highaltitude.log.20081124

[00:15] Laurenceb (i=83e3dd6d@gateway/web/ajax/mibbit.com/x-8dd6e037f416d60d) left irc: "http://www.mibbit.com ajax IRC Client"
[01:02] Laurenceb (i=83e3dd6d@gateway/web/ajax/mibbit.com/x-091aa7f71adc726b) joined #highaltitude.
[01:59] natrium42 (n=alexei@CPE000625d867e2-CM0014045885be.cpe.net.cable.rogers.com) joined #highaltitude.
[05:27] edmoore (n=edmoore@88-212-167-121.rdns.as8401.net) joined #highaltitude.
[05:29] edmoore (n=edmoore@88-212-167-121.rdns.as8401.net) left irc: Client Quit
[06:34] Laurenceb (i=83e3dd6d@gateway/web/ajax/mibbit.com/x-091aa7f71adc726b) left irc: "http://www.mibbit.com ajax IRC Client"
[08:07] Simon-MPFH (n=simon@phantom.mpfh.co.uk) joined #highaltitude.
[08:42] Laurenceb (i=83e34e2c@gateway/web/ajax/mibbit.com/x-df374dc07500ab60) joined #highaltitude.
[09:48] fergusnoble_ (n=fergusno@fn217.quns.cam.ac.uk) joined #highaltitude.
[09:49] fergusnoble (n=fergusno@fn217.quns.cam.ac.uk) left irc: Read error: 104 (Connection reset by peer)
[10:24] edmoore (n=edmoore@pomegranate.chu.cam.ac.uk) joined #highaltitude.
[10:46] <Laurenceb> hi edmoore
[10:46] <Laurenceb> edmoore: http://wiki.ukhas.org.uk/code:4_state_extended_kalman_filter_in_matlab
[10:46] <Laurenceb> I'm just constructing the page atm
[10:51] edmoore (n=edmoore@pomegranate.chu.cam.ac.uk) left irc: Read error: 113 (No route to host)
[13:16] edmoore (n=edmoore@chu-gw.churchillcambridge.co.uk) joined #highaltitude.
[13:27] <Laurenceb> hi edmoore
[13:34] <edmoore> yo
[13:34] <edmoore> sorry, having a work bash
[13:34] Nick change: edmoore -> edmoore|working
[13:34] <Laurenceb> ok
[13:35] <Laurenceb> dont mean to interrupt but I had a mad idea
[13:35] <Laurenceb> why not hack the lassen iq?
[13:35] <Laurenceb> it has rom
[13:35] <Laurenceb> parrallel rom on the board
[13:35] <Laurenceb> and there a c compiler for the baseband processor
[13:45] edmoore (n=edmoore@chu-gw.churchillcambridge.co.uk) joined #highaltitude.
[13:48] edmoore (n=edmoore@chu-gw.churchillcambridge.co.uk) left irc: Client Quit
[13:57] edmoore (n=edmoore@chu-gw.churchillcambridge.co.uk) joined #highaltitude.
[13:58] edmoore|working (n=edmoore@chu-gw.churchillcambridge.co.uk) left irc: Read error: 104 (Connection reset by peer)
[14:17] <Laurenceb> edmoore: still working?
[14:22] <edmoore> ok i just read through your stuff
[14:22] <edmoore> so
[14:22] <edmoore> first q: how do you get 5hz out of a gps?
[14:22] <edmoore> what does it transmit often enough?
[14:23] <edmoore> or is it just the very presence of a data and you know the time the data comes in on a given code
[14:23] <edmoore> relative to other data on different codes
[14:24] <edmoore> Laurenceb: ping
[14:25] <edmoore> back v briefly
[14:25] <Laurenceb> hi
[14:25] <Laurenceb> 5hz?
[14:26] <Laurenceb> where did I say 5hz?
[14:26] <Laurenceb> you cant get that out of the iq
[14:26] <Laurenceb> but you can go as high a 1KHz in the general case
[14:26] <edmoore> no, but the ublox, say
[14:26] <Laurenceb> as your correlators fire off every 1ms
[14:26] <Laurenceb> sorry I misunderstood
[14:27] <edmoore> so you know a correlator corresponds to a sat
[14:27] <Laurenceb> I think its the processing overheads
[14:27] <edmoore> you know the location of the sat
[14:27] <edmoore> you know the fire time of the correlator
[14:27] <Laurenceb> they tend to run the processors slowly
[14:27] <Laurenceb> and dont have an fpu
[14:27] <Laurenceb> you can leave the correlators coasting by them selves
[14:27] <edmoore> sure, but for out hypothetical tightly couples gps/ins running on a dsp fpga
[14:27] <Laurenceb> so yuo you can go really fast
[14:27] <Laurenceb> up to a KHz with ease
[14:28] <edmoore> cool
[14:28] <edmoore> so we could run the whole thing at 50hz then, conservatively
[14:28] <Laurenceb> anyway, I was thinking you could reflash the lassen
[14:28] <Laurenceb> yeah no probs doing that
[14:28] <Laurenceb> theres a c compiler for the baseband processor
[14:28] <edmoore> is it n arm?
[14:28] <Laurenceb> and it uses external flash
[14:28] <Laurenceb> no C33
[14:29] <Laurenceb> its a DSP made by epsom and Ti
[14:29] <Laurenceb> but Epsom appear to have a compiler and IDE
[14:29] <edmoore> ok
[14:29] <edmoore> interesting
[14:29] <Laurenceb> the one in the lassen is a custom job with correlators added
[14:29] <Laurenceb> but I think thats the limit of the mods
[14:30] <Laurenceb> it boots off external prom - looks like SOIC on the photos I've seen
[14:30] <edmoore> this is proper nerd hacking
[14:30] <Laurenceb> I havent tried unencapsulating a lassen yet :P
[14:30] <edmoore> is that definitely easier than starting from scratch?
[14:30] <Laurenceb> probably
[14:30] <Laurenceb> who knows
[14:31] <Laurenceb> the hacking element is just cool
[14:31] <edmoore> is the epson compiler free?
[14:31] <Laurenceb> looks like it
[14:31] <Laurenceb> I dont have admin acess here
[14:31] <Laurenceb> so cant try it out
[14:31] <Laurenceb> http://www.epson.jp/device/semicon_e/product/mcu/32bit/index.htm#ac10
[14:32] <Laurenceb> so, advantage is correlators for free
[14:32] <Laurenceb> and frontend all built
[14:32] <edmoore> cool
[14:32] <Laurenceb> diadvantage we bont know how its interfaced
[14:32] <edmoore> very cool
[14:32] <edmoore> indeed - that's quite a biggie
[14:32] <Laurenceb> but its sure as hell going to be onto the address bus
[14:32] <Laurenceb> so you could get some probe code running
[14:33] <Laurenceb> fairly easily I would guess
[14:33] <Laurenceb> theres a uart to dump your data over :P
[14:33] <Laurenceb> incidentally, theres a feature in the existing code to input rate gyros and accels
[14:34] <Laurenceb> but only for 2D filtering, its designed for in car systems
[14:34] <Laurenceb> seems to be common to all Trimbles firmware
[14:35] <edmoore> neat
[14:35] <Laurenceb> oh I got an "artificial horizon" working
[14:36] <Laurenceb> IMO euler angles are about as abstract as quaternions iif your trying to fly something
[14:36] <Laurenceb> you need to know elevation angles of the horizon
[14:36] <Laurenceb> to input into the control loops
[14:38] <Laurenceb> the maths was a pain
[14:39] <Laurenceb> anyway its up on the wiki now
[14:41] <edmoore> ok will have a look
[14:41] <edmoore> going to bury my head into this maths for now
[14:53] <Laurenceb> oh wow
[14:53] <Laurenceb> looks like EPSOM have ported GCC
[14:53] <Laurenceb> to their processor
[14:53] <Laurenceb> it wasnt listed on the wain GCC page
[14:53] <Laurenceb> *main
[15:20] <Laurenceb> http://www.besprovodka.ru/catalog/products_pictures/lassen_sq_2.jpg
[15:20] <Laurenceb> I did think the IC at the rhs corner was the flash
[15:20] <Laurenceb> but theres not enough pins
[15:22] <Laurenceb> oh it uses a bidirectional bus
[15:22] <Laurenceb> nvm that is almost certainly it
[16:21] <Laurenceb> oh wow
[16:21] <Laurenceb> http://www.datasheetcatalog.org/datasheet/Xemics/mXyyztzt.pdf
[16:21] <Laurenceb> that soic IC is the collosus
[16:22] <Laurenceb> holy &^%$ I thought it would be bga
[16:22] <Laurenceb> screw the SE4120
[16:23] <Laurenceb> you can see it more clearly here http://www.navgeocom.ru/oem/gen/lassensq/lassensq_big_2.jpg the BGA is the baseband
[16:28] <edmoore> Laurenceb: am super interested, will brb
[16:28] <edmoore> need to visit the dept
[16:28] edmoore (n=edmoore@chu-gw.churchillcambridge.co.uk) left irc:
[17:10] edmoore (n=edmoore@chu-gw.churchillcambridge.co.uk) joined #highaltitude.
[17:13] <Laurenceb> wb
[17:14] <Laurenceb> I'm confused by this processor, is it Harvard architcture or not? it says it is, but the instructions and data seem to exist in the same address space
[17:20] <Laurenceb> god I need to get out of here
[17:20] <Laurenceb> everyones playing spot the IC again
[17:20] Laurenceb_ (i=83e34f19@gateway/web/ajax/mibbit.com/x-25d6689e0b31e74f) joined #highaltitude.
[17:23] Laurenceb (i=83e34e2c@gateway/web/ajax/mibbit.com/x-df374dc07500ab60) left irc: "http://www.mibbit.com ajax IRC Client"
[17:28] fergusnoble_ (n=fergusno@fn217.quns.cam.ac.uk) left #highaltitude.
[17:45] <Laurenceb_> edmoore: there is a probably simpler solution: http://www.datasheetcatalog.com/datasheets_pdf/G/P/2/0/GP2021.shtml
[17:45] <EI5GTB> evening
[17:46] <Laurenceb_> something like that, but I like the idea of getting a collossus out of a lassen
[17:46] <Laurenceb_> hi
[17:46] <edmoore> ok
[17:46] <edmoore> so a 12ch gps just has 12 correlators
[17:47] <Laurenceb_> yes
[17:47] <Laurenceb_> it may be easier
[17:47] <Laurenceb_> you'd have to map out the correlators in the lassens baseband processor
[17:47] <edmoore> so ublox keep rabbiting on about 1000000 correlators - some kind of wierd pseudo-truth?
[17:47] <Laurenceb_> hehe
[17:47] <Laurenceb_> completely ott
[17:48] <Laurenceb_> on that photo of the main board I think the shiny square thing by the ant connector is a saw filter
[17:48] <Laurenceb_> then the TCXO is in the far corner
[17:48] <Laurenceb_> the flash must be on the other side of the bopard
[17:49] <Laurenceb_> you should be able to get an idea of the internal address structure by monitoring the address and data busses where they go through to the flash with a logic anyliser
[17:49] <Laurenceb_> but its a crazy job
[17:50] <Laurenceb_> theres a lot of vias around the baseband, some may be there as test pads
[17:52] <Laurenceb_> but yeah I'd say maybe try cannibalising a lassen for the collossus and SAW filter, then make a new pcb with a seperate correlator IC
[17:53] <Laurenceb_> alternatively, the RF solutions GPS modules use this setup http://www.datasheetcatalog.org/datasheet/Xemics/mXyyztzt.pdf
[17:53] <Laurenceb_> the serial interface is proprietory, but it may be easy to hack
[17:53] Ebola (n=Ebola@unaffiliated/ebola) joined #highaltitude.
[17:57] edmoore (n=edmoore@chu-gw.churchillcambridge.co.uk) left irc:
[18:11] Laurenceb_ (i=83e34f19@gateway/web/ajax/mibbit.com/x-25d6689e0b31e74f) left irc: "http://www.mibbit.com ajax IRC Client"
[18:59] edmoore (n=edmoore@chu-gw.churchillcambridge.co.uk) joined #highaltitude.
[19:24] natrium42 (n=alexei@CPE000625d867e2-CM0014045885be.cpe.net.cable.rogers.com) left irc:
[19:44] Laurenceb (i=83e3dd6d@gateway/web/ajax/mibbit.com/x-f9695dbd84bb90e4) joined #highaltitude.
[19:44] <Laurenceb> ello
[19:50] <edmoore> yo
[19:50] <Laurenceb> so, I spoke to the experts about a gps receiver on a HAB launched rocket
[19:51] <Laurenceb> the conclusion is some sort of front end like a Collossus or SE4120 then a GP2021 correlator
[19:51] <Laurenceb> its hard enough as it is
[19:51] <Laurenceb> dont want to make it too tricky
[19:51] <edmoore> i.e. not a from-scratch fpga effort?
[19:51] <Laurenceb> prob not
[19:51] <edmoore> does that give us all the info we need?
[19:52] <edmoore> like doppler?
[19:52] <Laurenceb> yes
[19:52] <Laurenceb> its a fully configurable set of 12 correlators
[19:52] <edmoore> ok
[19:52] <Laurenceb> thing is its designed to go onto a data bus
[19:52] <edmoore> something awesome for DSP?
[19:52] <Laurenceb> so you can memory map it to an ARM or similar
[19:53] <edmoore> well we can put stuff on the mem bus or whatever processor we end up using
[19:53] <Laurenceb> yeah
[19:54] <Laurenceb> I may be able to get you one of the ICs
[19:54] <edmoore> the correlator?
[19:54] <Laurenceb> but they are no longer in production
[19:54] <Laurenceb> yes
[19:54] <Laurenceb> so they have a large stash built up
[19:55] <Laurenceb> but http://www.gpscreations.com/NewFiles/GPS500_501%20Brochure.pdf might be better
[19:56] <Laurenceb> thats what my supervisor recommended
[19:57] <edmoore> cool
[19:57] <Laurenceb> so its a GP2021 with front end
[20:07] <Laurenceb> http://wiki.ukhas.org.uk/code:4_state_extended_kalman_filter_in_matlab <-- I've written a decent description (hopefully)
[20:08] <edmoore> will take a look in a bit - still bashing work
[20:41] edmoore (n=edmoore@chu-gw.churchillcambridge.co.uk) left irc:
[20:59] Hiena (n=Hiena@81.93.195.181.datatrans.hu) joined #highaltitude.
[21:17] Hiena (n=Hiena@81.93.195.181.datatrans.hu) left irc: "-=Halt! Hammerzeit!=-"
[21:23] Ei5GTB_ (n=Paul@78.16.132.169) joined #highaltitude.
[21:30] Simon-MPFH (n=simon@phantom.mpfh.co.uk) left irc: "Leaving"
[21:31] Laurenceb (i=83e3dd6d@gateway/web/ajax/mibbit.com/x-f9695dbd84bb90e4) left irc: "http://www.mibbit.com ajax IRC Client"
[21:31] Laurenceb (i=83e3dd6d@gateway/web/ajax/mibbit.com/x-ecf629cd2682c44f) joined #highaltitude.
[21:34] natrium42 (n=alexei@CPE000625d867e2-CM0014045885be.cpe.net.cable.rogers.com) joined #highaltitude.
[21:40] EI5GTB (n=Paul@78.16.153.30) left irc: Connection timed out
[21:52] Ebola (n=Ebola@unaffiliated/ebola) left irc: Remote closed the connection
[22:00] Nick change: Ei5GTB_ -> EI5GT
[22:00] Nick change: EI5GT -> EI5GTB
[22:37] jcoxon (n=jcoxon@host86-163-198-244.range86-163.btcentralplus.com) joined #highaltitude.
[22:41] <jcoxon> evening all
[22:44] jcoxon (n=jcoxon@host86-163-198-244.range86-163.btcentralplus.com) left irc: "Leaving"
[23:16] edmoore (n=edmoore@pomegranate.chu.cam.ac.uk) joined #highaltitude.
[23:50] edmoore (n=edmoore@pomegranate.chu.cam.ac.uk) left irc:
[00:00] --- Tue Nov 25 2008