highaltitude.log.20080819

[00:00] <jcoxon> not too bad
[00:00] <jcoxon> at some point over the weekend there should be minimal jetstream
[00:00] rcaron_ (n=rcaron@c-75-68-113-23.hsd1.nh.comcast.net) joined #highaltitude.
[00:01] <natrium42> good to hear
[00:01] <jcoxon> got capture to work
[00:01] <natrium42> neat, happy with quality?
[00:01] <jcoxon> yup
[00:01] <jcoxon> however i managed to break the camera :-(
[00:01] <jcoxon> lightened it a little too much
[00:01] <SpeedEvil> Dead?
[00:02] <SpeedEvil> Or resting?
[00:02] <natrium42> so what are you using now?
[00:02] <jcoxon> nah dead
[00:02] <SpeedEvil> :/
[00:02] <jcoxon> natrium42, ordered a new camera
[00:02] <jcoxon> well i bid on ebay for one
[00:02] <natrium42> ah, same model?
[00:02] <jcoxon> actually an A70
[00:03] <jcoxon> which is identical to the A60 but a little bit better quality
[00:03] <jcoxon> yeah the quality is much better
[00:05] <natrium42> you could use fergus' cam
[00:05] <natrium42> but then you won't get the videos
[00:05] <jcoxon> fergus cam?
[00:05] <natrium42> i mean, dougs
[00:05] <natrium42> sorry
[00:05] <jcoxon> oh
[00:05] <jcoxon> i did cross my mind
[00:06] <jcoxon> but
[00:06] <jcoxon> the point of the payload is to keep it seperate
[00:06] <jcoxon> they are actually 2 seperate insulated payloads on a single deck
[00:06] <natrium42> k, what i thought
[00:06] <jcoxon> now that i know what not to do it won't take very long to set up the new cam
[00:06] <jcoxon> and the code is all in place
[00:06] <jcoxon> just a matter of plugging it in
[00:07] <fergusnoble> SpeedEvil: http://www.srcf.ucam.org/~cuspaceflight/images/PICT0018.JPG
[00:07] <fergusnoble> SpeedEvil: the photo of the different geometries we tried
[00:07] <fergusnoble> SpeedEvil: i think we are going for the third from the left
[00:11] <SpeedEvil> Fun. Have you actually computed effort/lift? the rectangular one looks very easy.
[00:11] <SpeedEvil> I suppose it depends on size.
[00:12] rcaron (n=rcaron@c-75-68-113-23.hsd1.nh.comcast.net) left irc: Read error: 110 (Connection timed out)
[00:12] <fergusnoble> yeah, the rectangular one could be good also
[00:12] <fergusnoble> was worried a bit about stress concentrations
[00:13] <SpeedEvil> As I understand it, the corners would be very un-stressed in that desigh.
[00:14] <jcoxon> right beacon done
[00:14] <SpeedEvil> :)
[00:16] <jcoxon> so now if it all goes tits up we still might be able to find it
[00:17] fergusnoble (n=fergusno@fn217.quns.cam.ac.uk) left irc: Read error: 104 (Connection reset by peer)
[00:17] fergusnoble (n=fergusno@fn217.quns.cam.ac.uk) joined #highaltitude.
[00:27] <jcoxon> night all
[00:27] jcoxon (n=jcoxon@81-178-103-50.dsl.pipex.com) left irc: "Leaving"
[00:33] natrium (n=alexei@DNab423c96.Stanford.EDU) joined #highaltitude.
[00:36] <fergusnoble> jcoxon: that change shaved 25% off the run time for me
[00:36] <fergusnoble> oh hes gone :(
[00:40] Action: SpeedEvil ponders digital noise-resistant image transfer.
[00:44] natrium42 (n=alexei@DNab422b3d.Stanford.EDU) left irc: Read error: 110 (Connection timed out)
[00:48] natrium (n=alexei@DNab423c96.Stanford.EDU) left irc:
[00:52] edmoore (n=edmoore@37.124-84-212.ippool.ndo.com) left irc: Read error: 104 (Connection reset by peer)
[00:52] edmoore (n=edmoore@37.124-84-212.ippool.ndo.com) joined #highaltitude.
[01:02] edmoore (n=edmoore@37.124-84-212.ippool.ndo.com) left irc:
[01:03] icez (n=icez@unaffiliated/icez) left irc: Read error: 60 (Operation timed out)
[01:07] icez (n=icez@unaffiliated/icez) joined #highaltitude.
[01:15] Laurenceb (n=Laurence@host86-133-69-206.range86-133.btcentralplus.com) left irc: Read error: 110 (Connection timed out)
[03:46] natrium42 (n=natrium4@rescomp-conf-08-10043.Stanford.EDU) joined #highaltitude.
[04:00] natrium42 (n=natrium4@rescomp-conf-08-10043.Stanford.EDU) left irc:
[05:49] natrium42 (n=natrium4@rescomp-conf-08-10043.Stanford.EDU) joined #highaltitude.
[05:55] natrium42 (n=natrium4@rescomp-conf-08-10043.Stanford.EDU) left irc: Remote closed the connection
[05:55] natrium42 (n=natrium4@rescomp-conf-08-10043.Stanford.EDU) joined #highaltitude.
[07:01] <akawaka> http://blog.makezine.com/archive/2008/08/hp_20b_biz_calc_can_be_mo.html?CMP=OTC-0D6B48984890
[07:40] Ebola (n=Ebola@unaffiliated/ebola) joined #highaltitude.
[07:52] natrium42 (n=natrium4@rescomp-conf-08-10043.Stanford.EDU) left #highaltitude.
[07:55] natrium42 (n=alexei@rescomp-conf-08-10043.Stanford.EDU) joined #highaltitude.
[08:02] jcoxon (n=jcoxon@81-178-75-36.dsl.pipex.com) joined #highaltitude.
[08:02] Shanuson (n=Peter@p54A95148.dip.t-dialin.net) joined #highaltitude.
[08:03] <jcoxon> morning all
[08:09] <jcoxon> ooooo its bank holiday weekend this weekend
[08:09] <jcoxon> that gives us another potential day to launch
[08:12] edmoore (n=edmoore@37.124-84-212.ippool.ndo.com) joined #highaltitude.
[08:16] <jcoxon> morning edmoore
[08:16] <edmoore> hello sir
[08:17] <jcoxon> would you be free on monday?
[08:18] <jcoxon> as its bank holiday
[08:18] <jcoxon> as the jetstream at the moment looks a lot better then
[08:18] <jcoxon> (still a little early i know)
[08:22] akawaka (n=akawaka@external.treyarch.com) left irc: Read error: 110 (Connection timed out)
[08:22] <edmoore> yo
[08:22] <edmoore> sorry
[08:23] <edmoore> hrm, maybe
[08:23] <jcoxon> np
[08:23] <jcoxon> just checking
[08:23] <edmoore> I'll say yes
[08:23] <jcoxon> just gives me more options
[08:23] <edmoore> yep
[08:23] <jcoxon> fergus might have sped up the sstv code
[08:24] <natrium42> so monday looks better than weekend?
[08:24] <jcoxon> right now yeah
[08:24] <jcoxon> but its a bit early to tell
[08:24] <jcoxon> its a bank holiday on monday so people are around
[08:26] mc- (n=mfcastle@cpc1-glfd1-0-0-cust487.glfd.cable.ntl.com) joined #highaltitude.
[08:26] <jcoxon> morning mc-
[08:27] <mc-> morning jc
[08:27] <mc-> congratulations on passing exams
[08:27] <jcoxon> thanks
[08:30] <jcoxon> fergusnoble, you awake at this early hour?
[08:31] edmoore (n=edmoore@37.124-84-212.ippool.ndo.com) left irc:
[08:51] mc- (n=mfcastle@cpc1-glfd1-0-0-cust487.glfd.cable.ntl.com) left irc:
[08:55] edmoore (n=edmoore@smtp.vorticity-systems.com) joined #highaltitude.
[08:56] <edmoore> jcoxon: that actually works well for me
[08:56] <edmoore> as I have a mate working in cam staying over
[08:56] <edmoore> so I'll take him with me to home home (chichester) on Saturday night/sunday, have a roast and eat lots of food, then head up early monday morning
[08:56] <edmoore> he'd probs be quite interested in coming on the mission aswell
[08:57] <natrium42> grr
[08:57] <natrium42> i hope they have free wifi in the airport
[08:57] <jcoxon> edmoore, okay well will have to see how the weather turns out
[08:57] <edmoore> sure sure. I'm flexible
[08:57] <jcoxon> natrium42, you off on holiday?
[08:57] <edmoore> natrium42: where are you off to?
[08:58] <natrium42> i am in stanford, california ATM
[08:58] <natrium42> going back home on monday
[08:58] <natrium42> :P
[08:58] <jcoxon> but this launch will be 2am in the morn for you won't it
[08:59] <natrium42> if you do it on monday, then yes
[08:59] <natrium42> well, actually
[08:59] <natrium42> that might be a good time
[08:59] <natrium42> since my flight it at 6 am
[09:00] <natrium42> i am going to be in the airport
[09:00] <natrium42> and i could just sign up for wifi
[09:00] <natrium42> not much to do there anyway
[09:00] <jcoxon> hehe
[09:00] <jcoxon> well we'll know more in a few days
[09:00] <edmoore> what's happening in stanford?
[09:00] <natrium42> when would the launch be in GMT?
[09:00] <natrium42> edmoore, visiting my sister
[09:01] <natrium42> checking out the university and area
[09:01] <jcoxon> ummm i reckon around 10am BST
[09:01] <edmoore> cool. must be pretty nice there this time of year
[09:01] <natrium42> yeah, it's nice
[09:01] <jcoxon> so that'll be 9am GMT
[09:01] <natrium42> ok
[09:02] <natrium42> should be fine then :D
[09:02] <jcoxon> hehe
[09:12] icez (n=icez@unaffiliated/icez) left irc: "Lost terminal"
[09:16] Simon-MPFH (n=simon@phantom.mpfh.co.uk) joined #highaltitude.
[09:19] <jcoxon> haha
[09:19] <jcoxon> http://www.flickr.com/photos/spiritofknoxville/2198936062/in/photostream/
[09:20] <jcoxon> i thought they might have done an 'armless' but sadly they were sensible and put a gap between harm and less
[09:25] <jcoxon> bbiab
[09:25] jcoxon (n=jcoxon@81-178-75-36.dsl.pipex.com) left irc: "Leaving"
[09:26] EI5GTB__ (n=Paul@78.16.77.84) left irc: Read error: 110 (Connection timed out)
[09:33] natrium42 (n=alexei@rescomp-conf-08-10043.Stanford.EDU) left irc:
[09:49] jcoxon (n=jcoxon@81-178-75-36.dsl.pipex.com) joined #highaltitude.
[09:50] borism_ (n=boris@195-50-199-92-dsl.krw.estpak.ee) joined #highaltitude.
[09:54] borism (n=boris@195-50-206-56-dsl.krw.estpak.ee) left irc: Read error: 145 (Connection timed out)
[10:02] <fergusnoble> jcoxon: here now but only quickly
[10:03] <jcoxon> what changes should i make to the code?
[10:03] <fergusnoble> just woke up so i am very late fore work
[10:03] <jcoxon> hehe
[10:03] <jcoxon> (but if you need to go - we can sort this later)
[10:03] <fergusnoble> ok, change i from being an int to being a float
[10:04] <fergusnoble> (in the tx_pixel function)
[10:04] <fergusnoble> and then comment out the while loop below "fin back to old phase"
[10:04] <fergusnoble> and add this line instead:
[10:05] <fergusnoble> i = sstv->phase * sstv->samplerate / (2.0*M_PI*freq);
[10:05] <fergusnoble> and thats it
[10:05] <jcoxon> okay - i'll make some changes
[10:05] <gordonjcp> what was the bug?
[10:05] <jcoxon> its not really a bug
[10:06] <jcoxon> just a more efficent way of doing it
[10:06] <gordonjcp> oh, ok
[10:06] <gordonjcp> I was going to have a crack at that later too
[10:06] <jcoxon> :-)
[10:06] <fergusnoble> but just thinking, the gumstix doesnt have floating point in hardware, so might try to change it a bit more this evening
[10:06] <jcoxon> okay i'll hold off
[10:07] <fergusnoble> jcoxon: no, give it a go, it would be good to see if it helps on the gumstix
[10:07] <fergusnoble> and if it affects the image quality
[10:07] <jcoxon> okay
[10:07] <fergusnoble> in theory it should make the quality better as it matches the phase better
[10:12] <fergusnoble> ok, bye
[10:35] <edmoore> fergusnoble: here?
[10:35] <edmoore> where is the code to this software? I missed this bit
[10:35] <edmoore> jcoxon: ping
[10:36] <jcoxon> pong
[10:36] <jcoxon> http://fkurz.net/ham/sstvtx.c
[10:39] <edmoore> and what was up with it?
[10:39] <jcoxon> nothing in theory
[10:39] <jcoxon> it works
[10:39] <jcoxon> just its quite slow
[10:40] <edmoore> a lookup table or fast trig library wouln't go amiss in the tx_pixel function either
[10:40] <edmoore> you can get good approximations of trig functions with orders of magnitude faster functions
[10:43] <jcoxon> just made the changes that fergus suggesteed
[10:43] <jcoxon> just going to compile it so shall see it
[10:55] <jcoxon> i'm not sure what i'd do without scp
[10:55] <jcoxon> it makes my life so much easier
[10:56] <edmoore> lol
[10:56] <edmoore> just buying some parachute material. Going to have a go at building one
[10:58] <jcoxon> i'm so annoyed with this damn camera breaking
[10:58] <jcoxon> had it all sorted
[10:58] <jcoxon> typical me
[11:04] <jcoxon> ooooo fergusnoble's changes have made a difference
[11:05] <jcoxon> its a lot less blocky
[11:13] <jcoxon> and 10% faster
[11:26] <edmoore> that's good
[11:26] <jcoxon> updated code: http://www.pegasushabproject.org.uk/wiki/doku.php/missions:haps:sstvtx
[11:26] <edmoore> maybe try a fast trig routine?
[11:26] <edmoore> jcoxon: where's the syntax colouring :p
[11:27] <jcoxon> done
[11:27] <jcoxon> sorry only did <code> instead of <code c>
[11:28] <gordonjcp> what do you use for decoding sstv?
[11:28] <jcoxon> macrobotsstv
[11:29] <edmoore> jcoxon: it would please me greatly if you could change it. It would let me enjoy it so much more
[11:29] <jcoxon> its done!
[11:29] <jcoxon> tis fixed
[11:30] <edmoore> beautiful
[11:30] <edmoore> thank you :)
[11:35] Ei5GTB (n=Paul@78.16.77.84) joined #highaltitude.
[12:44] robert1971_ (n=rharriso@gateway.hgf.com) joined #highaltitude.
[12:44] <robert1971_> Hi all
[12:46] <robert1971_> edmoore: You mentioned getting some blue stuff from the builders merchants. Do you know the name of the product or what the builders merchants was called?
[12:47] <robert1971_> Getting to the point of building an interim flight box
[12:47] <edmoore> robert1971_: I'm afriad i don't
[12:47] <edmoore> it was just like polystyrene, but a bit more dense
[12:47] <edmoore> and consequently a bit nicer to machine
[12:55] <robert1971_> Cool I'll trawl the internet for a bit and see what turns up :0
[12:56] <edmoore> it'll probably be a b&q or homebase job
[13:10] <robert1971_> It's looking that way. I was hoping to get a polyurethane insulation
[13:16] Action: SpeedEvil misread that as polyurethra.
[13:16] Action: SpeedEvil has not fully woken up.
[13:17] <SpeedEvil> I think the blue stuff is also polystyrene - it's just denser.
[13:51] <edmoore> jcoxon: ping
[13:51] <jcoxon> pong
[13:51] <jcoxon> (just returned)
[13:51] <edmoore> i would be fascinated to see what happens if you would be willing to give this code at the bottom a try http://www.ganssle.com/articles/atrig.htm
[13:53] <jcoxon> into the sstvtx code?
[13:53] <edmoore> yeah
[13:53] <jcoxon> where would it go?
[13:53] <edmoore> replace the sin(phase) with that
[13:53] <edmoore> i mean, make a function call to that
[13:54] <jcoxon> just a direct swap in?
[13:56] <edmoore> not quite - sin(x) = cos((pi/2)- x)
[13:56] <edmoore> so replace sin(phase) with cos(1.570796 - phase)
[13:57] <jcoxon> and stick the cos function further up
[13:57] <edmoore> it may lead an an unacceptable degredation in quality, I'd be interested to see what you think
[13:57] <edmoore> or further down, wherever. it's just a function
[13:58] <edmoore> you'll have to stick the function prototype at the top
[13:58] <edmoore> so:
[13:58] <edmoore> double cosine(double x); int main() {blah} ; double cosine(double x) { from the website}
[13:59] <edmoore> so yeah, that should be cosine(1.570796 - phase) in the code
[14:04] <jcoxon> running it now
[14:07] <edmoore> notice any difference?
[14:10] <jcoxon> it took 72 secs
[14:10] <jcoxon> however it doesn't seem to have generated the right signals
[14:11] <jcoxon> just running the old version to check thats its not something else
[14:14] <edmoore> ok. will be back in 3 mins. ooi, how long did it take previously?
[14:14] <jcoxon> 93secs for the test image
[14:14] phatmonkey (n=ben@scooby.firshman.co.uk) joined #highaltitude.
[14:18] <jcoxon> nah ed, the quality in the signal means that my reciever software doesn't pick it up
[14:19] <jcoxon> back in a bit
[14:25] <edmoore> very peculiar
[14:25] <edmoore> maclaurin series are normally pretty sound for this kind of thing
[15:06] <jcoxon> might be worth running on a big linux box
[15:11] <edmoore> yeah, just to test
[15:11] <edmoore> well, I have one sitting around doing nothing
[15:11] <edmoore> dual core 2.4ghz and 4gb of ram ok?
[15:15] <edmoore> though really, if that's only about 20% faster, there are probably savings to be made elsewhere
[15:16] <edmoore> jcoxon: does dokuwiki have a site map?
[15:16] <edmoore> that's the one thing I find very useful in mediawiki
[15:21] <robert1971_> edmoore: Have found a local packaging firm that does about 10 different foam boards with different properties. They are interested in the project and may be able to supply suitable product for free. Will feed back if I find out anything interesting. I'm looking for polyurethene sheet with open cell structure to prevent expansion due to decreased pressure.
[15:25] <edmoore> robert1971_: that sounds really interesting, and good luck with the sponsorship :) on that point, I have some observations
[15:27] <edmoore> we've never had a problem with the blue foam as it seems quite rigid despite being closed cell. Also, normal polystyrene in fine with balloon flights, except it seems if you cover it in gaffa or similar, which seems to let the sir out on the slow ascent, but the fast descent seems to make the gaffa act as a pressurized impermeable membrane and crush the polystyrene that it's on
[15:35] <SpeedEvil> The blue stuff can prolly take 14PSI with no problems
[15:35] <SpeedEvil> Even if it depressurised at altitude.
[15:49] <robert1971_> Interesting bit about the gaffa tape almost worth leaving an air hole to prevent compression
[15:50] <SpeedEvil> You'll probably need a few holes
[15:51] <SpeedEvil> Just to be safe - you're relying on otherwise the tiny bit of polystyrene on the other side of the tape to allow all of the air for that area to flow through it.
[16:15] edmoore (n=edmoore@smtp.vorticity-systems.com) left irc:
[16:22] edmoore (n=edmoore@smtp.vorticity-systems.com) joined #highaltitude.
[16:27] <robert1971_> I was thinking more of drilling a 5mm hole from the outside through to the inside. I'm sure that would be enough to allow the pressure to equalize
[16:29] <edmoore> I get to machine metal. wahay
[16:29] <edmoore> gosh, it's been so long since I've actually made stuff
[16:29] <edmoore> just faffing about with pcbs and software of late
[16:39] <robert1971_> I did a bit of that to make my aluminum frame for my camera. Are you going to be using a lathe?
[16:48] <jcoxon> hey edmoore
[16:48] <edmoore> hi jcoxon
[16:48] natrium42 (n=alexei@rescomp-conf-08-10043.Stanford.EDU) joined #highaltitude.
[16:48] <jcoxon> hmmm a sitemap
[16:48] <edmoore> robert1971: sorry I drifted from irc. yes
[16:48] <jcoxon> like an index?
[16:48] <edmoore> though probably a mill mainly for this bit
[16:48] <jcoxon> http://www.pegasushabproject.org.uk/wiki/doku.php/start?do=index
[16:48] <edmoore> then finish some facses off in the lathe
[16:49] <edmoore> jcoxon: brilliant
[16:49] <edmoore> that's proper useful
[16:49] <edmoore> to me, that's better than a table of contents
[16:49] <jcoxon> it relies of people keeping everything organised
[16:50] <edmoore> well seeing parentless random pages on it would be a good incentive, I imagine
[16:50] <jcoxon> hehe
[16:50] <jcoxon> both mine and ukhas aren't too disorganised
[16:50] <jcoxon> won that replacement camera
[16:50] <edmoore> no it's true, but they're vaguely difficult to navigate I find, sometimes
[16:50] <jcoxon> hopefully they will send
[16:50] <edmoore> oh cool - when's it arriving?
[16:50] <jcoxon> soon
[16:59] <natrium42> RealSoon (TM)
[17:01] <jcoxon> haha, i love weather forecasts
[17:01] <jcoxon> so this morning i was saying monday was best
[17:01] <jcoxon> now sat or sun
[17:02] edmoore (n=edmoore@smtp.vorticity-systems.com) left irc:
[17:05] <robert1971_> jcoxon: There seems to be a problem with the construction pics for Pegasus II http://www.pegasushabproject.org.uk/wiki/doku.php/missions:pegasus2
[17:05] <robert1971_> Launch ones are fine
[17:05] <jcoxon> yeah i've noticed that
[17:05] <jcoxon> its not the only place that happens
[17:05] <jcoxon> seems to be something to do with the jpgs and hte gallery code
[17:05] <jcoxon> will have a lookinto it later
[17:07] <jcoxon> bbiab
[17:12] robert1971_ (n=rharriso@gateway.hgf.com) left irc:
[17:12] natrium42 (n=alexei@rescomp-conf-08-10043.Stanford.EDU) left irc: Remote closed the connection
[17:15] Ei5GTB_ (n=Paul@78.16.169.36) joined #highaltitude.
[17:22] edmoore (n=edmoore@37.124-84-212.ippool.ndo.com) joined #highaltitude.
[17:23] Ebola (n=Ebola@unaffiliated/ebola) left irc: Remote closed the connection
[17:30] <edmoore> jcoxon: looks like saturday mid to late afternoon is the best bet atm
[17:32] Ei5GTB (n=Paul@78.16.77.84) left irc: Connection timed out
[17:58] <jcoxon> edmoore, yeah
[17:58] <jcoxon> i have to be home saturday evening but it'll come this way anyway
[17:59] <edmoore> yeah, still too far out to be particularly meaningful too
[18:15] <edmoore> jcoxon: who is a good domain name provider?
[18:15] <jcoxon> phatmonkey?
[18:16] <jcoxon> he'd know
[18:16] <edmoore> phatmonkey: ping
[18:16] <phatmonkey> I use 123-reg.co.uk
[18:16] <phatmonkey> not sure if they're the best, but they certainly get the job done
[18:17] <jcoxon> i use streamline.net
[18:17] <phatmonkey> and don't charge fees for anything simple (full DNS control, email forwarding etc is included)
[18:17] <edmoore> subdomains?
[18:17] <edmoore> i guess so if we have wiki.ukhas
[18:33] <phatmonkey> yeah, they allow full control of the DNS
[18:33] <phatmonkey> you can do almost everything
[18:33] <phatmonkey> when I say almost, I host my own name servers because I don't think they support TXT records and things like that
[18:38] <edmoore> well I will use afriad.org I think for stuff too
[18:40] akawaka (n=akawaka@external.treyarch.com) joined #highaltitude.
[19:20] <edmoore> jcoxon / phatmonkey: it's not unexpected that it won't register and start forwarding instantly, isn't it?
[19:20] <phatmonkey> 123-reg.co.uk takes 5-10 mins iirc
[19:20] <phatmonkey> might take longer
[19:20] <edmoore> they claim up to 24hrs
[19:20] <edmoore> fun
[19:39] kamaji (n=kamaji@handtomouse.demon.co.uk) joined #highaltitude.
[19:41] <edmoore> oh well, i got infiniteorbit.org
[19:41] <edmoore> I'll leave it for a few hours
[19:49] ball (n=ball@client-216-176-80-183.consolidated.net) joined #highaltitude.
[19:52] <edmoore> hi ball
[19:52] <jiffe99> how much problem is there with heat dissipation at 30km?
[19:53] <SpeedEvil> jiffe99: little generally
[19:53] <SpeedEvil> Given the temp of -50Cish
[19:54] edmoore_ (n=edmoore@37.124-84-212.ippool.ndo.com) joined #highaltitude.
[19:54] edmoore (n=edmoore@37.124-84-212.ippool.ndo.com) left irc: Read error: 104 (Connection reset by peer)
[19:55] <edmoore_> jiffe99: Whilst cold, also consider than the air density is <1% of sea level, so convective losses make up a much smaller fraction of total losses
[19:55] <SpeedEvil> jiffe99: if you want to do actual high power stuff, then convection and forced cooling are lots less efficient. Due to the low thermal capacity of the stuff.
[19:55] <SpeedEvil> jiffe99: but barring actual high power, getting too hot isn't the problem typically.
[19:57] <jiffe99> yeah, well I'm just wondering if at the high altitudes the voltage regulators won't lose very much of the heat they are generating
[19:57] <fergusnoble> yo'
[19:57] <ball> hello edmoore
[19:58] <jiffe99> in all my reading about high altitude instrumentation I haven't heard anymore mentioning overheating issues
[19:59] <SpeedEvil> Mainly as people aren't doing high power stuff up there typically.
[19:59] <SpeedEvil> Batteries are heavy.
[20:00] <ball> It would be nice if a baloon could be made of a very thin PV material
[20:00] <ball> ...generate its own power.
[20:01] <jiffe99> yeah, this isn't too high powered, the transmitter is drawing directly from the battery, I have the gps and camera coming off the 3.3v rail though
[20:03] <edmoore_> hi fergusnoble
[20:03] Nick change: edmoore_ -> edmoore
[20:03] <jiffe99> I have no clue how much current the camera draws, I'm going to have to test that out
[20:04] icez (n=icez@unaffiliated/icez) joined #highaltitude.
[20:04] <fergusnoble> brb
[20:05] <ball> hello icez
[20:08] <edmoore> ball: have you an interest in making a HAB?
[20:09] <ball> I'm interested in learning more about what's involved
[20:10] <edmoore> well you've come to the right place :)
[20:10] <ball> I had help finding you ;-)
[20:11] <edmoore> from who?
[20:11] <ball> Are starter balloons available off the shelf?
[20:12] <ball> edmoore: a couple of people. I see familiar nicks in here.
[20:12] <edmoore> ball: the balloons themselves are all off the shelf
[20:12] <ball> I have to go, my daughte just woke up
[20:12] <ball> I shall return.
[20:12] <edmoore> the electronics is up to you, though there's tonnes of info around (and specifically here) on building payloads
[20:12] <edmoore> ok
[20:12] soneil (n=soneil@89.100.243.212) joined #highaltitude.
[20:12] <edmoore> hi soneil
[20:12] ball (n=ball@client-216-176-80-183.consolidated.net) left irc: "leaving"
[20:13] <soneil> hey
[20:14] <edmoore> all well?
[20:14] <soneil> yup .. just received my gps module today, so toys to play with
[20:15] <edmoore> fun fun
[20:15] <edmoore> which one?
[20:16] <soneil> the larsen(?) one from sparkfun .. no-one seemed to have any complaints with it
[20:17] <soneil> ah, lassen IQ. my head's kinda unreliable lately
[20:17] <jiffe99> I have a lassen
[20:17] <jiffe99> seems to work ok
[20:18] <jiffe99> mostly been inside tests, I've seen it acquire 8 satellites in the window sill
[20:18] <jiffe99> altitude jumps around a lot but I understand that is common to all gps's
[20:19] <fergusnoble> soneil: lassen was used on novas 6 & 7 with no probs
[20:20] <soneil> have to wait for the wife to go back to work before I can break out the iron 'n figure out how to wire it up. that's a fairly dainty connector
[20:20] <fergusnoble> jcoxon: hows the sstvtx thingy going, did the change work?
[20:21] <fergusnoble> jcoxon: might have a look at making it non floating point in a bit
[20:22] <edmoore> fergusnoble: I suggested james try it with a fast approximate trig library, but i think it was too approximate. James says the signal was too degreaded
[20:22] <fergusnoble> jcoxon: although is this a critical speed component of the system?
[20:23] <fergusnoble> surely it takes longer to send the pic than to process the next one?
[20:23] <fergusnoble> edmoore: good idea
[20:23] <fergusnoble> be careful though, some of those libraries use taylor approximations which are only valid for 0 - 2pi
[20:24] <fergusnoble> and that code puts values into the sin function not in that range
[20:25] <edmoore> indeed, it had quadrant fuding
[20:25] <edmoore> fudging*
[20:38] <jcoxon> hey guys
[20:38] <fergusnoble> jcoxon: hi
[20:38] <fergusnoble> jcoxon: did the change work?
[20:39] <jcoxon> yeah it sped it up by about 10%
[20:39] <jcoxon> and made teh quality a lot better
[20:39] <fergusnoble> schweeet
[20:39] <fergusnoble> i was going to ask, is making it any faster a big issue?
[20:40] <jcoxon> hmmm perhaps
[20:40] <fergusnoble> as surely its done processing the next frame by the time the last one is sent
[20:40] <jcoxon> its a little bit more linear then that
[20:40] <jcoxon> as in
[20:40] <jcoxon> it takes, processes, sends
[20:41] <jcoxon> in the mean time its sending packets
[20:41] <jcoxon> after 7 1/2 mins it sends it
[20:41] <fergusnoble> ok, how long does it take atm on the gumstix?
[20:41] <jcoxon> well thats the thing
[20:41] <jcoxon> it seems to vary depending on the image
[20:42] <jcoxon> the test image takes 92 seconds
[20:42] <jcoxon> however a real image takes a bit longer
[20:42] <jcoxon> say 3 mins
[20:42] <fergusnoble> i see
[20:42] <jcoxon> (need to test when i get the new camera)
[20:42] <jcoxon> sometimes it gets stuck and doesn't process a picture
[20:42] <fergusnoble> ok, when i get time i could look at removing the dependence on floating point
[20:42] <jcoxon> so after a set time i kill the process just in case
[20:43] <fergusnoble> thats what will be slowing it down mainly as the gumstix has no hardware fp
[20:52] <jcoxon> its not that vital to tell you the truth
[20:53] <jcoxon> actually finding out why it sometimes doesn't work would be more important
[20:53] <jcoxon> really need to stress test it with the new camera
[20:55] <SpeedEvil> Isn't the sin() function only used for generating the output waveform?
[20:55] <SpeedEvil> In which case a much faster small lookup table would be quite adequate to get 8 bit precision.
[20:56] <SpeedEvil> (diddn't look at code)
[20:57] robert1971 (n=rharriso@80.176.172.227) left #highaltitude.
[20:59] <fergusnoble> SpeedEvil: yeah that would work just dandy
[20:59] <jcoxon> fergusnoble, i've stuck the updated code on my wiki
[21:00] Ei5GTB_ (n=Paul@78.16.169.36) left irc: "Leaving"
[21:21] <edmoore> back in 5 mins
[21:24] edmoore (n=edmoore@37.124-84-212.ippool.ndo.com) left irc:
[21:26] edmoore (n=edmoore@37.124-84-212.ippool.ndo.com) joined #highaltitude.
[21:30] Simon-MPFH (n=simon@phantom.mpfh.co.uk) left irc: "Leaving"
[21:31] G8KHW (n=Steve@217.47.75.27) joined #highaltitude.
[21:36] <jcoxon> evening G8KHW
[21:37] <G8KHW> hi jcoxon - just having a look at the sstv code again
[21:37] <jcoxon> fergus has improved it
[21:37] <jcoxon> just grabbing the link
[21:38] <G8KHW> ah - good whats it run like now?
[21:38] <jcoxon> http://www.pegasushabproject.org.uk/wiki/doku.php/missions:haps:sstvtx
[21:38] <jcoxon> it runs about 10% faster
[21:38] <jcoxon> and its better quality
[21:39] <G8KHW> spiffing - I'll start with that then
[21:40] <jcoxon> just to say that the gumstix doesn't have floating point hardware
[21:40] <jcoxon> and occasionally it doesn't work - chokes on a picture
[21:42] <G8KHW> I doubt that there is any need to use floating point - its probably better converted to ints/longs then that will make it much faster
[21:47] <fergusnoble> G8KHW: using a lookup table or somesuch for the sin() call could help a lot too
[21:48] <G8KHW> yeah probably - I see phse is the only fp variable
[21:48] <G8KHW> phase
[21:50] <edmoore> i was just compiling a lookup :)
[21:50] <edmoore> though I read that for phase modulated signal applications, the lookup has to be pretty good
[21:51] <edmoore> it's one of the more demanding apps anyway
[21:51] <G8KHW> what I sustect is
[21:51] <G8KHW> suspect is happenong is
[21:51] <fergusnoble> i dont know about how much ram the gumstix has but yo could affort to make a huge lookup table
[21:51] <G8KHW> that a frequency is being generated for each pixel
[21:51] <fergusnoble> like several megs if need be
[21:52] <G8KHW> each grey level (or colour)
[21:52] <G8KHW> is a particular frequency
[21:52] <G8KHW> so at worst 256 frequency tables
[21:53] <G8KHW> since ther are 256 grey/colour levels
[21:53] <G8KHW> since the output sample rate is fixed
[21:54] <G8KHW> then each table needs to include the number of samples for one bit period
[21:55] <G8KHW> the only issue is making the bits join together smoothly in phase
[21:56] <G8KHW> so if a bit (pixel) ends at 230deg at one frequency the next starts at 230deg for the next frequency
[21:56] <fergusnoble> G8KHW: yeah, the basis of the improvement to quality in my change was making the starting phase shift for the pixel a float not an int
[21:56] <G8KHW> just a matter of jumping beteeen the tables
[21:57] <fergusnoble> but you could change back to using an int and the quality would be the same as the software was otb
[21:58] <fergusnoble> but that sounds like a very good idea, having precomputed tables for the different bit values
[21:59] <G8KHW> since there are so few levels
[21:59] <akawaka> http://blog.makezine.com/archive/2008/08/blimpin_aint_easy.html?CMP=OTC-0D6B48984890
[22:03] <akawaka> jcoxon: how was is the sstv now?
[22:04] <jcoxon> using the float its better quality now
[22:04] <akawaka> still takes 5 minutes to generate?
[22:04] <jcoxon> the test image takes 92secs
[22:04] <fergusnoble> ok, got to go, bye
[22:05] <jcoxon> however it takes longer with a real pic
[22:05] <jcoxon> however i'm waiting for a new camera
[22:05] <jcoxon> then i'll properly test everything
[22:05] <akawaka> weird that it takes so long
[22:05] <jcoxon> whats more annoying is that occasionally it chokes
[22:05] <G8KHW> ha - looks like each table will need to be at most 32 samples (for 48Khz sample rate of the wav)
[22:06] <G8KHW> so thats 32 x 256 = 8K table entries
[22:06] <akawaka> where does it die?
[22:06] <jcoxon> it does about 30 lines of the image then gives up
[22:06] <jcoxon> i need to kill the process
[22:07] <jcoxon> but i really need to test this again
[22:08] Nick change: fergusnoble -> fergus|away
[22:09] phatmonkey (n=ben@scooby.firshman.co.uk) left irc:
[22:09] <soneil> I swear I haven't done a good soldering job since the damned europeans took the lead out of solder
[22:10] <G8KHW> G8KHW stocked up with non ROHS solder before it dissapeared
[22:10] <G8KHW> enough to see me out
[22:11] <soneil> I'm going nuts. 75W temp-controlled iron at 400C .. and it won't melt
[22:12] <G8KHW> wow it should do
[22:14] phatmonkey (n=ben@scooby.firshman.co.uk) joined #highaltitude.
[22:15] <G8KHW> jcoxon: have you got the jpeglib.h stuff
[22:15] <jcoxon> ummm its in the jpeg lib i think
[22:20] <edmoore> soneil: gold a flux pen?
[22:21] <edmoore> live and die by those things
[22:21] <edmoore> got a*
[22:21] Laurenceb (n=Laurence@host86-133-69-206.range86-133.btcentralplus.com) joined #highaltitude.
[22:21] <Laurenceb> hi everyone
[22:21] <edmoore> ello
[22:22] <Laurenceb> I've been decorating since 8am :-/
[22:22] <Laurenceb> knackering
[22:22] <Laurenceb> right time to play with kalman filters :D
[22:27] fergus|away (n=fergusno@fn217.quns.cam.ac.uk) left irc:
[22:28] <Laurenceb> hmm there seems to be memory issues
[22:30] <Laurenceb> http://wiki.ukhas.org.uk/code:parafoil_tsip?do=show the dataprint function ocassionally has outputbuffer equal to "il"
[22:30] <Laurenceb> any ideas? does it look like a stack overflow?
[22:30] Ei5GTB (n=Paul@78.16.169.36) joined #highaltitude.
[22:35] <akawaka> how big is your stack?
[22:35] edmoore (n=edmoore@37.124-84-212.ippool.ndo.com) left irc: Read error: 104 (Connection reset by peer)
[22:35] <Laurenceb> well there is 1K or ram
[22:35] <Laurenceb> *of
[22:35] <Laurenceb> I have about 340 bytes of .data
[22:35] edmoore (n=edmoore@37.124-84-212.ippool.ndo.com) joined #highaltitude.
[22:38] <Laurenceb> the kalman filter seems to work ok until sprintf buggers up
[22:38] <Laurenceb> its kind of odd
[22:38] <Laurenceb> need to investigate further
[22:39] <Laurenceb> thing is, I have reduced my ram use from stuff in .data and the error condition doesnt seem to have been affected
[22:41] <Laurenceb> ok, I've chenged it to only use printf
[22:41] <G8KHW> jcoxon: that would be a linux library then?
[22:41] <akawaka> sprintf needs a lot of stack space sometimes
[22:41] <akawaka> G8KHW: libjpeg
[22:42] <akawaka> especially converting floats to strings
[22:42] phatmonkey (n=ben@scooby.firshman.co.uk) left irc:
[22:46] <Laurenceb> yeah, what I was doing
[22:47] <Laurenceb> so, I put checksum+=sendchar; in the put_char function
[22:47] <Laurenceb> then I can use printf and calculate checksums on the fly
[22:48] <Laurenceb> now to try it... if it gives the same error it'll be very weird
[22:48] Shanuson (n=Peter@p54A95148.dip.t-dialin.net) left irc: Read error: 104 (Connection reset by peer)
[22:50] kamaji (n=kamaji@handtomouse.demon.co.uk) left irc: "sleepytiems"
[22:52] <Laurenceb> hmm would appear to be fixed :P
[22:53] <Laurenceb> but the kalman filter still blows up if I move the gyro violently... I suspect its a dodgy gyro possible
[22:53] <Laurenceb> as its works as expected if you go slowly
[22:58] Action: Laurenceb hacks up some gyro test code
[23:00] phatmonkey (n=ben@scooby.firshman.co.uk) joined #highaltitude.
[23:01] phatmonkey (n=ben@scooby.firshman.co.uk) left irc: Remote closed the connection
[23:04] <Laurenceb> aha the gyro is f*xored
[23:04] <Laurenceb> which implies that everying works firmware wise bar minor issues tike cutdown timing/message formatting ect :D
[23:05] <Laurenceb> edmoore: do you know much about configuring the lassen iq ?
[23:06] <edmoore> not since you asked me about 2 days ago :p
[23:06] <Laurenceb> I want to send it some floats, but I've forgotten if its little or big endian
[23:07] <Laurenceb> so say I want to send -1 (leaves a variable unchanged) do I give it 0xBF800000 ?
[23:07] <Laurenceb> or the reverse?
[23:09] <Laurenceb> this gyro damage is quite interesting... appears my cooldown test fractured either a trace or a join to Vdrive, but the gyro continued operating intermittently
[23:10] <Laurenceb> I've soldered on a jumper to supply it properly, and the analogue output appears to be ok, but any shaking and the SPI output is screwy
[23:15] <edmoore> we're keeping our stiff warm for that reason now
[23:16] <edmoore> a pcb can probably take only so many 80K amplitude cycles
[23:40] fnoble (n=fnoblef@fn217.quns.cam.ac.uk) joined #highaltitude.
[23:40] <fnoble> hi all
[23:50] <G8KHW> hi fnoble
[23:50] <G8KHW> just been looking at the sstv code
[23:50] <G8KHW> I see what you have done - its more accurate as well as faster
[23:51] <jcoxon> yeah the image quality is a lot better
[23:51] <G8KHW> also the stuff below that is just as dodgy
[23:52] <G8KHW> it would be better to calculate a phse increment given the frequency trying to be generated and the wav sample frequency
[23:52] <G8KHW> phase
[23:52] <G8KHW> instead it keeps re-calculating the next sample phase
[23:53] <G8KHW> lots of floating point multiplys rather than fp adds
[23:53] <G8KHW> so it should improve loads (and be more accurate)
[23:54] <G8KHW> I ony wish i had somthing to complile it on
[23:54] <jcoxon> G8KHW, i can give you a shell account on my linux box?
[23:55] <G8KHW> I'll boot ubuntu tomorrow
[23:55] <jcoxon> just need to ssh in
[23:55] <G8KHW> yeah that would be good
[23:57] <jcoxon> sent you an email
[23:58] <G8KHW> also the code seems to assume that a pixel is an integer number of wav samples - which aint the case
[23:58] <G8KHW> ta - night guys
[23:59] <jcoxon> night
[23:59] G8KHW (n=Steve@217.47.75.27) left irc: "Leaving"
[00:00] --- Wed Aug 20 2008