You would be well served to learn to read and comprehend that which is written, rather than launch into tirades.
You clearly lack the abilities of discernment and logic.
Now, pull your head in!
If to my post that (let me quote): âSadly, 10 III power consumption on 4.5.0.18 seems to be even worse than on 4.4â youâre answering that âI am not seeing that at all. It is identical with 4.5.0.16â then obviously it is YOU who has severe problems with reading, and even worse with understanding, because 4.5.0.16 â 4.5.0.18 was a small hotfix mainly addressing video playback issues introduced by switching to ffmpeg5, so it obviously did not bring any power consumption changes and no one ever said it did.
So first yourself write something that brings anything meaningful to this discussion (other than that you âdidnât observeâ what others did) and only then expect others to discover any sense in it, or feel entitled to scold people, but even then not in such a boorish way.
Over and out.
I go away for 1-2 weeks without power. I need my phone to be in flightmode and GPS and take photos a bit. It must still be working after more than a week. I just went away for 3 weeks. I had to take the Android phone because of standby current. At home I can charge every day or two, and I wouldnât really care.
Now, lets look at your case.
So lets say 1%/45min = 1.33%/hr
My Android X10iii (see previous reports above) runs at 0.55%/hr cell+wifi.
So if we took my SFOS standby (0.8%) away from your figure
1.33% - 0.8% = 0.53%/hr for cell+wifi
This is same number as I get with Android.
My conclusion: It is likely to be the standby/flightmode current that is the major difference from Android.
Edit: Just tested idle with Cell,Wifi,BT ON. 0.75%/hr i.e. same as with them off.
No.
Now, I donât watch videos, browse lots of web etc etc on my phone, so I simply donât have any idea if SFOS is worse than Android if you do. I am only really interested in long battery life under light usage or flightmode. (In Android I usually have it set in âStaminaâ mode)
I do play podcasts and music via BT, and it doesnât seem to add noticeable drain, nor does skype running in AD. (but I havenât done measurements)
No, it still runs for 2-3 days - I just need better than that. Doesnât mean you need it too.
You are right, after one week of logging with different setups I come to the same conclusion, but I also get heavy battery drain with mobile data if my VPN is on. In fact, that is even worse than WLAN. Without VPN it is OK.
1 Like
Thatâs the SailfishOS native VPN solution?
1 Like
I always turn WiFi off when I go to sleep (and most of the time mobile data as well). Android too and the only native thing running being Battery Buddyâs background service. It still anihilates 1% per each hour of doing literally nothing.
2 Likes
Same here. Just like turning off the lights here. I generally have nothing running at night. Thatâs what the servers are for 
The GS5 with 4.5.0.18 with all comms off has been using about 3-5 % battery over the course of 7-8 hours of the night. Thatâs ok. So it âdoesâ look like some significant hardware diffs.
1 Like
There was a change to WiFi a long time ago, that made low power wifi idle possible. Before that, your phone burned up power on WiFi. The router firmware had to be ungraded. That was back in the N900 days. There may be an issue related to your specific router.
I just tried my X10iii with wifi and cell, at 0.75%/hr i.e. exactly the same as with them off.
1 Like
0.75%/hr in case of a 4500 mAh battery equals less than 34 mA of average power draw. Are you really sure that you get such a low power consumption on the 10 III? Because itâd be the lowest Iâve ever heard of. In my case, 30-35 mA is the lowest I ever see in my logs but only for short periods of time, with average power consumption being in the 50 mA range.
Anyway, even such hard to reach values of 34 mA that youâre reporting are still more than 3 times higher than my XA2 Ultraâs 10-11 mA (see the screenshot in one of my previous posts), i.e. with its 3580 mAh battery actually less than 0.3%/hr, which once again takes us back to the sad conclusion that thereâs a lot to be done about the X10 III hunger for power.
No. I canât measure the actual current (and also I donât really believe battery ratings eitherâŚ)
Iâm not using software tools to report diagnostic currents, as they:
a) require the cpu to run (thus will increase the reading)
b) are sampling the current, not giving an integrated value
c) are not measuring at all when the phone is actually sleeping, and thus cannot measure the idle currents like ram refresh, cpu leakage etc, etc.
d) canât run the same s/w on my Android phone.
e) didnât have much luck getting stable reproducible results when I tried it
My methodology is to charge the phone to 80% (happy with the end-charge setting. Yay!)
Then I set the phone up how I want and leave it unused in a standard place for some hours e.g. overnight.
I then take (InitialBatt%-FinalBatt%)/NumHours = %/hr
These numbers have remained fairly consistent at different battery levels, so I consider the battery gauge readings to be roughly valid.
This requires no software running, and lets me easily compare my Android and SFOS X10iiiâs
There are three good things about this:
- The Android version is ~4x better, so the X10iii can do it.
- The XA2 didnât have this issue in SFOS. This means that unlike the GPS issue, it does not appear to be a long term Sony AOSP issue across all devices, but rather is a 10iii specific issue.
- The biggest issue appears to be standby current, not the cellular and wifi running current. The standby current is not dependent on signal strength, basestations etc. It should be much easier to test , find a fix, and prove it.
%-based approach may also be quite unreliable, because the battery % indication depends on its correct calibration. The % indicator is purely virtual, calculated from the batteryâs voltage. If the cell isnât from time to time nearly fully charged/discharged, the % scale may get quite de-calibrated, causing that e.g. 1% in the upper range doesnât equal 1% in the lower part (% values may change with different âspeedâ). In such case, any readings based on it arenât worth much. Also, as battery wears, its % indication is more and more susceptible to load because higher current draw causes temporary voltage drops - the higher the more worn a cell is. So I wouldnât trust %-based values for any precise measurements.
Thatâs why I prefer to read the actual current draw, especially if I compare two devices (in this case 10 III and XA2 U). No matter if I use Battery Buddy to log it, or directly read from /sys/class/power_supply/battery/current_now e.g. via ssh or to a file, I get the same readings.
I donât think the impact of the software logging the values is of any significance if the XA2 Ultra with the software or the terminal command running is still able to settle at as low current draw as 9-11 mA and stay in such a low state without any spikes for as long as you donât start doing anything with it (see my screenshot from the XA2U)⌠If there was any significant impact of the logging software on those 9-11 mA figures, then it would mean that the XA2 Ultra actually doesnât draw nearly any current without it 
Definitely.
The XA2 seems to have a completely different power management setup. Different governors, switching cores between online and offline depending on load (which is not used on the 10 III at all where all cores run all the time), and possibly more.
Thatâs right. But when I see that during phone calls it sometimes consumes 1% every 8-10 minutes then it also really frightens me as it means talk time of 800 minutes (i.e. 13:20 hours) whereas benchmarks/reviews of the 10 III with genuine Sonyâs Android give it talk time of 31,5 hours, i.e. once again nearly 3 times higher.
Ok, thatâs not the case on the GS5. Itâs almost double the drain overnight in sleep mode with wifi on. Drain without comms overnight was 4% in 11 hours, so aprox. .36%
1 Like
Iâm going to do this after finishing the current tests. Perhaps I could setup a repo where we can dump log runs of the one or the other sort? EDIT: log:
POWER_SUPPLY_VOLTAGE_NOW=
?
anecdotal evidence only iâm afraid:
i tend to get 2 days of battery life with mobile-data, wifi, and nordvpn configured.
probably an hour of so of heavy data use a day.
1 Like
All this cpu ower saving stuff like disabling unused cores doesnât change anything noticeable about the idle power consumption of the 10iii.
Great idea!
Did you really try on the 10 III switching some cores to OFFLINE, like the XA2U does, or were you only playing with its governors and clock speeds?
If the former, then how? Because I canât see the online/offline switching mechanism being implemented on the 10 III.
1 Like
Here are the needed sysfs variables for the core load settings:
echo 1 > /sys/devices/system/cpu/cpu0/core_ctl/enable
echo 1 > /sys/devices/system/cpu/cpu6/core_ctl/enable
echo 85 > /sys/devices/system/cpu/cpu0/core_ctl/busy_up_thres
echo 75 > /sys/devices/system/cpu/cpu0/core_ctl/busy_down_thres
echo 85 > /sys/devices/system/cpu/cpu6/core_ctl/busy_up_thres
echo 75 > /sys/devices/system/cpu/cpu6/core_ctl/busy_down_thres
echo 0 > /sys/devices/system/cpu/cpu6/core_ctl/min_cpus
3 Likes
I just checked and I have an unused repo that was intended for debug data: GitHub - poetaster/SFOS-debug: A small repo to keep track of debugging Sailfish OS problems I could set up directories thereâŚ