New Hantek DSO2X1X models?

[ars_ha3hz]

This is  FEL button, read more here: https://www.eevblog.com/forum/testgear/hantek-dso2x1x-firmware-updates-and-best-use-practices/

[DavidAlfa]

I need someone who can test a flash tool I'm working on, it's able to backup and restore the whole flash from FEL (Bootloader) mode.
It's important to understand it's in testing stage and you might need to recover it using an external spi programmer.
I managed to lock my flash twice by sending corrupted data, platform-tools didn't work anymore, had to remove the flash and erase it using a CH341A to make it work again.

It didn't happen again since I fixed the code, but I must give a serious, proceed-at-your-own-risk warning!
I'm doing my lasts tests, if it keeps working everytime without any issues, I'll release the tool for the brave.
It's not really required, as platform-tools + Backup Builder will do the job, but it's always nice to have a complete 1:1 backup.
Reading takes 50 minutes and writing 25.

[vistorik]

I need someone who can test a flash tool I'm working on, it's able to backup and restore the whole flash from FEL (Bootloader) mode.
It's important to understand it's in testing stage and you might need to recover it using an external spi programmer.
I managed to lock my flash twice by sending corrupted data, platform-tools didn't work anymore, had to remove the flash and erase it using a CH341A to make it work again.

It didn't happen again since I fixed the code, but I must give a serious, proceed-at-your-own-risk warning!
I'm doing my lasts tests, if it keeps working everytime without any issues, I'll release the tool for the brave.
It's not really required, as platform-tools + Backup Builder will do the job, but it's always nice to have a complete 1:1 backup.
Reading takes 50 minutes and writing 25.

OK David, I could test your flash tool. I have all to do it without any problem. So, where could I get the tool?

[DavidAlfa]

Update:
Added DSO Flash Tool to Drive and updated the FAQ section.

[morgan_flint]

Morgan, all it takes is installing and testing, don't be lazy, just try and check yourself and report the result
The worst that could happen is having to remove it and install the older version.

OK, after asking also in Hantek's semi-official forums, I installed Keysight IO Libraries Suite 2022 and I can confirm it works correctly .

I preferred to ask because apart from being lazy (I recognize I am... ), I'm not very fond of installing and uninstalling things, as I think (maybe it's a conspiracy theory) they tend to leave garbage in my not very powerful computer during the process, so if I can go to the best option at the first try, I'd better do that.

By the way, while experimenting with SCPI, I found another way to set the internal clock of the scope alternative to David's method. The details are at the previous link, but I'll post them also in the hacking thread.

[DavidAlfa]

I extracted the scpi commands long time ago here

[morgan_flint]

I extracted the scpi commands long time ago here

Yes, I quoted your post when describing the method in the hacking thread. Very interesting list, especially the PRIvate set, though a bit scary to try them without documentation (impossible to find, I imagine).

Also curious the result of "SYSTem:RAM?" that, according to the SCPI Remote communication control manual for DSO4000BC, reports the number of analog channels, 4 in that case (is a 4 channel scope) but also reports 4 in this one

[DavidAlfa]

No idea, but anyways I expect a lot of bugs on that side.
I really doubt you can break anything as long as you made your backups (builder+Eeprom).

[vistorik]

I need someone who can test a flash tool I'm working on, it's able to backup and restore the whole flash from FEL (Bootloader) mode.

Hi David. I've tested your tool. Something is wrong with the memory size of the flash. Your tool defines the memory size as 138412032 bytes (132MB) (see disoflash_begin.jpg and disoflash_end.jpg), while the NeoProgrammer with CH341A defines the memory size as 132217728 bytes (128MB)(see NeoProgrammer.jpg). In the original memory dump that was read by the tool at address 06e007f0 there are 16x00 and then at address 06e00800 - 16xff, and in the backup.bin file at address 06e007f0 16xff and then at address 06e00800 16xff.

[DavidAlfa]

It's not wrong, this tool includes the ECC area, that's why the final size is 132MB, and only a backup of that size will be allowed when writing.
If the file and flash size differ, in will show an error and block the write.
I might add checksum verification in the future, but I'm way tired of this device.

You can get the same reading with the CH341 selecting "Read OOB area".

By restoring the original ECC, you would have a chance if the backup was corrupted, but without it it'll be gone.

[vistorik]

It's not wrong, this tool includes the ECC area, that's why the final size is 132MB, and only a backup of that size will be allowed when writing.

Hi David. Using your tool I wrote the backup.bin file renamed as restore.bin to a flash memory, then read the contents this flash memory using the programmer CH341A, saved the read to a file and compared this file with the original file with the damp that was written to the flash memory before reading it with your tool. The files are identical. So everything works. Than you for your tool.

[vistorik]

You can get the same reading with the CH341 selecting "Read OOB area".

Where (which programmer-software) did you find this option?

[DavidAlfa]

Check NeoProgrammer.
It's also uploaded in the Drive folder (Hacking/CH341...)

[vistorik]

Check NeoProgrammer.
It's also uploaded in the Drive folder (Hacking/CH341...)

I've got it In the NeoProgrammer 2.2.0.10. You need to check the box against "Read/Write with spare area". Then the NeoProgrammer reads/writes 138412032 bytes. Did you mean this with selecting "Read OOB area"?

[DavidAlfa]

Yes, my bad. Spare area, OOB(Out Of Band) area, ECC area...it's the same thing.
That area is reserved for ECC when enabled, or can be used as regular storage when disabled.
Anyways, remember it's a writable filesystem, two backups will probably be different if the scope was started between them.

[vistorik]

Yes, my bad. Spare area, OOB(Out Of Band) area, ECC area...it's the same thing.
That area is reserved for ECC when enabled, or can be used as regular storage when disabled.
Anyways, remember it's a writable filesystem, two backups will probably be different if the scope was started between them.

In continuation of testing of your dsoflash-tool I've got two files of 132MB from the same flash memory, one using CH341A programmer with "Read/Write with spare area" option and the second one, using your tool from my scope. The both files are identical. So it works.

In my opinion it would not be superfluous to mention in the tool-Readme.txt file about this Spare area, OOB(Out Of Band) area, ECC area, to explain why the tool reads 132MB from flash memory but not 128MB and to say about the "Read/Write with spare area" options of the NeoProgrammer.

[sagias]

hello im giorgos from greece today arrived my hantek and hacked to 2d15 works ok
just a question how i can apply overclock
download the file from gdrive but how im going to alter the multipliers

[DavidAlfa]

Isn't clear in the readme?

The system will execute the file overclock/oc_manager from the usb drive. If there's no drive, overclock won't be applied.
Now you can try rising the CPU_OC_PLL_N multiplier in that file, rebooting each time to apply the new settings.

overclock/oc_manager

Code: [Select]

# Effective frequency is (BASE_FREQ*(N+1)*(K+1))/(M+1)
# Possible values:
# PLL_N: 1-31
# PLL_K: 0-3
# PLL_M: 0-3

# ------ USER ADJUSTABLE ------
# This example: (24*(26+1)*(0+1))/(0+1) = 648MHz
CPU_OC_PLL_N=26
CPU_OC_PLL_K=0
CPU_OC_PLL_M=0

# This example: (24*(13+1)*(0+1))/(0+1) = 336MHz (168MHz DDR)
MEM_OC_PLL_N=13
MEM_OC_PLL_K=0
MEM_OC_PLL_M=0


[sagias]

done thank you oc to 668 336 mhz

[Ben7128]

Don't buy this crap. Seriously.
I've trying to follow 50MHz signal pulses lasting 50-100us.
Larger Time division than 10-5us shows nothing or garbage.
In the zoom view, you can't scroll through the whole screen, so another useless feature.
SPI decoding keeps failing. No *** way to decode anything.

*** Hantek designers. Never again. Not even free, damn.
This scope is just crap.

DavidAlfa, thanks for all your work on this. I recently retired and was thinking about getting 2D10 after using Hantek DSO8060 at work for many years; but now turned off due to the bug reports here and elsewhere. What's your recommendation for a reliable and accurate budget 50/100MHz scope? I mainly work on audio gear and general electronic design/repair work but have some STM32F4DISCOVERY and other micro-controller projects.
https://www.eevblog.com/forum/testgear/new-hantek-dso2x1x-models/875/

[DavidAlfa]

It was a quite good offer at the release price, around $200, but now people seems to get it for almost $400!

What can I recommend having a 2D15 myself? All I can say is get anything but Hantek!

I'd make a new thread asking for options given a price tag, for sure you'll get a lot of nice feedback from experienced people.

[Ben7128]

It was a quite good offer at the release price, around $200, but now people seems to get it for almost $400!

What can I recommend having a 2D15 myself? All I can say is get anything but Hantek!

I'd make a new thread asking for options given a price tag, for sure you'll get a lot of nice feedback from experienced people.

Thx for the reply. 2D10 still available for AUS $320 (US $240) in Australia, but I want something trustworthy.  I will take up your suggestion to start a new thread.

[DavidAlfa]

I just got a very weird crash that managed to brick the scope!
Changed the memory depth to 4M, after few single shots the scope suddenly crashed with the plot area full of random green lines.
On reboot, all it showed was an empty grid, completely unresponsive to any button press.
Had to use platform-tools to recover it!

[DavidAlfa]

Edit: I zoomed until finding the smallest step. What Hantek is showing on screen is definitely BS.

Testing with 50MHz SPI signal in 1us/div
- 4Kpt shows 250MSa/s, waveform is acceptable, missing details due the lower sample rate, resolution is 4ns.
- 40Kpt shows 500MSa/s, shows garbage. Resolution is 8ns, that's 125MSA/s, bast***s! 
- 4Mpt shows 125MSa/s and is 8ns, looking as expected due sampling limitations.

- Testing with 1.56MHZ SPI signal, 50us/div
- 4Kpt shows 5MSa/s, waveform lacks a lot of details but is never missing clocks, resolution is 200ns.
- 40Kpt shows 50MSa/s, but clearly isn't. Resolution is mostly 400ns. That's 2.5MSA/s.
   Strangely, sometimes there're 200ns steps. What the hell is going on?
- 4Mpt shows 125MSa/s. Should I laught? Resolution is 400ns. That's 2.5MSA/s, liars! Looks almost the same as 40Kpt,

[pcprogrammer]

Don't know the Gowin FPGA, but having done some playing around with the Anlogic FPGA in the FNIRSI-1013D I think running sampling at 500MSa/s is not that simple. Also the phenomena mentioned by others that when selecting more sample memory things get messed up might be a problem with pushing the FPGA to its limits.

It might be interesting to do measurements on the ADC clock to see if it is actually sampling as fast as they say it is. Maybe whenever I feel like doing it, I will hook up my DL9705. Still only a bandwidth of 500MHz so no square wave, but it should be able to show if it is running at 500MHz clock.