FNIRSI-1013D "100MHz" tablet oscilloscope

[pcprogrammer]

@Atlan,

I answer your questions here, as others might find it useful too.

How does FEL mode actually work?

FEL is something of Allwinner technology and build in to (some of the) the MCU's they make. When there is no boot program found on either the SD card or FLASH memory it starts this FEL code. This opens an USB connection with a host and allows for all kind of actions. See https://linux-sunxi.org/Sunxi-tools#sunxi-fel

does it have any advantage over uploading to an sd card?

It certainly does, because with the scope it becomes possible to load the compiled firmware directly to the memory and execute it. This way there is no wear of the SD card, and it allows for a bit faster testing of new development.

The default FEL mode does not allow loading to the DRAM because it does not enable it, but with the starting FEL via the bootloader the DRAM is enabled and therefore allows loading to it.

does it allow reset via PC?

No, you will have to power cycle the scope to reenter the FEL mode. You could add a menu option to the main menu though, that does this for you and with that avoid power cycling.

is a specially compiled firmware required?

No, the binary firmware as it rolls out of the .elf file can be uploaded and executed. For this you need the sunxi-fel tool found here. This one is build by me from the sources I found on the net and is intended for Linux only. Don't forget to set it executable.

I have used this during the development of both the 1013D and the 1014D firmwares. Very easy. Check the instructions in this file.

I load the binary to a non existing address 32 bytes (size of the BROM header) before the start of the memory to have it in the right location for the vectors. The original firmware binary does not need this.

[Atlan]

Why does it work when I have the number 220, but not if I use "variable"?

[pcprogrammer]

Why does it work when I have the number 220, but not if I use "variable"?

That is a very good question to which I have no answer.

Are you sure the variable is set to a correct value?

[Atlan]

Just changing the boot screen.
The time to select the option to start the firmware is approx. 2s, if you touch anywhere on the screen other than the buttons, the reaction time is extended to 9s. Added the remaining time counter.

[engineer.r152]

Good afternoon! In the latest version 0.025c, the screen was displayed with an offset, position correction via Loader LCD/TP Config does not help, it is in an offset position. This problem has been solved before!

[Atlan]

Also the boot screen is shifted?

[pcprogrammer]

When not mistaken the original boot loader loaded the configuration file for the display shift and the touch panel transform, so you probably have to add that again to the new one.

[Atlan]

v0.025d with botloader 0.02 it should be ok. After the first start it will be moved, but after the second it will be OK

[engineer.r152]

yes, the start screen is shifted, but the original firmware is not. I haven't tested the new version yet. Thanks!

[Felipe Lacerda]

v0.025d with botloader 0.02 it should be ok. After the first start it will be moved, but after the second it will be OK

Thanks a lot!
You made it very useful!

Is this difference between channels normal in AD comp?
Calibration seems ok to you?

[engineer.r152]

Now the main firmware and the alternative are displayed correctly, and the bootloader is still shifted.

[Atlan]

ADC comp is a bit high, but I saw that some people have it that way.

152. Try this, boot fw 0.03.
It initializes the display, prints error messages if they exist, loads data for image shift if they are correct, makes changes and displays the main boot menu.

[engineer.r152]

It's okay! Version.003 operates without bias in all modes.

[Atlan]

In the dc shift calibration menu, the calibration values ​​are displayed. V+ and V- enables their change.

v0.025e

[engineer.r152]

Atlan, thank you for the next firmware mod! Could you remind me which parameters are configured in countries 1, 2 and 3?
On page 1 and 3, changing the settings does not affect anything. On page 2, the signal is adjusted relative to 0.

[Atlan]

Turn on both channels of DC mod and measurement, bring a suitable signal to both inputs. (eg rectangle 1V peak to peak), press auto button. The option "dc shift calibration" is activated in the diagnostic menu. In the oscilloscope mode, the CTRL key is pressed. The menu to the right allows you to change the value of dc_shift_center, place the signal in the center. Press the next page button. The menu on the right allows you to change the value of dc_shift_size, set the size of the signal so that it graphically corresponds to the size of the signal. The menu on the right allows you to change the value of dc_shift_value, set the measured value in "measurements" to correspond to the size of the supplied signal. Press the end cal button. The values ​​are saved and the oscilloscope goes into normal mode.

I don't have time to test it, I deleted the calibration values ​​during bootloader development.  So that I would have to carry out the angular calibration, I don't have time.

[Atlan]

First, it is necessary to have the input divider calibrated.  After switching on the oscilloscope, wait 3 minutes, perform base calibration.  And after connecting the signal, press autoset, activate DC shift calibration in the menu and perform the setting.  The change of the first value will be reflected when you press autoset again!

The first value (116) should be such that after pressing autoset, the signal on the screen is placed symmetrically about the central axis.

The second value, set the position of the signal relative to the pointer to the left

Third value, set the measured value of Vmin to 0



It will require a small program modification to keep the dc shift calibration menu enabled.

[Evi]

Are we betting on the right horse?

“Better Fewer but better.”
Vladimir Lenin


“As we have evolved, we have encountered more and more problems.”
NASA Annual Report

[Evi]

1. Stop oscillating
2. Define really existing problems.
3. Separate problems reasons: SW or HW
4. Solve only one task at the moment

or stay in endless loop

[Atlan]

How about you try it out to see if it works as I described.  Do the waveforms then correspond in the DC mode.?  That would help matters.  We have known for a long time that it is a hardware failure.