CFW for KSGER/Quicko STM32 Soldering Stations

[DavidAlfa]

Yep I might have forget that in the last changes.
Just replace with +50 values (Cal at 250, 350, 450)
Will look at that later.

Edit: Fixed the debug screen (Tested, uses +1.5KB RAM)

[AndyC]

Well, we should enable the pullup by default then
Oh, were you talking about the the JCD version?
Yes, it's a bit stupid to put resistors when you can use the internal pullups/puldowns and save some cents.

I was referring to  the KSGER 2.1s having no pullup resistor. I got it from here: https://www.banggood.com/V2_1S-T12-Digital-Temperature-Controller-Soldering-Station-Electric-Soldering-Iron-Tips-T12-K-907-Handle-p-1338117.html

Debug screen is actually divided in 2,
The first shows internal pid values, ADC reading, pwm output.
And the second just lets you set the setpoint, but instead based on temperature, it's the desired ADC value. That's how it works internally.
Not really useful for most people and uses plenty of space, it will not fit in 10KB RAM devices.
If you want to enable it just go to screen.h, and uncomment the line that has something like:
"//#define ENABLE_DEBUGMENU"
I don't remember the exact name but it's seen clearly.

I tried a build with the debug screen, but it looked like the attached photo.
The good news was that we can add the JCD station to the compatibility list.
Worked straight away with a '2.1,STM32F101R8' build.
Had to adjust the settings quite a bit though. ADC delay 20->30. Still not 100% with the PID settings, but usable.

[DavidAlfa]

That's true, I didn't remember that.
As you see, I just quickly fixed the errors, but didn't flash it.
I abandoned the debug screen when switched to u8g2 library, as there was nothing relevant there and it was taking a lot of work.
I will check that on the weekend.

Remember that new iron tips are usually unstable when new and will jump like crazy. Let it run for 30 minutes or so.
It's strange that it requires that much time, usually 5mS is enough.
The genuine hakko tips had a lot of delay, that's why the final value was increased to 20mS.
Are you using EMA filtering and factor 2?
And I repeat that the little overshoot doesn't appear at the tip.

[darkfox]

Hello everyone, i want to buy stm32 soldering station and install CFW in it I have some questions. As far as i checked Chinese sellers there are 4 pin, 5 pin and 6 pins versions. Which one do you suggest? What are the pros and cons of them? I am also interested in with using C245 tips as a future option.

Thanks!

[DavidAlfa]

What you need is a supported board, check the GitHub readme.
As far as I know all the supported boards use 5 pin connectors. Most will work with JBC tips.
C210 will work without board modifications (TC is also in series with the heater, as T12s).
C245 needs with minor mods, as the TC has it's own pin.

[darkfox]

I will follow the compatibility part on github. I was wondering for 6 pin model after i saw this video on youtube. https://youtu.be/XDQdpYJrET0

Thanks!

[DavidAlfa]

I will follow the compatibility part on github. I was wondering for 6 pin model after i saw this video on youtube. https://youtu.be/XDQdpYJrET0

Thanks!

Ouch sorry, that was a slip. Yes, Quicko has 6 pin, while KSGER 5.
That's my station indeed 

All are similar in the real world.
But I prefer Quicko due better power supply design, ok-ish quality analog parts (amplifier, analog frontend). Overall, better design.
Also the display is SPI and it's connected to the dedicated hardware pins in the STM32, making it much, much faster, while barely needing CPU power.
KSGER v3.1 also has SPI display connected to the hardware pins.
KSGER v2.1 uses I2C (much slower), and not connected to dedicated hardware (even more slower and taking a lot of CPU power).
Ksgers in general tried to save as much as possible in parts, some people have stability issues with theirs.
But also there a lot of KSGER users happy.
If the price is close, I'd go for the Quicko. But let others share their experience.

I tried a build with the debug screen, but it looked like the attached photo.

The debug screens are fixed! The space is really tight there.

I found something in the internal PID settings: "PID.minI = -50;"
I think, since the heater can't cool down actively, there's no sense in setting a negative factor.
I set that value to 0 and now it reacts much faster when the temperature falls down.
Both tips (BC3,D24) are doing great with these PID values: 52-25-14 (I set this by default now).

The PID values are stored in the profile data upon initialization, so flashing the firmware won't change them if the profile already exists.
In that case, the profile must be resetted form the settings menu.

[darkfox]

I will go for spi I found another T12 soldering iron which has F103CBT6 mcu, It looks very interesting seems 2.1s style and i2c.

I texted Quiko store for MCU model, they said they are using F101C8T6, on customer reviews there are photos with F103CBT6  Seems it is related to luck.

F101C8T6 covers the requirements of 10KB RAM and 64KB flash with 36 Mhz. Do you think its good to go?

[DavidAlfa]

Don't worry about the MCU. In the worst case, you can get a 103 for little and replace it.
That board doesn't look too nice, more like done on the cheap.
Yes, pretty similar to the KSGER v2.1. I like it has the bridge for the JBC.
I do that bridge at the handle connector, so I can just swap the handle on the fly.

I don't have a build for the Quicko with F101, but won't be a problem, it just needs a new profile.
Ask for board pictures to ensure they didn't change everything.

[tommi-muc]

Hi guys,

after about 3 month I received my station: It is a T12-955 from Queckoo which I bought in the Quicko store. Is there an easy possibility for the firmware upgrade? I can reverse-engineer the pin-out if needed.

CPU is a STM32 F103CBT6 48 Pin. Edit: Just seen that it has 128kB of flash.

Attached 2 pictures and here the Ali link: https://de.aliexpress.com/item/1005001608493682.html?spm=a2g0s.9042311.0.0.63264c4dtXm7T0


Cheers
Tommi

[DavidAlfa]

Edit: It's similar to the KSGER v2.1 but seems to have some differences?

The display, encoder, PWM seem to be the same as the KSGER v2.1, but there are pins with unclear connections.

I didn't find anything about the CEIAH IC.
Seems  to be an independent voltage regulator for the analog? Or at least it sems to power only the op-amp and the stm32 VDDA.
I don't see the NTC anywhere, it's probaby in the handle like the KSGER. Or it might use the internal stm32 sensor.

I see the earth (PE, R25) is joined to the internal ground with a solder bridge. At least in mine, this causes a big reading offset > 70ºC.
I left the original 100K resistor there.

As a reference, check the v2.1 64 pin schematic.
https://github.com/dreamcat4/t12-t245-controllers-docs/blob/master/controllers/stm32-t12-oled/v2.1s-and-Ve2.1S/KSGER%20STM32%20Ver2_1S%20schematic%20rev%202_floobydust.pdf

Don't compare the pin numbers, but the pin names.
Ex.  PA2 is pin 12 in 48-pin package or pin 16 in 64-pin. If the same signal go to to PA2 in both, the firmware would work right away.

[AndyC]

The debug screens are fixed! The space is really tight there.

Cool!

Meanwhile, I've been testing the JCD station and it behaves odd with the standard settings:

And yes, that temperature reaches the tip.

Also, I get this when shaking the handle:

This made me look into the TC wiring and opamp. It's different from KSGER stations. Tried to draw something in EasyEDA and came up with this:


Is this a feature or a bug? Time for some hardware mods?

[DavidAlfa]

I was talking about a single overshoot.
But yours oscillate a lot!
If it was a small 1 second overshoot it wouldn't actually appear there.
Did you try with the last PID settings I wrote in the last posts?
When it oscillates like that, decrease "I" factor in steps of 10.
I guess that doesn't happen in the Ksger?

Let me check it more thorougly, but the amp schematic seems the same.

The second picture is not bad, that variation is just 1-3 °C.
But the first one is way too much.

Are you using the last fw? And resetted the tip profile?

[AndyC]

I was talking about a single overshoot.
But yours oscillate a lot!

Yes, indeed!

If it was a small 1 second overshoot it wouldn't actually appear there.
Did you try with the last PID settings I wrote in the last posts?

Yes, it oscillates even more with the new settings. Erased the flash and used the latest version.

I guess that doesn't happen in the Ksger?

No, the KSGER works fine.

Let me check it more thorougly, but the amp schematic seems the same.

The notable difference is that the JCD version has a capacitor on the opamp output and the led cathode is connected to the inverting input(?)

The second picture is not bad, that variation is just 1-3 °C.

This is with ADCdelay set to 33mS and the PID 'calmed down' a bit.
I'm going to test it with some hardware mods and see what happens.

[DavidAlfa]

As I said I had no time yet for fully checking it....
Yes, I didn't notice that! The cap is ok, but the led is not.
That circuit will let the 24V reach the amp inputs.
It's not good to let the clamp diodes work during normal operation, they're there for protection, but the extra voltage before the diodes start conducting can affect a lot.
I've seen that in other MCUs ADCs, the diode protected the input, but the extra 0.6V caused weird measurements.

Since the max input voltage is 13mV before saturating the amp output (gain 250), I clamped to the lowest voltage possible.
Quicko has a 1N4148 diode, which will clamp to 0.6V, I replaced it with a ultra low voltage schottky diode, which clamps at 0.15V
Didn't see a noticeable difference, but I think it might help to recover from saturation faster.
In your case, I would put a 1N4148 at the non inverting input,  straight to gnd.
Anyways that circuit is interesting, I will simulate it.

Lower the PID I to 0. It shouldn't spike now. Then increase it slowly. Maybe 5 each time, cooling the tip every time before checking the result.
If you use water, don't fully submerge it, only the part where the solder sticks.
I already lost some tips because water got inside and messed up, causing the measurements to go crazy, didn't fix itself by letting it hot for a long time, I guess it caused internal corrosion and that affected the TC output.
I use some cleaning alcohol when I want to do this.

If you have oscilloscope, measure the ADC input, you could discover something.

[postman]

Beware that the stock fw can't handle C245 cartridges.
You will probably burn the cartridge as it's set for T12 only and doesn't allow calibration.

A big problem with the Quicko is that the connector is soldered and to have access to the back of the pcb you must unsolder it.
That is calling for trouble, the board will get damaged for sure.

I found a simple way to connect the C245 without much hassle. This is how I did and it worked very well.
Only a trace must be cutted and a wire soldered.
It uses the pin normally used for the handle NTC, not used with JBC, neither in this firmware.
The existing circuitry shouldn't be a problem.
I found best performance PWM settings with 50mS Period and 5mS Delay. Less delay will catch random noise from conmutation and affect the temperature reading.
Keep in mind that calibration is a must! Select the C245 profile, and use a temperature sensor for calibration, be careful to not do anything wrong that could end burning the tip!

Hello!
Thank you all for those work and sharing it.
I have the same board showed avobe. I’ve flashed the cfw from @DavidAlfa with no problems, but I think I’m doing something wrong with the connections.
The trace from the mosfet is cutted and the join C7-C19 is done, but when I connect the C245 cartridge, I get a short circuit (the current is limited to 1A in the PSU to prevent smoke), and the tip doesn’t heat up.
I think I’m not understanding what you want to say with VIA HERE! DON’T CUT! What’s the meaning of that?
Than you very much!

[DavidAlfa]

Do you see the small hole? That is a via.
It connects to the bottom trace, and goes to the amplifier.
If you cut it, it will stop working.

What do you expect with 1A.? A C245 will draw 9Amps at 24V!
Are you using C245? Not C210?

[postman]

Hello.

Thank you for teaching what is a via to an ignorant.

I didn’t expect so much performance with only 1A, but the tip should heats up even slowly, shouldn’t?

I’ve tried with a genuine C245 and a clone, with the same results.

Best regards.

[DavidAlfa]

No worries! No one is born already educated 
These cartridges have a resistance of ~2.6ohms. At 1amp limit it will draw 2.6W at most.
Being lucky, it'll get barely warm to the touch. But the mosfet wont't work at 2.6V, and I also doubt the stm32 does.
The circuit likely will be resetting in a closed loop, turning on the mosfet, drawing too much current, current limiting trigerring, voltage goes down, circuit resets again...
If the screen isn't working, the heater won't either.

Set the voltage to 8V, 3amp limit.
That's still pretty safe, but enough to make everything work, and the power will be 24W, heating the tip for sure.

[postman]

Ok, I will try with 8v 3A and bring the feedback.
Thank you.

[postman]

As you supposed, the cartridge heats up with the configuration you suggested 

Nevertheless, I have had to do the test without handle (just with cocodrile wires) cause I must have something wrong in my handle (T245 clone), because when I connect it with the cartridge inserted, I have total short and the screen of the controller goes off.

So, some post ago I saw how you dismantle your T245 handle, how do you do? I'm doing some force between the green and black part, but I can't pull apart.

Thank you very much!

[DavidAlfa]

Check the wiring in the connector, maybe that's the issue. Either wrong wired or touching something.
Without cartridge inserted there shouldn't be a short anywhere.
The C245 handle only uses 3 wires: Gnd, Heater and sensing (TC).
Ensure that heater and TC  connections aren't reversed.
Once inserted, Heater-Gnd is 2.6ohms,  TC-Gnd ~0.5ohms, Heater-TC ~3ohms.

About dismantling the handle, if it's the cheap aliexpress clone, just pull hard the green part out.
It just has a drop of glue holding it, no fancy locking mechanisms.

[postman]

Hello.
I don’t know why, but I have been trying the system (whitout handle), and the mosfet achieve so high temperatures that has desoldered itself.
I’m suspecting from the psu, seems to do weird things while the cartidge is connected.
I’ll come back with the results.
Thank you.

[DavidAlfa]

Use 24V, or at least 12, without current limit.
8v was only to test for shorts, the mosfet will not work correctly and heat up.

[DavidAlfa]

What's up guys? Are you using the latest fw? ( 11-3-2021, "SW: git c612c32" )

I think I damaged something when replacing the OLED. However I didn't use the station for few weeks, until now.
Yesterday I noticed that it was having a really hard time melting the solder.
When I resetted the calibrations, they were are almost 100ºC less (for 250, 350, 450 I measured 160, 265, 390).
I doubt it's a software issue because I don't remember changing anything, but who knows.

I just wanted to know if it's doing well for you? Because maybe I could have screwed it again without noticing

Removing the GX12 connector it's really hard, too tight (They should have done it with higher clearance), so it requires a lot of heat and the cheap ass PCB will just fall apart.