Marvell AQC113C FourPartID Firmware updater edit

Hi,
The latest version being 42 that I managed to put on my X670 Proart without problem via a key

AQC113

Can confirm that 1.5.42 works for the Fenvi AQC113C, too.

And it comes with a "ASPM_Enabled" image!
I somehow missed that before and so didn't try that version again after I got the updating working...

any chance of getting a backup of your card on 1.5.42

Here is the generated image from flashing mine:  www.dropbox.com/scl/fi/zzq7hm4e03k5dtmz5...ywr&st=z6izgm55&dl=0
You can directly flash it with the CH341 programmer. Just remember to erase the flash chip first (and verify the content afterwards).

Hi,
I have to warn you about this firmware!

VERSION 1.5.42 IS COMPATIABLE ONLY WITH THE LATEST REVISION OF THE CARDS WITH AQUANTIA AQC113...
IT IS ALSO POSSIBLE THAT THIS FIRMWARE VERSION IS INTENDED TO USE IN PARTICULAR DEVICE ONLY!

This version bricks majority devices. There should be a warning on a download site.

This is also why someone ask for help on the forum, after bricking his card by flashing this version.

@admin confirms working on the X670 Proart motherboard. For now, DO NOT flash any another device unless someone would confirm the proper working. Maybe this version is only for new devices with the latest revision of the hardware.


All firmwares for AQC113 has specific values for every different device. These values are idenctical for every firmware version. They prepare PHY and MAC regions for proper working. The flashing tool uses them to authenticate/signing and build the final flash file..

The 1.5.42 version and only that version has other values for PHY and MAC regions. From my observation, those values brick all older devices at all. The Chinese (1d6a/14c0/0001/1d6a) cards as well.


All firmwares contain the same PHY and MAC values:

 


Only 1.5.42 version is different:

 


We can see that for almost all devices, MAC region has 63d8ca value instead of db9ca1 included in all previous firmwares.
In PHY we can see 2 changed bytes. All of them increase by 0x2 (0x86->0x88, 0x81->0x83, 0x85->0x87, 0x82->0x84, 0x80->0x82, 0x7F->0x81)





Next warning!
I've read that some of you dumped fimware from the Chinese AQC113 card using CH341A.

I found the video and saw that this card has MX25V80066 memory chip. This chip works with 3,3V.
The CH341A programmer has a bug. The voltage can be changed by a jumper to 3,3V but on all data lines there are still 5V!
Without a simple fix of the programmer, using it for 3,3V chips can quickly destroy them. Futhermore, the dump usually contains corrupted data. MX chips are vulnerable for the higher voltage. Only first few times the chip can resist higher voltage and reads the dump correct. If you dumped the firmware only few times using the unfixed CH341A, never do it again, Find the manual in the Internet and fix your programmer in 5 minutes.


BTW. These chip from the Chinese card is 8Mbit size (1MB). I thought that Aquantia uses 32Mbit memory (4MB). AQC107 does. Am I missing something?



PS. I need a huge favor from you @Shonk. I sent you a PM. Did you read it?
Could I ask you for sending to me the dump your Chinese AQC113 card? I bricked my card flashing that 1.5.42 firmware and your help would allow me to fix it.

I also ask you @hu4567. I understood that you have a working dump as well. Could I ask you to send me it?

I can pay for the help. Thank you.
My email is i225v@outlook.com




EDIT

---

Holy... a lot happened when I was writing the post.
1.5.42 really works on your Chinese card?
2 weeks ago I flashed my card just like you described but the card never started again. Totally bricked.
When did you buy the card? Could you show some photos of your card? Maybe you have got somekind of a newer revision.
I saw so many bricked cards after flashing 1.5.42...
Thank you for sharing your dump. I understand that the rest of 1MB are zeros?

I'am really impressed. Firmware is really signed with 63d8ca value from 1.5.42.