TJC HMI 2.8inch red screen SYS ERROR!

Hi, on one of my 2.8inch TJC HMI’s the winbond 25Q32 memory chip was physically damaged, so I ordered some new memory chip, copied the memory contents of the chip from the working 2.8inch TJC HMI to the new memory and soldered to other HMI board, after inserting the SD card everything is OK, the project is copied from the SD card to HMI and verification is OK, unfortunately after removing the card and reconnecting the power supply, all I can see on the HMI is a red screen and SYS ERROR!
did any of you have such a problem or know if it can be brought back to life??
THANKS
Piotr.

Problem fixed!
Topic closed.

Hi @Elektronik.pcb,

Glad you got rid of the issue. Since obviously no one else knew the solution it would be great if you could share it.

Thanks,
Max

I have got the same issue with my NX4827T043_011R. When I solder the w25q128 from a working display to the brick one, it is running. So I try to copy the whole content from the working Flash to another. Solder it in and get SYS ERROR! Every think looks fine. Compared a few pages… Looks good… Compared status register… Looks good. I’m really confused why a exact copy from the Flash is not running. Someone an idea?

Worst case would be that they get chips with specific manufacturer ID or device ID range. I’m no flash chip expert but I guess that would be hard to overcome.
In any case my next step would be to hook up a logic analyzer and see what’s the difference between the working and non-working case. There must be some difference in the IO - and that will very likely tell you what’s going on.

Kind regards,
Max

Did a few tests today. First of all, I wrote a little code which compares every page from the working flash with my copied flash. All pages are copied 100% correct.

I also compared two working flash memories with the same tft loaded. There are a lot of pages unequal.

The flash memory got a register called “unique ID”. This is a 64bit unique id. The manufactor recommended in the datasheet, using this unique id for a software copy protection.
I think this is what nextion did.

I don’t have a logic analyzer and doing so deep resarch would cost me too much time.
I put the brick displays away from my table :slight_smile: . Maybe someone else will find a solution one day.

Thanks and greets,
Andy

I totally agree. If time is a factor, reverse engineering is not the way to go.

I also compared two working flash memories with the same tft loaded. There are a lot of pages unequal.

This however makes me really curious and I’d love to take a look at those differences myself. Would you agree to send me the dumps via PM or something similar?

Kind regards,
Max

I dumped two unused nextion displays which come with a Demo file loaded. There are not so many pages different. Seems downloading a new TFT file will not clear the whole flash. Probably my last dumps have loaded different TFT files before.

Anyway here is a dumps of two unused displays:
https://filebin.net/om17sg0tpmtri5kn
Password for rar file is: NX482

Good luck!

The file link seems to have expired.
Can you upload it again?