[Request] ROG Strix G15 Advantage Edition G513 (AMD)

[Sml6397]

Hello crew,

I will respond to all three of you

@jeanlegi
Unfortunately, both of your BIOS dumps contained just FF bytes (ie: FF FF FF FF FF ....... FF FF FF was the whole file). They are unfortunately not viable as they contain no actual data. Don't worry, I totally understand your frustration. I've been there!

I think if you are getting something that looks like "id 9f, id 90, id AB, ID15", it is identifying *a* chip. When you got to this, did you get a popup window saying that the chip was identified as "Winbond XXXXXXXX" or something like that (it might not be Winbond, though I'd imagine it is the same as Geonode's)? If not, try getting it to give you those IDs again and clicking IC>SPI>Winbond XXXXXXXX at the top left corner of the AsProgrammer window, where XXXXXX is whatever the model number is on your BIOS chip. If you can't find the model number in the list, choose one that is very similar looking.

@Geonode
Kudos for your valiant efforts! I've been here before with flashing problems and they are frustrating for sure!

@CoolBeans
I took a look at your BIOS dump.

It appears that XMP does not exist in your BIOS image. However, there are many overclocking, timing, and voltage settings related to RAM in the AMD CBS Menu that do the same thing, I believe (the same CBS Menu that has been mentioned throughout this thread). Your BIOS also contains the "mythical" AOD Menu, which usually has every kind of overclocking setting a technical enthusiast could dream of. As I stated earlier, BDMaster and I are figuring out how to unlock the CBS, PBS, AOD, and PciDynamic Menus. We think we have a lead right now. As soon as we figure this out, I will unlock these 4 menus for you and everyone else on this thread. All four of these menus most likely need to be unlocked in the exact same way, meaning that if we figure out one, we have all four.

Right now, I can offer you the same preliminary mods I am offering jeanlegi and Geonode, which contain a second very fully featured Advanced Menu, a new Chipset Menu, and a better Boot Menu. I can offer this mod as soon as you can get me a reliable dump using the CH341A Programmer.

[jeanlegi]

No popup but i check the different entries in ASProgrammer for the winbond chips with 1.8V to find maybe one which is compatible to mine winbond chip W74M12JW.

Edit:
It doesn't matter what i try from alle the links and what ever the chip is unreadable.
i try now the 3rd programmer software...

if i do a mistake i can't see it on my own currently...

[jeanlegi]

i gave me a break for some days and then i try it again .

[Sml6397]

@jeanlegi

When you are ready to have a go with the SPI programmer again, could you take several high quality pictures of (1) all the connection points between the CH341A and the clip with it attached (2) the BIOS chip without the clip attached (3) the entire setup with the clip on the chip exactly as if you were about to flash it?

You probably have the connection right, but I just want to make absolutely sure. It is so easy to mix up.

[jeanlegi]

I create a new folder in the link i shared with you. don't know why i can't upload it here.
18MB should fit from the size of 20MB.

I try it last night again several hours but no dump writing possible. i change position of the clip, try different chips from the programmer with 1.8V and 16MB/128 Bit etc...

[Sml6397]

Hello jeanlegi,

With the exception of the 1.8v adapter (I can't comment on this because I have not used it), everything looks like it is plugged in properly. Good work.

My advice would be to do one of three things:
1.) Purchase a clip that gets a better connection to the BIOS chip, such as the Pomona 5250 clip. I had to do this in order to get a good BIOS dump from my test Phoenix UEFI machine. CoolBeans is going to be trying out the better clip as well.

2.) Keep trying to get BIOS dumps after connecting the programmer. Download a hex editor (such as HxD) and open each dump to view the hex data. If the entire file isn't FF or 00 bytes, then your dump might be good (the beginning of the file might be lots of FFs, which is fine). In the case of a not-all-FF/00 dump, take a second one and either use HxD to compare them or upload them to your OneDrive and I can compare them to see if they are the same. Keep in mind that dumps taken after different system boots will be different because of the NVRAM region, but I can tell if this is the case. I'd most recommend this step after 1, as it sounds like you've already thoroughly exhausted the possibility of the cheap included clip being viable for BIOS reading/writing.

3.) Contact BDMaster via PM for more help. He is way more experienced than I am with hardware programmers.

[jeanlegi]

i wrote bdmaster via pm.
in the datasheet i found this:
Operations of Authentication Method
The W74M12JW is equipped with four 32-bit Monotonic Counters. Each set is accessible by the 8-bit
Counter_Address that is HMAC-signed by the appropriate secret key. The SPI Flash Host controller uses
the Monotonic Counter value to validate authenticity of the attached W74M device or modules which the
W74M device resides.

i understand that it can be possible that i need a secret key but to less knowledge about this topic from my side.

[jeanlegi]

i think i got something ... i really hope it.

you can find it on onedrive

[Sml6397]

i think i got something ... i really hope it.

you can find it on onedrive

Okay, we are getting somewhere now! I checked this image and it *is* corrupted, but the result is exactly 16MB in size, which is good. From the looks of it, there is a relatively small amount of data in the dump compared to what I'd expect normally. My guess is that the clip is still getting a poor connection on the chip (I believe this even more now). If you can, getting two dumps in a row could help shine light on this.

I will do some further research regarding the authentication check to see if it is relevant to the case of using a hardware programmer (It might just be for CPU access or some other type of access, I haven't gotten a chance to read the chip documentation yet).

[jeanlegi]

second dump is online as well.
i do both in a row.

add test 2 and 3 to onedrive