out varies by platform
[01:03] <hallam> sometimes 13, sometimes 20, sometimes both
[01:03] <Laurenceb> I set it to recognise ascii 10
[01:03] <hallam> er, or was it 10
[01:03] <hallam> yeah
[01:03] <hallam> 10 is LF I think
[01:03] <hallam> 13 is CR
[01:04] <Laurenceb> when I put on auto Cr+LF it works
[01:04] <Laurenceb> and it works when installed on the rogallo
[01:04] <Laurenceb> I thought there was a bug in the code
[01:05] <Laurenceb> od that without auto cr+LF a character comes through that causes fldigi to give a newline
[01:05] <Laurenceb> maybe fldigi thinks ascii 13 is newline
[01:05] <Laurenceb> this is interesting, I've left the radio on for 3 hours
[01:05] <Laurenceb> but only just turned on fldigi
[01:06] <Laurenceb> and its drifting a fair bit
[01:06] <Laurenceb> cartainly several %
[01:06] <Laurenceb> in audio frequency
[01:06] <fergusnoble> http://www.authorhouse.com/BookStore/ItemDetail~bookid~32221.aspx
[01:06] <Laurenceb> the radiometrix transmitters arent that unstable
[01:06] <fergusnoble> "This is a text on Hamiltionian and Lagrangian Dynamics aimed at the student who has just finished the calculus sequence."
[01:06] <Laurenceb> so looks like the sound
[01:06] <hallam> nice
[01:06] <fergusnoble> lol
[01:07] <Laurenceb> I cant wait
[01:07] <Laurenceb> a relaxing evening read
[01:07] <hallam> that looks like a pretty good book
[01:07] <hallam> Sondy has the FORTRAN Colouring Book
[01:07] <fergusnoble> but you will never read it because you need to finish summing an infinite series
[01:08] <hallam> arrrrrrgh
[01:08] <hallam> that deserves punishment
[01:08] <Laurenceb> why the hell would my soundcard drift?
[01:08] <hallam> because it's cheap crap
[01:09] <Laurenceb> :P
[01:09] <Laurenceb> I know that much
[01:10] <Laurenceb> but why is it stable to start off with
[01:10] <Laurenceb> and then starts drifting and giving clock glitches after about 10 minutes of recording
[01:10] <Laurenceb> I'm 100% sure its not the radio now
[01:11] <Laurenceb> hmm I need a ferrite on the audio cable
[01:11] <Laurenceb> as well
[01:11] <hallam> Laurenceb: maybe it's a software audio thing
[01:11] <hallam> does it depend on CPU load?
[01:11] <Laurenceb> ooh interesting
[01:11] <Laurenceb> well
[01:11] <Laurenceb> it doesnt drop out atall
[01:12] <Laurenceb> actually... maybe theres hardware to the speakers
[01:12] <Laurenceb> when you load up other applications the audio in fldigi drops out
[01:12] <Laurenceb> and playing audio files on here under linux is a pain
[01:13] <Laurenceb> hang on, how do you mean?
[01:15] <hallam> eh I wasn't really thinking very hard
[01:15] <hallam> was thinking about software modems but that's totally different
[01:16] <hallam> but I still like to blame linux wherever possible
[01:18] <fergusnoble> Laurenceb: fiddle with alsa
[01:18] <Laurenceb> its using portaudio
[01:18] <fergusnoble> its sometimes doesnt like to share
[01:18] <fergusnoble> ok
[01:19] <fergusnoble> but portaudio will probably use alsa at some level
[01:19] <Laurenceb> guess I've got enough info to blame the soundcard
[01:19] <Laurenceb> need to get a usb one by the look of it
[01:20] <Laurenceb> alternatively I could go with the plan of an AVR based icom controller and receiver
[01:24] <hallam> do the decoding on the AVR?
[01:25] <hallam> I'm making some progress on the position stuff
[01:25] <hallam> I think perhaps I need to make separate corrections for clock error and carrier
[01:25] <hallam> i.e. separate the carrier aiding from the sample time thing
[01:25] <hallam> Laurenceb: Do you know with any certainty whether the IF is 38400 or 38000MHz?
[01:25] jiffe20 (n=jiffe@209.159.247.189) left irc: Read error: 60 (Operation timed out)
[01:27] <Laurenceb> didnt you work it out?
[01:28] <hallam> I think it's 384
[01:28] <Laurenceb> yeah
[01:28] <hallam> welll
[01:28] <Laurenceb> Hz
[01:28] <hallam> argh confusing
[01:28] <hallam> yeah Hz sorry
[01:28] <Laurenceb> you worked it out as slightly lower
[01:28] <Laurenceb> erm no
[01:28] <Laurenceb> ignore me
[01:28] <Laurenceb> yes, you worked out 38.4
[01:29] <Laurenceb> but there going to be a fair bit of drift
[01:29] <Laurenceb> TCXOs are typically 0.5ppm
[01:30] <hallam> the crystal change should make the LO at the mixer 1571.2896MHz instead of 1571.328
[01:30] <hallam> so that gives an IF with L1 of 4.1304MHz
[01:30] <hallam> but then the ADC sampling does something
[01:31] <Laurenceb> yeah
[01:31] <Laurenceb> converts that down
[01:31] <hallam> how?
[01:31] <hallam> ADC sample rate is 8.1838MHz
[01:32] <Laurenceb> yeah, 4.1304 is more than half that
[01:32] <Laurenceb> hmm I get 38.5KHz
[01:34] <hallam> ok I get that too
[01:34] <hallam> I think that's the IF
[01:36] <hallam> ok now I'm trying a separate clock correction and carrier aiding
[01:36] <hallam> have to go back to the matlab code for this, takes ages
[01:37] <hallam> but the reason I'm doing this...
[01:37] <hallam> I tried my nav code on the results of an older matlab tracker that didn't use carrier aiding
[01:37] <hallam> it gave results with a fuckton of noise, but centered on the right place
[01:38] <hallam> obviously noisy because the DLL had to have lots of gain with no aiding
[01:38] <hallam> but it did have a separate sample clock correction
[01:38] <hallam> so now I've put the aiding back in but using the real IF, and kept the clock correction
[01:39] <Laurenceb> right...
[01:39] <hallam> let's see how it goes
[01:39] <Laurenceb> yeah I had issues with that
[01:39] Action: hallam waits for matlab
[01:40] <Laurenceb> surely you say the sampling freq is 8.184
[01:40] <Laurenceb> i.e. correct
[01:40] <Laurenceb> then IF=0
[01:40] <hallam> that's what I did
[01:40] <hallam> and it didn't work
[01:40] <Laurenceb> you have large dopplers
[01:40] <hallam> yeah
[01:40] <hallam> it works to keep the lock
[01:40] <Laurenceb> and correct the tracker correctly
[01:40] <Laurenceb> yeah, thats what I did
[01:40] <Laurenceb> with borres tracking code and acquisition code
[01:51] <Laurenceb> cant get a bit of sence out of the radio now
[01:52] <Laurenceb> if the drift on the fft is all down to the soundcard it must have drifted by about 8% or more
[01:52] <Laurenceb> radiometrix are quite stable right?
[01:52] <hallam> to a degree
[01:52] <Laurenceb> I mean we have to retune in flight
[01:52] <Laurenceb> but thats the temperature shift
[01:52] <hallam> yeah but that's probably temperature
[01:52] <hallam> I wouldn't expect one to shift by several kHz on the ground
[01:53] <Laurenceb> yeah
[01:53] <Laurenceb> guess theres the icom as well
[01:53] <Laurenceb> which is quite toasty now
[01:54] <Laurenceb> still
[01:54] <Laurenceb> doesnt change the fact I cant get a single bit through the link
[01:54] <Laurenceb> wheras before it was working fine
[01:54] <Laurenceb> - ~ 30 minutes ago
[02:03] <hallam> so
[02:03] <hallam> what kind of spread were you getting on your "fern"?
[02:06] <Laurenceb> 150x100m
[02:06] <hallam> yeah I think I can beat that
[02:07] <Laurenceb> I got it down to a 35m radiusx 10m thick disc
[02:07] <Laurenceb> using pseudorange correction
[02:07] <SpeedEvil> with 1ms?
[02:07] <Laurenceb> with DLL
[02:07] <Laurenceb> yes, at 1ms
[02:08] <hallam> hm
[02:08] <hallam> ok yeah I'm on the order of the 35m x 10m
[02:08] <hallam> with the center pretty close to the true position
[02:08] <Laurenceb> cool
[02:08] <Laurenceb> how far?
[02:08] <hallam> ~2m
[02:08] <Laurenceb> nice
[02:09] <SpeedEvil> hallam: do you have a second dataset from the same place?
[02:09] <Laurenceb> looks like you've cracked it
[02:09] <hallam> yes
[02:09] <hallam> haven't tried processing it yet
[02:09] <hallam> well, there are some caveats
[02:09] <hallam> this is with a matlab tracker
[02:09] <SpeedEvil> :)
[02:09] <hallam> need to make the same changes to the C one
[02:09] <Laurenceb> yeah
[02:09] <hallam> and it was at 8MSps but with decimated data, so that oughtn't to make a difference
[02:10] <Laurenceb> you used carrier aiding?
[02:10] <Laurenceb> ad pseudorange based on dll
[02:13] <hallam> yeah
[02:15] <Laurenceb> ok
[02:16] <Laurenceb> I'm not sure how 1Hz receivers do it...
[02:16] <Laurenceb> maybe just one solution
[02:18] <hallam> they could average a bunch of pseudoranges and put that into the solution
[02:19] <Laurenceb> pseudoranges change
[02:19] <Laurenceb> over the course of a second
[02:20] <hallam> true
[02:20] <hallam> maybe they're just artificially limited for economic reasons
[02:20] <Laurenceb> maybe a linear fit would be ok
[02:21] <Laurenceb> theres certainly more noise as you increase the update rate
[02:23] <Laurenceb> bet its a linear fit
[02:26] <SpeedEvil> doppler only changes a couple of m/s/s or so
[02:26] bfirsh (n=ben@host-137-205-75-156.res.warwick.ac.uk) left irc: Remote closed the connection
[02:26] <SpeedEvil> so the dominant is actually braking/accel
[02:31] <Laurenceb> the sats move at 4Km/s
[02:32] <Laurenceb> so over 1 second you'll be up to 4Km
[02:32] <hallam> no
[02:32] <Laurenceb> differnet
[02:32] <hallam> 1. they don't move that fast (IIRC)
[02:32] <SpeedEvil> the sats move at 4Km/s
[02:32] <Laurenceb> 3.85
[02:32] <hallam> 2. the range rate doesn't change at the same speed the sats move
[02:32] <SpeedEvil> but for trig reasons, they never move over about 700m/s WRT earth
[02:32] <Laurenceb> sure
[02:32] <Laurenceb> ok
[02:32] <Laurenceb> but still
[02:33] <hallam> yeah you're right though, they prob don't just average
[02:33] <SpeedEvil> and 700m/s sinewave with a 12h period doesn't change very fast
[02:33] <Laurenceb> yeah, so a linear fit
[02:33] <SpeedEvil> sorry - couple of m/s a minute I meant - not second
[02:35] Action: SpeedEvil ponders real world. You don't generally approach even 1G in a car.
[02:36] <Laurenceb> SpeedEvil: it seems fastrax have stolen your idea
[02:36] <Laurenceb> for arm based gps
[02:36] <SpeedEvil> Oh Noes!
[02:36] <Laurenceb> :-/
[02:37] <SpeedEvil> yeah - well...
[02:37] <SpeedEvil> I'm not interested in making GPS modules
[02:37] <SpeedEvil> that's _way_ too commoditised.
[02:37] <SpeedEvil> I'm hoping to make niche stuff.
[02:38] <SpeedEvil> Want a USB missile launcher with JDAM?
[02:38] <Laurenceb> lulz
[02:39] <SpeedEvil> cheap CPUs have gotten fast enough that the really simple digitising chips are good enough
[02:39] <SpeedEvil> don't need anything more.
[02:40] <SpeedEvil> ARM what?
[02:40] <Laurenceb> 9
[02:40] <Laurenceb> I was pondering avr32
[02:40] <Laurenceb> ngw100 gps :P
[02:42] <SpeedEvil> Thinking of a bit smaller than that
[02:43] <SpeedEvil> Tracking isn't a big problem - lock is fun.
[02:44] <SpeedEvil> Preliminary thoughts are 20ms to RAM, preprepare it, and crunch on that data for a while.
[02:44] <SpeedEvil> seems to work OK.
[02:45] <SpeedEvil> This pops out approx doppler and phase, and maybe some nav-bit edges.
[02:46] <SpeedEvil> then do the tracking loop thing based on the few seconds old data
[02:47] <Laurenceb> you have to be sure the crystal hasnt drifted in the meantime
[02:47] <SpeedEvil> don't really care.
[02:47] <SpeedEvil> It's 20-30s - it's not going to have drifted that badly.
[02:47] <hallam> I was able to jump 35 seconds into the data file based on carrier-aided code phase prediction
[02:48] <SpeedEvil> And worst case, can do a second acquisition frame that'll take .1% of the time
[02:48] <Laurenceb> yeah
[02:48] <Laurenceb> it wont have drifted too much
[02:48] borism (n=boris@195-50-211-250-dsl.krw.estpak.ee) left irc: Read error: 145 (Connection timed out)
[02:48] <Laurenceb> at least with a tcxo
[02:49] <SpeedEvil> For some of the stuff I'm pondering I'd actually like to be able to wake the GPS from off for 2ms, and sleep again.
[02:49] <SpeedEvil> but I imagine that's not going to work for a number of reasons.
[02:51] Action: SpeedEvil lobbies for 4 easy acquire sats in GEO.
[02:52] <Laurenceb> WAAS
[02:54] <hallam> tcxo could cost more than your micro
[02:54] <Laurenceb> ha
[02:54] <Laurenceb> a crystal would drift too much
[02:55] <Laurenceb> to be useful past a second or
[02:55] <Laurenceb> ignore that
[02:55] <Laurenceb> most of the drift is temperature
[02:55] <Laurenceb> so it shouldnt actually matter
[02:57] <Laurenceb> http://s2.b3ta.com/host/creative/46933/1232497226/DONCHENEY.gif
[02:57] <hallam> thanks for that
[02:58] <SpeedEvil> I was assuming I only cared about voltage and aging drift
[02:58] <SpeedEvil> and tempco not so much - I planned to calibrate when testing
[03:00] <Laurenceb> yeah
[03:00] <Laurenceb> when you find the dopplers, you accomodate any drift
[03:00] <SpeedEvil> only caring about delta over the board
[03:00] <Laurenceb> so as long as change in temperature over a few seconds is small, you're ok
[03:01] <SpeedEvil> don't want to be too large though - otherwise you increase search space
[03:19] <Laurenceb> avr32 is quite slow
[03:19] <Laurenceb> only 200MHz
[03:19] <Laurenceb> but thats prob fast enoug for this
[03:19] <Laurenceb> its has dma and even mmu
[03:35] <hallam> Laurenceb: how did you make your square wave carrier?
[03:35] <hallam> in your mods to Borres
[03:36] <Laurenceb> yes
[03:36] <Laurenceb> just rounded it
[03:36] <hallam> how?
[03:36] <hallam> carrier=sign(carrier)?
[03:37] <Laurenceb> yes
[03:37] <hallam> I think that's why yours works and mine doesn't
[03:38] <Laurenceb> I thought yours was working
[03:38] <hallam> carrier=sign(carrier) keeps the phase and just makes it magnitude 1
[03:38] <hallam> i.e. it doesn't change it
[03:38] <SpeedEvil> I was wondering about dithered square wave
[03:39] <hallam> you need to use carrier=0.7071*(sign(real(carrier)) + i*sign(imag(carrier)))
[03:39] <hallam> SpeedEvil: yeah, thinking about that too
[03:39] <Laurenceb> oh
[03:39] <SpeedEvil> instead of 0000000011111111 0100010110111001 or something
[03:39] <hallam> just waiting on a tracking run to see if this is the answer to my mysteries
[03:39] <Laurenceb> it doesnt change oit
[03:39] <hallam> SpeedEvil: look up "magic sine waves"
[03:40] <SpeedEvil> that sort of thing, yes.
[03:42] <Laurenceb> you have to be careful
[03:42] <Laurenceb> or you will correlate with the signal
[03:43] <SpeedEvil> yeah
[03:43] <SpeedEvil> that too
[03:44] <hallam> yeah it looks like the square wave carrier is quite punishing
[03:44] <Laurenceb> hmm
[03:44] <Laurenceb> I'll try
[03:47] <SpeedEvil> it'll have maximal signal - but also excess noise - and wierd response as it slips
[03:52] Action: SpeedEvil ponders two bit sequences - one for 512us...
[04:11] <Laurenceb> hallam: I've rerun it with a square wave
[04:12] <Laurenceb> no huge difference
[04:12] <Laurenceb> <1m
[04:12] <Laurenceb> carrier=0.7071*(sign(real(carrier)) + i*sign(imag(carrier)))
[04:12] <Laurenceb> stuck that in
[04:12] <hallam> yeah nm I did something stupid
[04:12] <SpeedEvil> prolly more important in teh presence of multipath
[04:12] <Laurenceb> most of the points are inside a 40m radius ellipse
[04:13] <hallam> http://www.mibbit.com/pb/tvAqnT spot the deliberate error
[04:13] <Laurenceb> +-15m through the ellipse
[04:14] <Laurenceb> on a 45 degree to vertical ellips
[04:16] <SpeedEvil> It'd be interesting to evolve the proper function
[04:17] <SpeedEvil> is a sinewave best - or a distorted sine, or a triangle, or ...
[04:26] <Laurenceb> hmm how do I invert x/(0.25+x^2)
[04:26] <Laurenceb> I'm trying to invert the discriminator to improve the accuracy of pseudorange estimation
[04:40] <Laurenceb> hmm thats a pain to invert
[04:40] <Laurenceb> I have something approx, going to rerun
[04:41] <Laurenceb> hmm frequencies shift by about 2Hz over the entire 40s
[06:36] natrium42 (n=alexei@CPE000625d867e2-CM0014045885be.cpe.net.cable.rogers.com) left irc: Remote closed the connection
[06:50] akawaka (n=akawaka@66-215-97-198.dhcp.malb.ca.charter.com) joined #highaltitude.
[06:53] akawaka (n=akawaka@66-215-97-198.dhcp.malb.ca.charter.com) left irc: Read error: 104 (Connection reset by peer)
[07:15] Simon-MPFH (n=simon@phantom.mpfh.co.uk) joined #highaltitude.
[07:56] G_SQ5FNQ_Marcin (n=G_SQ5FNQ@235-tar-1.acn.waw.pl) joined #highaltitude.
[08:13] Laurenceb (n=laurence@dyres221-140.surrey.ac.uk) left irc: Remote closed the connection
[08:36] borism (n=boris@195-50-211-250-dsl.krw.estpak.ee) joined #highaltitude.
[08:48] hallam (i=836fc8c8@gateway/web/ajax/mibbit.com/x-eced4d68af989fe0) left irc: "http://www.mibbit.com ajax IRC Client"
[08:57] Laurenceb (i=83e34f19@gateway/web/ajax/mibbit.com/x-8ec43c64cce4e257) joined #highaltitude.
[08:57] <Laurenceb> ello
[09:03] <Laurenceb> http://www.sige.com/uploads/briefs/SE4150L_brief_Rev_2p3_Sep-22-2008.pdf
[09:07] edmoore (n=edmoore@pomegranate.chu.cam.ac.uk) joined #highaltitude.
[09:09] <Laurenceb> http://www.electronicstalk.com/news/faa/faa105.html
[09:09] <Laurenceb> love the radiometrix advert at the top
[09:09] <edmoore> yeah :)
[09:09] <Laurenceb> everything I'm working on on one page :P
[09:09] <edmoore> ironic
[09:10] <Laurenceb> looks like it runs under an os
[09:11] <Laurenceb> guess it must use a DMA controller
[09:16] <Laurenceb> ooh active robots carry sparkfun stuff
[09:16] <Laurenceb> overpriced, but in the uk at least
[09:16] <edmoore> right, lectures
[09:16] <edmoore> bbl
[09:17] edmoore (n=edmoore@pomegranate.chu.cam.ac.uk) left irc:
[09:20] josepharmbruster (n=josephar@45.144.119.70.cfl.res.rr.com) joined #highaltitude.
[09:23] <josepharmbruster> anyone around?
[09:24] <Laurenceb> me
[09:24] Action: gordonjcp
[09:24] <josepharmbruster> random question, ever used an XTend serial modem?
[09:24] <Laurenceb> nope
[09:25] <josepharmbruster> talked to someone on here once before who had... forgot who it was :-(
[09:25] Action: gordonjcp has used RDT modems
[09:26] Action: Laurenceb is trying to use a radiometrix module
[09:26] <josepharmbruster> i have an XTend-PKG and one of the portable XT field modems
[09:26] Action: gordonjcp is determined to write some code this morning before firing up the radios
[09:26] <josepharmbruster> but the range is complete crap
[09:27] <Laurenceb> I see
[09:27] <josepharmbruster> just got the units in yesterday... did some testing with it
[09:27] <Laurenceb> gordonjcp: I think my onboar dsound is screwing things upp
[09:27] <josepharmbruster> one guy on here said he pulled 50 mile range or so... i wanted to follow up
[09:27] <Laurenceb> I'm determined to get a usb sound "card"
[09:28] <Laurenceb> really
[09:28] <Laurenceb> we've been using radiometrix modules
[09:28] <josepharmbruster> link?
[09:29] <Laurenceb> http://wiki.ukhas.org.uk/communication:mihab2_radio_system
[09:31] <Laurenceb> I'm working on an avr system
[09:31] <Laurenceb> but its not working, pretty sure its my soundcard
[09:31] <Laurenceb> you need a stable clock
[09:32] <Laurenceb> problem is I literally cant find one - I'm looking for usb dongle with line in
[09:32] <Laurenceb> they all have mic in
[09:49] <Laurenceb> I dont really weant to use an attenuator
[09:50] <Laurenceb> its only going to add noise
[10:36] Xenion (n=robert@p579FC8B1.dip.t-dialin.net) left irc: Read error: 101 (Network is unreachable)
[10:36] Xenion (n=robert@p579FC067.dip.t-dialin.net) joined #highaltitude.
[10:54] josepharmbruster (n=josephar@45.144.119.70.cfl.res.rr.com) left irc: Read error: 113 (No route to host)
[10:54] hallam (i=836fc8c8@gateway/web/ajax/mibbit.com/x-3f672929f7749de0) joined #highaltitude.
[11:17] <Laurenceb> hey hallam
[11:17] <Laurenceb> hallam stand by for awsomeness
[11:17] <hallam> hi
[11:17] Action: Laurenceb awsomeness in 5 minutes
[11:17] Action: hallam stands by
[11:17] <hallam> aw, 5 minutes
[11:17] Action: Laurenceb gets back to ms paint
[11:17] <Laurenceb> to edit graphs
[11:18] <hallam> dude, paint.net
[11:18] <Laurenceb> I have position noise down on the order of 3 meters
[11:18] <hallam> not bad
[11:18] <hallam> how did you manage that?
[11:19] <Laurenceb> around a cm at high frequencies
[11:19] <Laurenceb> muhahahaha
[11:19] <Laurenceb> just a sec
[11:20] <hallam> do you use carrier phase?
[11:20] <Laurenceb> yes
[11:20] <Laurenceb> grr this is complicated
[11:20] <hallam> didn't realise that was useful for non-differential applications
[11:20] <Laurenceb> I should keep notes
[11:20] <Laurenceb> need to read the logs...
[11:24] <hallam> from when?
[11:24] <hallam> just checked todays
[11:26] <Laurenceb> http://imagebin.org/37035
[11:26] <Laurenceb> 6.5m total 3D error
[11:27] <Laurenceb> position is stable to +-2.5m
[11:28] <hallam> what did you filter with what?
[11:28] <Laurenceb> trackResults(channelNr).absoluteSample(loopCnt) =(ftell(fid)-(remCodePhase*16))/dataAdaptCoeff;
[11:28] <Laurenceb> was the key
[11:29] <Laurenceb> carrier aiding helped
[11:29] <Laurenceb> its insanely accurate
[11:30] <Laurenceb> thats got to be the limit
[11:30] <Laurenceb> Borre says about 4m
[11:30] <hallam> this is with shitty satellite geometry and crap signal on one of them
[11:30] <Laurenceb> but thats with a good DOP
[11:30] <Laurenceb> yeah exactly
[11:30] <Laurenceb> but the noise is so low
[11:30] <Laurenceb> its awsome :P
[11:30] <hallam> btw I checked the new file I sent you, couldn't get any sats out of it
[11:30] <hallam> well done!
[11:30] <Laurenceb> :-/
[11:31] <hallam> er sorry, could only get 2
[11:31] <hallam> but still useless
[11:31] <Laurenceb> right
[11:31] <hallam> which of the approximations did you use for this?
[11:31] <Laurenceb> try a park or something
[11:31] <hallam> square wave carrier?
[11:31] <Laurenceb> yes
[11:31] <Laurenceb> 8Msps
[11:32] <Laurenceb> hallam: http://maps.google.co.uk/maps?f=q&source=s_q&hl=en&geocode=&q=52+12%27+7.4943%27%27+N,+0+7%27+5.4632%27%27+E&sll=52.201993,0.118285&sspn=0.000582,0.001207&layer=t&ie=UTF8&ll=52.201825,0.128739&spn=0.00028,0.000603&t=h&z=21
[11:32] <Laurenceb> try there :P
[11:32] [1]G_SQ5FNQ_Marc (n=G_SQ5FNQ@235-tar-1.acn.waw.pl) joined #highaltitude.
[11:32] <hallam> yeah I was going to
[11:32] <hallam> not right this minute though
[11:32] <Laurenceb> open space with very well defined landmark
[11:33] <Laurenceb> :P
[11:33] <hallam> Laurenceb: can you try running it with decimation on?
[11:33] <Laurenceb> how do you mean?
[11:33] <Laurenceb> $Msps ?
[11:33] <Laurenceb> $
[11:33] <Laurenceb> 4
[11:33] <SpeedEvil> And a survey grade GPAS too
[11:34] <SpeedEvil> Maybe irc.oftc.net - #osm - mAY BE abble to help with a good GPS
[11:34] <SpeedEvil> I think there are some people around cambridge
[11:34] <hallam> Laurenceb: set every other sample to 0
[11:34] <Laurenceb> got you
[11:34] <Laurenceb> yeah...
[11:34] <hallam> in your tracking function or in whatever loads it from the file or wherever
[11:36] <Laurenceb> its interesting the position form an arc in 3D space
[11:36] <Laurenceb> over about 20m path lenght
[11:38] <Laurenceb> the botop of the arc must only be 1.5m or so away friom the true pos
[11:38] <hallam> yeah I've noticed that
[11:38] <SpeedEvil> What is carrier aiding - /me hasn't read the book or code. I assume not arrestor wires
[11:38] <SpeedEvil> And I would suapect 8m is a bit low for a 3rd floor window unless it's floor-ceiling
[11:38] <hallam> and it's swept over time
[11:38] <Laurenceb> but the average ends up some way away
[11:39] <hallam> SpeedEvil: you use the doppler from the PLL to aid the DLL tracking loop, by estimating (pretty accurately) the code phase rate
[11:39] <hallam> in fact for one run I accidentally set the DLL gain to zero and it still tracked because of the carrier aiding
[11:40] <SpeedEvil> ah
[11:40] <Laurenceb> http://imagebin.org/37036
[11:42] <hallam> Laurenceb: are you filtering the nav solutions or are these raw ms solutions?
[11:42] <Laurenceb> raw
[11:43] <Laurenceb> it reminds me of what I got out of a ublox4
[11:43] <Laurenceb> almost identical behaviour
[11:43] <SpeedEvil> the KHz sample-sample RMS noise is what?
[11:43] <SpeedEvil> I mean psuedorange
[11:43] <Laurenceb> ~40cm
[11:43] <Laurenceb> oh
[11:43] <Laurenceb> 6m
[11:44] <SpeedEvil> It would be interesting to go one antenna -> splitter -> ublox, garmin, survey GPS, ...
[11:44] <Laurenceb> yeah we have such a rig here
[11:44] <SpeedEvil> oh - right that pic was smoothed then
[11:44] <SpeedEvil> the previous previous imaGEBIN
[11:45] <Laurenceb> hmm maybe I want to modify the file...
[11:45] <Laurenceb> SpeedEvil: no
[11:45] <Laurenceb> no smoothing
[11:45] <Laurenceb> I dont quite understand why the noise is so low
[11:45] <Laurenceb> maybe its mostly comping from TCXO jitter
[11:45] <SpeedEvil> I have trouble with that myself.
[11:45] <Laurenceb> and cancles a lot
[11:46] <Laurenceb> but yeah theres no filtering
[11:46] <Laurenceb> its just a dequantisation step
[11:46] <SpeedEvil> well - that's eassy to test
[11:46] <SpeedEvil> there should be a straight correlation between errors of different sats
[11:46] <Laurenceb> based on the DLL filter to get floating point pseudoranges
[11:47] <SpeedEvil> if it's timing jitter that is
[11:47] <Laurenceb> yeah
[11:47] <Laurenceb> sorry I'm too lazy to add that now
[11:47] <Laurenceb> but I'm guessing its something like that
[11:49] <SpeedEvil> would be nice to nail down all the sources of (in) accuracy
[11:50] <Laurenceb> how do I read in an entire data file in matlab?
[11:50] <Laurenceb> hang on... bad idea
[11:50] <Laurenceb> ok, how do I detect eof ?
[11:50] G_SQ5FNQ_Marcin (n=G_SQ5FNQ@235-tar-1.acn.waw.pl) left irc: Read error: 110 (Connection timed out)
[11:50] Nick change: [1]G_SQ5FNQ_Marc -> G_SQ5FNQ_Marcin
[11:51] <SpeedEvil> feof()
[11:52] <SpeedEvil> you also probably want to read 8K or so at a time for better read performance
[11:52] <SpeedEvil> oh
[11:52] <SpeedEvil> ignore the above
[11:54] <Laurenceb> hallam: about?
[11:54] <hallam> yeah
[11:54] <Laurenceb> howe do I reduce the data down in matlab?
[11:55] <Laurenceb> I'm going to create a new file
[11:55] <Laurenceb> need to loop until I reach eof
[11:55] <hallam> feof
[11:55] <Laurenceb> while ..
[11:55] <Laurenceb> arg
[11:55] <hallam> isn't it a big hassle to change all your code to operate at 4MS/s?
[11:56] <hallam> while ~feof(fp)
[11:56] <Laurenceb> Im a bit clueless aboput file handling in matlab
[11:56] <Laurenceb> ok
[11:56] <Laurenceb> thanks
[11:56] <hallam> I'd have thought it would be much easier just to change the code that reads the file to set every other sample to be 0
[11:56] <Laurenceb> hmm
[11:57] <Laurenceb> I can create a new file, then edit the config file
[11:57] <Laurenceb> to 4Msps sampling
[11:57] <Laurenceb> then theres no editing required
[11:57] <hallam> sure
[11:57] <hallam> enjoy
[11:57] <hallam> it'll be faster to run too, probably
[12:01] <SpeedEvil> I wonder how much you can decimate it
[12:02] <SpeedEvil> Shoyuldn't more decimation simply lower S/N?
[12:02] <SpeedEvil> Even if below nyquist.
[12:03] <Laurenceb> yeah
[12:03] <Laurenceb> theres a filter on I and Q
[12:03] <Laurenceb> so the bandwidth of the background noise is limited
[12:04] <Laurenceb> if we sample more, we should increase the signal to noise
[12:04] <Laurenceb> if there were no filters, we'd pick up more noise at the same time...
[12:04] <Laurenceb> I think that how it works
[12:05] <Laurenceb> hmm no
[12:05] <Laurenceb> the noise would be mapped back by nyquist
[12:05] <Laurenceb> confusing
[12:06] <Laurenceb> hmm its got to help
[12:06] <Laurenceb> well our prn and carrier acts as a selective filter
[12:07] <Laurenceb> so if we have more datapoints we can filter out more signal
[12:07] <Laurenceb> but we wont filter out an equivalent amplitude increase of noise
[12:08] Simon-MPFH (n=simon@phantom.mpfh.co.uk) left irc: Read error: 54 (Connection reset by peer)
[12:08] Simon-MPFH (n=simon@phantom.mpfh.co.uk) joined #highaltitude.
[12:10] <Laurenceb> yeah your signal amplitude will rise with the sample rate
[12:12] <Laurenceb> hmm I think the filter formed by your prn and local osc becomes more selective as you increase sample rate
[12:12] <Laurenceb> as a side effect, a square wave local osc allows through noise, due to the harmonics
[12:14] rjharrion_quiet (n=rharriso@gateway.hgf.com) joined #highaltitude.
[12:17] Simon-MPFH (n=simon@phantom.mpfh.co.uk) left irc: Read error: 131 (Connection reset by peer)
[12:17] simon__ (n=simon@phantom.mpfh.co.uk) joined #highaltitude.
[12:19] [1]G_SQ5FNQ_Marc (n=G_SQ5FNQ@235-tar-1.acn.waw.pl) joined #highaltitude.
[12:22] <Laurenceb> http://www.mibbit.com/pb/wyyfVL
[12:22] <Laurenceb> hallam: ^ is that code ok?
[12:24] edmoore (n=edmoore@pomegranate.chu.cam.ac.uk) joined #highaltitude.
[12:24] <hallam> sure, should be fine
[12:24] <hallam> it might be pretty slow, reading in more bytes at a time might be faster
[12:24] <hallam> but try and see before bothering to rewrite it
[12:25] <Laurenceb> yes
[12:25] <Laurenceb> as I'm discovering
[12:26] <Laurenceb> http://hackadaycom.files.wordpress.com/2009/01/handsfree.jpg?w=450&h=338
[12:27] <hallam> yeah I saw that, stylish
[12:27] <edmoore> hallam: cool
[12:28] <edmoore> just going to dive into the shower
[12:28] <Laurenceb> edmoore http://imagebin.org/37035
[12:29] <hallam> edmoore: be careful, you might hit your head
[12:30] <Laurenceb> http://hackaday.com/2009/01/24/road-sign-hacking/
[12:31] <hallam> been there done that
[12:31] <hallam> it's not exactly difficult when the box is unlocked, the password is abcd and written on it in permanent marker
[12:32] Action: hallam has intimate knowledge of http://hacks.mit.edu/Hacks/by_year/2008/sign_factory/
[12:32] <Laurenceb> wtf stupid firewall
[12:32] <Laurenceb> that url is blocked
[12:33] Action: Laurenceb gets high gain wifi antenna
[12:33] <hallam> hah
[12:35] <Laurenceb> right lunch time
[12:35] <Laurenceb> bbl
[12:36] G_SQ5FNQ_Marcin (n=G_SQ5FNQ@235-tar-1.acn.waw.pl) left irc: Read error: 110 (Connection timed out)
[12:36] Nick change: [1]G_SQ5FNQ_Marc -> G_SQ5FNQ_Marcin
[12:36] G_SQ5FNQ_Marcin (n=G_SQ5FNQ@235-tar-1.acn.waw.pl) left irc: " HydraIRC -> http://www.hydrairc.com <- Organize your IRC"
[12:54] simon__ (n=simon@phantom.mpfh.co.uk) left irc: Read error: 104 (Connection reset by peer)
[12:55] Simon-MPFH (n=simon@phantom.mpfh.co.uk) joined #highaltitude.
[12:55] <Laurenceb> hallam: dont they use ip ?
[12:55] <Laurenceb> the signs
[12:55] <hallam> ip?
[12:55] <Laurenceb> internet
[12:56] <hallam> not portable ones
[12:56] <Laurenceb> ah ok
[13:09] <SpeedEvil> Laurenceb: I was actually wondering about further tradeoffs between noise and CPU usage
[13:10] <SpeedEvil> If you've got 'plenty' of signal, morwe than you require to hit the accuracy budget, decimate harder, removing more samples.
[13:11] borism (n=boris@195-50-211-250-dsl.krw.estpak.ee) left irc: Read error: 145 (Connection timed out)
[13:15] <Laurenceb> yeah
[13:24] borism (n=boris@195-50-211-250-dsl.krw.estpak.ee) joined #highaltitude.
[13:27] <Laurenceb> cant find any sats at 4Msps
[13:30] Simon-MPFH (n=simon@phantom.mpfh.co.uk) left irc: Read error: 104 (Connection reset by peer)
[13:31] Simon-MPFH (n=simon@phantom.mpfh.co.uk) joined #highaltitude.
[13:34] <SpeedEvil> is this rewritten, or externally decmiated?
[13:35] <Laurenceb> externally decimated
[13:35] <Laurenceb> I'm not quite sure it was decimated right
[13:35] <Laurenceb> I should just be able to reconfigure setting.samplingFreq
[13:37] <Laurenceb> hmm it doesnt work grr
[13:37] <Laurenceb> this should be easy
[13:40] <hallam> Laurenceb: you could actually externally decimate it rather than downsampling it
[13:40] <hallam> then the settings would really be unchanged
[13:40] <SpeedEvil> that was what I meant
[13:40] <SpeedEvil> simply puncture it
[13:40] <hallam> set = 0
[13:40] <Laurenceb> yeah
[13:41] <Laurenceb> I'll just check this first
[13:42] <SpeedEvil> http://img.4chan.org/b/res/113144998.html#113144998 sort-of-related.
[13:42] <Laurenceb> its so annoying working with code you didnt write yourself
[13:42] <Laurenceb> erm... 4chan
[13:42] <SpeedEvil> (not work-safe - just the wikipedia page on kalman filters popped up)
[13:42] <Laurenceb> why didnt you say that first
[13:42] <Laurenceb> what is it then?
[13:43] <SpeedEvil> It's a purported list of pages you go to from a random wikipedia page to a specific page.
[13:43] <hallam> that's quite funny
[13:43] <hallam> takes a special kind of genius
[13:43] <Laurenceb> NSFW?
[13:44] <SpeedEvil> 7 links from 'kalman filter' to the nsfw target page
[13:44] <SpeedEvil> yes, very nsfw
[13:45] <SpeedEvil> well - not very for 4chan
[13:46] <hallam> no pictures
[13:46] <hallam> yeah I'm not as good as those guys, took me 8 links from a random page
[13:46] <SpeedEvil> The key is knowing hte likely back-chains that are popular
[13:47] <hallam> protip for matlab: do_some_stuff_that_takes_ages;beep;
[13:48] <SpeedEvil> hallam: I can do it in 1! /me edits.
[13:50] <Laurenceb> this is really confusing
[13:50] <hallam> SpeedEvil: the subtler misanthrope changes key digits in physical constants on wikipedia pages
[13:50] <Laurenceb> it refuises to edit the entire file
[13:52] <SpeedEvil> My only vandalism was relatively mild. http://en.wikipedia.org/w/index.php?title=Speed_record&oldid=247654617
[13:52] <Laurenceb> http://www.mibbit.com/pb/mqeZB6
[13:53] <Laurenceb> oh do I need to augment the file?
[13:53] <hallam> wtf
[13:53] <hallam> what's that code supposed to do?
[13:54] <Laurenceb> stick zeros in
[13:54] <hallam> you don't get any speed advantage by reading in large chunks if you then proceed to iterate through the whole chunk sample by sample
[13:54] <Laurenceb> yes its a bodge
[13:54] <Laurenceb> hmf
[13:55] <Laurenceb> I can never work out the weird matlab tricks
[13:55] <Laurenceb> how do I do it with vectors?
[13:55] <hallam> data(3:4:n)=0;
[13:55] <hallam> data(4:4:n)=0;
[13:55] <hallam> replaces that whole loop
[13:55] <Laurenceb> I see
[13:55] Action: Laurenceb sucks at matlab
[13:56] <hallam> now if you were cunning you'd read it as int16 rather than int8
[13:56] <Laurenceb> ah yes
[14:00] <Laurenceb> ontput file is only half the size...
[14:01] <Laurenceb> I'm really failing at this
[14:06] <hallam> that's why I suggested putting it in the read routine, you can hardly go wrong with that
[14:06] <Laurenceb> it gets a bit complex as its reading blocks
[14:07] <Laurenceb> http://www.mibbit.com/pb/HvLglc
[14:08] <hallam> should be fine
[14:08] <Laurenceb> it halves the datasize
[14:11] <Laurenceb> its like theres a limit to file write size
[14:15] <SpeedEvil> main(){FILE *fp;int n;unsigned char buf[8193];fp=fopen("file","r");out=fopen("out","w");if(fp)while(!feof(fp)){fread(fp,8192,buf);for(n=0;n<4096;n++)buf[n*2]=0x00;fwrite(out,8192,buf)}}
[14:15] <SpeedEvil> Or something
[14:15] Simon-MPFH (n=simon@phantom.mpfh.co.uk) left irc: Read error: 54 (Connection reset by peer)
[14:15] <Laurenceb> yeah maybe I should just use c
[14:15] <Laurenceb> matlab isnt liking this
[14:16] <Laurenceb> I tried opening and closing the file
[14:16] <Laurenceb> but same deal
[14:16] simon__ (n=simon@phantom.mpfh.co.uk) joined #highaltitude.
[14:17] Simon-MPFH (n=simon@phantom.mpfh.co.uk) joined #highaltitude.
[14:17] <SpeedEvil> Though I probably got the fread args wrong - I always have to look up the order
[14:17] <hallam> er, works fine for me, I think
[14:18] <hallam> the code you pastebinned 10 mins ago
[14:18] simon__ (n=simon@phantom.mpfh.co.uk) left irc: Read error: 131 (Connection reset by peer)
[14:18] <Laurenceb> the output ias 64MB ?
[14:18] <hallam> 640
[14:19] <hallam> and alternate samples are zeroed
[14:19] <Laurenceb> it sticks at 320 here
[14:19] <SpeedEvil> stupid question. You're not out of disk?
[14:19] <SpeedEvil> thouhg exactly 320 is suspictions
[14:20] <Laurenceb> hmm I'll check
[14:20] <hallam> fclose all
[14:20] <hallam> may help you
[14:20] <hallam> or change the out filename
[14:21] <Laurenceb> oh hang on
[14:21] <Laurenceb> I have a gig on here
[14:21] <Laurenceb> hmm think I'm out oif disk
[14:21] <Laurenceb> dam,n
[14:22] <Laurenceb> have toi mover everything onto my machine
[14:22] <Laurenceb> this will take a while
[14:22] <Laurenceb> arg should have zipped first
[14:22] <hallam> you could always decimate it as you read it
[14:22] <hallam> just sayin'
[14:23] <Laurenceb> wow this is fast
[14:23] <Laurenceb> about 100Mbps
[14:23] <Laurenceb> must be a fibre line across to the campus
[14:24] Action: SpeedEvil sighs.
[14:25] Action: SpeedEvil finds that his boot disk has been offline since 5AM.
[14:25] Action: SpeedEvil ponders trying to fix it live.
[14:28] <Laurenceb> http://www.kenable.co.uk/product_info.php?referer=froogle&products_id=749
[14:28] <Laurenceb> think I'll get that for my radio
[14:28] <Laurenceb> hopefully has a stable clock
[14:28] speedevi1 (n=user@mauve.plus.com) joined #highaltitude.
[14:28] <speedevi1> /server irc.freenode.com
[14:29] <hallam> aw man
[14:29] <hallam> arrgh
[14:30] <hallam> 24 hours of debugging
[14:30] <hallam> and it's operator precedence
[14:30] <Laurenceb> :-/
[14:30] <speedevi1> :/
[14:30] <hallam> not general relativity
[14:30] <Laurenceb> oh the position drift
[14:30] <hallam> dll_disc=-corr_eml_i*(corr_prompt_i > 0)?1:-1;
[14:30] Action: speedevi1 hands hallam ()
[14:30] <hallam> is not the same as
[14:30] <hallam> dll_disc=-corr_eml_i*((corr_prompt_i > 0)?1:-1);
[14:31] <Laurenceb> ah
[14:31] <Laurenceb> yeah those things are annoying
[14:31] <Laurenceb> I've been stuck for ages on similar
[14:31] <Laurenceb> any idea if that usb thing is trying to say it uses a pll for the adc?
[14:32] SpeedEvil (n=jfkjdjkf@mauve.plus.com) left irc: "[BX] Become a BitchX Certified Systems Engineer today! Apply within!"
[14:32] <Laurenceb> "Single 12MHz crystal input with on chip PLL"
[14:32] Nick change: speedevi1 -> SpeedEvil
[14:32] Simon-MPFH (n=simon@phantom.mpfh.co.uk) left irc: Read error: 131 (Connection reset by peer)
[14:32] <Laurenceb> hmm prn32 has gawn
[14:33] Simon-MPFH (n=simon@phantom.mpfh.co.uk) joined #highaltitude.
[14:35] <SpeedEvil> USB2 needs 48MHz cock
[14:35] <SpeedEvil> so maybe it means that
[14:35] <SpeedEvil> (typically)
[14:36] <Laurenceb> how would it sample?
[14:36] <Laurenceb> the adc?
[14:36] <SpeedEvil> Sorry - what usb thing
[14:36] <Laurenceb> http://www.kenable.co.uk/product_info.php?referer=froogle&products_id=749
[14:37] <SpeedEvil> Ah
[14:37] <Laurenceb> hallam: was 19 visable?
[14:37] <SpeedEvil> Gotta love the 'copy the chip blurb and stick it on the box' style of docs
[14:37] <Laurenceb> in that dataset
[14:37] <Laurenceb> hehe
[14:38] <Laurenceb> I need a stable time reference for radio use
[14:38] Simon-MPFH (n=simon@phantom.mpfh.co.uk) left irc: Read error: 113 (No route to host)
[14:38] Action: SpeedEvil sees if the cheapish sound card he got from ebuyer is still tehre
[14:39] <hallam> Laurenceb: in the original dataset, or in the new one I sent?
[14:39] <Laurenceb> origional
[14:39] <Laurenceb> I've got smaller peaks for 32 and 19
[14:39] <hallam> not to my knowledge
[14:39] <Laurenceb> about equal for each
[14:39] <Laurenceb> hmm ok
[14:39] <hallam> I'll try and detect 19
[14:39] <hallam> would be cool if it's there
[14:39] Simon-MPFH (n=simon@phantom.mpfh.co.uk) joined #highaltitude.
[14:40] <hallam> nah 19's not even above the horizon
[14:41] <hallam> 11, 13, 23, 32 are definitely there
[14:41] <Laurenceb> ok, I've blocked it
[14:41] <Laurenceb> 19 is very slightly higher than 32 now
[14:42] <hallam> some hope of 17 and 20 although I don't think I ever successfully acquired either, let alone tracked them
[14:42] <Laurenceb> me neither
[14:42] <hallam> I think I had glimmers of 20 but it was lost in the noise
[14:42] <Laurenceb> interesting frequency mirroring
[14:42] <hallam> oh?
[14:42] <Laurenceb> due to the zeros
[14:42] <hallam> really
[14:42] <Laurenceb> yeah, view the entire sample in freq space
[14:43] <Laurenceb> but it shouldnt make an difference to carrier tracking
[14:43] <Laurenceb> - the zeros mean you pick up high frequencies in the fft
[14:44] <Laurenceb> it seems to track 11 and 23 ok
[14:49] Simon-MPFH (n=simon@phantom.mpfh.co.uk) left irc: Read error: 104 (Connection reset by peer)
[14:49] Simon-MPFH (n=simon@phantom.mpfh.co.uk) joined #highaltitude.
[14:50] <Laurenceb> http://news.bbc.co.uk/1/hi/technology/7858498.stm
[14:50] <Laurenceb> what the hell
[14:51] <Laurenceb> I just understand the government
[14:51] <Laurenceb> either nationalise it or leave it to the private sector
[14:51] <Laurenceb> they seem to want to control private sector industries like they are part of government
[14:56] <Laurenceb> hallam: I dont think substituting in zeros is quite the same thing
[14:57] <Laurenceb> as the prn is still free to shift by one 8Msps sample
[14:59] <hallam> I did think about that
[14:59] <hallam> but it's close enough
[14:59] <hallam> oh man I'm so pleased the C tracker gives good results now
[14:59] <Laurenceb> yopu have the c working on a compressed smaple right?
[14:59] <hallam> yeah
[14:59] <Laurenceb> cool
[15:00] <Laurenceb> i.e. actual 4Msps
[15:00] <hallam> start to position fix in 22 seconds, of which 15 are spent in acquisition (trying all 32 sats)
[15:00] <hallam> yes
[15:00] <Laurenceb> nice
[15:00] <hallam> the tracking takes less than 2 seconds for all channels
[15:00] <Laurenceb> what does the fix look like?
[15:00] <hallam> apparently 60cm mean error
[15:00] <hallam> but I think that's just luck
[15:00] <Laurenceb> hehe
[15:01] <hallam> std deviation is 7 meters
[15:01] <Laurenceb> are they clustered in a cicle or what?
[15:01] <hallam> ellipsoid
[15:01] <Laurenceb> interesting
[15:01] <Laurenceb> sounds like the pseudorange dequantisation isnt quite on
[15:01] <Laurenceb> thats what I was getting before
[15:02] <Laurenceb> but not quite that small
[15:02] <Laurenceb> suggesting you've only got a very small problem
[15:03] <hallam> yes I think if I make the DLL 2nd order it will get better
[15:03] <hallam> now that I have the discriminator right
[15:03] <Laurenceb> I just used first order
[15:03] <hallam> does the output of your DLL adjust code phase or code phase rate?
[15:03] <Laurenceb> oh phase rate
[15:04] <hallam> right, mine goes straight for the phase
[15:04] <Laurenceb> it sets phase rate using doppler
[15:04] <hallam> except carrier aiding gives the rate
[15:04] <Laurenceb> then corrects using the discriminator
[15:04] <hallam> ok
[15:04] <Laurenceb> I use the phase + file position to give very accurate pseudorange
[15:10] <Laurenceb> ok, got a position
[15:10] <Laurenceb> quite interesting
[15:10] <Laurenceb> starts 25m off
[15:10] <Laurenceb> then moves in over about 1 second
[15:11] <Laurenceb> then moves around in an area about 4m square
[15:11] <Laurenceb> it'll be the dll tuning in
[15:11] <Laurenceb> which takes about a second
[15:12] <Laurenceb> for some reason the response time is shorter now
[15:12] <Laurenceb> so I'll try and find the position
[15:14] <Laurenceb> actual noise is of thge order 20cm
[15:15] <Laurenceb> the wobbles are probably ionospheric
[15:15] <Laurenceb> of the order 50cm
[15:16] <Laurenceb> it oscillates a bit around the target, but I think its stabilising
[15:16] <Laurenceb> gived about a minute you should have a better result
[15:16] <Laurenceb> this is amazingly good
[15:19] <Laurenceb> about 2m vertical error
[15:23] <Laurenceb> hmm about 7m horizontal
[15:23] <Laurenceb> obviously some ionospheric
[15:23] <Laurenceb> unless google maps is wrong...
[15:23] <Laurenceb> hallam ?
[15:24] <hallam> ionospheric will dominate I think, I don't think Borres applies even the correction that's included in the nav message
[15:24] Action: hallam is trying to get a digital filter to work
[15:26] <Laurenceb> http://imagebin.org/37041
[15:34] <Laurenceb> http://www.mibbit.com/pb/GPZBcF
[15:34] <Laurenceb> from the ephemeris code
[15:34] <Laurenceb> looks like its not used
[15:34] <Laurenceb> that would explain things
[15:35] <Laurenceb> wow wonder how accurate we can get with ionospheric
[15:40] Nick change: edmoore -> edmoore|away
[15:43] <hallam> what's h, and why does it keep failing to approximate it to the desired precision?
[15:48] <Laurenceb> height
[15:54] <rjharrion_quiet> Anyone listening out for the ISS tomorrow at 13:30
[15:59] <Laurenceb> hallam: if you want a really good book on gps
[15:59] <Laurenceb> pratap misra and per enge
[16:00] <Laurenceb> "global positioning system, signals, measurement and performance"
[16:02] <Laurenceb> hallam: ionospheric excess range should have been eround 10m
[16:02] <Laurenceb> so its very significant
[16:05] <hallam> rjharrion_quiet: am I likely to pick it up on my DC-daylight job?
[16:05] <hallam> Laurenceb: yes, I get the impression that some of the very good average errors we've been getting are luck
[16:06] <hallam> Laurenceb: FYI, newer c code up at http://www.srcf.ucam.org/~cuspaceflight/websvn/listing.php?repname=CUSF&path=%2FHenryGNSS%2F
[16:06] <hallam> if, like me, you prefer your tracking to finish in 2 seconds rather than 20 minutes
[16:07] <Laurenceb> :P
[16:07] <hallam> right bedtime for me
[16:07] <hallam> see you later
[16:07] <Laurenceb> hehe
[16:07] <Laurenceb> cya
[16:07] hallam (i=836fc8c8@gateway/web/ajax/mibbit.com/x-3f672929f7749de0) left irc: "http://www.mibbit.com ajax IRC Client"
[16:15] <Laurenceb> nighttime for hallam
[16:16] <rjharrion_quiet> http://www.csbf.nasa.gov/antarctica/ice0809.htm
[16:16] <rjharrion_quiet> Is anyone following this
[16:17] <rjharrion_quiet> Apparently it's still up in the sky.
[16:17] <Laurenceb> http://news.bbc.co.uk/1/hi/magazine/7847351.stm
[16:17] <Laurenceb> yeah
[16:17] <Laurenceb> ^ is that an autonomous parafoil?
[16:39] <SpeedEvil> I know the military were testing those, yes.
[16:40] <SpeedEvil> There is an easy way to find out though.
[16:40] Action: SpeedEvil hands Laurenceb an AK47 and a speedboat.
[16:40] <Laurenceb> haha
[16:40] <Laurenceb> wow this book is amazing
[16:40] <SpeedEvil> book?
[16:40] <Laurenceb> entire chapter on gps aided IMUs
[16:41] <Laurenceb> http://www.amazon.com/Global-Positioning-System-Measurements-Performance/dp/0970954409
[16:43] Action: SpeedEvil adds to wishlist.
[16:43] <SpeedEvil> I plan on doing my GPS without more books.
[16:44] <SpeedEvil> Your next task.
[16:44] <SpeedEvil> Write a book.
[16:49] <SpeedEvil> Okaaay. morphic resonance on Material World.
[16:56] Laurenceb (i=83e34f19@gateway/web/ajax/mibbit.com/x-8ec43c64cce4e257) left irc: "http://www.mibbit.com ajax IRC Client"
[16:59] <SpeedEvil> Next weeks Material World - high altitude balloon telescopes
[17:08] natrium42 (i=natrium4@CPE000625d867e2-CM0014045885be.cpe.net.cable.rogers.com) joined #highaltitude.
[17:09] Hiena (n=Hiena@81.93.195.181.datatrans.hu) joined #highaltitude.
[17:11] <rjharrion_quiet> Hi natrium42
[17:12] <rjharrion_quiet> How's it going over there?
[17:12] <natrium42> hey
[17:12] <natrium42> pretty good
[17:13] <natrium42> what about you?
[17:15] jcoxon (n=mirggi@212.183.134.65) joined #highaltitude.
[17:15] <jcoxon> Hey all
[17:15] <natrium42> hi jcoxon
[17:15] <jcoxon> Anyone launching this weekend?
[17:16] <natrium42> YOU
[17:16] <jcoxon> Hey natrium42
[17:17] <natrium42> :D
[17:17] <jcoxon> Not me!
[17:17] <natrium42> bah
[17:19] <jcoxon> Hehe, soon
[17:21] Laurenceb (i=83e34f19@gateway/web/ajax/mibbit.com/x-3af4c0032c010315) joined #highaltitude.
[17:21] <Laurenceb> http://www.navcen.uscg.gov/pubs/gps/sigspec/gpssps1.pdf
[17:24] jcoxon (n=mirggi@212.183.134.65) left irc:
[17:38] <rjharrion_quiet> hi natrium42 sorry got called away
[17:38] <natrium42> no prob
[17:38] <natrium42> :)
[17:38] <rjharrion_quiet> Looking to launch w/e after next
[17:38] <natrium42> cool, what's the mission?
[17:38] <rjharrion_quiet> Though the weather looks great
[17:39] <rjharrion_quiet> First test of the competition setup for two way coms
[17:39] <rjharrion_quiet> Then to test the listener and that my code fixes work
[17:40] <rjharrion_quiet> Will have both internal and external sensors
[17:40] <rjharrion_quiet> temp ones
[17:41] <natrium42> nice
[17:41] <rjharrion_quiet> Any know what pressure level one should look at on wundergorund for winds
[17:41] <rjharrion_quiet> I'm looking at 300mb ATM
[17:45] <rjharrion_quiet> Is CUSF doing anything this w/e on the launch fromt given the nice weather?
[17:45] Nick change: rjharrion_quiet -> rjharrison
[17:45] <rjharrison> ls
[17:46] <natrium42> they better
[17:47] <rjharrison> I hope so
[17:47] <rjharrison> But I doubt they will
[17:47] <rjharrison> I'm ready for a bit of tracking
[17:47] <rjharrison> natrium42
[17:47] Laurenceb (i=83e34f19@gateway/web/ajax/mibbit.com/x-3af4c0032c010315) left irc: "http://www.mibbit.com ajax IRC Client"
[17:48] <rjharrison> natrium42: You had any chance to have a look at the tracker?
[17:48] <natrium42> no :(
[17:48] <natrium42> but i did send the code to a friend who works at google
[17:49] <natrium42> he might accelerate the map display
[17:49] <rjharrison> Hehe cool
[17:49] <rjharrison> That's a good contact to have esp. if they are in the coding department
[17:55] <natrium42> Hello, the CSC is hosting a talk entitled "The Free Software Movement and the GNU/Linux Operating System" given by Richard M. Stallman. It will now occur at 6:30 PM in the Modern Languages Theater.
[17:55] <natrium42> :(
[17:59] rjharrison (n=rharriso@gateway.hgf.com) left irc:
[18:37] borism_ (n=boris@195-50-211-166-dsl.krw.estpak.ee) joined #highaltitude.
[18:42] borism (n=boris@195-50-211-250-dsl.krw.estpak.ee) left irc: Read error: 145 (Connection timed out)
[19:03] Bluenarf (n=Paul@apollo.paulsnet.org) joined #highaltitude.
[19:03] EI5GTB (n=Paul@apollo.paulsnet.org) left irc: Nick collision from services.
[19:03] Nick change: Bluenarf -> EI5GTB
[19:15] Laurenceb (n=laurence@dyres221-140.surrey.ac.uk) joined #highaltitude.
[19:15] <Laurenceb> hello
[19:39] Simon-MPFH (n=simon@phantom.mpfh.co.uk) left irc: Read error: 104 (Connection reset by peer)
[19:39] Simon-MPFH (n=simon@phantom.mpfh.co.uk) joined #highaltitude.
[19:41] <edmoore|away> natrium42: what's wrong with that?
[19:43] <natrium42> moved to 6:30pm
[19:43] <natrium42> but i can still make it :P
[19:43] <edmoore|away> have you been to one of his talks before?
[19:43] <natrium42> no, that's the first one
[19:43] <natrium42> you?
[19:43] <edmoore|away> yes
[19:44] <edmoore|away> he will start off saying 'Hello Hackers'
[19:44] <natrium42> haha
[19:44] <edmoore|away> and go on a totally incomprehensible rant
[19:44] <edmoore|away> but
[19:44] <natrium42> so vi or emacs??
[19:44] <edmoore|away> i issue you a challenge
[19:44] <edmoore|away> if he starts trying to draw a diagram, see if you can make sense of it
[19:44] <natrium42> haha
[19:45] <edmoore|away> because he illustrated one of his points with a diagram at the talk i want to, and it was about as unilluminating as a diagram can possibley be
[19:45] <edmoore|away> something about media distribution vs time
[19:45] <edmoore|away> be he sort of switched the axes half way through
[19:45] <natrium42> i think it's going to be fun
[19:46] <natrium42> listening to oninionated mostly-nutcases :)
[19:46] <edmoore|away> and then just started tracing lines for no reason at all. It was quite incredible how little information he encoded in that diagram. It was like maximum entropy on a sheet
[19:46] <natrium42> *opinionated that is
[19:46] <natrium42> XD
[19:46] <edmoore|away> oh and he had bare feet
[19:46] <edmoore|away> and an OLPC with an external keyboard from about 28BC
[19:46] <edmoore|away> I'm glad I went to it though :)
[19:47] Nick change: edmoore|away -> edmoore
[19:47] <natrium42> yeah, i heard about the bare feet
[19:47] <natrium42> isn't that what hippies do? :P
[19:47] <edmoore> cool 1960s unix hippies
[19:47] <edmoore> the best kind
[19:48] <natrium42> bbl food
[19:49] <Laurenceb> natrium42: you have a visit from stallman?
[19:56] rjharrison (n=rharriso@80.176.172.227) joined #highaltitude.
[19:57] <edmoore> Laurenceb: he went for food
[19:57] <edmoore> but yes, Stallman talk for him tonight
[19:57] <edmoore> not a personal visit though I assume
[19:57] <edmoore> I can't help imagining he'd be quite a fussy guest
[19:57] <rjharrison> edmoore: Is CUSF taking adv. of the weather this w/e
[19:57] <edmoore> pethaps
[19:58] <edmoore> you should know by now that no decision gets made more than 48hrs in advance :p
[19:58] <rjharrison> BTW on wunderground is it 200mb or 300mb where I want to model?
[19:58] <edmoore> both
[19:58] <rjharrison> True
[19:58] <rjharrison> But that's about it too
[19:58] <edmoore> the jetstream is somewhere between the two
[19:58] <edmoore> mainly at 300mB
[19:58] <rjharrison> ok
[19:58] <edmoore> if i could only do one, it'd be that
[19:58] <rjharrison> Now I understand what pink is
[19:58] <edmoore> pink = smiliey people
[19:59] Action: Laurenceb is trying to work out how to do ionospheric correction
[19:59] <rjharrison> Child needs computer for a sec
[20:00] <Laurenceb> edmoore: did you see my traces?
[20:00] <edmoore> show me your traces
[20:01] <Laurenceb> http://imagebin.org/37041
[20:04] <edmoore> nice
[20:04] <Laurenceb> I'm really impressed with the low noise
[20:04] <Laurenceb> the position it end up at is about 6m off
[20:04] <Laurenceb> it need ionospheric correction
[20:05] <Laurenceb> it moves in towards the fix position over a few seconds as the dll converges
[20:05] <edmoore> how do you do ionospheric correction?
[20:05] <edmoore> how do you model the interference?
[20:05] <Laurenceb> well its just an extra path lenght
[20:06] <Laurenceb> so you have 6 coefficents from the ephemeris
[20:06] <Laurenceb> that model a clipped cos function
[20:06] <Laurenceb> using a polynomial expansion
[20:06] <Laurenceb> this gives you vertical electron count as a function of time
[20:06] <Laurenceb> over a 24 hour cycle
[20:07] <Laurenceb> then you work out the geometery
[20:07] <Laurenceb> and the extra path
[20:07] <Laurenceb> typically 10m or more
[20:07] <Laurenceb> its rather complex, but luckly in the official gps docs they go through all the equations
[20:08] <Laurenceb> its all optimised as well - for uC with no fpu
[20:09] <edmoore> awesome
[20:10] <Laurenceb> its still rather complex to integrate it in
[20:10] <Laurenceb> you need page 18 of subframe 4
[20:11] <Laurenceb> then theres 6 s08 variables in there
[20:11] <edmoore> this is super impressive btw
[20:11] <edmoore> great job
[20:11] <edmoore> am going to a quick bite to eat
[20:11] <edmoore> bbl
[20:11] <Laurenceb> me too, bbl
[20:11] Nick change: edmoore -> edmoore|away
[20:24] <Laurenceb> we're just coming out of a solar minimum
[20:25] <Laurenceb> so total vertical delta should be around 3 meters or something
[20:25] <Laurenceb> http://en.wikipedia.org/wiki/File:Solar-cycle-data.png
[20:25] <Laurenceb> http://waas.stanford.edu/~wwu/ahansen/papers/ionntm00/node3.html
[20:27] <Laurenceb> henry recorded a while after 2.30pm local time
[20:28] <Laurenceb> about 4.30, so it would have been past the peak, therfor ~3m zenith
[20:28] <Laurenceb> interestingly altitude is about 3m too low, and th eposition is about 6m too far north, when its using prn13 which is approx pue south near the horizon
[20:29] <Laurenceb> it looks very hopeful
[20:30] <Laurenceb> hopefully within 2meters...
[20:39] i_am_ed (n=ed@93-97-226-168.zone5.bethere.co.uk) joined #highaltitude.
[20:55] G8KHW (n=Steve@217.47.75.27) joined #highaltitude.
[21:30] Simon-MPFH (n=simon@phantom.mpfh.co.uk) left irc: "Leaving"
[21:40] i_am_ed (n=ed@93-97-226-168.zone5.bethere.co.uk) left irc: Remote closed the connection
[21:41] i_am_ed (n=ed@93-97-226-168.zone5.bethere.co.uk) joined #highaltitude.
[21:41] i_am_ed (n=ed@93-97-226-168.zone5.bethere.co.uk) left #highaltitude.
[21:41] Simon-MPFH (n=simon@phantom.mpfh.co.uk) joined #highaltitude.
[21:56] <fergusnoble> edmoore|away: you there?
[22:05] Simon-MPFH (n=simon@phantom.mpfh.co.uk) left irc: "Leaving"
[22:16] <fergusnoble> guys, got dominoex working :)
[22:18] <fergusnoble> and it can decode the stream with no antenna on either end and the volume on zero
[22:18] <fergusnoble> i cant even see anything in the waterfall lol
[22:30] <edmoore|away> fergusnoble: clue was in the name :)
[22:30] Nick change: edmoore|away -> edmoore
[22:30] <edmoore> cool - potential divider?
[22:30] <fergusnoble> yeah, but i thought it was worth a go
[22:30] <fergusnoble> seeing as i was excited
[22:30] <fergusnoble> yeah, its nasty
[22:30] <edmoore> lol
[22:31] <edmoore> well cool that it works at least
[22:31] <fergusnoble> all i could find in my room was two 470ks
[22:31] <edmoore> after the dac hiccup
[22:31] Hiena (n=Hiena@81.93.195.181.datatrans.hu) left irc: "-=Halt! Hammerzeit!=-"
[22:31] <edmoore> daccup
[22:31] <fergusnoble> the input impedance is 100k so using that as one half of the divider worked
[22:31] <fergusnoble> yeah, the code worked first time :)
[22:32] <fergusnoble> also even with the divider i had to up the speed to 22 to get the resolution
[22:32] <fergusnoble> and it only works for one dav value
[22:32] <fergusnoble> so its borderline
[22:33] <fergusnoble> the 16bit dac will do it fine though
[22:34] josepharmbruster (n=josephar@45.144.119.70.cfl.res.rr.com) joined #highaltitude.
[22:34] <fergusnoble> josepharmbruster: hi
[22:35] <fergusnoble> hows your project going?
[22:37] <edmoore> varicode is so suboptimal given our telem string is mainly numbers
[22:37] <edmoore> it's a shame
[22:38] <fergusnoble> edmoore: we could vary the code :)
[22:38] <edmoore> ho ho
[22:38] <edmoore> well designing the optimal one would be completely trivial to do
[22:38] <edmoore> given we have so much previous telem data
[22:38] <edmoore> but it would stop other people listening in, without recompiling fldigi :)
[22:39] <fergusnoble> well, its very easy, we just make the numbers and the comma ones that only need one symbol
[22:39] <fergusnoble> and then start on the letters starting with badger etc
[22:40] <edmoore> well, just do a histogram of all our previous data and do it in the standard way. you are right that commas would probably have the highest probability
[22:40] <edmoore> then probably 5,2,0 (lat MSB and lon MSB)
[22:40] <fergusnoble> there are 18 symbols that are only one symbol long
[22:40] <edmoore> oh and decimal points
[22:41] <fergusnoble> so you can fit all the numbers and some more
[22:41] <fergusnoble> oh sorry, actually only 9
[22:41] <edmoore> but it's probably not worth it in the name of keeping it easy for others to decode
[22:41] <fergusnoble> because the msb is used for "expect another symbol"
[22:42] <fergusnoble> yeah agreed
[22:43] <edmoore> what do you think to a sat launch?
[22:44] <edmoore> winds are ok (got a lot worse though in last 24 hrs) and tomorrow is a complete write-off for me
[22:44] <fergusnoble> what are the winds doing?
[22:44] <edmoore> christian staying over friday and sat, and paula's 21st b-day tomorrow eve
[22:44] <edmoore> + lectures 9-10, 12-1, supervision 1-3, HoM 4-6
[22:45] <fergusnoble> yeah, sat is ok for me if need be
[22:45] <fergusnoble> what does wyoming say?
[22:45] <edmoore> not looked
[22:45] <edmoore> I have a splitting headache and the screen isn't helping. I only sat down to check emails :)
[22:45] <fergusnoble> oh dear, still going on?
[22:46] <edmoore> i'm finind it hard to budge
[22:46] <edmoore> am getting 11pm - 7am sleep, 4 ibuprofen a day, 1 berocca a day, 3 lemsip a day
[22:46] <edmoore> it's not budging
[22:46] <edmoore> getting pretty annoyed
[22:47] <fergusnoble> sucks
[22:47] <fergusnoble> the cold is making everyone ill
[22:47] <fergusnoble> ive got a nast cough back again :(
[22:47] <edmoore> I'm loosing my work hours too (evenings) with both difficulty concentrating and going to bed early
[22:47] <edmoore> too many late nights spoil the broth
[22:48] <edmoore> tbh I don't think Tess's party helped matters
[22:48] <edmoore> i was already feeling a bit squiffy before that
[22:48] <fergusnoble> hehe
[22:48] <fergusnoble> before you go, what should we do about the rotator in the shack?
[22:48] <fergusnoble> want to go down and check it out sometime soon?
[22:48] <fergusnoble> like sat if the launch doesnt happen
[22:49] <edmoore> yes
[22:49] <edmoore> i would have done already but for feeling a bit grim
[22:49] <edmoore> sat sounds like it has potential
[22:50] <edmoore> not sure how much christian is expecting me to entertain him
[22:52] <fergusnoble> is he a geek too?
[22:52] <fergusnoble> or to rephrase will he enjoy seeing a launch?
[22:56] <edmoore> hmm
[22:56] <edmoore> he was a mathmo
[22:56] <edmoore> do you remember him?
[22:56] <edmoore> Tall guy, short dark hair
[22:56] <edmoore> HoM
[22:56] <edmoore> graduated last year from Part III
[23:00] <fergusnoble> oh, yeah i think so
[23:01] <edmoore> he is coming to HoM, as is miss Lough
[23:01] <fergusnoble> ok cool
[23:01] <fergusnoble> ill try and make it
[23:02] <edmoore> I been thinking
[23:02] <edmoore> once B2 is sent off to china
[23:03] <edmoore> the idea of just doing some fun individual stuff (not too expensive) that doesn't require much consultation with everyone else
[23:03] <edmoore> for instance, I want to make a basic tracker with an AVR, partly just as an exercise to see quite how much I can get onto an AVR before it starts to choke
[23:03] <edmoore> just for personal interest and to help with intuition
[23:04] <fergusnoble> yeah, would be good to see if we could make a whole tracker that fits inside the radiometrix can
[23:04] <edmoore> I'm asking you rather than just doing it for myself anyway as i think there's be an arguement for CUSF paying for it if it's useful to us
[23:05] <edmoore> if not a radiometirx can, I want to make one that fits on the back of a 2xAA holder
[23:06] <edmoore> if you can make one fit on a radiometrix can, I will buy you a beer
[23:06] <fergusnoble> with gps?
[23:06] <edmoore> yes :)
[23:06] <fergusnoble> deal
[23:06] <edmoore> ok
[23:06] <edmoore> am I allowed to enter this comp?
[23:06] <fergusnoble> actually, i might not use a radiometrix
[23:06] <edmoore> actually, can we constrain to 2-layer? or is that cheating
[23:07] <edmoore> ust it'll be stupidly expensive otherwise
[23:07] <fergusnoble> but basically beer if you can make one the same size or smaller
[23:07] <edmoore> can we specify FSK too?
[23:07] <fergusnoble> yeah, should be 2 layer for cheapness of manufacture
[23:07] <edmoore> a keyed TX with a class E amplifier is much easier but not so useful
[23:07] <fergusnoble> yup fsk sounds sensible
[23:08] <edmoore> ok, just had my 2 ibuprofen
[23:08] <edmoore> last 20 mins before sleep overcomes me
[23:09] G8KHW (n=Steve@217.47.75.27) left irc: "Leaving"
[23:12] <edmoore> you looking at the GR-10 datasheet to see if it'll do high alt aswel as me? :p
[23:21] <fergusnoble> yeah - it doesnt
[23:22] <edmoore> yep
[23:22] <edmoore> shame
[23:22] <fergusnoble> neither does the 5010
[23:22] <edmoore> sirfIII
[23:22] <fergusnoble> still think its worth it?
[23:23] <edmoore> hey, it's your challenge :)
[23:23] <edmoore> but actually, there's something to be said for similicity. i.e. not doing rf routing on the board
[23:24] <fergusnoble> meh, i think doinf the rf is ok
[23:24] <fergusnoble> i was thinking for a bit of extra size you could put a chip scale antenna on the board
[23:25] <edmoore> the antenna needn't be counted in the size restriction if u don't want
[23:25] <edmoore> I think that's fair
[23:25] <fergusnoble> is the trimble copernicus any good?
[23:25] <edmoore> yes
[23:25] <edmoore> bill borwn just tested it i think
[23:25] <fergusnoble> yeah, i wasnt thinking of including the antenna
[23:25] <fergusnoble> in the restriction
[23:26] <fergusnoble> but the antenna would be like twice the size of the rest of the board, so by relaxing the size restrictions and doing something with an onboard ant you might end up with a better solution
[23:26] <edmoore> that's my thought
[23:26] <edmoore> something like the paparazzi autopilot but with much less of the 'not ground plane' area
[23:27] edmoore (n=edmoore@pomegranate.chu.cam.ac.uk) left #highaltitude.
[23:27] edmoore (n=edmoore@pomegranate.chu.cam.ac.uk) joined #highaltitude.
[23:28] <fergusnoble> also the copernicus doesnt need a 1.8v supply
[23:28] edmoore (n=edmoore@pomegranate.chu.cam.ac.uk) left #highaltitude.
[23:28] edmoore (n=edmoore@pomegranate.chu.cam.ac.uk) joined #highaltitude.
[23:28] <edmoore> whoops
[23:38] <fergusnoble> had a really good practical today
[23:38] <fergusnoble> playing with microwave waveguides
[23:39] edmoore (n=edmoore@pomegranate.chu.cam.ac.uk) left #highaltitude.
[23:39] edmoore (n=edmoore@pomegranate.chu.cam.ac.uk) joined #highaltitude.
[23:39] <edmoore> the window seems to close if I click on it
[23:39] <edmoore> odd
[23:40] <edmoore> fergusnoble: what were your last 2 lines?
[23:40] <edmoore> missed them
[23:40] <fergusnoble> had a really good practical today - waveguides
[23:40] edmoore (n=edmoore@pomegranate.chu.cam.ac.uk) left #highaltitude.
[23:40] edmoore (n=edmoore@pomegranate.chu.cam.ac.uk) joined #highaltitude.
[23:40] josepharmbruster (n=josephar@45.144.119.70.cfl.res.rr.com) left irc: Read error: 113 (No route to host)
[23:40] <edmoore> this must be a bug
[23:40] edmoore (n=edmoore@pomegranate.chu.cam.ac.uk) left #highaltitude.
[23:40] edmoore (n=edmoore@pomegranate.chu.cam.ac.uk) joined #highaltitude.
[23:41] <edmoore> yes
[23:41] <edmoore> oh well i am off to bed
[23:41] <edmoore> tell me about pracs tomorrow
[23:41] edmoore (n=edmoore@pomegranate.chu.cam.ac.uk) left irc:
[00:00] --- Fri Jan 30 2009