Open Hour Chameleon ROM for R89 ?

[BTR11]

Warning.
To my surprise hisound didn't set lower ARM_Logic voltage from astronomically high 1.45V. Be warned. As per RK3288 datasheet Absolute Maximum Ratings for this part of SoC is 1.4V. And standard operation voltage is 1.1V. I think that if you OC RAM to 1600MHZ DDR (800MHz base frequency) then ARM_Logic can be overvoltaged to about 1.2V. More is completely unnecessary (more heat) and potentially dangerous. Also notice that RAM voltage is soft locked at 1.2V on ACT8846 switching regulator DCDC1 output (/sys/class/regulator/regulator.1/max_microvolts = 1200000). Which is equivalent to 1.5V for DDR3 memory (Feedback input of DCDC1 is connected by voltage divider as follows: 5kOHm to DDR_VDD and 20kOhm to ground). It can be unlocked by modding resource file and act_dcdc1 DTB leaf. But setting 1.45V for act8846 DCDC1 output because of its feedback input voltage reduced by 1.25 you will get you about 1.81V for DDR3 modules. It is quite high for just nominal frequency of MIRA P3P4GF4BLF-GGN 1.5V DDR3-1600 CL11 memory modules.

I have modified resource_R89.img and kernel_R89.img (unlocked CPU OC above 1.6GHz ) to lower this voltage and make some CPU (max 1920MHz from real 1.6GHz and "fake" 1.8GHz) and GPU OC (max 600MHz from 500MHz). Resource file attached below.

Kernel is too big. You can get it from my Google drive.

Gormar your advise is to replace the resource.img from stage3 with this one modified by you to let the device work on"safe" voltage values for the components?.

I am a little worried with your statements, it´s confuse to me the different opinions between you and hisound, he had said not to see any danger at all with those values.

After one of my beelink R89 (2G/16G-AP6335- PCB v3) had been unstable using wasser 3.0 rc 3 (lollipop) I had decided to test ugoos_stage2 released by hisound, and for the first time I had seen this device working properly without "stutter" during videoplayback (but still with a little micro-freezes on UI operations and mouse pointer movements). I had been working a couple of hours playing some 1080p videos with that rom and the device been stable without any overheating.
The change of the led colors from blue to red is lost on this rom (the led stay on blue in any condition, "on", suspend or sleep, and just off when the device is turned off) and the web cam (logitech c170 & c525) work but can not record videos; but the rest for me is the best performance obtained with this device, not reached with any of the previous rom tested; so that jump is obtained for the raise of voltage value going to the components of the board? is that condition really unsafe; why the device can work several hours without be burned working on those conditions? why the device is not burned immediately as soon that internal voltage limit of 1.4 V is exceeded?.

Now I´m usnig in the problematic R89 the ugoos_stage3 rom by hisound, and the performance is as good as on stage_2 but with improvements on youtube 1080p - 60 fps comparatively to stage_2.

On the second unit of R89 (as well PCB v3), that I had not find problems, that work with very good performance with lollipop wasser 3 rc3, so, definitely the R89 tv boxes had been released with different hardware problems each individual device, surely, assembly and productions problems.

[gridlock]

i wasn't aware that i was spamming.... Never did and never will... you were on line when i was texting so i thought you would be like wasser , mo123 and bluesmanuk ( i mean being kind enough to reply back )...
i would rather keep the system bar forever on than to ask how to disable it from someone with this sort of answer...
But you know what , it's not your mistake....
For you to know , your first line was enough for me to read. Keep it up.

[gridlock]

You can add (at the end of the existent lines can be a good position) the following line to "build.prop" file located on /system/

qemu.hw.mainkeys=1

This will hide permanently the bottom system bar.

You can do that using ESFile explorer, turning "on" the options "root explorer" and let "hidden system files" be visibles.

Be careful do not chage the name of the file build.prop and keep the original file permissions (rw-r-r), and do not change anything else.

thanks a lot bro.... truly appreciate it...

[gridlock]

BTR11....it worked. Thank you so much again bro and a very good day ahead.....

[gridlock]

BTR11 , i seem to struggle now with the upper notifications (status) bar which is appearing in browser and youtube... So , yet again , sorry for any inconvenience. Just wondering if there's anyway to disable the upper status bar too permanently... Actually , i never use both bars (upper and bottom) and i don't mind disabling them permanently.... Blessings

[BTR11]

gridlock you can hide the top bar - notifications with full¡screen apk by tsorn. Look for this apk on the play store, this is what I use. There are others. You must to "enable" the full screen mode and configure it to autostart on boot. I don´t about any command line to add into the build.prop for this, but the apk does the work.

Link:

https://play.google.com/store/apps/details?id=de.tsorn.FullScreen&hl=en_419

[BTR11]

Coming back to ugoos_stage3 on Beelink R89 (2Gb/16Gb - AP6335), LAN - ethernet performance is bad with the original resource; and with the resource_alt img the ethernet performance is recovered, but youtube capabilities to play 1080p videos @ 60 fps it´s lost, just can be played @ 1080p - 30 fps, and the OTG port has the "connect to pc" set to "on" by default which made unconfortable to use the air mouse on that port, and this is required because many models only work properly on that OTG port, on the ohters USB aport are very laggy.
So I have to sacrifice the LAN performance and keep the device working on WiFi AC which give a very good, smooth playback on Kodi/spmc and Youtube 1080p - 60 fps, so I must to use the resource.img original and discard the resource_alt.img; hopefully hisound it´s monitoring the feedbacks here and provide some fixes for the probs, because the rom it´s very promising for the R89.

[el_easy]

Can't you get archive file with this ROM from Yandex drive? I might upload it to my Google Drive.

Thanks Gormar I already download your resource_R89.img and kernel_R89.img.. The reasons I ask for a mirror like Mega or other its because I have slow internet and downloading with yadi.sk its take me between 2-3 hours..

[gridlock]

BTR11....Your kindness is overwhelming.... truly thanks

[Gormar]

The reasons I ask for a mirror like Mega or other its because I have slow internet and downloading with yadi.sk its take me between 2-3 hours..

There is a MEGA link to UGOOS stage 3 firmware for Beelink R89/Tronsmart R28 made by hisound.

[Gormar]

Information to Jabader from 4pda.ru forum:
1. You found wrong DTB leaf describing DCDC1 ACT8846 output voltage software limit. You took data from RK808 PMIC, not ACT8846.

Voltage limits used for DCDC1 ACT8846 (regulator.1) output are described in DTS file in act8846@5a branch and look like that:

act8846@5a {
                        reg = <0x5a>;
                        status = "okay";
                        compatible = "act,act8846";
                        gpios = <0x4 0x1 0x1 0x95 0x6 0x0>;
                        act8846,system-power-controller;

                        regulators {
                                #address-cells = <0x1>;
                                #size-cells = <0x0>;

                                regulator@0 {
                                        reg = <0x0>;
                                        regulator-compatible = "act_dcdc1";
                                        regulator-always-on;
                                        regulator-boot-on;
                                        regulator-name = "act_dcdc1";
                                        regulator-min-microvolt = <0x124f80>;
                                        regulator-max-microvolt = <0x124f80>;
                                };

Note: 0x124f80 = 1200000

It can be confirmed that DDR3(L) voltage settings for DCDC1 output of ACT8846 regulator is limited to 1200000 uV (real 1.5V when DDR3 memory installed like in case of Beelink R89 device and similar) thanks to parameters presented by kernel through /sys/class/regulator/regulator.1/ directory.

2. You didn't understood what I wrote about voltage divider on DCDC1 FeedBack input. Therefore you wrongly concluded that ACT8846 will give on that output 3/4 of voltage configured by software. On the contrary. In its current hardware configuration when pired with DDR3 memory it will output 1.25x of that voltage because it reads back (feedback) 3/4 of real voltage generated on choke connected to DCDC1 ACT8846 output which is used by it as a reference.

[Gormar]

Gormar your advise is to replace the resource.img from stage3 with this one modified by you to let the device work on"safe" voltage values for the components?

Certainly yes.

I am a little worried with your statements, it´s confuse to me the different opinions between you and hisound, he had said not to see any danger at all with those values.

I thought that hisound understood my explanation. Mayby he just forgot to change that. In resource_alt.img all seems to be good and max is 1250000 uV.

Memory voltage setting for ACT8846 regulator (DCDC1 output) is soft locked to 1200000 uV which thanks to voltage divider on DCDC1 feedback input in reality provides 1.5V for DDR3 memory. But for some reason memory voltage setting is also used for ACT8846 DCDC3 output which supply power to ARM_LOGIC which is part of RK3288 SoC and one should know this.


Note: It seems that even exceeding 1.4V voltage to ARM_LOGIC isn't instantly deadly. But operating in that condition for longer time (maybe even months) might have negative consequences.

[BTR11]

Thank you Gormar, I had taken your advise, now I'm using the original resource modded by you thinking in a long run with this rom.
So, if you had changed only the voltage limits making lower to don't exceed 1.4 volts, that must have a positive influence on the ethernet because now it's stable and with 100% of speed, and Wifi & OTG port continue working fine as in the original resource.img. Thanks again.

Edit: even more good things with your resource image on my R89, the webcam logitech c170 it´s now able to record videos. So looks like the over-voltage used by hisound on the original resource image on stage_2 and stage_3 have immediate negative influence on the performance of the LAN & camera, taht had been solved whit your resource image, with the voltage limited to do not exceed 1.4 volts. The device is working very well with stage_3 & your modified image, Thanks Gormar. Only feature lost is the change of led colors from blue to red when change from status "On" to "sleep" or "rebooting", but this is minimal.

[hisound]

The firmware UGOOS_stage_3 always been recource_alt.img file that has a factory default frequency and voltage for all.
I always any theory underpins practice, a lot of discussion at the expense of excess voltage or frequency response is very simple, each instance has its challenges, do not rule out that some of the items can not work, even when the values ??recommended by the manufacturer of chips, in turn, shows the low quality of the components used and the whole device.
CONCLUSION: All future versions (if they are) will have only the default factory settings of voltage and frequency, everyone will be able to change them if you wish for as many devices allows this instance.

[BTR11]

Thanks hisound for this great work, you porting this Ugoos fw to R89 it´s a great boost for the device; the intermediate solution propose from Gormar to flash with his modded resource_img, only this image help to recover the proper performance of LAN & usb camera without sacrifice the performance of the firmware; the resource_alt.img however let this device function properly with LAN & WiFi, but sacrifice the benefits of your original reource_img or the Gormar's resource image, so, looks like this device need to exceed the default factory settings, but not to go the extreme of goign extremenly high the internal operating voltage for different components of the board, and we are not talking about of burning the components, it´s a matter of proper function of the parts, may be for the reason you had said, bad quality of hardware components on this Beelink R89. I´m happy with the performance of Ugoos_stage_3 with the modified resource image from Gormar, it´s the best performance that I had never been having from the R-89. The factory values contaiend on the resource_alt.img do not reach what it´s pretended which it´s to improve or boost the performance of the R-89. Those are my conclusions of my test. Thanks again for this amazing rom.