Hey all,
I have an idea for an application that I think could be really helpful. With more skill and experience, I might be able to write it myself but I lack in both of those fields
The general idea:
An application that can be used to measure the battery consumption, 2G/3G signal LEVEL, and data down/up SPEED over Wifi and cellular data modes, over periods of time. The device should not be moved or interfered with until the process has finished (i.e. an audio notification is used)
This could be useful for comparing various versions of radio and how they perform IN THE USERS GEOGRAPHICAL AREA.
Details:
The app will kill non-critical processes and services, and lock the phone for a short-ish space of time (e.g. an hour) so that it can test without disruption
In this time, it will read and record the cellular signal level (2G or 3G) every minute or so for the hour to find the average signal level and the same for Wi-fi
It could also download a small file using cellular and Wifi connections to test average download speed and upload could also be tested
There could be an 'overnight' mode or a longer time period where the battery level could be recorded over the space of several hours while the connection is idle or even kept alive. A decrease in battery level per hour could then be worked out
All this information is recorded and the radio is then rated in terms of battery consumption, connection strength and speed. The process can be repeated with other radio versions and the user can choose which radio suits them
What do you all think - viable or not?
A very good idea. Users can then publish their results in the forum and the radio choice will be far easier.
Good idea but so many variables, so what may be a good radio for one user may be awful for another...
dial ##DEBUG# from your phone or run the hidden file fieldtrial.exe from your windows folder. it has everything you could ask for regarding radio and consumption.
This is the follow up thread to:
Fast Dormancy -- Save Your Battery from 3G Drainage
http://www.youtube.com/watch?v=O8L50sCY7CI
This is a thread to discuss your findings and concerns about Fast Dormancy.
Feel free to leave questions for me, I will be constantly monitoring this thread.
****NOTE*****
There is an original app by Gokhanmoral, which effectively disables Fast Dormancy on the Galaxy S3 (ONLY TESTED ON i9300). Please refer to the description on market app link below for future compatibility.
https://play.google.com/store/apps/details?id=com.gokhanmoral.fastdormancytoggle.i9300
Supercurio is creating an app that disables FD, but also ads separate diagnostic functionality.
Yeah so Erica spoke about an application we developed when studying the Fast Dormancy and RRC states.
In order to experiment and understand all about FD, we had to developer a new tool also allowing to enable/disable Fast Dormancy on several Samsung phones, and apply changes in real times, including when tuning Fast Dormancy timers.
Requiring root depending on the device itself, its main purpose it to help you analyze if the network you're using supports fast dormancy or not, then allow you to set reasonable settings providing satisfying battery performance while keeping the network latency low.
Be sure you watch Erica's video as she will explain the technical challenges behind 3G power management better than I could.
Phase 1:
Today's application is very simple:
It provides an easy access to the useful Samsung proprietary ServiceMode page displaying RRC states and also advanced network SNRs.
Download preview app from Google Play
Voodoo RRC Tool source code on github
Note: the app kind of suck and as you'll notice it benefits from no UI design at all yet
Contributions are indeed welcome.
Phase 2:
Automatize the FD detection process in order to help you telling us which kind of network behavior you experience with your carrier.
Phase 3:
With fast dormancy carrier database, recommending automatically optimal settings.
Add support to enable/disable and choose Fast Dormancy timers on non-Samsung phones.
Phase 4:
Reverse-engineered Service Mode in order to build statistics of RRC state used, maybe to guess automatically if a network supports FD or not.
I'll probably not go this far but it might be interesting to develop this, not only on Samsung phones but on many more.
Once again contributions will be welcome here too.
I hope you'll like this joined XDA TV & development effort, which aims to investigate a topic, share knowledge and provide tools for everyone's benefit.
Hopefully its only the first collaboration among many. I'd like to encourage other developers as well to consider XDA TV as an extension of everything we're doing here, and interact with videos producers.
Guess i'll be the first to reply. Thanks for the info, i really learned alot. Keep up the good work erica.
Cheers!
video is no longer avalible WHy?
rehmankhan101 said:
video is no longer avalible WHy?
Click to expand...
Click to collapse
There was an unexpected glitch in one of the demo, the one explaining how to diagnose if Fast Dormancy is implemented or not by your network.
Concretely the actual demo was exhibiting a different behavior from our previous observations, so to be completely sure to guide you in the right direction we're re-investigating.
Except this part where there was doubt everything was perfectly accurate: Erica video shall be back soon on XDA TV once we figure it out
Thanx sc,nice tool and great work,even though I was expecting another app from u,hopefully that will come eventually
Good to see you're back.
Not working??
EDIT - the app by Gokhanmoral on the OP works perfectly now! (on my sgs2)
ignore the text below
The app doesn't show any output for me, is this :
1: Because FD is not available with my Carrier? (UK Vodafone)
2: Because I'm in the UK?
TIA
Irv
*Either reply in this thread or PM me if you need any further info/testing
EDIT : See - http://forum.xda-developers.com/showpost.php?p=28107797&postcount=8 and : http://forum.xda-developers.com/showthread.php?t=1742813
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ Codes don't work on AOSP/AOKP ROMS's (apparently)
I have always wondered if AOSP on for example the galaxy nexus supports FD, There is a general lack of information, it looks like it should, but I don't know.
It would be great to find a way to gather this sort of data on AOSP or other devices, no idea if it's doable though.
Great video series, FD is under documented and there is a lot of misinformation and unbacked assumptions
As far as networks go, unless the documentation is incorrect, my network (3UK) have FD enabled, I am unable to back this up with data, as I sold on my sgs2 a while back now
I can't get any information about FD in my Galaxy Nexus with JB
Yes good question about AOSP.
It's possible Samsung RIL is still reading fast dormancy activation status and timer values from somewhere in pure- Google devices.
Samsung devices running AOSP instead of Samsung firmware must still read the FD configuration parameters, as they're rocking the same RIL binary code.
This is definitely something to investigate, I'll take a look but feel free to report your findings, that might be useful features to add to RRC Tool.
Tip: you can search for strings in libsecril*.so in /system/lib and attach strace to the rild process : both give a lot of clues
On my Galaxy Note the code do access the service menu is *#*#0011#*#*
Does this depend on the radio version? model? network?
Awsome Three uk support Fast Dormancy
I have been poking the rild with strace, at the moment the output is refusing to be piped to a textfile, which is annoying, will work on that later.
Funny thing is that it runs for a bit then crashes it, have to toggle airplane mode to get network back, its most likely me not running strace with the right parameters though.
if i enable FD in the app and set the timer to 15 sec then nothing changed in *#9900#. there ist the curent status of my FD stil disabled.
okay so i enabled both and charged phone to 100% and toggle mobile data and was watching the battery. in half an hour I lost 2% of Bat.(did nothing .. only get few emails via push and wrote an message via whatsapp)
than i checked BBS and in the section kernle wakelocks i get an wakelock named : "sercril_fd-interface" with an time of 22m ( FD was anebled for 30 min) witch take an total wackelock of 50%?
now i kinda dont it what should i do ? will watch again 30 mins and will check than without FD .
/andy
Hello angelofmusic, I have a Samsung Galaxy Note GT-N7000 currently running on CheckROM Revo NoteHD v5. Upon accessing SysDump, there is an option - Enable fast dormancy(Current State : Disabled).
Thus, I suppose I would not require the prescribed app, would I?
:silly:
What is the appropriate stepwise testing procedure/methodology that you suggest I follow?
Thanking you and awaiting your response.
So will FD save me some battery life if I enable it? What about black crush how do we fix that?
You might want to note that Gokhanmoral's app was based on the guide here: http://www.androidnz.net/2012/06/at-your-own-risk-how-to-disable-fast.html
Telstra RRC states/timers
Telstra does the following on 850MHz (MCC/MNC 505/01):
DCH (T1 1s) -> FACH (T2 30s) -> URA-PCH
Three/Telstra does the following on 2100MHz (MCC/MNC 505/06):
DCH (T1 5s) -> FACH (T2 10s) -> URA-PCH
Didn't time the T3 timers (long).
If you're on Telstra, the fast dormancy timer will force the IDLE state (for both networks above), and you will miss out on the benefits of the PCH state.
Background info on Australia 3G networks:
Three + Telstra had a reciprocal network sharing agreement for years (Telstra 2G 900MHz + Three 3G 2100MHz)
Three + Vodafone merges.
Three network shares with Vodafone 2G (and wiith Vodafone 3G this month).
Three/Telstra network sharing agreement ceases end of this month.
I think Three network would likely stay as a seperate network (rather than merged), in which case the states/timers would likely be the same.
nCoder said:
On my Galaxy Note the code do access the service menu is *#*#0011#*#*
Does this depend on the radio version? model? network?
Click to expand...
Click to collapse
Confirm..yes me too on my Galaxy Note code is *#*#0011#*#* to enter the Service Mode
maybe it gonna be awesome if possible to direct from the Voodoo RC tool..btw that app is awesome for starting
keep a great work for you Supercurio :good:
JGuinan007 said:
So will FD save me some battery life if I enable it? What about black crush how do we fix that?
Click to expand...
Click to collapse
as long as your carrier provider support that FD ..
black crush ? hmm i think that a different problem..this thread is discuss about networking not the Screen Colour
....
finnaly ..
before trying your Voodoo RC tool i check the RRC state behaviour with default setting ( enable fast dormancy from *#9900# )
and the result is:
DCH to FACH pretty random timer ( some times 8s, 15s ) .. and very quick from FACH to PCH -/+ 5 sec and stay on PCH never goes to IDLE state
then i try voodoo RC with guide from erica video (thanks erica)
1. when i enable and first try the 1 sec FD with policy then back to test
DCH to FACH still ramdomly but never reach above the 10s ..and still very quick switch from FACH to PCH and stay on PCH state until the same process repeat PCH to FACH then DCH
2. then i try the 10 sec FD with policy and tested
DCH to FACH still ramdomly but sometimes lil bit longer the 10s ..and still very quick switch from FACH to PCH and stay on PCH state until the same process repeat PCH to FACH then DCH
.. then i decide to stay in that setting and see what happen..until more or less 5 h i check with BBS that secril fd_interface on kernel wakelock is 2 h more or less is about 51% and yes my phone cannot enter the deep sleep by that ( i forgot to capture it sorry )
my question is : is that mean my carrier provider deffenetly didn't support the FD or i did something wrong with the step .. coz last time i disable the fast dormancy ( with FastDormancy toogle by Gokhanmoral ) i never saw that secril_fd process so high..
but i'll try looking for and test again
again thx Supercurio and Erica
This is a great thing Erica is triying to do!
I wrote a guide in Spanish for people in South America or Spain:
http://loqueselocomparto.blogspot.com/2012/08/fast-dormancy-en-s3-colombia.html
Got bad Focus 1.3 GPS performance? Info below helped me. Not for the fainthearted, requires unlocking and registry edits, YMMV, I'm not responsible if you wreck your phone, I pulled some of this info from other posts, etc.
Symptoms:
GPS accuracy/ability to lock degrades. Updates & reflashes make GPS work great for a week or so, then it has trouble locking, then stops locking inside 350 meters at all.
Fixes:
These helped; not sure all are needed.
•Change settings
-WiFi: on (improves initial H accuracy)
-Location: on (obviously)
-Cellular: on (need for initial fix)
-Find my phone: connect and save both checked
•Diagnosis App
Enter "Test Mode" by dialing *#1575# in diagnosis app, then see three options in the "Change state" menu:
-MOLR CP/UP: Set to CP MOLR state
-3G MOLR request type: Set to LocEstimate type
-TLS Option: Set to Tls on
•Registry tweaks (must be fully unlocked, use WP7 Root Tools for reg edits)
-HKEY_LOCAL_MACHINE\System\CurrentControlSet\GPS Intermediate Driver\Drivers\GPSOneDevice\OpenStandalone
Change from 0 to 1
-HKEY_LOCAL_MACHINE\System\CurrentControlSet\GPS Intermediate Driver\Drivers\GPSOneDevice\QOSAccuracyMeter
Set to 100
•Reset GPS when it stops getting a lock
1-Remove battery
2-Remove sim card
3-Let phone sit with battery out for 10~30 minutes
4-Replace battery and restart without sim card
5-Shut down and replace sim card
6-Run and calibrate compass app before launching GPS app
•Use offline apps
At first I figured my problems came from lousy Samsung satellite GPS hardware. Then I read this on a MS dev site:
-"GeoPositionAccuracy.Default accuracy uses Wi-Fi or the Cell tower signal ..."
-"GeoPositionAccuracy.High accuracy uses the [satellite] GPS receiver ... built into the windows phone device."
-"Developers write apps to use GeoPositionAccuracy.High accuracy instead of GeoPositionAccuracy.Default ONLY WHEN NECESSARY, because it could ... affect the performance ... So, devs use Default accuracy level whenever possible."
So I started wondering if the phone would be more accurate with all data turned off.
Result: works much better with data off. Using MapSnap GPS with no data, there's no more location jumping between cell towers, just a steady track. Tracks me hour after hour on bike rides, hiking under trees, etc. Tracks exported to Google Earth are pretty accurate. Best thing is I can use the GPS offline week after week and it keeps locking down to ~10 meters.
Tip: Get your first fix with data on before going offline. I open MapSnap, get initial fix, then go into airplane mode.
Apps: MapSnap GPS is flawless, but you need to supply and calibrate your own maps. Haven't tested with Nokia Drive because I don't need turn-by-turn, but people say Drive works with no data.
EDIT: Just tested with Nokia Drive and works well with no data. Same drill: get GPS fix in Nokia Drive with data on, leave app running in background, turn on airplane mode, return to app.
Best Guesses for Focus 1.3 GPS Weirdness:
-Driver that juggles cell tower and satellite GPS info is poorly written.
-Cell tower location driver is poorly written.
Thanks for sharing, looks like good info.
I'm currently working on a geofencing library, which should work without depending on the Google Play services (or any other non open source services). Most of the code is already working pretty nicely (including location resolving of wifi and cell data), but what i'm having trouble about, is the dynamic scheduling of the next location update.
It shouldn't miss any of the geofences, but on the other hand it shouldn't be too power consuming.
I have access to the following information:
the last known cell location & time of measurement
the last known wifi location & time of measurement
the last known gps location & time of measurement
the current location (could come from any of those location providers)
the distance to the nearest geofence (or all if needed)
the user's current activity (moving or still) based on accelerometer data
The problem is, that these measurements don't use fixed update rates and I don't know the current velocity and direction. If I tried to calculate it from the last known locations, it would become rather inaccurate, since some if these measurements will be rather old.
Is there any other way to schedule the next location update (approximating the time it will take to reach the nearest geofence) without requesting velocity information via GPS?
Since I already continuously monitor the accelerometer it would be an ideal solution to just use the information from the accelerometer in combination with the already known location information.
Has anyone here an idea about an approach I could look into?
Hi,
I have an Honor Band 5 and love it.. I was wondering how long it can manage to store data if not connected to my mobile.
It would save power if I only connect it to the phone to transfer recorded data but I have no idea how much data it can store.
Anyone any idea ?
Cheers
Steve
I got my honor band 5 this wednesday, and yesterday i was testing my band without bluetooth conection for 8 hrs (with trusleep and heart options enable).
24 hours is my limit now. With all functions enable. (intelligent heart and trusleep enable)