CFW for KSGER/Quicko STM32 Soldering Stations

[DavidAlfa]

My station is a  Quicko, so I can't test other builds properly.
Try the release build and report the result.

It's strange to have only one user having this issue, for now I highly suspect a problem with that stm32

Is anyone running the latest ksger v2 firmware?

[anwfeb]

My station is a  Quicko, so I can't test other builds properly.
Try the release build and report the result.

It's strange to have only one user having this issue, for now I highly suspect a problem with that stm32

Is anyone running the latest ksger v2 firmware?

I don't understand what you mean do you want a all file in "SSD1306_Release" folder? if this is true I will upload it

i tested with 2 original chips with different code, i have 1 more original stm32 chip with different code batch, and result is same. do you want me to make firmware for your quicko? maybe you can try it in your quicko with built firmware from my laptop, so you can find the difference.

yes, I am using your latest code source on github. stm32_soldering_iron_controller-1.10.5.
with and without "Fix for CubeIDE 1.13.0 warning" both are the same. i have tested it.

[asupsp700]

Is anyone running the latest ksger v2 firmware?

I am using the latest ksger v2 firmware. Everything is fine. Works without glitches

[DavidAlfa]

anwfeb, I'm telling you to use the existing precompiled binary in the Releases section, here!

https://github.com/deividAlfa/stm32_soldering_iron_controller/releases

No compiling the latest source!

[anwfeb]

anwfeb, I'm telling you to use the existing precompiled binary in the Releases section, here!

https://github.com/deividAlfa/stm32_soldering_iron_controller/releases

No compiling the latest source!

that is working fine. your binary file is working, I have plans to add ID language, so I need to build the firmware independently.

I am using the latest ksger v2 firmware. Everything is fine. Works without glitches

please try my firmware on post #3574
and try menu -> SYSTEM -> boot, change SLP to RUN, and select. after it you you scroll to Btn wake, and don't change anything, don't touch encoder, and look at the screen, it will exit to the main menu by itself, and find an error.

[DavidAlfa]

Try CubeIDE 1.12.1. I have not yet tested 1.13.0.

[anwfeb]

Try CubeIDE 1.12.1. I have not yet tested 1.13.0.

on CubeIDE 1.12.1 i get warning

Code: [Select]

arm-none-eabi-gcc "../Core/Src/user_main.c" -mcpu=cortex-m3 -std=gnu11 -DUSE_HAL_DRIVER -DSTM32F101xB -DSSD1306 -c -I../Core/Inc -I../Drivers/addons -I../Drivers/CMSIS/Device/ST/STM32F1xx/Include -I../Drivers/CMSIS/Include -I../Drivers/generalIO -I../Drivers/graphics -I../Drivers/graphics/gui -I../Drivers/graphics/gui/screens -I../Drivers/graphics/u8g2 -I../Drivers/STM32F1xx_HAL_Driver/Inc -I../Drivers/STM32F1xx_HAL_Driver/Inc/Legacy -Os -ffunction-sections -fdata-sections -Wall -Wno-builtin-macro-redefined -fstack-usage -fcyclomatic-complexity -MMD -MP -MF"Core/Src/user_main.d" -MT"Core/Src/user_main.o" --specs=nano.specs -mfloat-abi=soft -mthumb -o "Core/Src/user_main.o"
../Core/Src/user_main.c: In function 'HardFault_Handler_cp':
../Core/Src/user_main.c:235:3: warning: implicit declaration of function 'HardFault_Handler'; did you mean 'HardFault_Handler_cp'? [-Wimplicit-function-declaration]
  235 |   HardFault_Handler();
      |   ^~~~~~~~~~~~~~~~~
      |   HardFault_Handler_cp
arm-none-eabi-gcc -o "STM32SolderingStation.elf" @"objects.list"   -mcpu=cortex-m3 -T"../_STM32F101CBTX_FLASH.ld" --specs=nosys.specs -Wl,-Map="STM32SolderingStation.map" -Wl,--gc-sections -static --specs=nano.specs -mfloat-abi=soft -mthumb -Wl,--start-group -lc -lm -Wl,--end-group
Finished building target: STM32SolderingStation.elf
 
arm-none-eabi-size   STM32SolderingStation.elf
arm-none-eabi-objdump -h -S  STM32SolderingStation.elf  > "STM32SolderingStation.list"
arm-none-eabi-objcopy  -O ihex  STM32SolderingStation.elf  "STM32SolderingStation.hex"
   text    data     bss     dec     hex filename
  88212     304   12332 100848   189f0 STM32SolderingStation.elf
Finished building: default.size.stdout
 
arm-none-eabi-objcopy  -O binary  STM32SolderingStation.elf  "STM32SolderingStation.bin"
Finished building: STM32SolderingStation.hex
 
Finished building: STM32SolderingStation.bin
 
Finished building: STM32SolderingStation.list
 

12:30:52 Build Finished. 0 errors, 1 warnings. (took 16s.558ms)

edit:
after several times of testing making the firmware, I found a problem when CUBE MX asked to migrate the project to a newer version version 6.9.0, when I clicked migrate it became a problem even though I was using CUBE IDE 1.12.1 finally I was able to make firmware with CUBE IDE 1.12.1 without Migrate Cube MX, I ignored the Migrate request and clicked continue. now the firmware is working fine.

if draw conclusions then the problem is in CUBEIDE v1.13.0. source code does not match the new version of CUBE IDE.

[DavidAlfa]

Stay with CubeIDE 1.12.1!

So I tried building in CubeIDE 1.13.0 again, CubeMX has changed, the build script no longer works.
Also it asks to login for downloading the firmware package... what the hell ST?
I will not spend time in this now, perhabs in a few days or weeks. Just use 1.12.1.

All I can do for now is adding a warning in the building instructions.

anfweb, just use the precompiled binary.
There's no reason to build yourself unless modifying the source code.
I simply don't want to waste my time explaining everyone how to use the IDE so they can compile the SAME thing already available in the Releases requiring no work!

[DavidAlfa]

So I compiled for my Quicko, nothing failed.
When I used the linker script for the stm32F101x (10KB RAM), then it got the Hard Fault.
I inmediately suspected of the RAM. Indeed, it uses 300 more bytes in 1.13.0.
10KB devices are really tight!
I could simply reduce the reserved heap or stack and hope for the best, but I won't risk getting an explosion of users having all kind of errors.
A possible approach will be to use a different flash sector for each profile, removing the necesity of backing up the flash before erasing, will easily free 2KB.
Until I have enough time to find a workaround, it'll be CubeIDE 1.12.1 !

Attaching the Build analyzer output. I inverted the colour, the black theme is terrible (White text over bright green  ).

[anwfeb]

A possible approach will be to use a different flash sector for each profile, removing the necesity of backing up the flash before erasing, will easily free 2KB.

I believe you have many ways to give more ram space. while waiting, I will continue to test the firmware with cubeide 1.12.1

and please add one more profile for C115, it looks like the Chinese clone has added TIP C115 in one device.

[ststefanov13]

My station is a  Quicko, so I can't test other builds properly.
Try the release build and report the result.

It's strange to have only one user having this issue, for now I highly suspect a problem with that stm32

Is anyone running the latest ksger v2 firmware?

I am using the latest ksger v2 and v3 firmware. Everything is fine. Works without glitches.

Stations Ksger V2.1s , Ksger Ve2.1s , Ksger V3 and Handskit . Everything is fine.

[DavidAlfa]

Anfweb, you don't need a new profile for that as long as you calibrate the tip.
The C115 cartridge isn't easy to interface, if I remember correctly the thermocouple is connected backwards, giving negative voltage output.
That won't work with existing stations out of the box.

[anwfeb]

Anfweb, you don't need a new profile for that as lo g as you calibrate the tip.
The C115 cartridge isn't easy to interface, if I remember correctly the thermocouple is connected backwards, giving negative voltage output.
That won't work with existing stations out of the box.

i saw in this link, TIP C210 C245 C115 PINOUT it has the same pins as C210, but i haven't tested it yet, i want to buy a TIP and Handle for C115 so i can try it.

actually the ksger V2.1S variant has the original chip code STM32F103CBU6 and that's 20KB RAM, there's still plenty of space available if the board variant has STM32F103 like Quicko and Ksger v1.5

[DavidAlfa]

Use the search engine, C115/nt115 has been discussed several times:

Apparently, this is the real pinout.
https://www.eevblog.com/forum/reviews/stm32-oled-digital-soldering-station-for-t12-handle/msg4435738/#msg4435738

Someone tried and never answered back.
https://www.eevblog.com/forum/reviews/nt115-jbc-handle-for-ksger-t12-with-cfw-by-davidalfa/

The flash is not a problem, only the ram.
As I said, currently I won't spend any time in this project else than small bug fixes, I can't spend entire afternoons in this.
If you buy the C115 tip don't expect me to sit in my desk for two days debugging problems, sorry!.

[anwfeb]

Use the search engine, C115/nt115 has been discussed several times:

Apparently, this is the real pinout.
https://www.eevblog.com/forum/reviews/stm32-oled-digital-soldering-station-for-t12-handle/msg4435738/#msg4435738

Someone tried and never answered back.
https://www.eevblog.com/forum/reviews/nt115-jbc-handle-for-ksger-t12-with-cfw-by-davidalfa/

The flash is not a problem, only the ram.
As I said, currently I won't spend any time in this project else than small bug fixes, I can't spend entire afternoons in this.
If you buy the C115 tip don't expect me to sit in my desk for two days debugging problems, sorry!.

Code: [Select]

The flash is not a problem, only the ram.yes, STM32F103C8/CB has 20KB of RAM

there is someone who has tested and i saved the video. I want to buy a TIP because I want to test, regardless of firmware source code available or not, because some configuration is already prepared in your firmware such as input resistance "Heater" for example 5 Ohm and I can try with a lower power supply of 9-12V. So don't worry, if you haven't had the time, I'm not going to rush you to make the firmware right away, I want to test with all the existing source code with whatever electronics skills I have.

Video: TIP Solder C115

[DavidAlfa]

  He's just applying raw power to the tip, there's no temperature control.
The problem comes wth the sensing voltage, whcih is I'm not mistaken, it's negative, that won't work with the current T12 stations without hardware modifications.

[anwfeb]

He's just applying raw power to the tip, there's no temperature control.
The problem comes wth the sensing voltage, whcih is I'm not mistaken, it's negative, that won't work with the current T12 stations without hardware modifications.

do you mean modify the hardware by cutting the R11 line? i will test the thermocouple with the COM pin and (-) or with (+) before i will remove the mosfet, it works only to read the temperature on the unheated TIP, it only reads through the OP AMPS, if that works, then i will start by installing the mosfet with low voltage, because C115 does not support 24V. I will try it.

this is just a test, how can the OP AMPS read the thermocouple, reverse the thermocouple PIN by testing all the PINs on the TIP to find the appropriate PINOUT location, so the mosfet is then installed, and put with low voltage.

succeed or fail, there's no harm in trying

[DavidAlfa]

No!
I repeat for 10th time, the C115 sensor output gives negative voltage, contrary to the positive voltage outputted by T12, T210, T245, T470.
You can't magically measure negative voltage without heavy hardware modification, not just cutting a trace.
This could be wrong, as I haven't tested myself.
You don't need to remove the mosfet, simply enable and enter the Debug screen, setting the setpoint to 0ºC.
Or disabling the button and shake inputs from waking up the station in the settings, thus once you set the station into sleep nothing will turn the output on.
There're really small C210 tips, you can use these with zero modifications, connection is the same as T12.

[anwfeb]

hi DavidAlfa,
i have tried to change .ioc cubeMX STM32F101 to STM32F103 for Ksger V2, all went well. however I encountered a problem when I changed the clock to 64MHz. but it works fine at 40-48MHz.

I have read:

Code: [Select]

#if !defined DISPLAY_DEVICE  || (defined DISPLAY_DEVICE && defined I2C_TRY_HW)
// This delay is calibrated for 36MHz (Ksger v2 software i2c). If increasing the cpu frequency, also increase the nop count
__attribute__ ((noinline)) void bit_delay(void){                  // Intended no inline to add further delay
  asm("nop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop");
}

but i don't understand how to increase the nop. when I add more nop code it still doesn't work on my i2c display, even though I try to add a pullup resistor.

My goal of adding the STM32F103 device is so that I have 20KB of RAM working, because I'm using a 103 and the original chip embedded in my ksger board is also a 103.

[DavidAlfa]

Simply duplicate the line containing the NOPs.
Then open the project properties, go to build settings, linker settings and change the linker script name from the 101 to 103.
That will be all.

[anwfeb]

linker settings have been changed previously,

i did this and it didn't work:

Code: [Select]

#if !defined DISPLAY_DEVICE  || (defined DISPLAY_DEVICE && defined I2C_TRY_HW)
// This delay is calibrated for 36MHz (Ksger v2 software i2c). If increasing the cpu frequency, also increase the nop count
__attribute__ ((noinline)) void bit_delay(void){                  // Intended no inline to add further delay
  asm("nop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop");
}

and try this code not working:

Code: [Select]

#if !defined DISPLAY_DEVICE  || (defined DISPLAY_DEVICE && defined I2C_TRY_HW)
// This delay is calibrated for 36MHz (Ksger v2 software i2c). If increasing the cpu frequency, also increase the nop count
__attribute__ ((noinline)) void bit_delay(void){                  // Intended no inline to add further delay
  asm("nop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop");
  asm("nop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop\nnop");
}


[DavidAlfa]

Then add a 3rd or 4th copy.
But increasing the RAM or the cpu speed won't change anything in the firmware, it's a waste of time.

[anwfeb]

Then add a 3rd or 4th copy.

didn't work until the 5th copy

yes, I just want to try. if that's possible using 64MHz, if not doable I'd stick with 36MHz it works fine.

[DavidAlfa]

Let me be clear about this, open source means you can use and modify the code as you wish, but that doesn't include free support.
Please understand that I'm not gonna sit an entire afternoon with the debugger, the scope and the code, to find out why it's not working, just because you updated the mcu to a 103 and want 64MHz for no reason beyond your personal whim.

What you're doing will mean absolutely nothing, 36MHz can already do many times faster than you could see, it's the tip response time and pwm noise what keep the limits at 5-20Hz.
All I see is a lost cause.

[anwfeb]

Sorry actually i'm not asking for free support for my own sake, i'm just trying anything to change my board ksger with your firmware, starting from HARD FAULT when i use cube ide 1.13.0, and here i found the answer that the cube ide 1.13.0 version takes too much a lot of RAM, actually this is not that important, but because i saw that my ksger was STM32F103 and had 20KB of RAM, then i tried to change the ioc cube mx to 103, for clock problems there was actually no problem using 36MHz, but because i saw that the F103 was capable of up to 64MHz then i just wanted to try. If can't then i'm not asking you to debugger all day. i will use the firmware as it is without changes.

Thank you, so far i have been able to use your firmware and it has been great.