KotCzarny | n9 is so dumb, 'you need sim card before you can use update time automatically' | 08:54 |
---|---|---|
KotCzarny | even if you have wifi connected | 08:55 |
KotCzarny | what idiot thought it was a good design | 08:55 |
sicelo | isn't N900 like that too? | 09:51 |
sicelo | because they both don't do ntp time ootb, but do nitz instead, hence sim card | 09:51 |
KotCzarny | still, bad design | 09:54 |
sicelo | :-) | 10:22 |
sicelo | not sure if it's just mine, but i noticed android also doesn't do ntp natively ... they do nitz as well | 10:22 |
KotCzarny | maybe depends on version/build | 10:29 |
KotCzarny | mine always syncs time whenever there is any working net | 10:30 |
bencoh | maybe they didn't want to mess with it because of the side-effect it might have on some networks | 10:42 |
KotCzarny | then just add configurability, and let user decide | 10:47 |
KotCzarny | at this moment it doesnt work if someone wanted to use n9 as a wifi only calling/watching | 10:48 |
sicelo | https://source.android.com/devices/tech/connect/time-source ... at least all along, android preferred nitz. | 10:49 |
sicelo | just tested that adb command on my android 11 device, and sure enough, it gets nitz time | 10:49 |
KotCzarny | try disabling gsm , desynch time and check if it gets time with wifi | 10:49 |
sicelo | KotCzarny: why not install ntpdate or similar on the N9? at least that's what i did on my N900, and added an ifup hook. so each time i connect to a network, i get update instantaneously | 10:50 |
KotCzarny | sicelo, i'm not saying it's not possible, but ootb it's a bad design | 10:50 |
KotCzarny | and requires user to search for working repos, install few packages, etc | 10:51 |
sicelo | it does, but it's wrong | 10:58 |
sicelo | :-) | 10:58 |
sicelo | honestly, that's why even on android i installed an ntp client | 10:58 |
KotCzarny | i havent touched any droid higher than 8-9, so hard to say what they are doing now | 10:59 |
sicelo | or maybe i should take out my sim ... | 10:59 |
KotCzarny | enable plane mode? | 10:59 |
KotCzarny | and turn on wifi, should be the same outcom | 10:59 |
sicelo | i was on plane mode | 11:00 |
sicelo | and got a time update, which is about 3 minutes behind the correct time | 11:00 |
KotCzarny | lol? | 11:00 |
sicelo | what? | 11:01 |
KotCzarny | 3 minutes desynch is a lot | 11:01 |
KotCzarny | unless you mean 'it got correct time with a delay of 3 min' | 11:01 |
sicelo | i'm sure it's nitz ... i know my operator's clock is out of sync (that's precisely why even on android i installed an ntp client) | 11:01 |
sicelo | just done it without sim. still time is wrong by almost 3 min ... i get correct date though | 11:03 |
sicelo | so yeah, no idea what's up with android and ntp | 11:03 |
KotCzarny | i think you can call 112 without sim | 11:04 |
sicelo | of course | 11:04 |
sicelo | so yes ... i'm likely being nitz'd either way :D | 11:04 |
sicelo | that link from android.com says ntp is first preference from Android 12 onwards | 11:05 |
sicelo | so i'm sure you're also not really getting ntp time :P | 11:05 |
KotCzarny | but n9 is stubborn and doesnt want to get time wiithout *working* gsm network | 11:05 |
sicelo | now i have to juggle the time thing again on my phone :'( | 11:07 |
sicelo | because the ntp application isn't root (and phone isn't rooted), it can't set the time directly | 11:07 |
KotCzarny | dont have spare phone for tests? | 11:08 |
sicelo | i was testing on my recent Android 11 phone - my daily. I know the behavior of my Android 8 one since I used that one for even longer | 11:09 |
joerg | sicelo: isn't 3 min the offset meanwhile accumulated on raw GPS time? | 12:59 |
joerg | oops nope >>such that there is now an 18 second difference between GPS time and UTC time<< | 13:11 |
joerg | 3 minutes must be a total glitch of your NTP or NITZ | 13:12 |
drathir_tor | joerg: keep on mind gps sync (cold start) could take a time... | 13:13 |
drathir_tor | joerg: however not sure if gsm network sync is possible in emergency (no sim) connection mode... | 13:14 |
drathir_tor | joerg: in 4 and especially 5G that could in theory works bc them as good remember heavely operate at time based sync with bts but not sure if that shared with user... | 13:17 |
joerg | that's a good question but alas impossible to answer since the radio stacks are never open source | 13:30 |
joerg | just occurred to me right now that the modem might still _receive_ signals even in no_SIM / airplane mode | 13:31 |
joerg | then, it's indeed plausible the modem searches for carriers to do a 911 call when needed | 13:32 |
joerg | and searching for carrier means decoding the data in carrier | 13:32 |
drathir_tor | joerg: hmmm... not sure of that if passive listen have any reliable usage purposes even if able to collect nerby bts info in such mode i assume, but w/o actively ping the station for extra data them could be not as accurate i assume... like there is possible made list of backup tower to switch into, but w/o active ping tover the data could be too inaccurate for use... | 13:38 |
joerg | to ping a tower you first have to get to know of the frequency and ID of the tower, so the data is highly useful to shorten / eliminate the scanning time when dialing 911. Old featurefones had "select network / available networks" function that took up to several minutes to find all networks available in your location | 13:43 |
joerg | this data got even stored to sim so on next power-up the time needed to scan could get shortened | 13:43 |
bencoh | n900 has that too :) | 13:46 |
joerg | ok, directly from Harald Welte: <LaF0rge> joerg: NITZ is in MM INFO, and that happens only over a fully-established bi-directional dedicated channel | 14:06 |
joerg | IOW no time data without SIM and login, or actually establishing a 911-call | 14:07 |
sicelo | maybe it wasn't nitz ... but whatever it is, it gives me time that's 3 minutes behind correct ntp time | 14:20 |
sicelo | anyway, i'd choose n900/n9 behavior any day, hehe. better to know you don't have correct time, than to think you have correct time, and wonder why your OTP codes don't work right | 14:20 |
joerg | :-) | 14:25 |
joerg | https://en.wikipedia.org/wiki/Mobility_management MM INFO | 14:25 |
joerg | haha! >> Additionally, unlike 3GPP2, which transmits GPS-sourced, millisecond resolution time via the sync channel, for NITZ, the "accuracy of the time information is in the order of minutes".[<< https://en.wikipedia.org/wiki/NITZ | 14:41 |
joerg | >><sicelo> just done it without sim. still time is wrong by almost 3 min<< embedded RTC? | 14:58 |
sicelo | mmm, i don't know, but i doubt | 14:59 |
joerg | simple to test: wrap phone tightly into alu foil before powering up ;-) | 15:00 |
joerg | and don't get fooled by false assumptions like I were: I bought a dedicated microwave oven as poor man's faraday cage | anechoic chamber. Turned out the phone still had signal inside the oven | 15:03 |
joerg | lambda/2 seals, haha. Only work for a particular frequency they are tuned to | 15:04 |
KotCzarny | lol | 15:05 |
KotCzarny | isnt gsm freq lower than mw ? | 15:06 |
joerg | yep, GSM is | 15:06 |
joerg | 900/1800 vs microwave 2400 | 15:07 |
joerg | or US: 850/1900 | 15:07 |
KotCzarny | yeah, so it should block it to some extent | 15:07 |
KotCzarny | maybe phone had quite good sensitivity | 15:07 |
joerg | well, a lambda/2 seal works by tuning the exact geometry of a gap and edges to match the wavelength of the particular frequency you want to block. All other ferquencies (except odd harmonics) may pass | 15:09 |
KotCzarny | well, foil the door then, i guess | 15:09 |
joerg | nowadays all microwave ovens use lambda/2 seals | 15:09 |
joerg | oh sorry, those are named lambda/4 seals | 15:25 |
joerg | aka quarter-wave choke | 15:32 |
joerg | https://patents.google.com/patent/US6867404B2/en | 15:35 |
joerg | also nice https://www.freepatentsonline.com/3956608.pdf same topic | 15:44 |
joerg | I could have known, learned it the hard way | 15:44 |
joerg | KotCzarny: indeed the signal was significantly attenuated, but not as much as would be needed to block out the phone from network | 16:01 |
joerg | actually if you own a microwave, it's very simple to test: place mobile phone into oven, DO NOT POWER OVEN UP, and - using a second phone - call the number off phone in oven | 16:05 |
KotCzarny | or even simpler, just look at the screen | 16:05 |
joerg | I just did and received a call without any indications of any problem | 16:05 |
KotCzarny | disabling screen blanking first though | 16:06 |
joerg | looking at the screen is a tad tricky inside microwave oven | 16:06 |
KotCzarny | mmm, flashlights | 16:06 |
KotCzarny | just kidding | 16:06 |
joerg | also doesn't tell you if the phone may transmit out of the chamber | 16:06 |
KotCzarny | scren is visible | 16:06 |
KotCzarny | wouldnt it lose conenctivity as it pings the tower periodically? | 16:07 |
joerg | phone starting to ring though is a clear indication of a two-way communication established between bts and phone | 16:07 |
joerg | phones don't ping bts very frequently, only up to once every 10 hours | 16:08 |
KotCzarny | still interesting fact, i would think that 0.5-1mm metal walls would block the gsm anyway | 16:08 |
joerg | the walls do, the o.5mm gap between door rim and oven body however only blocks 2400MHz | 16:09 |
joerg | I thought exactly like you, until I bought me an oven for exactly that purpose. 10 minutes after unboxing, I packed it back into its box ;-D | 16:10 |
KotCzarny | wouldnt just replacing the plastic seal with more metallic one help greatly? | 16:11 |
joerg | no, the opposite is true. The gap may be as small as a few micrometers. As long as there's no tight galvanic contact on 100% of area, such gap will leak mirowaves | 16:12 |
joerg | unless it's designed as quaterwave-choke, which you defeat by adding anything metallic there | 16:13 |
joerg | I seen very old comercial microwave ovens back in 1975 or somesuch, which indeed hat a metal fabric elastic seal and a plain stainless steel rim on oven body | 16:14 |
joerg | had* | 16:14 |
KotCzarny | so it would be possible to still convert it into a quite good faraday cage | 16:15 |
joerg | for this to work you need proper contact of fabric to door steel sheet 100% of length where it's mounted to door, and that fabric directly touching the metal body of oven when door closed | 16:16 |
joerg | and even then it's tricky to get those seals tight | 16:17 |
joerg | alternative: steel can (cookie can, coffee, sugar, whatever) with steel lid. May or may not work | 16:18 |
KotCzarny | might be too thin | 16:18 |
KotCzarny | but i guess phone ring test is easy to do | 16:19 |
joerg | if the lid has electrical contact to the can, it might work. No matter how "thin" the wall of the can | 16:19 |
joerg | what always works: alu foil bag, edges of foil carefully folded a 2 or 3 times | 16:21 |
joerg | or, if you're curious, use wire mesh, with mesh size <2mm | 16:23 |
joerg | instead of the alu foil | 16:24 |
joerg | yet to get tested: using rescue blanket, you can see through that one a maybe 0.1% | 16:25 |
joerg | probably too dark to read the display | 16:25 |
joerg | anyway rescue blanket has only one electrically conductive side, if at all | 16:25 |
joerg | so you'd need to make a real sack of one seamless piece tied together with a zip tie in one point | 16:27 |
joerg | silver side on the inside, hoping the zip tie will create some electrical contact between the pleats | 16:28 |
joerg | multi-layer, i.e. wrapping a sheet of foil several times around the phone, first one direction to form a tube, then rolling up the 2 ends of that tube several turns, should work too | 16:31 |
joerg | _should_, I'm not sure | 16:32 |
joerg | may only steel can I have here to test is filled with sugar :-/ | 16:36 |
joerg | my* | 16:36 |
KotCzarny | i use plastic jars, with quite strong walls | 16:36 |
KotCzarny | it's nice to see what is inside | 16:37 |
KotCzarny | :) | 16:37 |
joerg | of course, but such stuff is almost unobtainium here meanwhile | 16:37 |
KotCzarny | put the phone inside pc case | 16:37 |
KotCzarny | :) | 16:37 |
joerg | well, at least not found frequently "in the wild" | 16:37 |
KotCzarny | j/k | 16:37 |
joerg | indeed a PC case is _supposed_ to be RF-tight. But usually isn't | 16:38 |
KotCzarny | yeah, lots of holes | 16:39 |
drathir_tor | sicelo: oh and in theory You should normally get ntp sync over wifi too... | 20:19 |
sicelo | theory is the keyword ;) | 20:22 |
sicelo | whatever the phone is using to get time sync when i'm on wifi with plane mode on or sim out, i get time that's 3 min behind. i guess that's ntp. as for what servers the thing uses, i'm not even sure i want to know | 20:23 |
* drathir_tor now wonder if fridge would block better than microwave... ^^ ;p | 20:24 | |
KotCzarny | fridge has a rubber seal on the door, ie. wide open for the em waves to go | 20:24 |
KotCzarny | https://source.android.com/setup/start/android-12-release | 20:25 |
KotCzarny | Starting in Android 12, the framework prioritizes the Network Time Protocol (NTP) time source over the Network Identity and Time Zone (NITZ) source by default. | 20:25 |
sicelo | at least using a real ntp client has been working reliably for me, so i'm good. android isn't really something i have patience to tweak a lot | 20:25 |
KotCzarny | https://source.android.com/devices/tech/connect/time-source | 20:26 |
KotCzarny | if you have adb you can try some commands | 20:26 |
KotCzarny | To check how the device time is set, run the following command: | 20:26 |
KotCzarny | adb shell dumpsys time_detector | 20:26 |
sicelo | i did :p | 20:27 |
sicelo | i pasted this link earlier | 20:27 |
KotCzarny | ahm | 20:27 |
KotCzarny | i just got back home | 20:27 |
KotCzarny | and didnt scroll far enough | 20:27 |
KotCzarny | anyway, it seems droid12 prefers ntp over nitz | 20:28 |
drathir_tor | sicelo: more crazy idea fm rds sync ? | 20:28 |
KotCzarny | adb shell settings put global ntp_server <new-ntp-server> | 20:28 |
sicelo | yes, i have 11 and 8, and 4 on the droid 4 | 20:28 |
KotCzarny | adb shell settings get global ntp_server | 20:29 |
KotCzarny | check what it is configured to | 20:29 |
sicelo | and really, i'm happy with my setup on android with a third party ntp client. we were on N9/N900 :-) | 20:29 |
KotCzarny | :) | 20:29 |
sicelo | drathir_tor: sync time from fm? :D | 20:32 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!