CFW for KSGER/Quicko STM32 Soldering Stations

[elektryk]

Nope, they're completely opposed? How is ERROR=0 in AC6 when it's the same ST libaries?

Yeah these enums are opposite, they've made a real joke on me.

[DavidAlfa]

In the tip editor I can only select the active tip for editing.
Please correct the information about the firmware version.

Bad version number - Yeah happens when making 25 builds in a day!

Tip editing - Ah, indeed! Happens if doing it from the tip list menu - FIXED.

Use this for now until next build is out (This is what I was using all the time):
1 - Highlight the tip in the main screen.
2 - Long-click to enter tip editor.

Also, based on your video, remember you can return to the main screen any time by long-clicking (While not editing any field).
In the tip editor this will discard any changes, but you can do it inmediately after clicking on SAVE/CANCEL/BACK, removing the need of navigating to BACK / EXIT in each screen.

In fact I was thinking in completely removing that screen, adding a NEW option below COPY.
- COPY->copies the current name to the new tip.
- NEW->clears the name.

[asupsp700]

In the tip editor I can only select the active tip for editing.
Please correct the information about the firmware version.

Bad bersion number - Yeah happens when makign 25 builds in a day!

 
Here's my test... (Attachment Link)

[DavidAlfa]

I think the people having those issues have a 64-pin STM32, but not R8 or RB... maybe RC, RD, RE? (Or relabeled by some chinese factory)
This will cause all these issues!
So, as I'm pretty sure the problem is coming from the flash page size being different, I made some changes to automatically detect and handle this, also few other fixes and enhancements.

Please try v1.12.3.

And if something is not working, remember I'm not a fortune teller, so please explain properly the issue and how to cause it.

[asupsp700]

Thanks David,  версия 1.12.3 Oled2. Everything works for me! I'll test it.

[DavidAlfa]

So only one person ran the flash test and tested the new FW version, reporting the results...

[ststefanov13]

Everything as functionalities works.
Two cosmetic bugs remain. The motion sensor icon slightly pushes the tip name when the last letter is in bulk, for example O, R or P (TIP name DESOLDER) . The SW and HW version information is not aligned equally.
David, now can the number of soldering tips be raised to 50 without becoming a problem?

[DavidAlfa]

Do you really have 50 tips? They won't fit right now, but I have some ideas for the future.
(Not the current priority)
What I really need to know if it works now on controllers that failed before.

Hmm I tested what you say when designing it, filling the tip name with ”F" looked fine, maybe those letters are a bit larger.

[ststefanov13]

For me, the problem with the Line 964 settings.c error disappeared even with version 1.12.2, the MCU with the error was STM32F103CBT6  [CPU ID (DBGMCU_ID) = 20036410], there are no problems with Handskit with MCU CKS32F103CBT6.

[oleg401169]

Of course, I have fewer stings in use - 11, I immediately add them to the list of station stings (in the photo there are those in the holder and two on the stations). Now I’m using firmware version 1.12.3 - everything works fine, there is no hard vault screen, which popped up for me on the 11th series of firmware (in some versions 1.11.*), purely subjectively, the response of the station when exiting sleep mode improved very much, calibration data is entered on all the tips that I use (in some versions 1.11.* the hard vault screen and the station popped up reset to factory settings (for me this was not particularly critical, so I did not inform you). There is not enough backup of the station settings, but I understand perfectly well that this is probably impossible to implement. I continue testing the firmware, if problems arise, I will inform you.

[anwfeb]

I saw a difference in the EDIT TIPS menu, then i selected TIPS T12-BC3, i saw that the value in Cal 250 was not the same as the value in CALIBRATION => SETTINGS.

for example, i calibrate manually, via the CALIBRATION menu => SETTINGS, then i change the value in Cal250 -> 1800, then save it. the result of this calibration change should be = the Cal250 value in TIPS, then what is the purpose of doing calibration and saving the calibration results if in the end EDIT TIPS menu displays the default numbers "1900" instead of the values ​​resulting from the calibration that has been saved?

[DavidAlfa]

No. The value in settings is for the automated calibration procedure, so it doesn't drift too much.
Read the manual!

Also, you're all ignoring me completely, I've said several times that the priority now is to know if the problems with certain controllers  (flash error, hard fault, black screen) is gone, also that I needed reports of the flash size check program posted few messages back...
So far it's been like talking to rocks, if it keeps like this I will drop the development, enjoy the same zero effort back...  .

[asupsp700]

No. The value in settings is for the automated calibration procedure, so it doesn't drift too much.
Read the manual!

Also, you're all ignoring me completely, I've said several times that the priority now is to know if the problems with certain controllers  (flash error, hard fault, black screen) is gone, also that I needed reports of the flash size check program posted few messages back...
So far it's been like talking to rocks, if it keeps like this I will drop the development, enjoy the same zero effort back...  .

People! Let's be more active! David is doing a great job! And absolutely free! IT'S NOT DIFFICULT to download and test the firmware for a few hours....

[anwfeb]

People! Let's be more active! David is doing a great job! And absolutely free! IT'S NOT DIFFICULT to download and test the firmware for a few hours....

yes I'm testing it, I just opened github today there's a big update, I'm going to test some functions so I can report if I encounter errors.

[DavidAlfa]

It's simple!
1 - Fix these problems making some controllers useless.
2 - Once  working, focus on the little things and features.

[AndyC]

I have tested v1.11.2 and v1.12.3  + the test BIN on two different controllers:

KSGER v2:
Size: 128
Page: 1
STM32F103CBT6, came with the board, unsure about if it's genuine or not.
v1.12.3 seems to work on this controller.

KSGER v3:
Size: 65535
Page: 2
Clone test: 20036410
STM32F103RBT6, came with the board, unsure about if it's genuine or not.
v1.12.3 gives "Flash error"
v1.11.2 seems to work on this controller.

EDIT:
Found another KSGER v3 board:
Size: 64
Page: 1
Clone test: 20036410
STM32F103R8T6, IIRC, this one is from Digikey.
v1.12.3 seems to work on this controller.

[DavidAlfa]

Thanks AndyC, that's exactly what I wanted.
That flash size report is definitely weird, I'll need some additional  testing in that controller.

[sst311212]

BOARD: KSGER GX V2.1S
MCU: STM32F103C8T6
Size: 64
Page: 1
BOARD: Quicko F103 (DXCHEMI T12)
MCU: CKS32F103CBT6
Size: 128
Page: 1
No Hard Fault with 18b4558 git version. (Customed Build)

[DavidAlfa]

What I really need is the flash test result for those controllers that failed completely before, or failing now in 1.12.3 like AndryC's.

If working in any version, like mine (Genuine stm32), that's the expected, no need to report anything.

Definitely there's some flash memory issue going on.

[DavidAlfa]

OK, so next try will be to assume all controllers have 1KB sectors *unless* they report 256K - 1M flash size (Assume it's genuine, follow datasheet values).

Anyone having a hard fault / black screen / checksum error bootloop in older releases, please try this new version and report if working now.
AndyC your clone/fake should be working now?

Release v1.12.4-beta.


Once working, I'll proceed with further enhancements.
My next target is to isolate each profile tips in its own 2KB block, this will raise each profile tip limit to ~80 (Probably going to limit this to something lower like 50?), while reducing the storing function heap usage by 50%, as it'll only need to backup the block being modified.
The tip block will be isolated from the profile and system settings, so if these change and the system is restored to defaults, the tips should always remain, the tip structure has not changed for 2 years at least.
Also a write-protect bit to prevent tips from being wiped when performing a factory reset, or even setting the Options bits to protect the flash areas in hardware, not even the programmer should be able to erase these areas accidentally.
(But still could if going to Option bits and manually resetting the protected pages)
But this won't make any sense unless the flash issue is completely fixed.

[AndyC]

AndyC your clone/fake should be working now?

Yes, at least the basic stuff worked in the short time I had for testing the previously non-working KSGER v3.
Saving settings, adding/editing tips, calibration etc seemed to work without getting "flash error".
Will do some more testing later.
Many thanks!

[ppptran]

Is this the regulator?
(Attachment Link)

When I find the time I'll try to figure out how hot it gets.

In my personal opinion, replace  that LM7805 / 78M05 as soon as possible . It does get hot, i mean pretty damnn hot.

I got burned one of the chinese / custom board ( 2in screen using nuvoton NUC029LAN) with the cheap 78M05 . And now the MCU is dead, i resolder the MCU and replace the 78M05, no more shorted. But i dont have the firmware or the chip. I am still tickering myseft should i order the chip, it's not that expensive 50 cents each.

Also  the voltage LDO for droping to 3.3V, it's just a complete mess / dissaster

Basically, the desgin just drop 24V -> to 5V , and then from 5V -> to 3.3V

[DavidAlfa]

v1.13.0-beta

[ststefanov13]

v1.13.0-beta

Everything as functionalities works.

[Hunghanh2009]

Hi David! Once again the hardfault error appears, PC Code:08004D9C in version KSGER_v3_OLED v1.13.0-beta. It appears after the new setup screen. then I turned the power off and turned it back on and a message appeared "BAD CHECKSUM! RESTORING THE TIPS"