[Algorithm]

Ok here are my findings.

I set bios to defaults and set all core offsets to -50

I installed a non-bloatware edition of windows 10 pro this time (containing all updates till the end of march). (rather than the default Win10 home that came with the device.
GFX was set to max performance in power settings in windows.

Upon installing all drivers (and not using windows update to update further) I inspected temperatures on Idle using hardware monitor. This was when it was approximately at 60% battery and not connected to the charger.

IDLE RESULTS
————
IA, GT LLC/Ring offset and system agent offsets were displaying as -0.051 which is around right as i set these on the bios

VID #0 and #1 was showing as 0.530 on average

Idle temperatures during the start approximately 36c

Power in total was 0.60 on average consisting of the below values on average
IA Cores=0.12, GT=0.00, Uncore=0.30. Dram=0.18

CPU utilisation was at around 2-3%. Clocks would be freely adjusted between 898-2596mhz
Current voltages showing as 7.702v

Then i decided to run the PPSSPP (PSP emulator running a good testcase that uses quite some resources from the system (god of war – chains of olympus)

These were the findings when set ingame to one of the scenes

PPSSPP – INGAME
—————

VID #0 and #1 was showing as 0.612 on average
Temperature was around 50-52c
Average total power usage was 3.29 on average consisting of the below values on average
IA cores=1.00, GT=0.90. Uncore=1.15, Dram=1.24

Cpu utilisation was at around 50% on average. Again clocks would change between 898-2596 where required.

Then i decided to use the maximum speed option in PPSSPP using the same scene, (This forces fastest emulation possibile unlocking the 60fps cap)

Results were as follows

PPSSPP – INGAME (60fps uncap – Max frame rate possible)
——————————————————–

VID #0 and #1 was showing as 0.850 on average
Average total power was 9.00 consisting of the below values on average
IA cores=5.00, GT=2.00, Uncore=1.00. DRAM=1.00

Cpu utilisation was at 100%. clock speeds would range between 989-2596.

I then decided to install the rest of the updates, this also automatically downloaded and installed display driver (28/02/18 v23.20.16.4973) and intel system drivers (27/10/16 v 1.1.1.22)
It was unable for some reason to install the April 2018 cumulative update. (Although i believe in my problem installation previously this was the same case – but it did install the march 18 cumulative update instead)

I had read somewhere that there was some stuck cpu speed issue on another core m tablet too (Yoga tablet) and it was mentioned that if the usb c charger is used when the battery is less than 25%, it deliberately throttles the cpu due to fast charging from that percentage but some glitch prevents the cpu speed from going back up later. Some solutions were to disable prochot via throttlestop or to charge until above 25% then go into bios, save settings without making any changes – which apparantly solves the cpu throttle stuck issue. I will keep all this into consideration if i experience any issues.

When the tablet was on 40% power, i decided to plug in my (Inateck 45w usb c charger with power delivery). and i noticed in HW info that all of a suddent the Package power went up to 11w. Quite some panic, but then it seemed that it had installed an update precisely at that point as the dialog box “windows has installed new features” had popped out. (I assumed that the cpu activity for that given few seconds may have been related to that, but did not want to risk my luck. (This was the first time i had plugged in this charger since i reinstalled the Windows 10 OS today). maybe some issues with initial voltage? I quickly decided to turn off the charge and some moments later placed it on charge again

And then it was not so bad (Package power was showing as 1.00)

Finally decided to drain the battery until 23%, to turn off, then connect the charger (USB C) and turn it on (I tried to replicate what i did the day before when the cpu cores were stuck at the low speed).

In HW info, everything seemed normal apart from the uncore being 0.90 on idle rather than 0.30. (Is this normal for the uncore to rise from 0.60 when it is on charge)

Thats it for now. No issues so far. Hopefully it will stay that way

[Algorithm]

It could very well be.  Some settings in the bios seemed to have definitely changed. (I guess the device being undercharged and not used for a few months was the culprit) – Date and time in the bios was also reset as well as previous undervolt settings.

No issue when I kept the same bios settings as above and installed Ubuntu (although as mentioned previously, I am sure during the installation process, the cpu was still locked to 399mhz as everything was sluggish and took a lot longer (x4 at least) then when I had installed it previously.   When Ubuntu did finish installing and rebooted, was back to its speed and working well.

With my previous windows install, It more than likely was windows getting mixed up with the changes and combined with windows update the problem was caused.

Will set the bios settings as they were (after clearing to defaults) and reinstall windows at some point today.

[Algorithm]

Definitely was not thermal throttling.  cpu temps were at around 60c.    Unless somewhat my bios settings got cleared (as I did not use the device for some time and I noticed my undervolting in bios was gone) and it confused windows somewhat.   But at 399mhz and using 4w in IA section for HW Monitor seems a bit too much.  Perhaps CPU voltage was too high by default? did not check that

On Ubuntu no issues at all, (But took a very very long time to install) so perhaps after installation it reinitialised and cores started scaling well after

[Algorithm]

Ok after a very long sluggish time installing Ubuntu, I got into the desktop, tested a few applications and emulators and checked the cpu scaling and all works perfectly

It seems there was a recent issue with a windows update causing some issues overload of power in a device causing the power limit throttling.  Anyone else had this issue on windows 10?

[Algorithm]

Maybe if its possible to get a specific charger which charges the battery slower, this would have less impact on components heating up (or/and) use a laptop cooler beneath the tablet/laptop to dissipate heat more

[Algorithm]

Touchscreen and Wacom enabled screens do indeed use more battery life than standard screens.      If you wish to keep the high performance of the device, then unfortunately battery life will be in the region of 4-5 hours on medium load.

However if you do not mind sacrificing some cpu/gpu performance, you can turn on power saving mode in the intel hd display control panel and set CPU Max to 99% in power settings (to disable turboboost).  Some bios settings to lower max power can also help.

Undervolting will also save a bit of power but difference will be negligible (maybe may give you an extra 20 minutes battery life or so).      Updating to latest windows 10 fall creators update also gives you a power slider when you click on the battery icon where you can select high performance or low performance modes

It is really more a tradeoff between price of the device and reliability.   I also own a surface pro 4 and that has none of these issues.  Has great battery life 8-10 hours, good performance and none of the battery calibration issues that plague these cheap Chinese brands.

Lowering display brightness and maybe even somewhat disabling Wacom option may somewhat increase battery life further.    Some users have reported better battery life in Linux on some systems (although not sure what the current compatibility is with Linux and this device)

 

[Algorithm]

I guess it probably then boils down to design flaw then.  (too much heat when charging and using the tablet) which may damage the device.  I tend to only use mine on battery and not while charging and so far it works fine (also applied thermal mod and undervolted)

[Algorithm]

For anyone that is experiencing issues with maximum charge being less than 100%,  Would it be possible if you can recall whether or not there was any stage where you had left it on charge after it reached 100%.    I think in one of the XDA posts, it was mentioned that the battery charge max calibration value in bios differed from that of the actual battery capacity which caused issues when charging fully (and system thinking it had charged to a value that is more than the actual charge capacity

[Algorithm]

Once you enable XTU interface in the bios, then XTU should show some further settings.

I have a zip file here which has some screenshots (This was from some other user who decided to go to the extreme, I advise this is not done, unless you want an exploding Cube) 🙂

https://www.dropbox.com/s/b91es1igujzma5i/BIOS%20SETTINGS%20OPTIMUM.zip?raw=1

[Algorithm]

It is in the following

power management control
Platform PL1 Power
Platform PL2 Power
Power Limit 4

DPTF Configuration
Minimum power limit default, Limit 0,1,2

Check the current settings and adjust accordingly

You can also use XTU for this and turn on XTU interface and Watchdog in bios (This will then enable you to change in XTU and will work on every reboot I think)

 

[Algorithm]

the EVE V has worse cpu performance.  GPU performance nearly half that of the cube mix plus and priced double+ the amount of the cube mix plus.  🙂  For the pricing of the higher end EVE V (i5/i7), its comparable in price to the high end dell XPS 13 devices 😐

[Algorithm]

Power limit throttling by default occurs when the total power used by the cpu and internal gpu exceeds specific amount (I think it is 7w by default).  This can be raised in the bios at the expense of more heat and less battery life when on load (but higher framerates etc for games)

[Algorithm]

Indeed, the device and charger should communicate and it should not give any more current than its designed to take.   (Exception being GPD XD device which blows the battery circuitry if using a fast charger – protocol just says “hey, give me all you got” 🙂

[Algorithm]

Some good points there.   Indeed, the thermal mod can spread out heat more (Heat is still mainly localised but spreads out slightly more to the area where the battery is.    I would assume that temperatures would not really be any less than 35c with or without thermal mod inside this tablet when on load.

Based on other users comments and issues, I would probably assume that there is some issues with the battery circuitry which may have an impact on the longevity of the hardware or battery itself.   One issue is that the reported battery capacity actually seems to increase over time (if overcharging) and this later only results in charging to 25 or 50% of its capacity after some time as one unfortunate person also mentioned in this forum  (Could also be some other damage)

With this particular device, my opinion is that its safer to just charge the device when fully off,  (being careful not to overcharge it)  then use it after.  Unlike other premium devices that usually should not have this issue.

 

[Algorithm]

That’s unfortunately the main issue with these cheap tablets (less quality control). You get the specs, but reliability is another thing.

Charging the device and using it causes additional stress and heat and constantly doing this may cause failure after a while perhaps.    Without the thermal mod, its actually quite easy to get it at temperatures >90c and prolonged high temperatures along with the battery charging can be an issue.

I normally charge my cube mix plus (via usb C charger) when the device is turned off  (and I am careful not to get it past 90% battery) as I think there is also some calibration bug when overcharging (for some users) where they end up with a battery that cannot charge more than a specific amount.    The supplied pin based charger I avoid as that seemed to cause some intermittent issues with the device (trackpad not working etc)

[Author]

Posts