installed 4k AUO screen in P775dm3-g - no g-sync

Just installed B173ZAN0 1.0 AOU UHD/4k screen in to p775dm3-g G-sync 1080 gtx machine and the G-sync option is not available in the Nvidia panel.

any idea why? driver installed, the GPU hardware ID did not change from FHD where G-sync works perfectly.

maybe @Prema or @Meaker@Sager knows?

 

Are you sure that panel is G-sync certified?

 

it was certified for p870dm1-g and p775m-g and was tested to work there with G-sync

also i know that system bios changes ID to a non-G one when wrong panel is installed, right?
and drivers would not install, however not the case here.

also i have asked Clevo if they are still using same panel and they said yes.

 

panels HW ID is AUO109B - seems to be a G-sync one according to what i read on the web.

or maybe different EDID is needed?

 

1080p and 4k p775dm3 uses different casing, that's why there's the differenece in the weight

Did you use stock FHD cable or bought a new one? May be you can flash latest clevo bios/ec as well


Sent from my SM-G930F using Tapatalk

 

off course i have used 40pin cable for 4k/UHD panel. what is the last bios version?

i have 1.05.06 installed here.

 

When you order a laptop with a Non-GSYNC screen, you also get a non GSYNC GPU

So now, you installed a GSYNC screen but your GPU is non GSYNC so you won't have a GSYNC option unless you swap the GPU as well.

 

my laptop is Gsync, please read above. it had a G-sync FHD screen that worked perfectly with G-sync.

my bios is G-sync, my GPU has G-sync ID.

 

something weird is going on, as this same panel works in p870dm-g with 980m, G-sync enabled!

 

ok, g-sync actually works, with driver from August, that i have downloaded from Sager site. hm...

 

ahhh then you need the new VBIOS which enables GSYNC

I hope @Prema can advice here

 

What if you overwrite (update) your august drivers with newer ones? Not completely remove with ddu

Sent from my SM-G930F using Tapatalk

 

what do you mean?

vbios not having G-sync option for 4k and having option for FHD?

actually just tried driver from August, from Sager website for p775dm3 and G-sync works

but after installing newest driver - G-sync is gone again. how come?

 

tried that, no G-sync.

 

Nvidia derping.
Same issue on quite a few desktops and laptops..

 

i ve read about that, but seems they still have option in the NV panel, i dont though

 

I've seen other users who ordered their laptop with a 4K screen also complain that they lose GSYNC if they update their driver. I don't know what the heck is up

 

do you have a link maybe?

 

@tanzmeister
maybe coincidence but your issues may have something in common with mine

 

What is your BIOS and VBIOS? What LCD cable are you using P/N?

 

@tanzmeister

That's the link I was trying to find for you. There you go

 

i ve got official response from CLevo, that new drivers lack support for AUO UHD G-sync

it's Nvidia messing, according to Clevo tech support

 

NVIDIA.. Fail fail and more fail!

 

Have you reached out to NVidia support about this by any chance?

 

hey budy I want to ask you about the p775dm3 because I'm thinking to pull the trigger for this
how is your temps vga and cpu??
your impresions?
I'm thinking the 1080 version
thank you a lot

 

so far i have tested all 3 videocards offered for p775dm3. i must say that gtx 1080 makes a sort of a hairdryer out of it, haha. but when it does, temps drop to mid 70's on GPU, also 1080 gtx acts very well to undervolting(1070 and 1060 not so good from my tests), that drops the temp even lower for same Mhz.

with 1070 gtx it becomes an average gaming laptop when it comes to noise and heat output,

with 1060 it's very quite and cool.

nope, funny thing though, nvidia missed p775dm3 out of the new driver completely. lol

 

Hi there. May I know from where you bought the screen? As I found several sellers at Ebay with wildly different prices for the Alienware. And do you know why Clevo listed the 4K version of the P775DM3 to be much heavier then the FHD version? Do you notice the weight difference?

 

Slightly off-topic but maybe here I'll be able to get what I need. I've did the same display upgrade as the topic starter did but I have a different laptop.
I have Acer Predator G9-792 and I've replaced the original FHD display with the AUO109B (B173ZAN01.0).
Here's the issue: it shows bios output, it can boot windows or linux without videodriver. For linux that means nomodeset key, for windows - booting without i530 driver. After win10 installs videodriver it goes to black screen while brightness controls start working. Same with linux. So basically it works untill a resolution changing attempt happens.
p.s. there is no bios update available for my model - the only accessible bios is 1.05.

Ideas/input appreciated. wrong / empty EDID? something else?

 

upd: I've figure out my EDID is wrong - checksum is incorrect.

Guys, could anybody share AUO109B EDID dump?
Its pretty simple to fetch that from your display with moninfo in windows or sudo get-edid > dump.bin in linux.
Please...

 

Guys, could anybody share AUO109B EDID dump?
Its pretty simple to fetch that from your display with moninfo in windows or sudo get-edid > dump.bin in linux.
Please...

I've got a failure there - checksum is wrong so it goes blank screen after any resolution/mode change:
0 1 2 3 4 5 6 7 8 9 a b c d e f 0123456789abcdef
00: 00 ff ff ff ff ff ff 00 06 af 9b 10 00 00 00 00 ........????....
10: 00 19 01 04 a5 26 15 78 02 23 25 a8 50 37 b4 26 .????&?x?#%?P7?&
20: 0e 50 54 00 00 00 01 01 01 01 01 01 01 01 01 01 ?PT...??????????
30: 01 01 01 01 01 01 66 d0 00 a0 f0 70 3e 80 30 20 ??????f?.??p>?0
40: 35 00 7e d6 10 00 00 18 00 00 00 0f 00 00 00 00 5.~??..?...?....
50: 00 00 00 00 00 00 00 00 00 20 00 00 00 fe 00 41 ......... ...?.A
60: 55 4f 0a 20 20 20 20 20 20 20 20 20 00 00 00 fe UO? ...?
70: 00 42 31 37 33 5a 41 4e 30 31 2e 30 20 0a 00 3b .B173ZAN01.0 ?.;
80: 98 00 00 a0 aa 98 00 00 a0 aa 00 00 00 00 00 00 ?..???..??......
90: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
a0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
b0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
c0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
d0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
e0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
f0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................

 

Yes, indeed. There's also some data beyond the 128-byte edid, yet it claims 'no extension block', so be careful not to overwrite that when flashing the new edid. Here's a good version of this screen:
AUO109B - B173ZAN01.0 (3B).bin

Spoiler: AUO109B - B173ZAN01.0, ok vs. corrupt

 

B173ZAN01.1 is the gsync version I think

 

t456, thank you very much!
Furthermore, it seems to be impossible to program it through eDP - will have to write eeprom directly.
i2cset is unable to write bytes to it. There are no errors on writing and i2cdump returns bytes like those are updated but after reboot it looks like those writes were fake

By the way, which h/w and f/w ver do you have?
Mine is H/W 2A, F/W 2

 

You can blame eVGA precision for that, they were writing to the panel and bricking displays so the connection was disabled.

 

Haven't the foggiest; don't have this panel, just gathered a lot of edids due to all the screen-brickings taking place. That version is the same as yours, no doubt; checksum and almost everything else matches. The only difference is a minor change in the colour coordinates, but even then not so much as to be 'wrong', per se. So apart from the checksum it is fine, which certainly was not the situation for those affected by the EVGA tool.

 

Can't change any byte with i2cset in my display. Tried several displays with no luck. I2cset shows no errors, then i2cdump correctly dumps contents thus it looks like its updated then after reboot i2cdump shows the initial dump so no actual changes has been made to the display eeprom. Any thoughts on how to fix that?

also eDP has no eeprom direct outputs to its main connector so can't flash it with a regular programmer without soldering out the chip. and I can't find that dumb eeprom chip (if it exists there, I guess its emulated by the display CPU)

 

Nope. That seems to be just a revision. The 01.0 works with Gsync just fine on my DM1 (trades brightness control) and DM3 (has both Brightness control and G-Sync).

 

Hm that is interesting.

The HP Omen 17 had the newer revision with brightness control and G-sync.

 

Yeah i was reading that some Clevo's dont have G-Sync on these even with G-Sync GPU's. Im not too sure about the HP's though.

 

IRC B173ZAN01. 1 has no hardware g-sync buffer (NVSR). Clevo uses B173ZAN01. 0, that's the model they ordered g-sync license for from NVIDIA.

 

I had a 1.1 revision and swapped it out for a 1.0 so I could get G-sync

 

There is a connection, but most eDP panels have these pins listed as 'NC' or 'Factory only'. For 40-pin panels it is pins #34 and #35, but that seems to do double-duty for g-sync on this panel:

Spoiler: AUO109B - B173ZAN01.0 - edid pins (34 and 35)

They must have implemented a switch somehow, enabling it to do either g-sync or edid read-out (and writing). Only have a preliminary specification sheet and it doesn't mention the method anywhere, unfortunately. Don't think it has gone awry though; the edid differences are certainly not random. More likely is that the changes in the colour coordinates were adjusted at the AUO factory, but they forgot to update the checksum to compensate for these changes (it should be 3D instead of 3B).

At least the switch is working, but the eeprom is likely write-protected. There's some soldering pads you could use, but it may be easier to disable wp and flash it with i2c-set. The pin they use for this would be pin #1 and it should be either grounded or set to 2.5 or 3.3V. Measure pin #1 when it is powered up and you know which option it is; no voltage, then set it to vcc and if there is ~3V then it needs to be grounded. To set voltage to it you can use any of the pins 18 to 21 (add a resistor of several Kohms) and for grounding simply solder the wire directly between #1 and any of the 23 to 26 pins.

Alternatively, you could do the same thing on the eeprom itself (or desolder it and flash it using a programmer):

Spoiler: AUO109B - B173ZAN01.0 - edid eeprom

Can also hook up the edid soldering pads to the programmer:

Spoiler: AUO109B - B173ZAN01.0 - edid eeprom solder pads

Spoiler: LQ133T1JW02 - edid eeprom solder pads

Problem here is that you have to know which pin is the write-protect, since it can be either #3 or #7. Would have to find the specification sheet or do some detectiving with a multimeter. This is also the most risky method unless you know how the eeproms should be written to; i2c eeproms can be daisy-chained (so writing can affect the main chip) and it also contains the real firmware of the panel (the edid being a small part of that), so using the wrong parameters risks corrupting the non-edid-section. A good full copy and flashing the entire eeprom would be required to fix that.

So ... options, options, options. None of these are very user-friendly, of course. Perhaps you can also sidestep the issue entirely and fix the edid copy in the registry (Windows) or force a good binary at boot-time (for Linux). The registry location is ' HKLM\SYSTEM\CurrentControlSet\Enum\DISPLAY':

Spoiler: edid copies in registry

Complication here is that usually you cannot edit these values directly (have to hack permissions first) and/or editing them causes a direct bsod due to Intel HD or Nvidia shenanigans. Could make an export to a .reg and change its edid values accordingly (checksum 3B -> 3D):

Spoiler: example

Code:

Windows Registry Editor Version 5.00

[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Enum\DISPLAY\SHP0456\4&2a00b1d&7&UID68092928\Device Parameters]
"EDID"=hex:00,ff,ff,ff,ff,ff,ff,00,4d,10,56,04,00,00,00,00,02,1b,01,04,a5,1d,\
  11,78,06,de,50,a3,54,4c,99,26,0f,50,54,00,00,00,01,01,01,01,01,01,01,01,01,\
  01,01,01,01,01,01,01,9a,5b,00,50,a0,a0,29,50,08,20,b8,00,26,a5,10,00,00,18,\
  55,73,00,50,a0,a0,34,50,08,20,68,00,26,a5,10,00,00,18,74,8b,00,50,a0,a0,3f,\
  50,08,20,18,00,26,a5,10,00,00,18,12,ac,00,50,a0,a0,4d,50,08,20,f8,00,26,a5,\
  10,00,00,18,00,6a

Then add an autoexec.bat to C:\ and boot using safe mode:

Code:

regedit /s fixededid.reg

Probably need to edit those permission first, but could try it 'as is' first and see whether it works.

 

T456, I highly appreciate your post and knowledge, thank you.
I had no idea that chip is eeprom since my is marked 2L550 and it can't be googled so I was not thinking its an i2c eeprom.
Regarding the clk-gnd-data connector - my board has two of those and none of them has direct connection to eeprom chip according to what I've figured out with multimeter.
Also you're saying its not user friendly... actually it is. Pretty much. Nothing complicated except lack of documentation - you're the only person in the world who shared this how-to. Thank you!

 

Allright, I've soldered a wire to pin #1, powered up the display and I see ~0.3V on pin 1. It is not looking like its in Z-mode.
I've tried both linking it to 3.3V and GND. No dice.

i2cset seems modifies some kind of a buffer since its read back as it was written successfully. But then its enough to do a hot unplug-plug-reread operation to see its not actually modified.

 

Software method is not working. Its working with nomodeset (backlight adjustment not working). Either there's black screen but backlight adjustment is working fine

 

I have been considering an MSI GT75VR laptop with a custom configuration, and the screen to be installed is
AUO B173ZAN01.0. I am wondering about this screen upgrade. I have been informed that it is not G-sync and that it is not 99% ADOBE RGB (It is supposed to be 72% Adobe according to the upgrade people). I have looked around on the web and I have found conflicting reports stating it is 100% ADOBE RGB, and G-Sync (the laptop is G-sync capable). Does anyone know anything about this, and also how the screen is to play video games, watch movies in 4K, and generally just to look at compared to the screens normally coming on the GT75 series 4K version laptop standard from MSI. The standard 4K model is supposed to be beautiful and good for content creators. Any comments?

ADOBE?

True Tech Color and Color Accuracy?

G-SYNC?[/QUOTE]

 

G-sync requires the panel to be certified as an original option. If you take a "G-sync" panel and put it in another G-sync machine that never shipped with it (with G-sync enabled) it would not work.

On the flip side you can install an exact replacement into a G-sync device and it will still be G-sync so long as the panel is identical.

 

Thanks! That is extremely helpful...and it reflects much of what I had already gathered.

 

Yeah, they locked down each step and then each config to stop simple unlocking.