• Visit https://www.embeddedcomputers.net/ for Hardware; Software and all other things related to FlashcatUSB

Trouble with MSI P67A-GD65

DeathKnight

DeathKnight

New Member
Been trying to get the Flashcat software to detect the BIOS chip on my MSI P67A-GD65 (Winbond 25Q64CVSIG x 2 (dual BIOS)) and am having no luck. I've got all the pins hooked up correctly as far as I know. Keep getting "no flash detected". When I swap the SO SI pins it erroneously detects it as Microchip (makes sense I guess). I've seen claims on the MSI forums that both GND's and VCC's on the JSPI1 connector need to be hooked up and/or the HOLD (pin 9) needs to be connected to VCC. Not sure how true this is.

Does support need to be added to the firmware to detect this chip properly? Thanks for any help!

Using build 390 of the software and the Flashcat has version 4.09 firmware (PCB 2.1)
 
Last edited:
When you get the connection as Microchip (formerly SST) what ouptut do you get in the console window ?
The Winbond chip is added in the software and has been tested before.
 
Here is the console output with the incorrect chip detected:

Device connected in SPI mode, firmware version: 4.09
Successfully opened device in SPI mode (8 MHz)
Connected to Flash (RDID:0x000000 REMS:0x8000)
Flash detected: Microchip 25AA160B (2,048 bytes)
Programming mode: SPI (high-speed)
Checking for a device specific script to automatically load
Loading device specific script: default.fcs
Loading script: default.fcs


The address readout is just all zeros.
 
Do me a favour and grab one of the older software versions from the software section and see if u have a similar output , i have found an issue with FCUSB v390 on some ejtag models , so just want to see if the issue is also with SPI
 
Ok, tried build 371 with same results. Build 365 detects a different chip (Atmel AT25010A, 128 bytes flash size). Build 355 detects the same Atmel chip but with a 16,384 bytes flash size. Build 350 no flash is detected by default, detects just "80 0000" as the connected to device when set to normal SPI mode in settings tab, and when set to nRF24LE1 SPI mode it detects SPI compatible device (Nordic RF24LE1 SPI port) with flash memory size of 16,384 bytes and a flash memory name "Main flash block".

Very strange, lol
 
On the old software if the Nordic is ticked it always detects something so that has to be unticked always .
In the documentation for FCUSB the developer has provided a schematic for connecting to MSI boards , i am assuming yours is connected like that ?
Weird that it doesn't detect your specific chip ... on the status tab does it give you the same JEDEC ID on each connection ?
 
When I have it connected per all the schematics I've seen for the MSI boards I got the no device detected in build 390. I've been running it with the SO and SI connections swapped for all these tests thus far. I swapped those back for how it's supposed to be and tried build 371 again and the connected device comes up as "FF FFFF" and says an SPI device was detected but not currently supported. Tried the same thing in build 365 and it scans for several seconds then says no device detected.

When a chip, albeit the wrong one, is detected the JEDEC ID is always "00 0000".

Very strange behavior. Would rolling back the Flashcat's firmware accomplish anything? I'm also starting to wonder if the BIOS chips on the board might be hosed somehow...
 
I can't see the chips themselves being damaged to the extent that you can't properly detect them , however the false detections you are getting confuse me , as they should show some form of JEDEC ID , the SI and SO have to be swapped on FCUSB for detection purposes ,,, there is a MSI thread on the forum from a while back where people had different issues in detecting , are you using power on the MSI Board when you attempt a connection ?
 
No power supply hooked up to the board. I did try though. The problem with this board is it does the power on power off cycling over and over due to a bad BIOS flash. I can't get constant power through it. I also did some voltage tests and the chips are getting 3.11 volts between VCC and GND. HOLD is also getting this same voltage.
 
Nope. The cable for MSI boards that got supplied with the Flashcat (purchased off of Ebay from an official reseller) isn't set up for splitting one pin out from the Flashcat board to two pins. Electrically the two VCC's on the JSPI1 header both lead to the same VCC pin on the flash chip (checked with continuity test) so I'm not sure what that would accomplish anyway other than effectively making it a larger gauge path. I tested voltage across the rest of the pins on the flash chip and am getting 3.11 volts on CS (chip select), SO/DO (serial data out), and WP (write protect). According to the PDF for this Winbond chip ( https://www.winbond.com/resource-files/w25q64cv_revh_052214[2].pdf ) CS is supposed to be brought low in order to read from it and write to it.

Also tested voltage on the Flashcat pinouts (hooked up to the motherboard)...

VCC: 3.15
SI: 1.07
SO: 3.13
CLK: 0
CS: 3.13

Is there a way to tell if the Flashcat board itself is functioning properly?
 
Last edited:
I have not done an MSI motherboard before so i am only going on what the developer has suggested as per connecting both GND and VCC , maybe make a jumper between them ?
I checked the voltage on my v2.1 and only differance i have is that i get 3.3 on my SI pin.
Without having another SPI device it is kind of hard to determine if your fcusb is at fault ..... Try setting your FCUSB to the 5v setting and see if your detection alters as well ( this will not cause any damage to the chip)
 
Are both chips corrupted on this board ? As I was under the presumption you could flick the switch to the alternative chip ?
 
Now I'm getting random results with build 390 (3.3v)... Bouncing around between Microchip 25AA160B, 80 000, C0 40C0, C0 6040, E0 60E0, F0 60F8 and no flash detected. I'm sure I could keep resetting to get more gibberish. Same random gibberish with 5V. Also similar results with build 371. Build 365 detects Atmel AT25010A every time regardless of voltage.

Tried some jumper wire to connect both VCC's and GND's with no difference.

AFAIK both chips are corrupted. Unfortunately there is no manual switch to swap which chip is used. It's supposed to be automatic.
 
Only thing I can think of to try is to manually solder to the chip itself and apply power to the chip and see what results you get.
 
Not sure I want to go that far. I didn't expect to run into this much trouble. It's supposed to just work. Again, all this finagling around leads to me to believe either the BIOS chips are fubar (or the motherboard) or the Flashcat is messed up. Like you said I'd need another SPI device to see what the problem is. Dunno if I have one lying around though.
 
Fiddling around some more and it's now detecting the chip properly.

Winbond W25Q64
JEDEC ID: EF 4017

Not sure what's different this time as I've tried all this before. I do have the power supply plugged in like before without the board on. Maybe something was loose previously? Beats me. Anyway, tried writing the BIOS file in 390 but it had write errors on every address. Tried again in 371 and it's writing fine as I type this.
 
No change. Still having verification errors. Don't think write-protect is enabled because the pin is reading 3V and it's active-low according to the documentation. I can't blank the chip either. It instantly says it completed the operation and there is no change in the data.

Thank you for all your help thus far by the way :)
 
You must log in or register to reply here.
Back
Top