Products > Test Equipment
Hantek MSO5202D doesn't boot
(1/1)
xrayman:
Hello everyone!
Back in 2014 I've bought Hantek MSO5202D; hw version: 1010×55778340; serial number: T 1G/012 019xxx. It has worked fine until recently. Then it started to boot once per 5-10 power-ups. I used it in such condition for some time. Then I had to switch power on and off like for a half an hour to get it booted up. And finally I've had enough and has started digging the information. First of all I connected to the oscilloscope using console port and obtained the boot logs - they're pretty much common for these kind of situations:
--- Code: ---Warning - bad CRC or NAND, using default environment
##### EmbedSky BIOS for SKY2416/TQ2416 #####
This Board: SDRAM is 64MB; LCD display size is: 800 X 480
Press Space key to Download Mode.!
SD:File System init failed!
Start Linux ...
NAND read: device 0 offset 0x300000, size 0x400000
4194304 bytes read: OK
Boot with zImage
Starting kernel ...
Uncompressing Linux... done, booting the kernel.
--- End code ---
I contacted Hantek support via email describing the situation and they sent me firmware update files. I updated the system and it didn't help. Tried to contact support once again and have never received an answer. Then after some time of investigation I found dnw.exe utility and proper drivers for my device. Also somewhere in this forum I found a link with dumps of the partitions from MSO5202D as it's stated in the title.
I ran the bootloader, selected the option `Test Linux Image (zImage)` and transfered kernel.bin from that link. And the oscilloscope booted up! Then I've reboot the bootloader and selected the option `Download Linux Kernel (zImage.bin) to Nand Flash` and transfered kernel.bin once again. Now I have the oscilloscope that boots up every time. The only issue is that those dumps are not for 5202D. Now in the System information I can see that the model is MSO5102B.
So the question is - could anyone please send me a dump of a kernel partition for MSO5202D?
My device has S3C2416 SoC and 4 partitions according to the boot logs (not 8). Boot logs of the oscilloscope before I downloaded the new kernel are attached.
xrayman:
The problem seems to be solved.
First of all I need to mention that all the experiments with flashing the kernel found on the Internet I was doing on a spare nand that I've bought. So I had the original nand without any modifications. The behavior of the oscilloscope with the original nand was as described - it booted up once per many power up cycles.
The idea of the repairing was that if the scope got booted up at least once per many tries it meant that on time when it actually booted up the info from the nand being read correctly. And if I using the programmer would repeatedly read the dump from the nand I would eventually get the copy that works.
I used Xgecu T48 programmer to read the nand, AutoIt app to automate the reading process, any tool that could generate checksum of a file (TotalCommander in my case), python script that removes OOB data from the dump and another python script that extracts the kernel into separate bin file (addresses of the kernel I've taken from the successful boot log).
Then I ran the automation script, obtained 10 dumps (it was the very first attempt just to be sure that automation worked), generated sha1 for each file, compared the checksums and found out that 8 of them were equal and 2 other were equal but different from the first 8. I concluded that those 2 were the good dumps. I extracted the kernel partition from one of them and ran it on the scope using bootloaders option `Test Linux Image (zImage)`. And the scope have booted up. Then I downloaded this kernel to nand using another bootloader option. And now I have fully functional oscilloscope with the original firmware.
Pilot65:
xrayman,
I have the same scope with the same problem and the same terminal dump.
Can I have a copy of your 4 disk images ?
Navigation
[0] Message Index
Go to full version