highaltitude.log.20070324

[00:39] edwardmoore (n=edwardmo@88.212.167.121) left irc:
[00:42] jcoxon (n=jcoxon@85-210-179-186.dsl.pipex.com) left irc: "Leaving"
[02:34] icez (n=icez@ip68-3-56-121.ph.ph.cox.net) joined #highaltitude.
[07:05] icez (n=icez@ip68-3-56-121.ph.ph.cox.net) left irc: Remote closed the connection
[07:28] jcoxon (n=jcoxon@85-210-179-186.dsl.pipex.com) joined #highaltitude.
[09:10] edwardmoore (n=edwardmo@88.212.167.121) joined #highaltitude.
[10:04] flowolf (n=flowolf@host81-187-static.13-87-r.retail.telecomitalia.it) joined #highaltitude.
[10:43] edwardmoore (n=edwardmo@88.212.167.121) left irc:
[10:57] edwardmoore (n=edwardmo@88.212.167.121) joined #highaltitude.
[11:07] edwardmoore (n=edwardmo@88.212.167.121) left irc:
[12:20] phatmonkey (i=nobody@81.2.121.150) left irc: Remote closed the connection
[12:42] phatmonkey (i=nobody@81.2.121.150) joined #highaltitude.
[12:46] edwardmoore (n=edwardmo@88.212.167.121) joined #highaltitude.
[12:51] Ebola (n=Ebola@host86-143-156-147.range86-143.btcentralplus.com) joined #highaltitude.
[12:54] laurence (n=Laurence@host81-154-84-2.range81-154.btcentralplus.com) joined #highaltitude.
[12:54] <laurence> Hi all
[12:56] flowolf (n=flowolf@host81-187-static.13-87-r.retail.telecomitalia.it) left irc: "Leaving"
[12:56] <edwardmoore> hi laurence b
[13:06] <laurence> Grr I've lost rack of whats what with yesterdays flight data
[13:06] <laurence> rack=track
[13:07] <laurence> got a huge txt file i'm trying to organise
[13:08] Nick change: laurence -> Laurenceb
[13:09] <Laurenceb> Have you seen the latest graph on the wiki?
[13:09] <edwardmoore> will just go and check to confirm
[13:10] <Laurenceb> The values of P,I,and D are messed up, just trying to sort that
[13:11] <edwardmoore> ok, am with you now
[13:12] <Laurenceb> okay G and K are correctly matched
[13:15] <Laurenceb> right got everything matched apart from L,m, and N but with those i'm totally confused
[13:16] <Laurenceb> Anyway what would you conclude from the behaviour?
[13:22] <Laurenceb> ok all sorted now
[13:23] <edwardmoore> whats the time period?
[13:27] <Laurenceb> well one division on the bottom is one second
[13:29] <Laurenceb> i'd say G, k, and M were the most promising, and they had P,I,d around 0.1, 0.01, 0.45
[13:30] <Laurenceb> but then so did other flights... could just be what the wind was doing
[13:32] <Laurenceb> looking at test J, which could have been unstable, I is quit
[13:33] <Laurenceb> - quite large wrt P
[14:04] <Laurenceb> working on the glider simulation :-)
[14:05] icez (n=icez@ip68-3-56-121.ph.ph.cox.net) joined #highaltitude.
[14:06] edwardmoore (n=edwardmo@88.212.167.121) left irc:
[14:08] edwardmoore (n=edwardmo@88.212.167.121) joined #highaltitude.
[14:20] edwardmoore (n=edwardmo@88.212.167.121) left irc:
[14:26] edwardmoore (n=edwardmo@88.212.167.121) joined #highaltitude.
[14:39] Nick change: edwardmoore -> edmoore
[14:40] mc- (n=mfcastle@cpc2-glfd1-0-0-cust995.glfd.cable.ntl.com) joined #highaltitude.
[14:46] <mc-> Hi Laurence
[14:46] <mc-> I had some thoughts for your simulator...you could start with a D term
[14:47] <mc-> it will give damped oscillations without D
[14:47] <mc-> sorry I meant start without D
[14:48] <mc-> And I don't think you need the I term, since the glider will drift off line anyway
[14:55] <icez> hey mc-
[14:55] <mc-> how's space access?
[14:56] <icez> henry told me they don't collect slides, so I'd have to ask individual people :p
[14:56] <icez> but Carmack is putting up a video on his website that he showed, really cool.
[14:56] <mc-> Have Masten presented yet?
[14:57] <icez> nah, 9:40 this morning
[14:57] <mc-> I guess everyone has been talking about spacex?
[14:57] <icez> mmm well
[14:58] <icez> the ITAR panel was talking about "what happens if you tell a canadian about a space technology"
[14:58] <icez> and someone yelled "shoot him"
[14:58] <icez> :/
[14:58] edmoore (n=edwardmo@88.212.167.121) left irc:
[14:58] <icez> Jim Muncy talked about the good/bad things that happened since the last conference
[14:58] <icez> not much spacex
[14:58] flowolf (n=flowolf@host81-187-static.13-87-r.retail.telecomitalia.it) joined #highaltitude.
[14:59] <mc-> there's not enough people actually building things,but I guess I'm guilty of that as well.
[15:00] <icez> bah, I don't have much knowledge about that stuff
[15:00] <icez> I don't think it'd be safe for me to make propellers and all :]
[15:01] <mc-> anyone talking about rockoons?
[15:01] <Laurenceb> sorry was away for a moment
[15:01] <icez> nah
[15:01] <icez> yesterday, pretty much only armadillo talked about "building stuff"
[15:02] <icez> other people were talking about how the whole thing is going, and space insurance, and FAA/ITAR regulations..
[15:02] <Laurenceb> sounds cool
[15:02] <mc-> Paul Breed is building something for the lunar lander challenge
[15:02] <Laurenceb> henry should have launched his rockoon by now in cambridge
[15:02] <Laurenceb> was sceduled for this morning
[15:03] <mc-> didn't know that..a real rockoon?
[15:03] <Laurenceb> unguided "model" rocket launch filmed from payload
[15:03] <mc-> wow, hope it goes well
[15:03] <Laurenceb> hopefully we will hear shortly
[15:04] <Laurenceb> the rocket has no electronics or anything, so they hope someone will find it
[15:04] <mc-> what altitude was he planning to launch at?
[15:04] <Laurenceb> not sure guessing 25km or something
[15:05] <Laurenceb> have you seen my glider test series?
[15:05] <mc-> so there's a small risk of someone having a rocket land in their garden/on their head?
[15:05] <mc-> Yes, been reading the wiki
[15:05] <Laurenceb> erm yes
[15:06] <Laurenceb> have you got any experience of PID control?
[15:07] <mc-> Whenever I've done it, I've ended up not having a D term, and letting it damp itself.
[15:07] <Laurenceb> Got the simulation sort of replicating the behaviour, but its way too slow
[15:07] <Laurenceb> oscillation wise
[15:08] <mc-> It should have smaller oscillation if you reduce P
[15:08] <Laurenceb> it treats it like a mass on a spring, calculating at 50hz, with a 0ne second old gps every second
[15:08] <Laurenceb> yes but if P is reduced it heads off in the wrong direction
[15:09] <mc-> shouldn't be in the wrong direction, just it will be slow to correct
[15:09] <Laurenceb> hmm I have limited test system ie hill
[15:10] <mc-> I've been caught out by - signs before when doing tan() in C
[15:10] <Laurenceb> it now works well enough to fly into the housing estate at the bottom, so I need better facilities
[15:11] <mc-> if it goes in roughly the right direction, that's a major improvement on no guidance.
[15:11] <Laurenceb> yes I had tons of problems with signs in the origional code, but the simulation doesnt need the complicated wind correction
[15:12] <Laurenceb> talking of which, what type of knots does gps use?
[15:12] <mc-> nautical knots AFAIK
[15:12] <mc-> I can check if you need?
[15:13] <Laurenceb> okay, I need to convert from "equatorial degree units"per second to knots
[15:13] phatmonkey (i=nobody@81.2.121.150) left irc: Read error: 104 (Connection reset by peer)
[15:13] <Laurenceb> during the ascent, by measuring the change in position every 100m ascent
[15:14] phatmonkey (i=nobody@81.2.121.150) joined #highaltitude.
[15:14] <mc-> I guess you're have a different conversion for lat and lon?
[15:14] <mc-> at 51 deg 1 deg lat = 110km
[15:15] <Laurenceb> well I made up "equatorial degree units" ie one degree along equator
[15:15] <Laurenceb> yes think your saying the same thing
[15:15] <mc-> 1 deg lon = 60km
[15:16] <Laurenceb> I use sin to convert into my made up units accounting for latitudes effect on longditude
[15:17] <Laurenceb> but... if the balloons is acending quite quickly, then the wind will not have moved it far between 100m intervals, so more error
[15:18] <mc-> I found it easier to hard code in a conversion from lat/lon to distance.
[15:18] <Laurenceb> probably best to use the gps heading and speed to find wind on the way up I think
[15:19] <Laurenceb> and take a few readings at each 100m mark to reduce error
[15:20] <Laurenceb> ie if presentgps-lastreading>100m then take 5 readings and average
[15:20] <mc-> if you have a $GPVTG msg, it has the speed in km/hr
[15:20] <Laurenceb> oh cool
[15:20] <Laurenceb> well im using $GPRMC so its all working in knots
[15:21] <Laurenceb> planning on a launch from 4 km in 2 weeks if I can iron out these instabilities
[15:22] <mc-> I checked a NMEA manual, and it doesn't say the speed is nautical knots, but I'm pretty sure it is.
[15:22] <mc-> since NMEA is a marine standard
[15:22] <Laurenceb> only need to write the wind calculation code for on the way up
[15:23] <Laurenceb> thanx, but thinking about it i'll just use the $GPRMC and so wont need to do any conversions
[15:23] <mc-> I guess you doing a wind calculation because you are worried the PID loop will be unstable if it's heading into wind?
[15:24] <Laurenceb> yes, the behaviour will be nonlinear unless you account for wind
[15:24] <mc-> if you could measure the heading directly, then you wouldn't need to account for wind?
[15:24] <Laurenceb> yep
[15:25] <mc-> hmm
[15:25] <Laurenceb> but digital compasses dont like to be nonlevel
[15:25] <mc-> but they will still give a reasonable reading.
[15:26] <Laurenceb> non really!
[15:26] <Laurenceb> -not
[15:26] <Laurenceb> just 10 degrees off horizontal can make the heading off by 30 or so
[15:26] <mc-> can you stablize a compass?
[15:27] <mc-> Hang it on a piece of string?
[15:27] <Laurenceb> hmmm, lol
[15:27] <Laurenceb> you could use a 3 axis one and thermopiles to find the orientation
[15:28] <mc-> if it's swinging from side to side, how about averaging the reading?
[15:28] <Laurenceb> the I2C ones from daventech can be used as 3 axis by directly reading internal registers over the i2c
[15:29] <Laurenceb> the center point of the swinging could change
[15:29] <mc-> or a crude method might be to have 4 sun sensors, then you have an approx heading.
[15:30] <mc-> and could interpolate between 2 sensors to get more accuracy.
[15:30] <Laurenceb> hmm that might not be bad
[15:30] <mc-> and use the GPS as the main source of heading, and the sun sensors as a sanity check
[15:31] <mc-> when the GPS has a slow speed.
[15:31] <Laurenceb> dont think the gps having a low speed should be a problem
[15:31] <Laurenceb> plane glides fast, and fly it when wind is low
[15:32] <mc-> I meant, when the glider is heading into wind, then the GPS speed will be low
[15:32] <Laurenceb> yes... but hopefully the wind will not be as fast as it is flying
[15:33] <Laurenceb> even if it is, the gps is good above about 3m/s, so the probability of it exactly cancelling is small
[15:33] <mc-> I guess if the wind is fast, then it doesn't matter which way the glider heads..
[15:34] <Laurenceb> ie wind cancels gliding speed to within 3m/s
[15:34] <Laurenceb> if wind is fast, dont fly lol
[15:34] flowolf (n=flowolf@host81-187-static.13-87-r.retail.telecomitalia.it) left irc: "Leaving"
[15:35] <Laurenceb> this thing glides at around 30mph, and appears to more or less recover from 20mph gusts
[15:35] <mc-> that's impressive
[15:35] <Laurenceb> yes, but the weather is a pain
[15:36] <Laurenceb> I cant seperate oscillation from gusts of wind on the graphs
[15:36] <Laurenceb> yesterday there were 10mph gusts according to forecast
[15:37] flowolf (n=flowolf@host81-187-static.13-87-r.retail.telecomitalia.it) joined #highaltitude.
[15:37] <mc-> you could have 2 sun sensors if you know the general wind direction, and aim them at the sun, so the glider flies a set heading.
[15:37] <Laurenceb> on test M there was a large gust of side wind just after launch, but you can see it recovering
[15:38] <mc-> sorry, gotta go, bbl
[15:38] <Laurenceb> cya
[15:38] mc- (n=mfcastle@cpc2-glfd1-0-0-cust995.glfd.cable.ntl.com) left irc:
[15:41] edmoore (n=edwardmo@88.212.167.121) joined #highaltitude.
[15:49] jcoxon (n=jcoxon@85-210-179-186.dsl.pipex.com) left irc: "Leaving"
[16:30] <Laurenceb> simulation is working ;-)
[16:39] <Laurenceb> seems the "carry forward" idea for approximating the current gps heading is about 3 times as efficient :)
[16:54] <Laurenceb> all up on wiki
[17:18] mc- (n=mfcastle@cpc2-glfd1-0-0-cust995.glfd.cable.ntl.com) joined #highaltitude.
[17:25] <mc-> Laurence, I've been looking at your simcode. Looks good
[17:25] <mc-> Are you filtering the heading?
[17:31] <mc-> Also I don't understand why you're working at 50Hz, when the GPS outputs heading at 1Hz?
[17:32] <mc-> How about a lookup table, which converts the heading offset to a rudder position? Then you can have a non-linear table
[17:32] edmoore (n=edwardmo@88.212.167.121) left irc:
[17:46] <Laurenceb> hi erm i'm simulating the planes yaw behaviour for 1 second at 50hz
[17:47] <Laurenceb> I think the rudder is pretty linear from the results with the plane
[17:47] <Laurenceb> its limited to +-30 degree displacement
[17:47] <Laurenceb> by the code on the glider
[17:50] <Laurenceb> simulating the plane at higher rate than 1hz shows how the slow 1hz update rate reacts with the faster response times of the plane
[17:52] <Laurenceb> I'm trying to make it self optimising, but the variable for "effectiveness" goes out of control grrr
[17:53] <mc-> Do you start the sim off with the heading away from the target heading?
[17:54] <Laurenceb> you can input the starting heading when you start it running
[17:54] <mc-> so you enter -40 or +40 as the initial heading?
[17:54] <Laurenceb> yep
[17:55] <Laurenceb> rudder constant=1.5 and damping constant=0.4
[17:55] <Laurenceb> for real plane like behaviour
[17:55] <mc-> does the rudder position get updated at 50Hz?
[17:56] <Laurenceb> rudder constant->size of rudder and damping->damping effect of aerodynamics
[17:56] <Laurenceb> no, 1 hz, same in the code
[17:57] <mc-> so what gets done at 50Hz?
[17:57] <Laurenceb> the 50hz is the heading of the plane
[17:57] <mc-> but you know the heading at 1Hz?
[17:57] <Laurenceb> could probably be done with an equation but i'm lazy and have a fast pc here
[17:57] <mc-> ok, now I understand
[17:58] <Laurenceb> basically its numerically integrating over 1 second at 50hz
[17:58] <mc-> got it
[17:59] <mc-> I'm going to try a table method..and see if it can stop the oscillation
[17:59] <Laurenceb> you can see that with the current heading approximation technique its better
[17:59] <Laurenceb> what would you look up in the table?
[18:11] <mc-> I was thinking of converting a heading offset to a rudder setting in the table
[18:19] <Laurenceb> yey autooptimising is working!
[18:20] <Laurenceb> D=-0.3389 for the unapproximated system
[18:20] <Laurenceb> when I set P=-0.15 and i=-0.03
[18:34] <Laurenceb> okay i'd better be off cya all
[18:36] <Laurenceb> from the autooptimiser it looks like D can be reduced a bit
[18:36] Laurenceb (n=Laurence@host81-154-84-2.range81-154.btcentralplus.com) left irc:
[18:41] <mc-> I thought you could reduce D...
[18:41] <mc-> I wrote some code with a single point as a table, it made a big difference
[18:54] <mc-> I've sent you an email
[18:54] mc- (n=mfcastle@cpc2-glfd1-0-0-cust995.glfd.cable.ntl.com) left irc:
[19:23] flowolf (n=flowolf@host81-187-static.13-87-r.retail.telecomitalia.it) left irc: "Leaving"
[19:32] jcoxon (n=jcoxon@85-210-179-186.dsl.pipex.com) joined #highaltitude.
[19:33] <jcoxon> any news about henry's launch?
[20:35] edmoore (n=edwardmo@88.212.167.121) joined #highaltitude.
[20:38] <edmoore> hi all
[21:13] rocketboy (n=steve@host86-140-238-35.range86-140.btcentralplus.com) joined #highaltitude.
[21:14] edmoor1 (n=ed@88.212.167.121) joined #highaltitude.
[21:15] edmoore (n=edwardmo@88.212.167.121) left irc: "toodle-pip"
[21:15] Nick change: edmoor1 -> edmoore
[21:47] flowolf (n=flowolf@host81-187-static.13-87-r.retail.telecomitalia.it) joined #highaltitude.
[22:45] LaurenceB (n=laurence@host86-133-64-30.range86-133.btcentralplus.com) joined #highaltitude.
[22:45] <LaurenceB> Hi all
[22:47] <rocketboy> hi - i see you have been making some progress with your pid - looks like it getting there
[22:48] <jcoxon> hey all
[22:48] <LaurenceB> yes.. its very tricky to seperate wind from oscillations
[22:48] <jcoxon> rocketboy, did you go to the launch today?
[22:48] <rocketboy> yep
[22:48] <rocketboy> nova5
[22:49] <jcoxon> henry said that gps was lost
[22:49] <rocketboy> still you are getting there laurence
[22:50] <rocketboy> jc - yep - it was working then it just sent 0 lat/long/alt - although time was working
[22:50] <jcoxon> oh dear
[22:51] <jcoxon> did the rocket launch then?
[22:51] <rocketboy> It looked like it happended abou the time the rocket should have launched - the telemetry said it had
[22:51] <rocketboy> abou = about
[22:52] <rocketboy> so its lost in suffolk/essex somewhere
[22:52] <jcoxon> radio signal still?
[22:53] <jcoxon> (i'll keep an eye out for it then :-D)
[22:53] <rocketboy> it was still transmitting - I put a post up in the ham radio newsgroup
[22:54] <rocketboy> its probably ok until tomorrow
[22:54] <rocketboy> pm
[22:54] <jcoxon> the payload is pretty large so i'm sure someone will find it
[22:54] <rocketboy> yep
[22:55] <jcoxon> i should really do some work on peg 5
[22:55] <jcoxon> haven't been very well in the last few days -hopefully the antibiotics will make a difference
[22:55] <rocketboy> me 2 xaben2
[22:56] <rocketboy> oh? - what is it?
[22:56] <jcoxon> the antibiotics? :-)
[22:57] <jcoxon> co-amoxiclav
[22:58] <rocketboy> thats a general anti-biotic isn't it - no whats the symptoms?
[22:59] <jcoxon> its a combination antibiotic
[22:59] <jcoxon> oh general stuff such as fever etc
[23:00] <rocketboy> o-key-dohhhh-key - hope u get better
[23:01] <jcoxon> I'll survive :-D
[23:03] <jcoxon> oh you were asking about the radio
[23:03] <rocketboy> the nova5 thing started off really well today - they had google earth projected with a live update from the payload
[23:03] <rocketboy> wayy-cooool
[23:04] <jcoxon> oh cool - novacom works then :-D
[23:04] <rocketboy> yep - but its not as good as truetty at the mo
[23:05] <jcoxon> oh right
[23:05] <jcoxon> basically for the radio i stole off henry
[23:05] flowolf (n=flowolf@host81-187-static.13-87-r.retail.telecomitalia.it) left irc: "Leaving"
[23:05] <jcoxon> i just connected up a ttl serial port tx line
[23:05] <rocketboy> I think all novacom was doing was creating a small KML each reading then getting gooogle earth to re-draw including it
[23:06] <jcoxon> over a divider circuit and "cat"ing data
[23:06] <rocketboy> yeah - sounds like that worked OK
[23:06] <jcoxon> henry managed to decode it on the other side
[23:06] <jcoxon> it was from catz to engineering with a lot of buildings in the way
[23:06] <rocketboy> sounds good - an no data corruption
[23:07] <jcoxon> for Peg V the plan will be
[23:07] <jcoxon> in a minute cycle - morsecode then some serial data
[23:07] <jcoxon> to test the range etc
[23:07] <rocketboy> I found that some characters were particularly suseptable to corruption without the shaping
[23:07] <jcoxon> i'm not sure how perfect it was
[23:08] <jcoxon> i think in the longrun it'll need shaping
[23:08] <rocketboy> welll it sounds like it needs to be tested (but I'd have a phone backup)
[23:08] LaurenceB (n=laurence@host86-133-64-30.range86-133.btcentralplus.com) left irc: Read error: 60 (Operation timed out)
[23:10] <rocketboy> BBL
[23:18] <jcoxon> right i'm off to bed
[23:18] <jcoxon> night all
[23:18] jcoxon (n=jcoxon@85-210-179-186.dsl.pipex.com) left irc: "Leaving"
[23:25] LaurenceB (n=laurence@host86-133-64-30.range86-133.btcentralplus.com) joined #highaltitude.
[23:25] <LaurenceB> Hi again
[23:25] <LaurenceB> so what happened at the launch
[23:26] <LaurenceB> my PC crashed , think the HDD is on the blink
[23:26] <LaurenceB> isnt the a dos thnk for checking for bad sectors?
[23:26] <LaurenceB> the-there
[23:28] <LaurenceB> did the rockoon launch?
[23:32] <rocketboy> the GPS on NOVA5 stoped reporting position (but time OK) about the same time as the rocket launched
[23:32] <LaurenceB> oh no did you recover it?
[23:32] Ebola (n=Ebola@host86-143-156-147.range86-143.btcentralplus.com) left irc: "gone to bed"
[23:33] <rocketboy> no - we don't know where it is
[23:33] <LaurenceB> dam thats not good
[23:33] <LaurenceB> what gps were you using
[23:34] <rocketboy> the last good position was about over the Mihab 2 landing spot
[23:34] <rocketboy> hedding east
[23:34] <rocketboy> hedding = heading
[23:34] <LaurenceB> I see
[23:35] <rocketboy> we think it will have landed in suffolk/essex
[23:35] <LaurenceB> maybe the rocket somehow messed up the gps...
[23:35] <rocketboy> thats a definate possibility
[23:36] <LaurenceB> after trying out this anataris chipset, I've becided I like sirf2, it seems very good for gps units a few years old
[23:37] <LaurenceB> always gets a lock within one minute in the open
[23:39] <LaurenceB> something is very wrong with my pc...
[23:39] <rocketboy> it probably just needs a good kicking
[23:39] <LaurenceB> is it chkdsk or something?
[23:40] <rocketboy> you mean the unix fdisk checker
[23:40] <LaurenceB> erm this is windows hence the problems
[23:41] <rocketboy> scan disk
[23:42] <LaurenceB> I would use google but that will crash it
[23:43] <LaurenceB> just typing too quickly it jams up
[23:46] <rocketboy> left click on the disk icon
[23:46] <rocketboy> select properties
[23:46] <rocketboy> tools tab
[23:46] <rocketboy> then check disk
[23:52] edwardmoore (n=edwardmo@88.212.167.121) joined #highaltitude.
[23:56] rocketboy (n=steve@host86-140-238-35.range86-140.btcentralplus.com) left irc: "Leaving"
[00:00] --- Sun Mar 25 2007