<qi-bot> [commit] Xiangfu Liu: new package: offrss and some depends pakcages http://qi-hw.com/p/openwrt-packages/1c605fe
<qi-bot> [commit] Xiangfu Liu: new package: timidity, MIDI files player http://qi-hw.com/p/openwrt-packages/49ec918
<kyak> does anyone know if Ben has a backup battery (or capacity?) for HW clock?
<kyak> i noticed it is reset to start of epoch when Ben is _fully_ discharged
<xiangfu_> kyak: no backup battery for HW clock.
<kyak> xiangfu_: ok, thanks
<kyak> though Ben can usually live throught the night without external power supply, i sometimes forget to connect it to laptop in the morning :)
<kyak> then i find it blacked out
<kyak> maybe i can make it produce sound signals via cron when the battery level is low
<wpwrak> roh: (sanded) good. i was already getting a little scared by the powers of your laser cutter :)
<kyak> this would be cool
<qi-bot> [commit] Xiangfu Liu: fixed section display not correct in ben nanonote http://qi-hw.com/p/gmenu2x/91d3fea
<qi-bot> [commit] Xiangfu Liu: gmenu2x: update http://qi-hw.com/p/openwrt-packages/f07d355
<wpwrak> interesting. this is a complete ben: http://downloads.qi-hardware.com/people/werner/wpan/rework-20110112/
<wpwrak> still need to get rid of the "mountain", but it seems that lcm may also play a role in making things a little messy
<wolfspraul> wpwrak: I introduced the cmdline patch on the kicad list, and someone replied. so that's a start...
<wolfspraul> what I need to do now is to move to the latest kicad version first.
<wolfspraul> do you mind if I just pick the latest and see whether it builds? (forward-port the patches)
<wolfspraul> hmmm. thinking about this maybe we should decouple the use of KiCad at Qi, and trying to upstream the cmdline patch.
<wolfspraul> I could make a subdirectory in kicad-patches, like kicad-patches/upstream
<wolfspraul> then work there to bring the patches into an upstreamable format
<wolfspraul> yeah I think that makes more sense than trying to sync this task with our own use of KiCad
<wolfspraul> let me know if you disagree...
<wpwrak> going to the latest sounds good to me. always better to stay close than to try to delay the inevitable for too long
<wolfspraul> wpwrak: yes but I think I need to decouple the upstream patches from the ones we are using.
<wolfspraul> there will be feedback, and I have to rearrange patches etc.
<wpwrak> if the latest version looks good with our designs, then we should update the version we generally use as well
<wolfspraul> and I may need to uptick very frequently
<wpwrak> anything else is just confusion ;-)
<wolfspraul> that will be a mess if it's the same base we are building our 'daily use' KiCad from
<wpwrak> sure, doesn't make sense to update every five minutes while merging
<wolfspraul> I got another quite encouraging reply on the list, so now I guess I'm obliged to actually do something :-)
<wolfspraul> at least it seems the KiCad community is functioning, that's good
<wpwrak> (checking the list) haven't caught of for weeks ...
<wolfspraul> sure, I will just try to get this upstream, then unsubscribe
<wolfspraul> unfortunately it's quite a bit of work, oh well
<wolfspraul> hacking codes is far easier than communicating, understanding, and reaching consensus
<wpwrak> well, the reactions are good so far. communication seems to be working ;-)
<wpwrak> hmm, so much about using a "clean" clock :-( http://downloads.qi-hardware.com/people/werner/wpan/rework-20110113/
<wolfspraul> oh wow, that's a lot worse than before :-)
<wpwrak> it does show that the chip seems to prefer a sinusoid clock, though. that's useful information.
<wpwrak> i kinda wonder why going from a complete ben to just the main pcb improved things so much, though. for comparison, here are yesterday's results (full ben): http://downloads.qi-hardware.com/people/werner/wpan/rework-20110113/
<wpwrak> (i had to go to the bare pcb because the board wouldn't fit into the 8:10 slot in the case anymore)
<buggus> hmm
<zrafa> Mauro_R: wpwrak : technology : http://www.youtube.com/watch?v=CiJeHHGqSSw
<Mauro_R> let me see ...
<Mauro_R> jaja :-) it's the last technology ... ecologic ....
<Mauro_R> you must keep safe of don't burn your parts ;-p
<lekernel> burning wood isn't that ecologic... and btw steam explosions are dangerous
<lekernel> afaik they were the main problem that prevented the development of the steam engine at the beginning
<Mauro_R> lekernel: I'm sure that it's more ecologic than the batteries ...
<wpwrak> zrafa: (bike) cool ! ;-)
<lekernel> maybe, maybe not. I haven't done the math, so I won't jump into saying which is more ecologic. it's not an easy question
<lekernel> but it sure is more dangerous than batteries
<wolfspraul> lekernel: I mentioned our emi tests to Adam, he was very happy to hear that such a complex board passed on most tests right away.
<wolfspraul> when I described the one failed esd/field test to him, he asked in which area the sink was, but I told him we were too lazy to track that down, going for the whole metal sheet instead
<lekernel> except that maybe DIY batteries might done the same way as this steam engine, i.e. with open acid and nasty chemicals just waiting to spill on you when you fall :)
<wolfspraul> we suspect it may be the dram area, right?
<lekernel> wolfspraul: yeah, probably
<wolfspraul> I don't want to do random layout changes in rc3 though, that sounds riskier to me than just going with the metal sheet for now
<lekernel> laziness is good. gets things done fast
<wolfspraul> one day if someone has time and energy, that person can track down the sink more precisely...
<lekernel> and attaching a metal shield to the DRAM alone can also be more expensive and complicated than the whole sheet
<wolfspraul> sure I'm not complaining. the metal sheet is the right approach now, not risky layout changes.
<wolfspraul> well we could try to make changes in the layout
<lekernel> metal sheet is only a few tens of cents, and simple assembly...
<wolfspraul> but it will be hard to make the right changes there that actually fix the issue, and be without regressions too
<wolfspraul> yes yes :-)
<wolfspraul> we are on the same page
<wolfspraul> I am just telling you about bringing Adam into this.
<wolfspraul> no layout changes in rc3 to try to fix this potential dram esd/field sink issue
<lekernel> no, sure not
<wolfspraul> you could add extra copper to the pcb
<lekernel> but we may want to fix the boot problem as early as rc3
<wolfspraul> it would not be too unusual in adam's experience. just fyi. we will not do it, no worries :-)
<lekernel> the pcb is already full of copper
<wolfspraul> yes, agreed
<wolfspraul> we are on the same page
<wolfspraul> at some point I need to decide the size of rc3
<wolfspraul> I'm thinking 50-100
<wolfspraul> tough decision again
<wolfspraul> I think we have to be brave and believe in the product :-)
<wolfspraul> I will start sourcing a few things already, anything with lead time > 4 weeks.
<wolfspraul> just fyi
<Mauro_R> lekernel: acid, nasty chemicals ... also, solid batteries are less eco than burn wood or coal ...
<Mauro_R> lekernel: but I understand your point, burn wood isn´t good enough ...
<LunohoD> re
<qi-bot> [commit] Xiangfu Liu: gmenu2x: add a new theme http://qi-hw.com/p/openwrt-packages/9186e37
<lekernel> wpwrak: are the openmoko gerber files available somewhere?
<lekernel> I can't find them
<wolfspraul> not available
<lekernel> really?!
<lekernel> wow...
<wolfspraul> yeah sounds like dark age from our perspective nowadays :-)
<wolfspraul> as some of my last official function at Openmoko, I released netlist, board outline and a few other small things under a cc license, to help jumpstart gta02-core.
<wolfspraul> but the full GERBER files, or PADS Layout files, were never published
<lekernel> amazing that project still got so popular
<kristianpaul> indeed
<lekernel> no features, lots of bugs, little openness...
<wpwrak> lekernel: only partoally available - the outside layers
<wpwrak> lekernel: it was still the most open at its time ...
<lekernel> probably has to do with the average hacker's interest for communication networks :)
<wpwrak> that too :)
<kristianpaul> the case was finally open isnt?
<wpwrak> kristianpaul: the CAD files have been published, yes
<kristianpaul> wonder how usable those are
<wpwrak> kristianpaul: me too ;-)
<wolfspraul> lekernel: it was popular for 2 reasons imho
<wolfspraul> one was the time of the announcement
<wolfspraul> the Symbian dark age
<wolfspraul> don't forget, this was _before_ the iPhone, before Android
<wolfspraul> I would have killed someone to get out of this Symbian hell.
<wolfspraul> I was working in the smartphone industry, and had been tortured by Symbian suits for years.
<wolfspraul> there was a lot of anger built up there, and Palm Treo was only partially able to benefit from that
<wolfspraul> so the promise of an 'open' platform was wildly welcomed
<wolfspraul> nowadays the iphone and android platforms are considered the 'open' ones, with 100k+ apps etc.
<wolfspraul> so that was the first reason
<wolfspraul> announcing an open phone platform in late 2006 was just perfect
<wolfspraul> second reason was gross brutal over-promising
<Jay7> bad that no one software stack was succeeded
<wolfspraul> back then I was clueless so I couldn't tell how bad it was
<wolfspraul> but today I know :-)
<Jay7> LiMo seems failed
<Jay7> MeeGo is half-life
<wolfspraul> it's like I make an announcement today that Sharism Ltd. will ship a competitor to the Airbus A380 in May this year.
<wolfspraul> so OM had a history of again and again and again over-promising
<wolfspraul> against all reason, almost insane
<wolfspraul> over-promising will always give you some popularity first :-)
<wolfspraul> those were the 2 reasons for me 1) timing 2) over-primising
<qbject> I had no idea, when I acquired an ebook of Gibson's Zero History to read on my BNN, that it would feature a Neo so heavily.
<qbject> (Just started reading it.)
<wolfspraul> Neo in fiction, maybe that's where its actual natural habitat was from the beginning...
<wpwrak> wolfspraul: well, sharism vs. a380 is exaggerating a little :) with a better hw team and a more realistic approach for the software, i think things could have turned out much better
<wpwrak> wolfspraul: and of course, management that is a lot less fond of u-turns :)
<wolfspraul> wpwrak: I stand behind sharism vs. a380
<wolfspraul> it was just gross, really.
<wolfspraul> there is a difference between what you want to do, and what you can do
<wolfspraul> but it was also the reason for the popularity, of course
<wolfspraul> popularity = delayed frustration
<wolfspraul> so that's exactly what happened later
<wolfspraul> anyway we are learning, and still around. :-)
<wolfspraul> I'm quite optimistic actually...
<wolfspraul> wpwrak: the dynamics of competition among the companies openmoko was compared with are very tough. so with those comparisons, it never had a chance, from day 1.
<wolfspraul> Palm tried the same thing with the Pre, they invested ca. 400 million USD, and even that was not enough.
<wolfspraul> to get into that league, you need multi-billion USD investments, and most importantly you need very strong partners
<wpwrak> wolfspraul: yeah, i was thinking more of the technical side - make a device that works properly, hw and sw. i think that would have been feasible.
<wpwrak> wolfspraul: maybe the competition would then have trampled openmoko anyway
<wolfspraul> you need scale, rock solid partnerships
<wpwrak> wolfspraul: yeha, no idea how/if it could have scaled
<wpwrak> wolfspraul: of course, had openmoko been successful in making the first usable open phone, maybe the partners would have come on their own
<qbject> I'm not using one any more, but my partner is certainly suffering through waves of scalability problems from Peek, Inc.
<qbject> Every time their marketing is successful, their infrastructure shudders.
<wpwrak> the "mountain" is rather persistent
<wpwrak> now i tried it with an external antenna. not much of a change. then with a cable straight to the usrp (having a bit of trouble keeping the power down without changing the gain)
<wpwrak> hm, looks better with lower gain. here it is.
<wolfspraul> yeah just wanted to say - even with external antenna looks similar
<wpwrak> yup. so it's not over-the-air self-interference
<wpwrak> still don't entirely trust the clock. maybe i'll try a board with a crystal.
<wpwrak> but first a bit more power bypassing
<wpwrak> there's also an interesting change between the complete ben (with lcm) and just the main pcb. the latter has a lot less noise. it could be just some field distortion caused by the extra metal objects, of course.
<wpwrak> alright. some more unlikely causes ruled out.
<kristianpaul> I would suggest the following:
<kristianpaul> -1 =>  00
<kristianpaul> -0 =>  01
<kristianpaul> +0 => 10
<kristianpaul> +1 => 11
<kristianpaul> wpwrak: ^^^^
<wpwrak> aha !
<kristianpaul> hum.... I am not entirely sure -0=+0. I would say that +0>-0. Remember
<kristianpaul> that the point is to use as much info as you can
<kristianpaul> from fabrizio
<wpwrak> maybe it's time to ask sige ?
<kristianpaul> hmm i want have something working with satellites at least before ask
<kristianpaul> or ask for 6-8 bits ADC... ;)
<wpwrak> well, you can pick one interpretation and just work with that. it shouldn't fail too badly if you're wrong
<kristianpaul> sure
<kristianpaul> my trouble now is converting imaginary data to real..
<kristianpaul> i need read more
<wpwrak> ah, entering the joyful part of signal processing ;-)
<kristianpaul> :D
<wpwrak> this looks like a step in the right direction. last two entries on http://downloads.qi-hardware.com/people/werner/wpan/rework-20110113/
<wolfspraul> indeed, slowly though. interesting stuff.
<wpwrak> intersting ... as soon as i connect the scope - on either side of the DC-blocking cap (!), the clock collapses :-(
<wpwrak> wolfspraul: (slowly) yeah, every single decibel puts up a fight :)
<qi-bot> [commit] Wolfgang Spraul: added cad/Makefile to automate svg2pdf and dxf2png http://qi-hw.com/p/m1/72c2bda
<kristianpaul> new gmenux theme is nice, but... some apps icon tell more than you can expect from a command line/ncurse interface.. talking from eye candy perspective
<wolfspraul> kristianpaul: you mean they look too good?
<wolfspraul> which ones in particular do you have in mind?
<kristianpaul> yes too good
<kristianpaul> mutt
<kristianpaul> sdvc may be
<kristianpaul> no more
<wolfspraul> hah, with mutt you picked the wrong one.
<kristianpaul> why?
<wolfspraul> I think it's the best email client in the world, the icon cannot possibly look too good :-)
<wolfspraul> (just kidding)
<kristianpaul> i use mutt
<kristianpaul> ah ok
<wolfspraul> if the icons are misleading, that's not good. I didn't look at them one by one yet, but I like that things look better now.
<wpwrak> wolfspraul: couldn't agree more :-)
<kristianpaul> sure looks good
<wolfspraul> looks 'too good' is a tough argument to make
<wpwrak> wolfspraul: (regarding the best email client)
<wolfspraul> but I will keep your feedback in mind and pay attention when I test the next image...
<wolfspraul> maybe all console/ncurses apps should have a common theme or 'hint' that they are console based?
<kristianpaul> yeap !
<kristianpaul> snownews is tricky i think is console app but  the icon make you think other thing
<kristianpaul> ah yes, thats newsbeuter style
<kristianpaul> mutt logo is actually a dog btw
<kristianpaul> snownews have right logo
<kristianpaul> interesting
<kristianpaul> or at least webpage favicon
<kristianpaul> any way, just saying :-)
<kristianpaul> he i love this logo http://wiki.mutt.org/mutt.gif
<kristianpaul> Int'l shipment release !!
<kristianpaul> i hope thats good, sounds good in spanish
<kristianpaul> wpwrak: are you aware of fedex procedures?
<kristianpaul> "El envío internacional fue liberado" means it passed customs isnt?
<wolfspraul> yes normally that means it's released from customs
<wolfspraul> very good :-)
<wolfspraul> although I have to say I never once had a customs problem in Colombia
<wpwrak> kristianpaul: passed customs. next, it should go to the local distribution center, and very soon thereafter (typically within an hour), on the delivery car
<kristianpaul> wpwrak: sure, i already pay customs ;)
<wpwrak> oh, how did you do that ? normally, fedex pay and you then pay them
<kristianpaul> hehe
<kristianpaul> i call
<kristianpaul> fedex webpage is slow
<kristianpaul> i pay them
<kristianpaul> (fedex)
<wpwrak> maybe their payment procedure differs from country to country
<kristianpaul> sure
<kristianpaul> if dont pay they dont move a finger
<kristianpaul> so this time i call soon and pay first to avoid delays..
<kristianpaul> this people is so lazy..
<wpwrak> and what will you be getting ?
<kristianpaul> MM1+, Nannote (for a friend) microSD -> 100mil header breakout cable,  Scope meter !!!
<kristianpaul> i hope pick it tomorrow if the package deliver to Cali (i worh there) but deliver is set to Buga, and they SLOW DOWN all up to 2 days because they dont cover intermediated cities...
<kristianpaul> s/worh/work
<kristianpaul> DHL do cover intermediate cities but i guess is expensive
<wpwrak> (lots of stuff) wow ! btw, i think the device you'll get is called just "scope", not a "scope meter"
<kristianpaul> ah ok
<wpwrak> (complicated delivery) hmm, seems quite limited indeed
<wpwrak> DHL and fedex should have about the same prices. but i don't know if they charge extra for remote locations
<kristianpaul> DHL is cool, they deliver the wikireader to my mother at her office with their OWN
<wpwrak> i had some bad experiences with DHL. stuff getting stuck at customs and dhl being very unhelpful about it.
<kristianpaul> fedex uses a third party to deliver to city where i live, i dont like that
<kristianpaul> hmm
<kristianpaul> well UPS is the other choice
<kristianpaul> but thats luxury here
<kristianpaul> ah, scopemeter is a hand oscilloscope?
<wpwrak> yup
<wpwrak> the one you're getting is a little bigger :)
<kristianpaul> oh, of course i have a multimeter !
<kristianpaul> *meter > hand then?
<kristianpaul> ;)
<kristianpaul> hmm snownews looks promising, i'll keep an eye on it
<kristianpaul> Hilbert Transform C implementation...