Acer Aspire TimelineX 3820TG mods/tweaks ONLY!

[Upabove]

thanks. Is there a mod bios for this laptop that add features.

[Just_a_Nick]

Hey,

could the free place U67 on the motherboard between the HDMI port and the Intel chipset be a place for a HDMI switch to witch the HDMI output port between the Intel and AMD graphic card?

[inteks]

Hey,

could the free place U67 on the motherboard between the HDMI port and the Intel chipset be a place for a HDMI switch to witch the HDMI output port between the Intel and AMD graphic card?

omg surfing with a quit intel vga on a big screen

edit: is there a chip missing ??

[inteks]

btw. i found the bug why reading EC doesnt work....

 
            public static byte ReadEC(byte iPort) {
            if (WaitWriteEC()) {
                ols.WriteIoPortByte(0x66, 0x80);
            }
            if (WaitWriteEC()) {
                ols.WriteIoPortByte(0x62, iPort);
            }
            if (WaitReadEC()) {
                return ols.ReadIoPortByte(0x62);
            }
            return 0;
        }

should be ->

 
            public static byte ReadEC(byte iPort) {
            if (WaitWriteEC()) {
                ols.WriteIoPortByte(0x66, 0x80);
            }
            if (WaitWriteEC()) {
                ols.WriteIoPortByte(0x62, iPort);
            }
            if (WaitWriteEC()) {
                WaitReadEC();
                return ols.ReadIoPortByte(0x62);
            }
            return 0;
        }

the Sleep(1) is not a problem (as long as rweverything is not running)





PS:
@_InKr_ and @prikolchik -> any news in KBC modding

[_InKr_]

PS:
@_InKr_ and @prikolchik -> any news in KBC modding

Actually I've stopped with following algorithm:
TZ0<60 - 0% speed
60-66 - fan is running at lowest possible speed F0 (lower that is used in original BIOS)
>66 - not modified.

Fan speed table has next values:
FF FF FF F0 81 6F 6F 6F 6F

I 99% satisfied with this algorithm. Next step on its improvement needs lots of playing with temps thresholds and I don't think it worth it.

Can share modded KBC if anybody interested.

[inteks]

Actually I've stopped with following algorithm:
TZ0<60 - 0% speed
60-66 - fan is running at lowest possible speed F0 (lower that is used in original BIOS)
>66 - not modified.

Fan speed table has next values:
FF FF FF F0 81 6F 6F 6F 6F

I 99% satisfied with this algorithm. Next step on its improvement needs lots of playing with temps thresholds and I don't think it worth it.

Can share modded KBC if anybody interested.

TZ0 is cpu temp?
does it work for gpu too?

if it does not have sideeffects pls share . . .

[_InKr_]

TZ0 is cpu temp?
does it work for gpu too?

if it does not have sideeffects pls share . . .

Now it works only for CPU Fan. Have no need to quiet GPU Fan since I only turn on ATI for gaming and noise is not a problem then

Of course one can modify GPU Fan tables if needed.

Modded KBC works fine for me and prikolchik so I'll share it later today.

[zsero]

Now it works only for CPU Fan. Have no need to quiet GPU Fan since I only turn on ATI for gaming and noise is not a problem then

Of course one can modify GPU Fan tables if needed.

Modded KBC works fine for me and prikolchik so I'll share it later today.

But what is TZ1? In HW Monitor its something like a system board temperature. Its definately not the AMD card. Is the AMD card in the ACPI table? I would be happy to use a more quiet AMD card, as I need to use it for HDMI all the time when docked.

Update:
Or maybe a much easier option: could you update KBC 1.13 with the thermal values from 1.19? The thermal settings in 1.19 are really quiet, but it has a bug which makes my AMD fan go into an infinite loop. I would need to send it into a warranty replacement to have it fixed.

[_InKr_]

btw. i found the bug why reading EC doesnt work....

 
            public static byte ReadEC(byte iPort) {
            if (WaitWriteEC()) {
                ols.WriteIoPortByte(0x66, 0x80);
            }
            if (WaitWriteEC()) {
                ols.WriteIoPortByte(0x62, iPort);
            }
            if (WaitReadEC()) {
                return ols.ReadIoPortByte(0x62);
            }
            return 0;
        }

should be ->

 
            public static byte ReadEC(byte iPort) {
            if (WaitWriteEC()) {
                ols.WriteIoPortByte(0x66, 0x80);
            }
            if (WaitWriteEC()) {
                ols.WriteIoPortByte(0x62, iPort);
            }
            if (WaitWriteEC()) {
                WaitReadEC();
                return ols.ReadIoPortByte(0x62);
            }
            return 0;
        }

the Sleep(1) is not a problem (as long as rweverything is not running) :cool

Actually no. With no other software that reads EC running, works even this code:

 
            public static byte ReadEC(byte iPort) {
            if (WaitWriteEC()) {
                ols.WriteIoPortByte(0x66, 0x80);
            }
            if (WaitWriteEC()) {
                ols.WriteIoPortByte(0x62, iPort);
            }
            if (WaitWriteEC()) {
                return ols.ReadIoPortByte(0x62);
            }
            return 0;
        }

But correct code, which is working on Linux is:
ec.c in trunk/src/ec/acpi ? coreboot

It corresponds to code which you have mentioned as buggy.

Also note udelay(10) - this couldn't be done on Windows. And this is a root cause.

[inteks]

Actually no. With no other software that reads EC running, works even this code:

 
            public static byte ReadEC(byte iPort) {
            if (WaitWriteEC()) {
                ols.WriteIoPortByte(0x66, 0x80);
            }
            if (WaitWriteEC()) {
                ols.WriteIoPortByte(0x62, iPort);
            }
            if (WaitWriteEC()) {
                return ols.ReadIoPortByte(0x62);
            }
            return 0;
        }

But correct code, which is working on Linux is:
ec.c in trunk/src/ec/acpi ? coreboot

It corresponds to code which you have mentioned as buggy.

Also note udelay(10) - this couldn't be done on Windows. And this is a root cause.

the missing "WaitReadEC();" is the problem (its the step 7.)

6. Wait IBF free ->WaitWriteEC()
7. Wait OBF full -> WaitReadEC()

since i fixed that i doenst have any acpi error in windowslog anymore !!

that it delivers allways 0 if RW.exe is running is maybe simply because the EC chip cant be used by two progs the same time ...


i have implemented in 3820tray if it returns "0" then EC is blocked and i retry later