#qi-hardware IRC log for Saturday, 2011-04-16

rjeffriesI assume Tabula uses the time dimension to accomplish its magic?  https://www.technologyreview.com/computing/37406/02:58
rjeffrieswpwrak speaking of SIMPLE networking protocols HsNeT is very simple and light http://www.hackeneering.com/Tiki/wiki/tiki-index.php?page=HsNet03:06
wpwrakrjeffries: (lighweight) oh sure, designing a non-standard protocol stack is always easy and fun. if it came to that, it wouldn't deny myself the pleasure of doing it entirely myself ;-)06:31
wpwrakrjeffries: (tabula) sounds as if they have a set of bits (8, apparently) for each config bit, plus a demux that picks one based on a central (?) clock. interesting approach.06:35
wpwrakrjeffries: such a thing may be able to increase the amount of available logic. won't help with the number of states, though.06:39
wpwrakrjeffries: you'd have to ask sebastien whether this is something that really helps. also, these are quite large chips, so they're targeting the very high end market.06:42
wpwrakah, nice ... finally found a good formula for visualizing position-dependent BER. hmm, the day i document all this, it will be a tome of fearsome mathematics06:54
wpwrakhmm, and bit_error_probability*packets needs to be > 100 before this statistical test even begins to look useful. grr ;-(08:24
wpwraki think i'll just hope atrf-path catches these as well :) (what i'm trying to detect is weak bypassing of RF power. if there's a problem with the capacitors, the signal strength will drop during a packet transmission. one way to find this is with fancy test equipment (e.g., an USRP). that measures the effect directly and is thus very accurate and fast. a less direct approach is to measure what this does on the probability of a bit goin08:29
wpwrakg bad. alas, this needs a lot of bad bits before any pattern emerges from the noise.)08:29
Action: whitequark is happy09:13
whitequarkbought a BIG 400ml can of FLUX-OFF09:13
wpwrakwhitequark: very good. if you splatter any flux on yourself, you can now bathe it off ;-))10:56
wpwrakwhitequark: now, what kind of flux is it for ? my guess would be RA/RMA resin.10:57
whitequarkwpwrak: they assure that it could remove _all_ kinds of flux12:33
whitequarkwell, that does not correlate well with reality :/12:34
whitequarkbut the 35W ultrasonic bath I've also got is truly awesome12:35
whitequarkit got rid of both RMA and that funky TT flux in just several minutes12:35
Fusinhi /me ;)12:47
FusinHow do I install Vista on nanonote? ;)13:01
Fusinmore srious question: how to install an 8gb microSD card13:04
Fusinwhich device is the sd-slot? how to partition an how to format?13:05
Fusinis ext3fs good, or is another fs more accurate?13:06
viric/dev/mmc*, for the SD block devices13:07
Fusinthx for this information :)13:08
viricfor fs, you will find many different preferences13:09
Fusini want to use the 8gb m ostly for music, pictures and maybe video. mostly static (seldom rewrite)13:10
viricI don't think fs are going to make any difference about that13:11
whitequarkwpwrak: I've finally did that. the seventh CP2102 finally works!! :)13:12
whitequarkwpwrak: probably that NC pin was the caus13:12
whitequarkwpwrak: also, it still heats, too much for a simple usb-uart bridge, but much less than before13:15
wpwrakwhitequark: (ultrasound) nice !14:05
wpwrakwhitequark: (#7 lives) even better ! ;-)14:05
Fusinone step forward ;)14:11
wpwrakultrasound sounds interesting. could be more efficient than all that scrubbing i do.14:18
rozzinHm. Seems my openwrt build-environment is being contaminated by things on the host machine.14:26
rozzinPygame finds an smpeg header, and wants to use it.14:29
whitequarkwpwrak: ultrasound is ultra-awesome14:29
whitequarkwpwrak: i can't ever say HOW it is freaking awesome14:29
wpwrakwhitequark: ;-) did you get one with heating ?14:30
whitequarkwpwrak: it is even more awesome when combined with flux-off. then all the dirt just goes away in beautiful stains14:30
wpwrakthey're cheaper than i thought. low end starts < USD 100, even in argentina.14:30
wpwrakwow. strains. your boards must be really dirty ;-)14:31
whitequarkwpwrak: yeah, my one is just a bit more expensive than $100, and it is really big, compared to some cheaper ones on the market14:32
whitequarkwpwrak: (strains) the dictionary said 'stains'. well, there was a lot of RMA residues. you know, they're nonconductive, but ugly14:32
whitequarkand that thing got all the board perfectly clean. no flux at all14:33
whitequarkand, according to my feelings, no fat (from fingers) too, through that may be the combination of flux-off with water14:34
wpwrak(non-conductive) you may find that they aren't really ... i had circuits that didn't come out of reset because the flux conducted well enough to partially cancel the effect of a pull-up.14:34
whitequarkhuh, now I don't have to worry about that at all =)14:34
rozzinOh, that might explain why dbus-x needed me to manually build libSM, too....14:36
whitequarkwpwrak: so, my opinion is that the device (CT-405) is the awesome essence of awesomeness14:38
wpwrakinteresting stuff indeed. i find the endless scrubbing quite inconvenient (and also a bit risky), so that may be something useful to try.14:38
whitequarkwpwrak: absolutely.14:40
Action: whitequark is very excited about it and ultrasound-washes everything he sees14:41
rozzinHm. Nope--dbus-x needing libSM but not specifying that looks like a bug.14:53
rozzinOh, nope--doesn't depend on libICE, either.14:55
Action: rozzin removes libice-dev from his host system, does a `make clean' on libICE in openwrt, and watches dbus-x build successfully without libICE or libSM.14:58
rozzinI wonder how many packages this is happening for....15:03
Action: rozzin starts rethinking autoconfiguration for his own packages....15:03
whitequarkanyone knows how to etch brass?15:37
C-Keenwhitequark: http://www.instructables.com/id/Etching-brass-plates/15:41
C-Keenwhitequark: they require registration but bugmenot usually has an account that works15:41
rozzinI wonder if dbus really needs to be using AC_PATH_XTRA rather than AC_PATH_X....15:45
whitequarkC-Keen: hm, looks like FeCl3 is fine for that15:48
C-Keenwhitequark: I have to say that I have never tried this myself15:53
whitequarkC-Keen: I'm trying that right now ;) so you'll soon have some real info about the process15:55
whitequarkI want to etch the brass plates to make solder paste masks. manually applying the paste for ~8 board is boring enough15:56
Fusinwhat's wrong with ubuntu? doesn't know /dev/usb0 (I see only /dev/usbmonX'es)16:38
FusinLaptop is an eeePC 1000 HG, internet connection is per GPRS (cellphone-chip)16:40
whitequarkFusin: that's a network interface, not a device file16:55
whitequarktry "ip link"16:55
Fusinright, internet goes over ppp0, but for connecting the nanonote i need an usb0 (as in booklet and wiki), but i can't ifconfig usb016:58
Fusinhow do i setup that one?16:58
whitequarkFusin: well, if you load a corresponding module on nanonote and plug it into usb, the interface will appear automatically16:59
whitequarkI don't have a NN, so you probably should ask someone else how to do that exactly on your distribution16:59
Fusinoh, i just bootet the nano, and voila: the usb0 is availlable in ubuntu :D17:00
C-KeenFusin: the usb0 network device will only be available when the device is detected17:00
C-KeenFusin: so only if you plug in your powered nanonote you will see it17:01
Fusinseems so17:01
C-Keenthat's how it works17:01
C-Keenyour kernel will see the even on usb, match the product / vendor id against a list and load up a network device driver that will act under the name of usb017:02
C-Keenwithout the usb events happening there will be no device17:02
wpwrakwhitequark: (clean everything) i hope you don't have pets ;-)17:13
wpwrakwhitequark: interesting: http://www.circuitnet.com/articles/article_40106.shtml19:04
wpwrakwhitequark: low power, high frequency, and if possible variable frequency19:05
--- Sun Apr 17 201100:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!