#qi-hardware IRC log for Tuesday, 2015-02-17

wpwrakxiangfu: nice pictures !  in Anelok-led-on.jpg, is that LED really on ?05:38
wpwrakalso, in Anelok-wrong-usb-connector.jpg the connector looks good. are you sure this it the wrong one ?05:38
xiangfuwpwrak: Hi. I mean the oled was on. :)05:39
wpwrakaaah ! ;-)05:39
xiangfuwpwrak: the connector pin was not 100% match the pad.05:39
wpwrakdid that cause a problem ?05:40
wpwrakit looks as if it should still work. though the leg hidden under the cable may be a bit close to the via05:41
xiangfuthe connection is right. but the pin is not 100% match the PCB pad.05:42
xiangfuno. it works fine. and pretty stable.05:43
wpwrakah well, details :) i'm not using the connector correctly anyway - there should be board cut-outs for the little "horns". of course, i noticed that only very late ...05:43
xiangfunext time I will find the ring connector.05:43
xiangfuah. Yes.05:43
wpwrakby the way, i made some changes that may break things. 1) i fixed a bug in touch screen polling that - ironically - made it work. i'm still not entirely sure why it mattered, but it did. i also changed how the input system handles time, so that it should now be independent from delays elsewhere. but this may introduce surprises. my devices work fine with the new code, though, so hopefully i didn't do too much damage.05:46
wpwrak2) i kicked out the non-open register definitions and linker script and replaced them with things i can legally redistribute. had to make some small changes for that, so i may have broken something. things look good here, though, so again there's hope that nothing bad will happen. just wanted to warn you.05:48
xiangfuI will test the new code. then let you know.05:52
xiangfulunch time.05:54
wpwrakin general, the touch sensor driver will see more changes in the future: right now it polls actively which a) wastes quite a lot of energy, especially in standby, and b) makes its timing fairly chaotic. the above changes should already bring some sanity to input.c, but touch.c still needs more work. the idea is to make it interrupt-driven and then let the CPU enter a sleep mode when there's nothing else to do. all this means that the 05:54
wpwrakcharacteristics of the touch sensor may change a bit. e.g., the threshold value and such may have to be adjusted at some point.05:54
DocScrutinizer05forget the particular schematics, but the service digikey offers here actually looks very smart:  http://www.digikey.com/schemeit/#wp611:16
DocScrutinizer05the idea is great anyway11:16
DocScrutinizer05EDA as a web service, offered with component lib already11:18
whitequark>EDA as a web service11:25
whitequarkand in a year they get acquihired and write you a message about "an incredible journey" you had together, and tell you they'll delete all your designs in a month11:26
viricI envision Xilinx software for the web11:28
whitequarkthey'll get there eventually11:29
viricnice, this incredible journey11:29
whitequarkbut it'll be just same old shit, now in a web browser11:29
whitequarkno innovation in EDA11:29
viricbut you will be able to design from your iphone11:29
eintopfvia iphone while sitting on the toilet!11:32
whitequarkso that's why none of the iot devices work11:32
Action: eintopf making iot stuff11:35
eintopfso my work doesn't work? Yea, it's mostly wip.11:35
whitequarkeintopf: does your iot stuff depend on a central server provided by $COMPANY?11:38
whitequarkdoes it stash a remotely exploitable ancient version of linux somewhere it's accessible by network?11:39
eintopfI improve the 802.15.4 /6lowpan stack in linux kernel to make it in some useable state11:41
eintopfall open source11:41
eintopfcommon solution with 6lowpan/802.15.4 is a contiki mcu as adaptation layer11:42
eintopfbut, we want remove the contiki stuff11:43
whitequarkyeah, contiki is terribad11:43
eintopfwhitequark: I also working on a contiki replacement. The goal is to have a easy development -> copy linux implemented IPv6 IoT application into the $CONTIKI_REPLACEMENT and just add it to the buildsystem11:45
eintopfI mean from userspace view, it should be the same interface11:45
eintopfalso contains, Kconfig as builsystem and such things... like linux11:46
eintopfbut this is my little hobby project11:46
eintopffirst I want only add a "generic" firmware application for atusb, so we can also run the firmware on the RZUSBSTICK (it's something like atusb, but different mcu and at86rf230)11:47
eintopfthen you need just replace two ".config" configs to making one atusb firmware and one RZUSBSTICK firmware11:47
eintopfbut I will keep the name atusb firmware :)11:47
eintopfso, that are big plans...11:48
Akiraahello, are you aware of ready-made solutions to implement a 'mass storage device' api for a device like 'banana pi' or olinuxino, boards that have USB-OTG11:48
Akiraaso that they may present themselves as 'removable storage' to computers or other USB hosts11:48
whitequarkeintopf: lately i have been doubting how wise it is to put linux in iot devices11:48
whitequarkheartbleed, right? basically none embedded devices with heartbleed will ever get patched11:49
whitequarkand if we find an exploitable vulnerability in the network stack, it's going to be an apocalypse, pretty much11:49
whitequarks/if/when/ really11:49
eintopfwhitequark: time will changes and for a IPv6<->6LoWPAN bridge is linux the best solution11:49
eintopfand thats a stationary nodes with power connection11:50
Akiraawhitequark: regardless, IoT devices are too low power to survive attacks if hooked directly to the internet, they require a separate box to act as a firewall, rate-limiter, load-balancer etc.11:50
whitequarkhow many routers get vendor firmware updates?11:50
whitequarkhow many vendor firmwares aren't horribly vulnerable by themselves?11:50
whitequarkAkiraa: well, most iot devices i have seen so far are pretty big. think nest thermostat or smoke11:51
whitequarkthat's a full-fledged linux board. the immediate future will be dominated by this form11:51
Akiraathey are expensive enough to have a fully powered CPU inside that can handle crypto, I guess, but wouldn't bet on it11:52
whitequarkwell, you /have/ to handle crypto in /any/ wireless device, it's not even a question, but this is not what i am talking about 11:52
whitequarkthey use arm boards running linux and openssl and all that11:52
whitequarkand that doesn't help11:52
eintopfthese devices does exists already, atheros sells arm socs with integradted 802.1111:53
eintopfI want something like that, but with 802.15.411:53
eintopfwith ~4MB flash11:53
eintopfsuch things, all is in one chip11:54
eintopfbut do really power managment with very low power and via battery...11:54
eintopfdon't know how possible that is currently11:54
AkiraaI would be interested even in some useful keywords for search if you can help (implementing USB "mass storage device" from a linux box with USB-OTG)11:56
wpwrakAkiraa: g_file_storage seems to be what you're looking for. (search term  linux gadget mass storage)13:21
wpwrakwhitequark: (web-based EDA) we also have mbed, so also sw engineers can work the modern way :)13:23
wpwrakand haven't we come a long way ? from SaaS (whatever-S-was-as-a-Service) to X-a-a-S *(Whatever-as-a-Service) now to YJaaS (Your-Job-...), all the way building up to YCaaS (Your Company)13:24
wpwrakthinking of it, the mafia selling protection and vandalism insurance did some quite pioneering work in that domain :)13:24
Akiraawpwrak: thanks13:30
DocScrutinizer054 new snapshots about BPW34: http://neo900.org/stuff/joerg/random-media/IR-photodiode/BPW34+1kR_electron-flash-1m-dist.jpg http://neo900.org/stuff/joerg/random-media/IR-photodiode/*14:25
DocScrutinizer05the funny stray signal at start of http://neo900.org/stuff/joerg/random-media/IR-photodiode/BPW34+1kR_electron-flash_detail2us_rise.jpg is prolly via inductive or capacitive interference from camera to the 5cm GND wire of the probe14:29
DocScrutinizer05most likely inductive, by the trigger inductor that creates high voltage to trigger electrode of electron flash tube14:30
DocScrutinizer05sidenote: the BPW provides an incredible 25mW power from a tiny electron flash fired ~10cm distant from it14:36
DocScrutinizer05even with the aleggedly "1m dist" which actually rather been 60cm, in http://neo900.org/stuff/joerg/random-media/IR-photodiode/BPW34+1kR_electron-flash-1m-dist.jpg14:37
wpwraknext: "wireless charging" :-)14:38
DocScrutinizer05for 10cm ( http://neo900.org/stuff/joerg/random-media/IR-photodiode/BPW34+1kR_electron-flash.jpg ) it's even 49mW14:39
DocScrutinizer05well, I guess you can calculate the distance ratio from power ratio more accurate than my guesstimations14:40
DocScrutinizer05I bet it's exactly the radiation-power / area  *  efficiency14:41
wpwrakthat would indeed kinda make sense :)14:41
DocScrutinizer05but the camera has a way to modulate flash power. Dunno how it does that14:42
DocScrutinizer05anyway SEVEN volt!!14:42
DocScrutinizer05now you see why they test photovoltaic panels in factory with strong electron flashlight14:43
DocScrutinizer05I guess on defective PV panels some traces simply will burn out14:44
wpwrakit's good to know that the panels will be able to resist the sun going supernova, just in case :)14:45
wpwrakthe aftermath will be messy enough. i guess the panel makers wouldn't want to have to worry about warranty claims on top of it, too14:46
DocScrutinizer05thunder lighting14:46
wpwrakyou must have quite some fireworks ;-)14:48
DocScrutinizer05there exists quite some fireworks, yes14:49
DocScrutinizer05some of those 20cm diameter bombs are filled with flash power14:50
DocScrutinizer05well, maybe 12cm only14:50
eintopfz/win 314:52
eintopfI meant win 30014:52
eintopfjust for clarify14:52
DocScrutinizer05hmmm http://www.elektronik-labor.de/Projekte/Alpha.html15:08
eintopfDocScrutinizer05: why do you need something like that?15:14
DocScrutinizer05just pondering synergies15:15
eintopfwahhh, the guy hold a radioactive clock15:15
DocScrutinizer05well, the girls who painted the phosphorescent paint on the clock hands formed their brushes by licking them ;-o15:16
DocScrutinizer05no surprise they _all_ died during max 10 years iirc15:17
eintopfDocScrutinizer05: do you think in moscow is the radioactive higher than in... germany?15:18
DocScrutinizer05no, why?15:18
eintopfdon't know, maybe because it's easier to get radioactive material15:19
eintopf(don't know if this is true)15:19
DocScrutinizer05well, the _total_ average radioactivity, incl all hotspots, even shielded ones, for sure is higher15:19
DocScrutinizer05for 99.999% of normal places though I don't think there's a significant difference15:20
eintopfah, ok.15:20
DocScrutinizer05in germany we have quite some vaiance in average environmental radioactivity too. Thanks to Chernobyl fallout which hit mostly Bayern15:22
whitequarkmoscow underground uses quite a bit of granite15:22
DocScrutinizer05yeah, there might be Radon15:22
pcercuei< DocScrutinizer05> in germany we have quite some vaiance in average environmental radioactivity too. Thanks to Chernobyl fallout which hit mostly Bayern15:23
pcercueinow everything suddently makes sense :)15:23
eintopfahh, I know radon. You can use it for measurement time?15:23
eintopfinstead of quarz15:23
whitequarkatomic clocks do not use radioactivity15:24
DocScrutinizer05they work like a microwave oven15:24
whitequarkthey use... basically resonance of the nucleus of some elements, cesium usually15:24
DocScrutinizer05filled with vapor of usually Caesium15:24
eintopfyes, caesium... but there exists something between caesium and quarz. Something with R...15:25
whitequarkso the thing is, the nucleus of caesium has spin, and that spin can flip, and flipping it takes a very small and also very consistent amount of energy15:25
eintopfother material15:25
whitequarkwhich happens to be in microwave range15:25
DocScrutinizer05aaah maybe15:25
whitequarkso you can use that difference as a frequency standard15:25
DocScrutinizer05some weird stuff, yes I seem to remember, Notr Raduim. Ruthenium?15:25
eintopfyeaa rubidium15:26
eintopfI know some guys sells that on ebay15:26
whitequarkit's just an alkali metal15:26
DocScrutinizer05well, it's maybe something pretty "normal"15:26
whitequarkit catches fire in air15:26
whitequarkotherwise nothing very interesting about it15:27
DocScrutinizer05like alkaline15:27
whitequarknow what /is/ interesting is chip-scale atomic clocks15:27
whitequarkfor use in GPS chipsets15:27
whitequarkcommercially available15:27
eintopfindoor gps comes possible15:28
whitequarknot really15:28
eintopfah, ok.15:28
whitequarkwhat this makes possible is instant fix15:28
whitequarkand more precision15:28
DocScrutinizer05indoor is not related to timebase15:28
whitequarki've seen a fun paper about indoor15:29
whitequarkusing MIMO receiver to fingerprint various locations15:29
whitequarki.e. NOT triangulation, that doesn't take into account multipath15:29
eintopfI know the solution to use RF and ultrasonic15:29
DocScrutinizer0535g. 17cm^3. I wonder how many bucks15:29
whitequarkrather this one /depends/ on multipath propagation to make fingerprints unique15:29
eintopfto measurement the ultrasonic wave traveling time15:29
whitequarkeintopf: that's not really used15:29
eintopfyea, but works in labor enviornment15:30
whitequarkwell, if you need to measure a meter or two, sure15:30
DocScrutinizer05whitequark: hehe, sounds like what I started in part1 of "GPS revised"15:30
whitequarkbut there are indoor positioning systems for warehouses15:30
eintopfRFID marks on the floor?15:30
whitequarkeintopf: that's a cheap way yes15:31
whitequarkbarcodes on the ceiling15:31
whitequarkif you are REALLY cheap15:31
whitequarkor you could use http://www.mit.edu/people/jhow/papers/Prigge_thesis_draft.pdf15:31
eintopfyou mean something like this?15:31
DocScrutinizer05>>Front of OM Ap. building, 3m away<<15:32
DocScrutinizer05that's the concept of highres-fingerprinting, done by me in 200815:33
eintopfI make a bookmark to the thesis, thanks15:33
DocScrutinizer05>>The max and min readings where obtained by moving the phone ~60cm!<<15:34
DocScrutinizer05once you got a "map" of your location fingerprints, you can even do pattern matching as soon as your receiver is moving15:36
DocScrutinizer05assuming there will never be jumps / teleports from ne location to another15:36
DocScrutinizer05assuming further that movement usually is mostly linear, you can check the sample pattern against plausible fottprint pattern sequences15:37
DocScrutinizer05but the most reliable and simple indoor location is U-TDOA done with e.g. WLAN15:39
whitequarkdon't you need many access points for that?15:40
DocScrutinizer05when you're the location owner and want to locate "arbitrary" visitors that happen to have a WLAN-enabled device with them15:40
whitequarkwith precisely synchronized clocks15:40
DocScrutinizer05wpwrak: yes15:41
DocScrutinizer05the E-OTD-alike "fingerprinting" by calculation of reflexion/multipath phase to original signal needs pretty specialized receivers15:44
DocScrutinizer05U-TDOA nees specialized APs15:44
DocScrutinizer05poor man's E-OTD like in http://lists.openmoko.org/pipermail/openmoko-kernel/2008-April/002434.html relies on interference between direct and multipath signal creating local standing waves that modulate signal aplitude over distances in the range a a few dozen or hundred wavelengths15:46
DocScrutinizer05you also need several distinct stationary transmitters for that15:47
DocScrutinizer05downside: such standing-wave-interferences are very unstable, they change a lot with every change in environment, like different temperature or air humidity15:49
DocScrutinizer05s/wprak/whitequark/ sorry15:49
whitequarkrelevant: http://hackaday.com/2015/02/17/mapping-wifi-signals-in-3-dimensions15:53
DocScrutinizer05U-TDOA is the method of choice. And for smartphones many carriers already offer location detection assisted by U-TDOA15:53
DocScrutinizer05UMTS--U-TDOA works fine indoors, and incredibly fast too. My N900 meanwhile gets a location fix during no more than 4s no matter where I am15:55
DocScrutinizer05thanks to the GPS controlled by the modem15:56
infobotrrlp is, like, the Radio Resource LCS (Location Service) Protocol as specified first in GSM TS 04.31, or http://security.osmocom.org/trac/wiki/RRLP15:57
DocScrutinizer05I didn't look into the RRLP detals but I guess there must be info from BTS to MT about the actual location of MT as seen by "the network" - in addition to the mere A-GPS "raw" data about which codes to check for at which frequency16:01
DocScrutinizer05so the N900 GPS gets enabled, the modem controlling the GPS asks network for RRLP assistance, network does a U-TDOA trilateration and sends the info with a estimated horizontal precision of maybe 10 or 20m back to the modem, which then decides this is already a nice "first fix" even when GPS has not yet correlated to the SVs16:05
DocScrutinizer05(this is _my_ semi-educated guess on what's going on)16:05
DocScrutinizer05quite possible that carriers do U-TDOA for _all_ MT they service, by default, nowadays16:07
DocScrutinizer05all the time16:07
DocScrutinizer05why waste a transmission? why delay stuff when you can do it in advance already, so to speed up LCS16:09
DocScrutinizer05and it's known the TLA agencies run their own independent U-TDOA networks, I just wonder about the bandwidth of RX bands they receive and record with timestamp for later computations when they are interested in a particular transmitter's exact position16:17
whitequarkthey seem to be using stingrays actually16:18
DocScrutinizer05what's a stingray?16:19
whitequarkfake BTS16:19
DocScrutinizer05well, yes, that too16:19
whitequarkused for intercepting calls / SMS / locating16:19
DocScrutinizer05just to crack encryption16:19
DocScrutinizer05when they want to eavesdrop on a particular communication16:19
whitequarki doubt they'd use stingrays, which are relatively easily detected by public, instead of proper U-TDOA if they could16:19
DocScrutinizer05U-TDOA only records, it doesn't do MITM16:20
DocScrutinizer05so encryption is effective in the U-TDOA records16:20
whitequarkwell you can do any MITM you want using the network as well as a stingray16:20
DocScrutinizer05as long as it's effective at all16:20
DocScrutinizer05the purpose of U_TDOA is completely different than that of an IMSI-catcher anyway16:21
DocScrutinizer05and actually I guess only the "poor folks" like regular law enforcement still use IMSI-catchers. The mighty TLA's use direct interception interface of the carriers16:23
DocScrutinizer05U-TDOA however is not limited to cellphones16:23
DocScrutinizer05they can trilaterate virtually every RF-device16:24
DocScrutinizer05and prolly not only real-time but alsio from records, for at least hours back16:25
DocScrutinizer05only need to know the signal signature16:29
DocScrutinizer05I wonder if they already record the complete spectrum from say 50kHz to 4GHz16:32
DocScrutinizer05prolly long wave is useless for U-TDOA16:32
DocScrutinizer05since U-TDOA by physics can't go significantly more accurate than a fraction of wavelength16:33
whitequarkwhat if you also record phase?16:34
DocScrutinizer05that's why I say "fraction of wavelength"16:35
DocScrutinizer05you can't record phase to arbitrary precision16:35
DocScrutinizer05sort of RF Heissenberg16:35
DocScrutinizer05anyway assume they'd sample 16bit at 8G-samples/s16:36
DocScrutinizer05that's a "ridiculous" 16GB/s16:37
DocScrutinizer05and somewhat covers 0...4GHz16:37
DocScrutinizer0516bit logarithmic would prolly also yield sufficient precision as well as sensitivity16:39
DocScrutinizer05if not, go 24 or even 3216:39
DocScrutinizer05place a grid of such "total receivers"  evenly spaced over a country, grid pitch of maybe 50 or 100km16:40
DocScrutinizer05record *all* transmitting devices, incl location down to 10m precision16:41
DocScrutinizer05data volume in the 1digit TB per minte range, per station16:42
DocScrutinizer05how many 2TB HDDs will they spend on that?16:43
DocScrutinizer05are a 1000 per station reasonable?16:43
DocScrutinizer05(long wave) even for 100kHz the wavelength is 2.7km if my math doesn't suck16:46
whitequark10PB per week per station16:46
whitequark1000 HDDs per station is a LOT16:46
whitequarkthat would be a substantial part of global production16:47
DocScrutinizer051000 a lot? why?16:47
DocScrutinizer05if somebody uses a significant fraction of global HDD production then that's the NSA data vcenter in... err Utah?16:48
DocScrutinizer05for germany you would prolly get away with some 2 digit number of stations, that would sum up to maybe 80k HDD. Samsumg sells more laptops every day16:49
DocScrutinizer05and I bet my scratchpad design can get significantly streamlined by some network engineers16:51
DocScrutinizer05and some smart math cracks to apply data compression and whatnot16:52
DocScrutinizer05nd maybe they do snapshots whenever something worrying happens16:54
whitequarkthe only way you can do snapshots with this amount of data is physically moving hard drives around16:55
DocScrutinizer05well, call it "freeze"16:55
whitequarknothing else comes close16:55
whitequarkin bandwidth16:55
whitequarkamazon offers a backup service where they record backup on HDDs and mail them you16:56
DocScrutinizer05and yes, I actually expect them to carry those HDD arounf in racks a 128 each16:56
DocScrutinizer05some guy pushes a button for station 2,3,4,18,19,20. A 2h later a lorry with spare racks drives by each of those 6 stations and some guys swap the new racks for the snapshotted ones16:58
DocScrutinizer05the poor noob is the one to carry the 8 new racks up to 3rd floor and the ones with the data down to the lorry16:59
DocScrutinizer05nothing comes close to the bandwidth of a lorry full of HDD17:02
DocScrutinizer05poor RTT but mega bandwidth ;-)17:02
DocScrutinizer05but hey, connect the stations with decent DSL to make a grid of distributed computing and analyze the data without moving it around in lorries17:04
DocScrutinizer05after all you're not interested in the complete set of raw data, only in a very specific negligible part of it17:05
DocScrutinizer05so as long as the stations have sufficient spare HDD for the next snapshot/freeze, no need to supply them with swap racks17:06
DocScrutinizer05a simple calculation: you want to save or swap or process the spare set of HDD with that snapshot before the HDD set in charge to store new data is filled up17:09
DocScrutinizer05unless you got a 3rd set of HDD, aka a 2nd spare17:10
DocScrutinizer05anyway consider *total* radio surveillance a fact17:18
whitequarka fact is way too strong. it's a possibility17:19
whitequarkit's entirely possible that it doesn't offer enough return for the (substantial) investment17:19
DocScrutinizer05in that context I consider everything faintly possible a given fact and already accomplished17:19
whitequarkso, where's our cold fusion plants?17:20
DocScrutinizer05the investment is negligible for the return 17:20
whitequarkor even hot fusion :]17:20
DocScrutinizer05how, that context?17:20
DocScrutinizer05to supply NSA with electricity?17:21
DocScrutinizer05and you forgot that fusion is not even theoretically stable yet17:22
DocScrutinizer05they hope to make it *work* in maybe 20 years or so17:22
DocScrutinizer05a 0..8GHz RF scanner is "available on ebay"17:22
DocScrutinizer05and expense for one station is prolly significantly < 1 million17:23
DocScrutinizer05that's peanuts for those guys17:23
DocScrutinizer05particularly when return benefit is "total TF surveillance". They'd pay 1000 times as much for that17:24
DocScrutinizer05s/ TF / RF /17:24
whitequarkhard drives: $80,00017:24
whitequarkyeah sounds about right17:25
DocScrutinizer05http://www.mit.edu/people/jhow/papers/Prigge_thesis_draft.pdf  >> I certify that I have read this dissertation and that, in my opinion, it is fully adequate in scope and quality as a dissertation for the degree of **Doctor of Philosophy.**<< Oh WOW, what PHILOSOPHY Doctors deal with, incredible17:44
whitequarkmeans Doctor of Philosophy17:45
whitequarkregardless of the actual domain17:45
wpwrakcf. "natural philosophy" :)17:49
DocScrutinizer05I guess I'm spoiled by the very precise german nomenclature regarding academic degrees, and the bad reputation _our_ philosophy graduates have, regarding their perspective and usefulness for real life18:02
DocScrutinizer05nevr heard of "natural philosophy"18:03
DocScrutinizer05(very precise) well until recently they introduced all that "bachelor" (SIC)and whatnot stuff18:06
wpwrakthey have perspectives ? :)18:09
wpwrak(nat. phil.) -> wikipedia. and it seems that you haven't read the fine works of Neal Stephenson :)18:11
DocScrutinizer05>>Deutschland: Dr. rer. nat. für lateinisch doctor rerum naturalium<<18:23
DocScrutinizer05anyway seems I confused PhD with Dr. phil.18:29
whitequarkno you did not18:31
whitequark"Doctor of Philosophy, abbreviated as PhD,"18:31
DocScrutinizer05Ohhh >>Dr. phil. nat. (philosophiae naturalis): Doktor der Naturwissenschaften, wie er an der Johann-Wolfgang-Goethe-Universität Frankfurt am Main anstelle des Dr. rer. nat. verliehen wird.<<18:31
wpwrakthe german and the anglosaxon system are different18:32
wpwrakthe anglosaxon system is basically based on the scientific culture of ancient greece, where all the "scientists" (as opposed to craftsmen) where "philosophers"18:33
DocScrutinizer05in Germany Philosophers are hardly scientists at all ;-)18:33
wpwrak(dr. phil. nat.) of course, no rule without exception ;-)18:34
DocScrutinizer05didn't know of that18:34
DocScrutinizer05anglo PhD == german Dr.18:34
DocScrutinizer05german Dr. phil. ~= PhD of Philosophy18:35
DocScrutinizer05or whatever18:35
wpwrakyes, they're equivalent. but the PhD comes from the origins of modern science in england, which in turn - that being the renaissance - adopted greek terminology18:36
DocScrutinizer05while Germany didn't (except Johann-Wolfgang-Goethe-Universität)18:36
wpwrakthe german "dr." probably just shed the "phil." at some point18:37
DocScrutinizer05whatever, puzzled me18:38
DocScrutinizer05here Doctors of Philosophy usually end as taxi drivers18:38
wpwraksurely a great place to discuss weighty matters of the universe - and all that transcends it - with a diverse audience18:40
DocScrutinizer05hehe, yes18:41
DocScrutinizer05some make it into an occasional TV show18:42
DocScrutinizer05talk shaw ;-D18:42
DocScrutinizer05industry anyway rarely ever pays for their living18:43
DocScrutinizer05and I'd feel scared when they now started messing with engineering topics18:44
wpwraksome of the more meandering standards could come from this sort of branch of society ...18:45
DocScrutinizer05yes, USB for example18:45
wpwrakthey provide the structural discourse, some engineering grunts get to fill in the numbers18:46
DocScrutinizer05the only plausble explanation for stuff like CARKIT18:47
DocScrutinizer05OSI also sounded like that sort of thing18:48
wpwraki'd attribute this more to the easy availability of cheap glue :)18:48
wpwrak(carkit, that is. osi, itu, etsi, .. yes)18:49
DocScrutinizer05oooh, now that took until the glue pun arrived. Oh yes, that's the better explanation18:50
Action: DocScrutinizer05 never again will get that glue out of the USB jack18:51
Action: DocScrutinizer05 might need to open up an Edding pen and see what it smells like18:52
wpwrakenjoy the trip ! :)18:52
DocScrutinizer05returning to previous monologue, I wonder if there already are digital frontend receivers available, i.e. stuff that digitizes directly after antenna and does all further signal processing digitally18:56
DocScrutinizer05I'm not even sure if some spectrum analyzers maybe work that way19:23
whitequarkI believe satellite receivers almost work that way19:27
whitequarkthey digitize the IF signal, not original19:29
DocScrutinizer05well, the IF still is some 2GHz, right?19:30
DocScrutinizer05I guess what's a major roadblock for doing this sort of stuff on plain direct RF is the AGC you find in any conventional receiver19:31
DocScrutinizer05a front end digitizer would need a dynamic range of... well some 80 or 90 dB19:32
DocScrutinizer05I wonder if the SAT LNBs have AGC built in19:33
DocScrutinizer05Low Noise Block(?)19:34
whitequark"low-noise block downconverter"19:34
DocScrutinizer05also SAT signals are relatively "clean"19:36
DocScrutinizer05i.e. no far off signals of high amplitude19:36
DocScrutinizer05the LNB already filters all that out19:37
DocScrutinizer05a - say - terrestrial digital frontend receiver would have to cope with signals that are maybe 80 below the "noise" from a nearby broadcast station19:39
DocScrutinizer05and since we don't do any filtering before ADC in our simple all digital design, all that "noise" (which actually is a valid signal as well, for above discussed total surveillance receiver) needs to get digitized as well, while not covering and garbling very weak signals19:41
DocScrutinizer05so yes, the ADC for such a receiver will need quite a dynamic range and will cost a few bucks19:42
DocScrutinizer05simply switching off the LO in RTL2832 dongle should turn it into a shortwave receiver with frontend digitizing. Would need to remove any highpass filters in input though19:46
DocScrutinizer05I think some guys already did this19:47
DocScrutinizer05anyway with such evidently feasible ~20Mhz upper freq digital frontend receiver, you could already receive _all_ shortwave broadcast with just one computer and store it to your HDD, and later on "tune in" to whatever broadcast station you like on that stored data19:55
DocScrutinizer05s/_all_/concurrently, simultaneously _all_/19:56
--- Wed Feb 18 201500:00

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