highaltitude.log.20090201

[00:00] <Laurenceb> maybe 100nm or something
[00:00] <Laurenceb> but they also talk about aluminium oxide... I dont get it
[00:01] <shellevil> isn't that just as an insulator?
[00:01] <shellevil> It's been a while since I looked at it.
[00:01] <Laurenceb> maybe...
[00:01] <Laurenceb> insulator where?
[00:03] <shellevil> Dunno.
[00:04] <Laurenceb> I think I get it now
[00:04] <Laurenceb> youre basically making a single crystal at 180C
[00:04] <shellevil> It's one of those huge class of nifty ideas that if you can go from one produced in the lab for a million dollars to one produced in a factory for a tenth of a cent, you can make obscene amounts of money
[00:05] <shellevil> the PET is just a filler because making huge flat crystals is insanely hard ?
[00:05] <Laurenceb> you have the nm sized single crystals, then enbed them in polymer, its not very useful
[00:05] <Laurenceb> the key is the 180C + HV
[00:05] <Laurenceb> which alignes the small crystals
[00:05] <shellevil> into effectively one big one
[00:06] <Laurenceb> so for all intents and purposes you now have one large single domain crystal
[00:06] <shellevil> like magnets.
[00:06] <Laurenceb> yeah
[00:06] <Laurenceb> its like an artifical polymer based magnet
[00:06] <Laurenceb> without th eB field
[00:07] Action: shellevil ponders why NiB polymer magnets aren't readily available.
[00:07] <shellevil> 80% NiB, 20% rubber binder.
[00:08] <shellevil> I suppose they'd corrode quite badly
[00:09] <shellevil> So what're you planning to do GPS wise - at the moment you're just playing with algorithms in matlab?
[00:10] <Laurenceb> and C
[00:10] <Laurenceb> hallam has a blackfin board
[00:10] <Laurenceb> the plan is to rig it up to the sige sampler
[00:10] <shellevil> Should be plenty
[00:11] <Laurenceb> yeah, it has DMA and a countones instruction
[00:11] <shellevil> I'm planning on _much_ smaller hardware - for a limited but useful GPS.
[00:11] <shellevil> The sort of hardware I can hopefully sell at 20 quid on ebay and make a profit.
[00:11] <shellevil> ASM is fun! :)
[00:12] <shellevil> probably rather lousy lock speed forex.
[00:13] <shellevil> If no ephemreis or time/position
[00:14] <Laurenceb> the surveyor board is quite cheap, might buy one for myself
[00:15] <Laurenceb> I'm thinking of sticking a port in the side of my sige sampler to get out raw data
[00:15] <shellevil> It also occured to me that a really modest FPGA can do this shit standing on its head.
[00:15] <Laurenceb> just need to get the case open without smashing it completely
[00:15] <shellevil> But that raises issues of mounting it.
[00:15] Action: shellevil loves hot-glue.
[00:15] <Laurenceb> yeah
[00:15] <Laurenceb> hehe
[00:15] <shellevil> two bits of wood, hot-glue the wood to the case.
[00:16] <shellevil> apply vice to force them apart
[00:16] <Laurenceb> erm
[00:16] <shellevil> apply heat to remove glue
[00:16] <Laurenceb> I'll probably drill in where the sockets going to be
[00:16] edmoore (n=edmoore@pomegranate.chu.cam.ac.uk) left irc:
[00:16] <Laurenceb> then use the hole to level the two halves apart
[00:16] <shellevil> you don't know the clearance to the board.
[00:17] <shellevil> Is it 'just' a USB dongle?
[00:17] <Laurenceb> I can see in around the usb
[00:17] <Laurenceb> it has a usb mini b socket
[00:17] <shellevil> 99.9% of those come apart if you force a sharp fingernail in along the crack, and work it along.
[00:17] <shellevil> check under labels for screws
[00:17] <Laurenceb> but is only slightly larger than a standard dongle
[00:17] <shellevil> the .1% are ultrasonically or solvent welded
[00:17] <Laurenceb> nah its glued together very well
[00:19] <shellevil> don't forget the ESD protection.
[00:20] Action: shellevil hugs 10K resistors.
[00:20] <Laurenceb> to gnd?
[00:24] <shellevil> series helps lots
[00:24] <shellevil> It makes a huge ESD current a much less huge one
[00:24] <shellevil> and means a very modest diode can cope with the current without much spike
[00:30] <Laurenceb> it also screws things up
[00:31] <Laurenceb> by adding 10k to the output impedance
[00:32] <shellevil> For digital signals, clocked at modest speeds into small capacitances, that's generally not really an issue.
[00:32] <shellevil> Though 10K in this case may not work so well.
[00:33] <shellevil> A TTL buffer is also good.
[00:39] <Laurenceb> shellevil: I'm thinking of making my icom portable
[00:40] <shellevil> what's it got - a serial or USB IF - I forget
[00:40] <shellevil> interface
[00:40] <Laurenceb> it needs 13v 550to 600ma, what
[00:40] <Laurenceb> do you think would be best
[00:41] <shellevil> is it specified for 13V?
[00:41] <shellevil> Or a range?
[00:41] <Laurenceb> I priced up NiMh AA cells but its a lot
[00:42] <shellevil> lead acid is cheap.
[00:42] <Laurenceb> 13 +-15%
[00:42] <Laurenceb> yeah, but you have to be careful not to use the full capacity?
[00:42] <shellevil> that's what...
[00:42] <shellevil> 11.2-15V or so
[00:43] <Laurenceb> I mean mah
[00:43] <shellevil> that's basically the 0% charge-fast charging volts of a lead-acid bat
[00:43] <Laurenceb> you only really want to use 50% of the rated capacity?
[00:44] <shellevil> how many hours are you planning to use this for - and I thought you'd got a radio with no external i/f.
[00:44] <shellevil> display, controls, ...
[00:44] <Laurenceb> ideally about 4 hours
[00:44] <Laurenceb> yeah its rs232
[00:44] <shellevil> got any laptop?
[00:45] <Laurenceb> yes
[00:45] <Laurenceb> but I dont want to deplete the battery that much
[00:45] <shellevil> do you care about weight?
[00:45] <Laurenceb> only 3 hours max
[00:45] <Laurenceb> not a great deal
[00:45] <Laurenceb> I'd hope for <1kg
[00:46] <shellevil> Ok - that's 18Wh
[00:47] <Laurenceb> wheres a good place ot buy mini deep cycle cells?
[00:47] <shellevil> http://cgi.ebay.co.uk/12V-2-3Ah-sealed-lead-acid-battery-Low-Postage_W0QQitemZ250365163704QQcmdZViewItemQQptZUK_ConsumerElectronics_Batteries_SM?hash=item250365163704&_trksid=p3286.c0.m14&_trkparms=72%3A1683%7C66%3A2%7C65%3A12%7C39%3A1%7C240%3A1318
[00:48] <shellevil> 2.2Ah - so it'll go about 3.5 hours
[00:49] <shellevil> you either need to compromise on weight, or go NiMh or lithium
[00:49] <shellevil> I
[00:50] <shellevil> I'd expect that to last >50 cycles to 80%
[00:50] <shellevil> you don't have a car do you?
[00:51] <shellevil> this is for payload location?
[00:51] <Laurenceb> nope
[00:52] <Laurenceb> yeah, for groundstation use
[00:52] <Laurenceb> lead acid is a lot cheaper, NiMh works out about £25
[00:57] <shellevil> what - wander up a hill?
[01:01] <Laurenceb> or take to launches
[01:22] fergusnoble (n=fergusno@fn217.quns.cam.ac.uk) joined #highaltitude.
[02:05] fergusnoble (n=fergusno@fn217.quns.cam.ac.uk) left #highaltitude.
[03:11] <Laurenceb> natrium42_: about?
[03:11] <natrium42_> sup?
[03:12] <Laurenceb> hi, I'm trying to get a makefile for avr to run on ubuntu
[03:12] <Laurenceb> runs fine on winxp
[03:12] <Laurenceb> http://pastebin.com/d3d30fe74
[03:13] <Laurenceb> its giving me "no rule to make target radio.c"
[03:13] wickerwaka (n=akawaka@66-215-97-198.dhcp.malb.ca.charter.com) joined #highaltitude.
[03:15] <natrium42_> yeah, i don't see any rules to build .o files from .c files
[03:15] <natrium42_> oh
[03:15] <natrium42_> nvm
[03:16] <natrium42_> why is there a space between the lines?
[03:16] <natrium42_> 111-113
[03:17] <Laurenceb> thats something to do with the c & p
[03:18] <wickerwaka> ?
[03:19] <Laurenceb> what aboit 111 to 113?
[03:21] <wickerwaka> whats c & p?
[03:22] <Laurenceb> copy and paste
[03:22] <Laurenceb> I think I got it - its a problem with file names
[03:22] <wickerwaka> makefile problem?
[03:26] <Laurenceb> nope fixed :P
[03:30] <wickerwaka> makefiles are a bucket of fun
[03:40] <Laurenceb> http://wiki.ukhas.org.uk/code:radio
[03:40] <Laurenceb> finito
[03:44] <Laurenceb> 1.4KB
[04:12] Laurenceb (n=laurence@dyres221-140.surrey.ac.uk) left irc: "The day microsoft make something that doesnt suck is the day they make a vacuum cleaner"
[06:58] Xenion (n=robert@p579FCD56.dip.t-dialin.net) left irc: "Verlassend"
[08:14] rjharrison (n=rharriso@80.176.172.227) joined #highaltitude.
[08:17] Simon-MPFH (n=simon@phantom.mpfh.co.uk) joined #highaltitude.
[08:30] wickerwaka (n=akawaka@66-215-97-198.dhcp.malb.ca.charter.com) left irc: Read error: 110 (Connection timed out)
[09:50] Hiena (n=Hiena@81.93.195.181.datatrans.hu) joined #highaltitude.
[10:06] jcoxon (n=jcoxon@host86-158-31-172.range86-158.btcentralplus.com) joined #highaltitude.
[10:12] <jcoxon> morning all
[10:15] <Hiena> Where?
[10:19] <jcoxon> where what?
[10:22] <Hiena> The morning. I can't see anything. Everything is white....
[10:23] <jcoxon> hehe
[10:23] <jcoxon> no snow here...
[10:23] <jcoxon> yet
[10:47] SpeedEvil (n=kfkfkfkf@tor/regular/SpeedEvil) joined #highaltitude.
[11:04] G8KHW (n=Steve@217.47.75.27) joined #highaltitude.
[11:04] SpeedEvil (n=kfkfkfkf@tor/regular/SpeedEvil) left irc: "[BX] Oral sex makes your day"
[11:06] SpeedEvil (n=jfjfjf@tor/regular/SpeedEvil) joined #highaltitude.
[11:11] hallam (i=836fc8c8@gateway/web/ajax/mibbit.com/x-acb2430675c20163) joined #highaltitude.
[11:13] edmoore (n=edmoore@pomegranate.chu.cam.ac.uk) joined #highaltitude.
[11:30] Simon-MPFH (n=simon@phantom.mpfh.co.uk) left irc: Read error: 104 (Connection reset by peer)
[11:31] simon_ (n=simon@phantom.mpfh.co.uk) joined #highaltitude.
[11:41] <rjharrison> morning al
[11:41] <rjharrison> all
[11:48] <hallam> Hi rjharrison
[11:48] <hallam> edmoore: think that email is likely to net me a JTAG?
[11:48] <edmoore> possibly
[11:48] <edmoore> working on it.
[11:48] <edmoore> will report back
[11:49] <edmoore> i'm fairly cool with the jtag - though is there one for that money (or maybe a bit more) that will do everything?
[11:49] <edmoore> i.e. FPGAs, ARMs, XScales, AVRs etc
[11:50] <hallam> I believe that one will do other JTAGgy things
[11:50] <hallam> certainly it talks about daisy-chainign to other devices
[11:50] <edmoore> cool
[11:50] <hallam> I think the really high end ones cost in the high 3 digits or more though
[11:50] <edmoore> sure they do, yeah
[11:50] <edmoore> and they're always ethernet devices or something
[11:53] <hallam> sainsbury's time
[11:53] <hallam> ttyl
[11:53] Nick change: hallam -> hallam|out
[11:59] Action: SpeedEvil starts a campaign that all devices have to have the option of being bought unprogrammed with SDK.
[12:00] Tigga (n=chatzill@pc-232-235-103.magd.cam.ac.uk) joined #highaltitude.
[12:03] <jcoxon> hey rjharrison
[12:03] <gordonjcp> morning
[12:03] <edmoore> morning
[12:03] <jcoxon> hey guys
[12:04] <edmoore> all well?
[12:05] Action: SpeedEvil had a great flood!
[12:05] <edmoore> ?
[12:05] <SpeedEvil> Well - a little one.
[12:05] <SpeedEvil> Power went out last night.
[12:05] <SpeedEvil> Air pump to fishtank went off.
[12:06] <SpeedEvil> Air in lines cooled and drew water back out of fishtank to air junction box.
[12:06] <SpeedEvil> Water dissolved tape I'd used to seal junction box.
[12:06] <SpeedEvil> Water siphoned out over carpet.
[12:06] <SpeedEvil> :/
[12:07] <SpeedEvil> Other than that, not much.
[12:07] <SpeedEvil> actually
[12:07] <edmoore> eeeek
[12:07] <SpeedEvil> it's more likely that the pipe blew off due to high pressure due to the trapped water and no outlet.
[12:08] <SpeedEvil> As I don't recall any tape.
[12:09] <gordonjcp> much water loss?
[12:12] <SpeedEvil> 30 gallons ish.
[12:15] Bluenarf (n=Paul@apollo.paulsnet.org) joined #highaltitude.
[12:16] Nick change: Bluenarf -> EI5GTB
[12:22] <gordonjcp> SpeedEvil: eep
[12:22] <gordonjcp> fish okay?
[12:23] <SpeedEvil> yeah.
[12:23] <SpeedEvil> Carpet too.
[12:24] <SpeedEvil> Just an annoying ages with a wet/dry vac and some carpet shampoo.
[12:34] shellevil (n=user@mauve.plus.com) left irc: Read error: 113 (No route to host)
[12:40] shellevil (n=user@mauve.plus.com) joined #highaltitude.
[12:41] fergusnoble (n=fergusno@fn217.quns.cam.ac.uk) joined #highaltitude.
[12:55] Tigga (n=chatzill@pc-232-235-103.magd.cam.ac.uk) left irc: Read error: 104 (Connection reset by peer)
[12:57] Tigga (n=chatzill@pc-232-235-103.magd.cam.ac.uk) joined #highaltitude.
[13:06] simon_ (n=simon@phantom.mpfh.co.uk) left irc: Read error: 104 (Connection reset by peer)
[13:06] simon_ (n=simon@phantom.mpfh.co.uk) joined #highaltitude.
[13:09] G8KHW (n=Steve@217.47.75.27) left irc: Read error: 104 (Connection reset by peer)
[13:24] jcoxon (n=jcoxon@host86-158-31-172.range86-158.btcentralplus.com) left irc: "Leaving"
[13:36] simon_ (n=simon@phantom.mpfh.co.uk) left irc: "Leaving"
[13:42] hallam (i=836fc8c8@gateway/web/ajax/mibbit.com/x-614c4d195db309ed) joined #highaltitude.
[13:44] hallam|out (i=836fc8c8@gateway/web/ajax/mibbit.com/x-acb2430675c20163) left irc: "http://www.mibbit.com ajax IRC Client"
[14:09] edmoore (n=edmoore@pomegranate.chu.cam.ac.uk) left irc:
[14:48] G8KHW (n=Steve@217.47.75.27) joined #highaltitude.
[15:06] <fergusnoble> natrium42_: did you say you lived in Kitchener?
[15:06] Tigga (n=chatzill@pc-232-235-103.magd.cam.ac.uk) left irc: Read error: 104 (Connection reset by peer)
[15:22] <hallam> Hey fergusnoble
[15:22] <fergusnoble> hallam: hi
[15:22] <hallam> holy cow, SNOW
[15:23] <hallam> also, any chance you could download ubuntu for me?
[15:38] <fergusnoble> hallam: sure, what for?
[15:38] <fergusnoble> i thought your little fling with linux yesterday was enough to put you off
[15:38] <hallam> since I updated ubuntu it won't boot
[15:39] <hallam> so I need to reinstall
[15:39] <fergusnoble> hehe
[15:39] <hallam> I have to get this blackfin toolchain working somehow
[15:39] <fergusnoble> all part of the fun
[15:39] <fergusnoble> and its not working on linux?
[15:39] <fergusnoble> sorry, not working on windows?
[15:40] <hallam> still haven't been able to get the libraries to work
[15:40] <hallam> and they say that the toolchain support isn't as good on windows as it is on linux
[15:40] <hallam> e.g. you can build most apps for uclinux but not the kernel
[15:41] <fergusnoble> ok
[15:44] jcoxon (n=jcoxon@host86-158-31-172.range86-158.btcentralplus.com) joined #highaltitude.
[15:45] <fergusnoble> hallam: ok, its downloaded
[15:47] <fergusnoble> do you want to come over?
[15:47] <hallam> brill
[15:47] <hallam> sure
[15:47] <hallam> cheers
[15:48] <fergusnoble> the cds are in the ifm still :(
[15:48] <hallam> did you get the DVD or CD version?
[15:48] <fergusnoble> the cd version
[15:48] <hallam> I have DVDs here
[15:48] <hallam> ok
[15:48] <hallam> can you burn a CD iso to a DVD?
[15:48] <fergusnoble> ed got some cdrs
[15:48] <fergusnoble> oh, maybe you can, i dont know
[15:49] <hallam> eh I probably shouldn't waste one, I'll swing by ifm
[15:49] Nick change: hallam -> hallam|away
[15:49] <fergusnoble> ooh, do you want 64bit?
[15:50] Laurenceb (n=laurence@dyres221-140.surrey.ac.uk) joined #highaltitude.
[15:50] <Laurenceb> hello
[15:50] <hallam|away> no 32 please
[15:50] <Laurenceb> ?
[15:50] <hallam|away> hi Laurenceb - was just leaving
[15:51] <Laurenceb> hi
[15:51] <Laurenceb> hows it going?
[16:05] <Laurenceb> wheres this snow then
[16:17] Action: SpeedEvil passes Laurenceb a '.
[16:17] Action: SpeedEvil has no snow.
[16:17] <SpeedEvil> Well.
[16:17] <SpeedEvil> I have a freezer I need to defrost.
[16:17] <SpeedEvil> But that's not quite the same.
[16:20] G8KHW (n=Steve@217.47.75.27) left irc: "Leaving"
[16:28] edmoore_ (i=836f0142@gateway/web/ajax/mibbit.com/x-e99055dd2443f073) joined #highaltitude.
[16:31] <jcoxon> afternoon edmoore_
[16:34] <edmoore_> hi
[16:35] G8KHW (n=Steve@217.47.75.27) joined #highaltitude.
[16:36] <jcoxon> interesting report from bill brown about dominoex
[16:36] <jcoxon> http://home.hiwaay.net/~bbrown/
[16:36] <jcoxon> he was really impressed
[16:45] <G8KHW> yeah - tis a nice mode and easy to generate (easier than BPSK-31)
[16:46] edmoore_ (i=836f0142@gateway/web/ajax/mibbit.com/x-e99055dd2443f073) left irc: "http://www.mibbit.com ajax IRC Client"
[16:47] <SpeedEvil> yeah
[16:47] <SpeedEvil> by the time you get to caring about PSK31/RTTY signal levels, you're probably screwed anyway.
[16:57] <Laurenceb> G8KHw: your technique for shaping logic transitions isnt strickly raised cosine
[16:57] <Laurenceb> G8KHW: you've convolving with a boxcar function
[16:59] <Laurenceb> in fact its not quite that either, as only transitions are shaped
[17:00] <Laurenceb> raised cosine convolves with a unit impulse +- depending on the data bit
[17:01] <G8KHW> your probably right Laurenceb - strictly speaking I thing you need a shape
[17:01] <G8KHW> that matches the communications channel
[17:02] <Laurenceb> yeah
[17:02] edmoore (i=836f0142@gateway/web/ajax/mibbit.com/x-1e74cf18fad230a5) joined #highaltitude.
[17:02] <Laurenceb> but we have several KHz of bandwidth
[17:02] <G8KHW> matched filter techniques
[17:02] <Laurenceb> so I dont see why shaping helps
[17:02] <G8KHW> its a technique to reduce bandwidth
[17:02] <Laurenceb> as if theres so shaping, youre sinc function will be very low at the boundaries
[17:03] <Laurenceb> the total area under the curve will still be the same
[17:03] <Laurenceb> so if the receiver is matched and we have white noise its optimal
[17:04] <Laurenceb> it will be more vulnerable to RFI if its narrow band and very high intensity
[17:05] <G8KHW> sure - but the chances that RFI is in a narrow bandwidth is also reduced
[17:06] <G8KHW> if its narrow band RFI is probabilities
[17:06] <Laurenceb> sure yeah
[17:06] <G8KHW> if is wide band rfi then there is proportionaly less too
[17:06] <Laurenceb> pulse shaping protects you from rfi
[17:07] <Laurenceb> I'm not sure what technique fldigi uses for receiving
[17:07] <edmoore> the code is there if you want to have a look
[17:07] <G8KHW> ah I think you have the wrong idea - I'm not talking about RFI when I say reduced bandwidth = less noise
[17:07] <edmoore> we nicked it to impliment DominoEX on the badger
[17:08] <G8KHW> i'm talking about thermal noise etc.
[17:08] <Laurenceb> which is white - approx
[17:08] <G8KHW> (also sky noise etc)
[17:09] <G8KHW> yes
[17:09] <Laurenceb> for white noise pulse shaping makes no difference
[17:09] <fergusnoble> yeah, i implemented dominoex on the badger and it seems to work nicely
[17:09] <fergusnoble> the only problem is you need very high resolution in setting the voltage to the radiometrix
[17:09] <fergusnoble> 10bit isnt nearly enough
[17:09] <G8KHW> I think it does because it reduces the bandwidyh of the signal
[17:09] <Laurenceb> fergusnoble: ont you have to generate multiple tones at once?
[17:10] <edmoore> no
[17:10] <Laurenceb> G8KHW: it only reduces the bandwidth at some fixed fraction of total power
[17:10] <fergusnoble> Laurenceb: no
[17:10] <fergusnoble> it uses one tone per symbol
[17:10] <edmoore> you jump between one of 18 tones and the info is encoded in the frequency difference between adjacent tones
[17:10] <Laurenceb> the area under the spectrum stays the same
[17:10] <Laurenceb> I see
[17:11] <edmoore> tomporally adjacent tones, to be clear
[17:11] <edmoore> temporally*
[17:11] <fergusnoble> yeah, thats the basis of it
[17:11] <G8KHW> I thought the dominoex tones were equi-spaced
[17:11] <fergusnoble> G8KHW: they are
[17:12] <Laurenceb> G8KHW: you cant reduce the impact of white noise with pulse shaping
[17:12] <fergusnoble> but then encoding makes sure that succesive symbols never use adjacent tones
[17:12] <fergusnoble> or the same tone
[17:12] <Laurenceb> G8KHW: consider how spread spectrum works
[17:12] <G8KHW> well then a 5 bit DAC should do
[17:12] <fergusnoble> which improves the performance under frequency drift a lot
[17:12] <G8KHW> with a linear V 2 F function
[17:13] <edmoore> sure, a 5-bit is fine if you can limit it to the fairly tight voltage range required
[17:13] <Laurenceb> fergus: fldigi can track RTTY
[17:13] <fergusnoble> G8KHW: in theory yes, our dac outputs over the whole range of the radiometrix though which is 25khz
[17:13] <G8KHW> may I suggest a variable resistor
[17:13] <fergusnoble> of which you only use a few hundered hz
[17:14] <fergusnoble> G8KHW: thats how i did it in the end, although the radiometrix goes a bit wierd near zero volts and stops being linear
[17:14] <edmoore> the solution for badger 2 will be a 16-bit dac, then we can stick different modes anywhere within the 25khz radiometrix b/w
[17:15] <Laurenceb> I think dominoex will be less robust to noise
[17:15] <fergusnoble> G8KHW: also fldigi seems to be very sensitive to precise tone spacing, if your just a little bit off then it doesnt work at all
[17:15] <G8KHW> sure - run it at about 1.5V +/- 100mV or so (or whatever gives +/- 100Hz)
[17:15] <edmoore> Laurenceb: what makes you say that?
[17:15] <Laurenceb> as you have many frequency bins to compare
[17:15] <Laurenceb> not just two
[17:15] <fergusnoble> G8KHW: how do yo easily shift up the zero point?
[17:16] <edmoore> sure but for a given data rate you get a muich longer integration time
[17:16] <Laurenceb> yes
[17:16] <G8KHW> humm hang on i have a diagram
[17:16] <fergusnoble> i mean, more easily than just using a dac
[17:16] <edmoore> as you have a lower baud rate
[17:16] <Laurenceb> but its a bit wasteful
[17:16] <edmoore> not really
[17:16] <Laurenceb> can it transition to any frequency?
[17:16] <edmoore> it's maybe 300hz b/w
[17:16] <Laurenceb> given f_n
[17:17] <Laurenceb> can f_n+1 be any of the frequencies?
[17:17] <edmoore> i don't know what you're asking
[17:17] <edmoore> any of which set of frequwencies?
[17:17] <G8KHW> This is an ascii art schem form the code - thats as much as i have written down at the mo
[17:17] <Laurenceb> theres a set f of frequencies
[17:17] <G8KHW> ;
[17:17] Last message repeated 2 time(s).
[17:17] <G8KHW> ; 3V
[17:17] G8KHW (n=Steve@217.47.75.27) left irc: Excess Flood
[17:17] <Laurenceb> woops
[17:18] <Laurenceb> set of possible transmit frequencies
[17:18] <Laurenceb> for each time period
[17:18] G8KHW (n=Steve@217.47.75.27) joined #highaltitude.
[17:18] <edmoore> for dominoEx, specifically?
[17:18] <Laurenceb> yes
[17:18] <Laurenceb> argg
[17:18] <Laurenceb> :P
[17:18] <edmoore> ok
[17:18] <G8KHW> how much of that did you get?
[17:18] <edmoore> 4 lines
[17:18] <Laurenceb> 4 lines
[17:19] <edmoore> so the bandwidth is contained within a range of freqs
[17:19] <Laurenceb> edmoore, so it can transmit on any frequency band for each time interval?
[17:19] <Laurenceb> are there any rules?
[17:20] <edmoore> you couldn't get some runaway freq encoding that would start marching across your spectrum
[17:20] <Laurenceb> like it has to make a transition or something
[17:20] <Laurenceb> sure theres 18 freuqncies
[17:20] <edmoore> take a look at some documenmtation for it
[17:20] <Laurenceb> yeah ok
[17:20] <edmoore> bill brown launched yesterday and said he had it long after the rtty and hellscgreiber went
[17:21] jcoxon (n=jcoxon@host86-158-31-172.range86-158.btcentralplus.com) left irc: "Leaving"
[17:22] <Laurenceb> ah /i see
[17:22] <Laurenceb> so its sending a hex character
[17:23] <Laurenceb> but it uses syncronising techiques are variable lenght words, so it should be better than rtty
[17:23] <edmoore> well it's the psk varicode charcter set
[17:23] <Laurenceb> yeah
[17:23] <edmoore> just a basic huffman job
[17:24] <edmoore> as fergus and I were saying, we could do a lot better by encoding our balloon character set, but it would rule other people out from listening without modded receiver software
[17:24] <SpeedEvil> Just modulate with 100+lat Hz + 1000+(lon*5)
[17:24] jcoxon (n=jcoxon@host86-158-31-172.range86-158.btcentralplus.com) joined #highaltitude.
[17:24] <Laurenceb> hehe
[17:25] <Laurenceb> edmoore: or send raw floats ints ect down
[17:25] <Laurenceb> just dump a status struct into the radio
[17:25] <edmoore> we'd do freq modulation if we could! but there's a VCO in the way
[17:26] <fergusnoble> jcoxon: cant find the report, where is it?
[17:26] <Laurenceb> yeah dominoex looks better than rtty
[17:26] <Laurenceb> also it downt suffer from the bit sync problem
[17:26] <edmoore> or drift problems
[17:26] <Laurenceb> as a corrupted stop or start bit can cause you to lose a load of data
[17:26] <Laurenceb> and you have to put in tune up pulses
[17:27] <Laurenceb> well fldigi seems very good at autotracking drift
[17:27] <fergusnoble> yeah the drif performance is awesome, i can tune it out by almost half the width of the stream in the waterfall and still get good copy
[17:27] Tigga (n=chatzill@pc-232-235-103.magd.cam.ac.uk) joined #highaltitude.
[17:27] <Laurenceb> but it should give about 3 times the data rate with my setup on the mini rogallo
[17:27] <edmoore> fergusnoble: are you coming tonight?
[17:27] <Laurenceb> hmm
[17:28] <Laurenceb> I'm going to have to code this :P
[17:28] <fergusnoble> Laurenceb: for some reason my fldigi isnt properly tracking the drift, you still have to click and recentre it
[17:28] <fergusnoble> edmoore: yes
[17:28] <edmoore> Laurenceb: fldigi source makes it trvial
[17:28] <edmoore> fergusnoble: mine tracks fine too
[17:28] <Laurenceb> fergusnoble: works here, I compiled from source
[17:28] <fergusnoble> Laurenceb: you can have my code iuw, ive cut all the chaff out of the fldigi code
[17:28] <edmoore> d/l the bin
[17:28] <Laurenceb> but my sound card is a bit glitchy
[17:28] <edmoore> this is on the eee, not tried it on the mac
[17:28] <jcoxon> fergusnoble, http://groups.yahoo.com/group/Ballooning/message/2752
[17:28] <Laurenceb> and rtty is hardly reliable
[17:28] <fergusnoble> im running the mac bin
[17:29] <Laurenceb> fergus: I dont follow you
[17:29] <Laurenceb> you've rewritten fldigi?
[17:29] <edmoore> i bet fergus is pleased about that
[17:30] <edmoore> Laurenceb: no, for the micrcontroller
[17:30] <Laurenceb> ok cool
[17:30] <edmoore> copy the dominoex sending source
[17:30] <Laurenceb> yeah thatd be good
[17:30] <Laurenceb> if you could pastebin it
[17:30] <Laurenceb> Huffman doesnt look too hard :P
[17:30] <edmoore> and get rid of the crap, and replace their tx functions with your own (maps value to your DAC or whatever)
[17:30] <Laurenceb> oh
[17:30] <Laurenceb> ok
[17:31] <edmoore> huffman is completely easy
[17:31] <Laurenceb> that one way to write the tx code :P
[17:31] <edmoore> you can't go wrong
[17:31] <Laurenceb> yeah
[17:31] <Laurenceb> right I'm off to get some food
[17:31] <edmoore> and optimal
[17:31] <Laurenceb> bbl
[17:34] <fergusnoble> http://www.srcf.ucam.org/~cuspaceflight/websvn/listing.php?repname=CUSF&path=%2Ffc1+software%2Ftasks%2FRadio%2F
[17:34] edmoore (i=836f0142@gateway/web/ajax/mibbit.com/x-1e74cf18fad230a5) left irc: "http://www.mibbit.com ajax IRC Client"
[17:35] <fergusnoble> Laurenceb: as you will see, in code its not very difficult
[17:36] <fergusnoble> the only cleverness is in this line: tone = (txprevtone + 2 + sym) % NUM_TONES;
[17:37] <fergusnoble> which makes sure it never uses adjacent tones or consecutive tones
[17:38] <G8KHW> fergausnoble: are you shaping as you slide from one frequency to the next?
[17:38] <fergusnoble> G8KHW: nope
[17:39] <fergusnoble> it would be a lot better if i did though
[17:39] <fergusnoble> the tones are quite smudged out in spectrum because of the lack of shaping
[17:39] <G8KHW> probaly - I don't know it makes much difference at slower symbol rates
[17:40] <fergusnoble> yeah, im loosing the most by not being able to get the spacing precicely right
[17:41] <G8KHW> I guess its gonna be better - the spectrum generated may make whong decisions in the decoder (ISI)
[17:41] <fergusnoble> yeah, i think thats the case
[17:42] <fergusnoble> the copy was perfect but the fldigi "signal strength" thingy was quite low down even though the tx was very strong
[17:43] <fergusnoble> by "signal strength" i mean the thing which indicates how well it is decoding the signal, probably some measure of signal to noise between the bins
[17:43] <fergusnoble> sorry, rx was strong
[17:44] <G8KHW> FergusNoble: does dominoex have a code whereby adjacent frequencies are avoided in the next symbol?
[17:44] <fergusnoble> yup
[17:44] <G8KHW> lattice coding i think its called
[17:44] <G8KHW> its a way of adding an extra dimension to the codind
[17:45] <G8KHW> coding
[17:45] <fergusnoble> its quite primitive, it just acumulates the symbol number as it goes and then transmits the tone corresponding to the accumulated number modulo 18
[17:45] <G8KHW> makes the "space" between synbols greater and therefore more noise immune
[17:45] <fergusnoble> yup
[17:48] <fergusnoble> G8KHW: on another note, im trying to design a pcb for the software gps with a passive antenna
[17:48] <fergusnoble> the antenna is 50ohm out and the gps chip input is 50ohm
[17:49] <fergusnoble> i dont need any matchin netwrks then do i?
[17:49] <fergusnoble> if i use a 50ohm transmission line on the pcb
[17:49] <G8KHW> nope - but you will need to make 50ohm tracks on the board
[17:50] <G8KHW> unless the socket is right next to the input pin
[17:50] <fergusnoble> ok, thats what i though, but the reference design has an L network
[17:50] <fergusnoble> cant figure out why
[17:50] <G8KHW> filter?
[17:50] <fergusnoble> maybe
[17:50] <G8KHW> or DC feed?
[17:51] <fergusnoble> it looks like it has a dc biasing thing for use with an active antenna but it looks like there is another matching network after that
[17:51] <G8KHW> have you got a link to the diagram I'll take a deco
[17:52] <fergusnoble> no, its im just looking at the pcb
[17:52] <fergusnoble> so i dont have the component values either
[17:53] <G8KHW> :-(
[17:53] <fergusnoble> i think ill just go with the datasheet and not bother with any of that
[17:54] <G8KHW> probbaly best - at least you know it should work
[17:54] <G8KHW> BB in a bit
[17:58] <gordonjcp> G8KHW: ping?
[18:03] <hallam|away> fergusnoble: IIRC the 4120 does *not* have a 50 ohm input to the LNA
[18:03] <hallam|away> it does have 50 ohm output of LNA and input to mixer
[18:03] Nick change: hallam|away -> hallam
[18:03] <fergusnoble> hallam: i know why the cache isnt working
[18:04] <fergusnoble> ill phone you
[18:06] <hallam> ah, the old level one core static random access memory cache protection translation lookaside buffer setting register table
[18:07] <hallam> should have known it was the L1CSRAMCPTLBSRT
[18:07] Nick change: hallam -> hallam|lab
[18:11] hallam|lab (i=836fc8c8@gateway/web/ajax/mibbit.com/x-614c4d195db309ed) left irc: "http://www.mibbit.com ajax IRC Client"
[18:22] <Laurenceb> feregus
[18:22] <Laurenceb> its a dc nertwork
[18:23] <Laurenceb> erm maybe...
[18:23] <Laurenceb> just a sec
[18:27] <G8KHW> gordonjcp: pong
[18:34] <rjharrison> Hi steve
[18:34] Action: G8KHW hails rjharrison arrival
[18:34] Action: G8KHW welcome fellow team emeber
[18:35] <rjharrison> Hi G8KHW
[18:35] <rjharrison> Picked up the 2e0 yesterday
[18:37] <rjharrison> I'm just going to polish code and payload ready for next possible launch
[18:37] Action: G8KHW evil paln to turn eveyone in #highaltitude is working - maniacal laughter
[18:37] <G8KHW> into hams that is
[18:37] <Laurenceb> I'm a muslim, no ham for me
[18:37] <rjharrison> we'll you are doing quite well so farr
[18:38] <rjharrison> hahah
[18:42] <Laurenceb> To allow the lowest possible sample rate, the
[18:42] <Laurenceb> downconverted IF signal is further filtered to re-
[18:42] <Laurenceb> shape the noise spectrum and remove any
[18:42] <Laurenceb> significant degradation from aliasing. Both
[18:42] <Laurenceb> frequency conversion and filtering functions are
[18:42] <Laurenceb> performed digitally, avoiding the tolerance, offset
[18:42] <Laurenceb> and imbalance imperfections associated with
[18:42] <Laurenceb> analogue converters.
[18:42] <Laurenceb> woops
[18:43] <Laurenceb> fergusnoble: thats from the datasheet, quite impressive (sorry to splat loads of stuff onto here)
[18:43] <gordonjcp> G8KHW: which licence category does G8 give you, these days?
[18:44] <G8KHW> full
[18:44] <gordonjcp> coo
[18:44] <gordonjcp> l
[18:44] <gordonjcp> did you do a morse test?
[18:44] <G8KHW> when they took away the morse test they moved all class Bs up to class A
[18:45] <gordonjcp> ah, righty
[18:45] <G8KHW> I could never find the time to get beyond 8WPM - so gave up with morse
[18:46] <gordonjcp> koch method...
[18:48] <Laurenceb> G8KHW: its find to us radiometrix modules without an antenna right?
[18:59] <G8KHW> well probably - the worst than
[18:59] <G8KHW> that can happen is an extra 10mW of power is dissipated in the PA
[19:01] <Laurenceb> yeaha
[19:02] <Laurenceb> fergusnoble: the diagram on p18 is for impedance matching
[19:02] <Laurenceb> I'd try and keep it within a couple of mm or so of the Se4120
[19:03] <Laurenceb> then use a 50ohm line to the antenna
[19:18] Bluenarf (n=Paul@apollo.paulsnet.org) joined #highaltitude.
[19:19] EI5GTB (n=Paul@apollo.paulsnet.org) left irc: Read error: 104 (Connection reset by peer)
[19:22] Nick change: Bluenarf -> EI5GTB
[19:55] <Laurenceb> fergusnoble: ping
[20:27] <Laurenceb> wow its snowing here at last
[20:28] <Laurenceb> I'm considering MFSK-64
[20:28] <Laurenceb> it has FEC and higher datarate than domineo
[20:29] rjharrison (n=rharriso@80.176.172.227) left irc: Read error: 110 (Connection timed out)
[21:10] Hiena (n=Hiena@81.93.195.181.datatrans.hu) left irc: "-=Halt! Hammerzeit!=-"
[21:35] wickerwaka (n=akawaka@66-215-97-198.dhcp.malb.ca.charter.com) joined #highaltitude.
[22:28] borism_ (n=boris@195-50-199-27-dsl.krw.estpak.ee) joined #highaltitude.
[22:34] borism (n=boris@195-50-201-33-dsl.krw.estpak.ee) left irc: Read error: 145 (Connection timed out)
[22:37] <Laurenceb> g|= (arg|(1<<(g|0xF0))) ^ (arg|(1<<((g-=16)|0xF0)));
[22:37] <Laurenceb> ^ is that valid c code?
[22:42] <gordonjcp> looks valid
[22:42] <gordonjcp> it's badly written though
[22:42] <G8KHW> I concur on both counts
[22:42] <Laurenceb> whys is badly written?
[22:44] <gordonjcp> what was your original question?
[22:44] <Laurenceb> nvm
[22:44] <G8KHW> its not easy to understand
[22:45] <Laurenceb> its designed to get down ram use
[22:45] <G8KHW> try breaking it up and letting the compiler do the optimization
[22:45] <gordonjcp> how is it meant to do that?
[22:45] <Laurenceb> oh god nvm
[22:45] <Laurenceb> its as bad as a language channel on here
[22:46] <gordonjcp> Laurenceb: even if you wrote it verbosely, the compiler would sort it out
[22:46] <Laurenceb> guess so
[22:46] <G8KHW> I think I see the error
[22:47] <G8KHW> which side do you think it will evaluate first g-=16 or g|= ...
[22:47] <G8KHW> or g |
[22:48] <Laurenceb> hmm
[22:48] <Laurenceb> yeah its open to interpretation
[22:49] <G8KHW> I guess your trying to save one variable by not breaking it up
[23:00] natrium42_ (i=natrium4@CPE000625d867e2-CM0014045885be.cpe.net.cable.rogers.com) left irc:
[23:02] natrium42 (i=natrium4@CPE000625d867e2-CM0014045885be.cpe.net.cable.rogers.com) joined #highaltitude.
[23:03] <fergusnoble> Laurenceb: hello, you ping'd me?
[23:04] <natrium42> hi
[23:04] <fergusnoble> natrium42: hi!
[23:04] <natrium42> finally finished watching tennis
[23:04] <fergusnoble> im applying to the perimeter institute this summer, am i right in thinking you live in Kitchener?
[23:05] <natrium42> ooh, nice
[23:05] <natrium42> yeah, i have been at PI on opening day
[23:05] <natrium42> attending some lectures
[23:06] <fergusnoble> i think its really competitive so i doubt ill get in but it would be cool
[23:07] <natrium42> is this for graduate studies?
[23:14] <fergusnoble> no, an undergrad placement between my second and third year
[23:14] <fergusnoble> would be 3-4 months
[23:15] <fergusnoble> would love to go there as a grad student, they do the coolest stuff
[23:16] <Laurenceb> hi
[23:16] <Laurenceb> PI ?
[23:16] <Laurenceb> yeah I was going to ask you why domineox? I think MFSK looks better
[23:16] <fergusnoble> http://www.perimeterinstitute.ca/
[23:17] <fergusnoble> domino is easier to implement i think and apparently has better drift performance
[23:17] <fergusnoble> but i dont know much about it
[23:17] <fergusnoble> i mean, domino is an mfsk mode
[23:18] <natrium42> fergusnoble: very cool, good luck
[23:18] <natrium42> we can do a few launches if you get in :)
[23:19] <Laurenceb> fergusnoble: good luck, but theoretical physics?!
[23:19] <fergusnoble> Laurenceb: hell yeah
[23:19] <Laurenceb> ok...
[23:19] <fergusnoble> i couldnt think of doing anything else
[23:19] <Laurenceb> I wouldnt have imagined it
[23:20] <fergusnoble> sure, i pretty much cant live live without doing some practical stuff
[23:20] <Laurenceb> its not that related to comp sci or electronics
[23:20] <fergusnoble> but theoretical is where i want to be
[23:20] <Laurenceb> at least the way it was run at ox
[23:20] <Laurenceb> I see
[23:20] <Laurenceb> your one of those :P
[23:20] <fergusnoble> yeah :p
[23:21] <Laurenceb> ok yeah I've heard of the PI
[23:21] <Laurenceb> couldnt remember what it was
[23:21] <fergusnoble> ive always done programming and electronics, but in terms of the academic side i want to go into theoretical
[23:21] <Laurenceb> but yeah its very good for theoretical phys
[23:21] <Laurenceb> good luck
[23:21] <fergusnoble> thanks, i think ill need it
[23:22] <fergusnoble> so your going the way of becoming an experimental physicist?
[23:22] <Laurenceb> IIRC I know someone from ox who was thinking of applying, I'll facebook him to see what happened
[23:22] <Laurenceb> fergusnoble: maybe
[23:22] <fergusnoble> or are you going into engineering stuff after you graduate?
[23:22] <Laurenceb> may still drop out
[23:23] <Laurenceb> well I'd like to go into business of some sort
[23:23] <fergusnoble> i mean finish your phd
[23:23] <Laurenceb> yeah, prob go into business, I dont think I'd want to work down here
[23:23] <Laurenceb> lots of the phds use it as a way to get a good job at the company
[23:24] <fergusnoble> no business as in management and that crap?
[23:24] <Laurenceb> no no
[23:24] <fergusnoble> good good :)
[23:24] <Laurenceb> business as an independant design consultant or something
[23:24] <fergusnoble> that would be really cool
[23:24] <fergusnoble> if i was to not go into acedemia i think i would like to do something like that
[23:25] <Laurenceb> I was working for lafarge during 6th form, that was good fun
[23:26] <Laurenceb> - maxxive cement plant near where I live
[23:26] <Laurenceb> everything ran off bbc model B machines with econet
[23:27] <Laurenceb> hmm premier institute = PI veyr nice
[23:28] <Laurenceb> oh perimeter
[23:28] <Laurenceb> woops
[23:29] <natrium42> lul
[23:29] <natrium42> it's really nice inside
[23:29] <Laurenceb> shows how much I know
[23:29] Tigga (n=chatzill@pc-232-235-103.magd.cam.ac.uk) left irc: Read error: 104 (Connection reset by peer)
[23:29] <natrium42> lots of black boards
[23:29] <natrium42> fire places
[23:29] <natrium42> etc
[23:30] <Laurenceb> hehe fire places
[23:31] <natrium42> steven hawking is going to visit for a few months this year afaik
[23:32] <Laurenceb> didnt hallam famously almost run steven hawking over on a bike?
[23:33] <Laurenceb> :P
[23:33] <natrium42> lol, really?
[23:33] <Laurenceb> I heard some tales a while back
[23:37] <Laurenceb> maybe it was _just_ some other VIP
[23:37] <Laurenceb> I remember he had a letter from the proctors framed on the wall
[23:44] <fergusnoble> Laurenceb: thats actually true
[23:44] <fergusnoble> he was knocked off his bike by the impact
[23:44] <fergusnoble> he just picked up his bike and ran
[23:45] <Laurenceb> omg worse than I thought
[23:45] <natrium42> hahaha
[23:45] <Laurenceb> bike wheelchair collision?
[23:45] <fergusnoble> yup
[23:45] <Laurenceb> omg
[23:46] <natrium42> epic
[23:46] <Laurenceb> maybe he wished a black hole would swallow him up
[23:53] G8KHW (n=Steve@217.47.75.27) left irc: "Leaving"
[23:54] <shellevil> The deed has its own punishment. http://uk.youtube.com/watch?v=BGauuoYOlvw
[23:55] jcoxon (n=jcoxon@host86-158-31-172.range86-158.btcentralplus.com) left irc: "Leaving"
[23:59] EI5GTB (n=Paul@apollo.paulsnet.org) left irc: Remote closed the connection
[00:00] --- Mon Feb 2 2009