-
#1775 Reply
Posted by
DavidAlfa
on 24 Jan, 2024 17:42
-
It's just cutting 2mm of the shaft plastic... Why the overthinking?
-
#1776 Reply
Posted by
Aldo22
on 24 Jan, 2024 18:21
-
It's just cutting 2mm of the shaft plastic... Why the overthinking?
I'm not in a hurry.
The part is not broken, it works "as designed" by Hantek.
If I can find the exact part for the same shipping costs, then I'd rather take that than cutting and filing a PEC12R-4225.
But it's good to know that this option exists.
I'll look around a bit and then probably order the PEC12R-4025F-S0024.
Btw. is there anything to consider when removing the front cover?
There is only one ribbon cable, right?
Thank you.
-
#1777 Reply
Posted by
habl
on 29 Jan, 2024 15:21
-
Hey, everybody.
At some point in the firmware update, my oscilloscope calibration function broke. I have read that this happens due to incorrect FPGA firmware. I have a full dump of the oscilloscope from the time of purchase, I rebuilt it - it shows firmware 220210.00 and Firmware 3204, but when I try to calibrate I get error 511d6. Tried different FPGA versions - some have error 511d1, some have error 511d6.
Help me to understand, please.
-
#1778 Reply
Posted by
DavidAlfa
on 29 Jan, 2024 18:58
-
Try latest 320x or 3102, must be one of them
-
#1779 Reply
Posted by
habl
on 31 Jan, 2024 03:06
-
@DavidAlfa I found out that the latest version of dsoflash at the moment from your google disk does not allow to restore correctly my backup, taken by dsoflash a year and a half ago. I have recorded and read with the old version of dsoflash - the results are identical and recorded and read with the google disk version several times and the files are different. I uploaded the files to a temporary folder on google drive so you can analyze them.
-
#1780 Reply
Posted by
DavidAlfa
on 31 Jan, 2024 04:42
-
With no booting between writing and reading? Otherwise that will change the filesystem.
Except small areas in the UBI partitions, all the flash is identical.
(Edit: Because it wasn't erasing the flash due a bug in the code!).
-
#1781 Reply
Posted by
habl
on 31 Jan, 2024 04:59
-
Yes, without booting between writing and reading.
-
#1782 Reply
Posted by
DavidAlfa
on 31 Jan, 2024 05:12
-
The UBI partitions CRC is different, this can only be done by the scope itself, DSOFlash can't do this.
If data was bad, like wrogn CRC written, the scope wouldn't boot or work properly!
-
#1783 Reply
Posted by
habl
on 31 Jan, 2024 05:19
-
That's the thing, if I write the dump with the old version of dsoflash, the oscilloscope loads, but if I write the dump with the your google disk, the oscilloscope does not load - hangs on the second splash screen. I've checked several times.
-
#1784 Reply
Posted by
habl
on 31 Jan, 2024 05:24
-
Looks like my nand memory chip is dead. Because now it is not detected by dsoflash.
-
#1785 Reply
Posted by
DavidAlfa
on 31 Jan, 2024 07:49
-
Strange, but will check it later. Writing to the flash can't kill it in such way!
-
#1786 Reply
Posted by
habl
on 31 Jan, 2024 09:52
-
I soldered out the flash drive, it reads in the programmer. The problem is not the flash drive, but a badly soldered contact near the flash drive. Thanks for the advice!
But the problem with dsoflash from your google disk remains. The images filled with it do not help to boot the oscilloscope.
-
#1787 Reply
Posted by
DavidAlfa
on 31 Jan, 2024 15:52
-
That's the thing, if I write the dump with the old version of dsoflash, the oscilloscope loads, but if I write the dump with the your google disk, the oscilloscope does not load - hangs on the second splash screen. I've checked several times.
Sill very rare than the first ~15MB are perfect and then only few bytes of the UBI header change. Will check it, but it's extremely slow and time consuming so I can't tell when it'll be ready.
Actually you'll never really need DSOflash to restore anything, just run lastest platform-tools, then restore fw version, model/serial and AWG calibration from Backup Builder.
Edit:
Read my flash, wrote back, read again, exactly the same md5sum.
-
#1788 Reply
Posted by
DavidAlfa
on 31 Jan, 2024 23:12
-
@DavidAlfa I found out that the latest version of dsoflash at the moment from your google disk does not allow to restore correctly my backup
Well, indeed, your restore.bin didn't match the original when reading back.
Will make more tests...
-
#1789 Reply
Posted by
DavidAlfa
on 01 Feb, 2024 22:44
-
I was able to initialize the SDRAM (Not easy, zero documentation available for FEL
), it's much larger than the SRAM scratchpad (64MB vs ~4KB), so it will allow much longer queues, removing a lot of overhead.
It made a dramatic speedup to 300KB/s, it only takes 7 minutes now!
-
#1790 Reply
Posted by
habl
on 02 Feb, 2024 07:30
-
Yes, dsoflash is now correctly restoring the image on my oscilloscope. Thank you very much.
-
#1791 Reply
Posted by
DavidAlfa
on 02 Feb, 2024 18:44
-
Great! Thanks for reporting.
I added the final touches, including md5 generation and verification, updated the files in Google Drive.
-
#1792 Reply
Posted by
vistorik
on 06 Feb, 2024 21:52
-
-
#1793 Reply
Posted by
DavidAlfa
on 06 Feb, 2024 22:16
-
-
#1794 Reply
Posted by
vistorik
on 06 Feb, 2024 22:47
-
-
#1795 Reply
Posted by
vistorik
on 08 Feb, 2024 01:18
-
-
#1796 Reply
Posted by
DavidAlfa
on 08 Feb, 2024 04:43
-
If the memory wasn't detected properly It would complain showing unsupported memory error.
The scope rebooting is a completely new behavior that I've never seen before!
Are you using latest version? Or did you copy dsoflash to the 4pda thread? Because it's been updated recently to correct a bug not restoring the firmware properly, also got much faster.
I noticed a unnecessary step in the .bat files, since now dsoflash sets the usb to high-speed mode by itself, sunxi-fel is no longer required.
Updated the files. But this would not cause any rebooting, just set the usb mode twice.
-
#1797 Reply
Posted by
phmarek
on 15 Feb, 2024 19:42
-
Hi!
I reactivated my 2D15 after a few months of non-use, and noticed that there was no waveform display - which I thought was due to wrong trigger settings.
Then I noticed that only a few keypresses work (Trig Menu, or CH1 Menu, a few others), but then the UI freezes.
As the USB and SSH console still worked, I could reboot without my local modifications -- same problem. Even dropping the CPU overclocking doesn't help.
Remove the USB cable doesn't help.
dmesg or the serial console have no (error) output when pressing the keys; the UI freezes, but the console keeps working, and the phoenix binary looks like it's running normally, using 98% CPU like always.
As (some) keys work once or twice after reboot, I don't think it's hardware-related - can anyone offer ideas what is wrong?
Thanks!
-
#1798 Reply
Posted by
DavidAlfa
on 15 Feb, 2024 22:59
-
Press load defaults. 98% CPU is normal in this device.
-
#1799 Reply
Posted by
Aldo22
on 16 Feb, 2024 09:47
-
As (some) keys work once or twice after reboot, I don't think it's hardware-related - can anyone offer ideas what is wrong?
I don't know if this is helpful, but my unit (overclocked) only crashes in the first few minutes after startup, if at all.
It depends a bit on what is turned on from the last session.
Then when I turn it back on, it runs for hours without interruption.
It looks like it might be "too cold" when it first starts up.
Sounds strange, but I can't explain it any other way.
Maybe it has something to do with the new PSU?
It's best not to do anything to it for 2 minutes after starting, at least with my device. Then it runs reliably.
But that probably has nothing to do with your case.