[Martin Gorbush]

Marcus wrote:

Here are the register settings for a chuwi 12.3 lapbook

Thanks for the rapport. Sadly it seems that there are no more MSRs. Changing CPU voltage seems impossible right now :(.

Only noticeable difference is that your BIOS has never microcode for this SoC (version 24h vs most common 1Ch).
Newest is 2Ch for this SoC CPUID value 506C9h.

 

[Martin Gorbush]

Chris G wrote:

Same here, new V4 model has a Sandisk eMMC, but no powerlimit option. That’s locked

Is it locked only in BIOS or completely? Check with ThrottleStop it will show if Turbo Power Limit (MSR 610h) is locked (bit 15 set).

[Martin Gorbush]

If MSR 610h isn’t in locked state then do not worry :).

There is new ThrottleStop 8.50 version which allows to set TDP limit for Apollo Lake SoC (TPL button).

 

Notice: There is a need to collect more registers dumps for this SoC. Especially if you have laptop with LPDDR3 memory. Then please go to this thread and download from SendSpace tool called RegReport. If you do not want to create new account on TechPowerUp forum then attache txt output file here. I will deliver it to that forum if there will be something interesting to see for the author of ThrottleStop. On my setup there is a lack of few MSR registers which are described in Goldmont core SDM (Software Developement Manual). Strange.

[Martin Gorbush]

External USB keyboard will work for 100%. Provided that USB adapter is functional. I tried both USB-A to USB 3 mini and USB-A to USB micro adapters and that always worked.

[Martin Gorbush]

Chris, as I mentioned before PMIC driver in the Android sets higher maximum battery voltage (so called CV charging mode) than in Windows. Your observation is entirely consistent with this. Also what “brainvision” said in his last words;) is true even if bigger part of his comment is offensive :(.

If “battery univesity” information is true then rising maximum voltage of li-ion cell by 0.1V shortens its life span (charge cycles) by half even if it allow that device to hold about 15% more charge. IMO this isn’t profitable exchange especially if you have many power sources around you (car battery, power bank, electrical outlet).

[Martin Gorbush]

I thought so. Your battery cannot be charged as high as 4.35V which under Android it is interpreted as 100% charge. If you get 4.25V maximum then it will be something like 95% Android battery indicator. Even so those battery cell are rated for 4.2V as nominal. 4.35V is their maximum voltage. It shouldn’t be charged so much in daily use.

[Martin Gorbush]

S K wrote:

(…) is there any reason why the battery on this tablet would not recharge past 95%?

Can you check what maximum battery voltage is there?

Do you charge it under Windows or Android? I ask because there is different maximum voltage for battery (4.2V vs 4.35V).

[Martin Gorbush]

brainvision wrote:

Last but not least, about the battery hitting 100 and then discharging even if power attached: well it’s completely normal and NOT A CRAPPY HARDWARE! Imagine what would happen if battery is full but the charger goes on putting 1.240 mah to the tablet! Briefly, it would burn or at least destroy the battery in a few days! Modern devices have an integrated chip that take care of this: if the battery is pretty empty is tries to charge as soon as possible (on normal condition); When battery is charged it stops charging so to not brake the battery; please note that the system will most of the times tells you that the charge is still 100% while actually it went to 98% (this is the most common reason why often the first 5% of the battery seems to go down quicker when you remove the charger – it was not really 100% charged); once the battery went to lower levels (still unknown to me, may change depending from OEM and hardware and charger) the chip unlock the charge and the battery starts to change again. Believe me, this is NOT an issue, this is NOT crappy hardware, but it’s a safe measure and your battery will last much more thanks to that!

I wouldn’t be so sure. Of course it might depend on battery what will happen to it when constantly is tople charged. In my Toshiba A8 notebook which was almost 100% time connected to charger battery almost died in no more than a year (just after 6 month of usage). That’s why in better notebooks there is application or system setting thanks to it you can select to charge battery to only 50%.

Side note. This problem is bigger under Android because there PMIC is set to charge battery till 4.35V (which this battery often couldn’t reach) and not 4.2V like in Windows.

[Martin Gorbush]

Gray wrote:

Martin Gorbush wrote:

Factory Android kernel use CV charging like 4.35V which is definetely too much for built-in battery.

For 3.7V battery maximum CV charging voltage 4.35V is not “too much”, I would say a little, for your explanation of the some like +/- 4.45V….

“Too much” is a vague statement. Sorry for that. I will refine it. When I want to have 8000mAh li-ion battery I can get 7000mAh at 4.2V cell and charge it to 4.3V. Doing so will increase charge capacity by about 15% but not without consequences. And yes it will probably not explode but I will reduce usable charging cycles by about half. For me this is “too much” as a trade off.
There is one more thing but I’m not completely sure if I got it right. Battery in my unit doesn’t reach 4.35V in Android because charging process is terminated probably abruptly (at 20% of maximum current) but it shouldn’t happen outside of Constant Voltage Charging mode. Even so when battery voltage drops to 4.25V topping charging is activated. When I don’t disconnect tablet from power source it happen quite often (few times an hour).

[Martin Gorbush]

Gray wrote:

Let me give you an example of settings of the (dr. Buchmann) device Cadex: For 3.6V battery; Maximum charge voltage – 4.35V; standby voltage – 4.15V. Accordingly, for our batteries cell with 3.7V rates are even higher

Yes, that is quite right. Therefore for 3.7V nominal voltage li-ion battery standby voltage should be 4.25V not 4.35V.
Besides everything. What I wrote about increasing maximum voltage should be true for almost every nominal voltage. When you charge 3.6V li-ion battery to 4.15V it will last for twice more charging cycles than when you charge it to 4.25V.

Gray wrote:

2. Basic test I did at the stock ROM, and checked back on others. Changes in charge parameters have not noticed. Control of “Over Voltage Protection” (IMHO) depends more on the Power Manager IC. BR, Gray.

I’ve got quite extensive knowledge about this tablet PMIC. I even found linux’s kernel source code for PMIC AXP20x which should be similar to Android’s kernel. One thing that catched my eye was that CV mode in absence of ACPI configuration is 4.35V. Therefore PMIC is trying to reach that voltage but it fails to do so when charging current gets below 20% (~370mA) of its programmed value (mostly 1800mA). That might happend in quite different voltage level depending on particular battery. Maximum which I saw myself was ~4.32V. Then when voltage drops below 4.25V (0.1 below CV mode setting) charging process starts again. It is very annoying and rather devastating for li-ion battery. At least I think so.

All of this can be verified by looking for “Dolar Cove” kernel parameters. I will edit this post when I recall correct path. I forgot to do that in another thread: http://techtablets.com/forum/topic/very-slow-battery-charging/#post-48327

Edit: I found Dollar Cove (PMIC AXP288x) configuration path: /sys/class/power_supply/dollar_cove_charger. There is parameter called “max_charge_voltage” which in my device is set to 4350mV and max_charge_current is 1800mA.

[Martin Gorbush]

Dear Gray.
I’m quite aware that there is many types of lithium ion batteries. Regardless. What I know mostly from batteryuniversity.com site as a rule of thumb you can say that every increase for 0.1V of maximum voltage for li-ion battery reduces available charge cycles by almost half and only increase holded charge by about 15%. For me calculation is simple. I’ve got many power source (car battery, power bank) and I don’t need to charge too much (topped under the cork).

This 4.25V is under this REMIX OS?

[Martin Gorbush]

Can someone check what maximum charging voltage for battery is used by this firmware. You can check this with CPU-Z or Battery Monitor applications after you charged tablet battery to 100%.

Factory Android kernel use CV charging like 4.35V which is definetely too much for built-in battery. I search for firmware with much lower value. Preferably 4.1V.

Thanks in advance ;).

[Martin Gorbush]

Andrew W wrote:

Martin Gorbush wrote:

It is a shame that one cannot configure same charging parameters from the system as an user.

Even with root there is no way of changing the values? Would this require manufacturer’s source code or something?

If I recall correctly parameters exposed by PMIC driver (code name “Dollar Cove”) are read-only even for root user (I will check that again when I will have opportunity to connect external keyboard to this tablet). Without kernel source code it is almost impossible to change that.

[Martin Gorbush]

Andrew W wrote:

 I wonder if you could mod a cable with shorted data lines (with the right resistor) to supply more current.? Could this be risky? Maybe…

You can short data pins if you want. It will not brake the USB port. Even so it will not give you faster charging times at least if your tablet can draw more than 500mA from the power source. It is so because these data lines are probed to determine what kind of USB port is supplying power for charging the tablet. It it is data port then it will not draw more than 0.5A. If there is constant resistance less than 250Ohm between D- and D+ pins then it is dedicated charging port (DCP) which can supply more than 0.5A and PMIC will try to use so much as it is configured to do. In this tablet around 1.8A (I don’t remember if there is difference between Windows and Adroid OS). Even so when USB supplied voltage get below ~5.05V current drawn from the source will get lower than 1.8A.

[Martin Gorbush]

J wrote:

If charging was controlled by the kernel then what controls charging when it is turned off

In this tablet (and many more on Charry Trail platform) charging is handled by AXP288 PMIC made by X-Power. It can be configured by system or BIOS (or rather UEFI) as well it has some default values and can operate without enabling OS.

Strange thing with this tablet is that PMIC configuration is different for Android and Windows (mostly in ACPI). For example constant voltage (CV) charging mode on Android is used when battery voltage reaches 4.35V and under Windows it is only 4.2V. Notice that build in Li-on battery is only rated for 4.3V as maximum. It is a shame that one cannot configure same charging parameters from the system as an user.

Edit: Wrong PMIC name.

[Author]

Posts