CFW for KSGER/Quicko STM32 Soldering Stations

[Mic]

First of all, many thanks to the author for the excellent work.
Please comment on my idea (maybe it's stupid, maybe it has already been implemented)

We have a "heating element with a sensor", after "thermal resistance"  to "end of the soldering iron head".(regulator try keep temperature of the sensor an not "end of the soldering iron head") 
The "heat flow" creates a proportional "thermal gradient" and temperature of the "end of the soldering iron head" is less than the temperature of the measuring point.
But "thermal gradient" is proportional to the "Heat Flow" and proportional to the current through the "Heater" .
We can add an additional "DeltaTemperatura" request to PID-regulator (proportional to Current/heat_flow/cooling_to_PCB )
It will partially compensate for the cooling of the soldering iron head (because the regulator tries to keep the temperature of the sensor and not "end of the soldering iron head").

[DavidAlfa]

Looks like you're using the wrong tip for the job?
Good tips like JBC have the sensors properly placed, so the performance is great, you can't expect the same from a cheap T12.
Also the thinner ones won't solder anything larger than a 0805 pad even at 400°C.
Trying to compensate for this will likely kill the tips, heavily overloading the heater.

Every tip, seller, fabrication date, aging process, everything will affect the tip performance, this sounds like an absolute nightmare to implement.
Ideas are easy to say, but nowhere close when taking them to practice 
If you know how to do it, welcome! 

[Rixi]

@DavidAlfa and @ststefanov13 Thanks for your replies regarding the pinout/wiring! I really appreciate it!
I will be out of home for the next days so unfortunately I will not be able to review my wiring accordingly. I will definitely post here once I make any progress!
Cheers
Rixi

[DavidAlfa]

Touch the tip body with the multimeter, find the wire measuring 0 ohms (Complete short), that one will be the gnd (G), then measure the other ones against that wire.
The one measuring ~3ohms is T+, the other (0.5-1 Ohm) is the sensor (H).

There's nothing else about it!

[Rixi]

That's how I did it. My GND is PIN 3.
My measurements 1to3 and 2to3 seam to be okay. The only strange thing is that I measure about 2.9Ohms from 1 to 2. That should be around 2.9 + 1.0 = 3.9 Ohms if I'm not wrong?

Here is what I re-measured on the connector with an inserted (used) C245 cartige a moment ago:
Pin 1 to 2: 2.9 Ohm
Pin 2 to 3: 1.0 Ohm
Pin 1 to 3: 2.9 Ohm
With no cartige inserted I have no connection/resistance between any connector pins.

I measured it multiple times, but will re-measure once I'm back home.

EDIT: I just realized that I probably wired the thermocouple and heater the wrong way around. According to this...

David, the HANDSKIT uses a different arrangement of the GX15-5 connector. See my last picture.
H is the thermocouple, T+ is Plus on the soldering tip, SW is the vibration sensor, GND is the signal ground, PE is the protective ground. With the Handskit, the NTC is soldered to the board.

...I should measure the 1Ohm from PIN 1 (marked H on the board) to 3 (marked with ground symbol on the board).
And 3 Ohm from PIN 2 (marked T+ on the board) to 3.

Another question: Do I need to add a diode here? (Attachment 1)

[DavidAlfa]

The max. expected voltage coming from the sensor is under 50mv, but 24V will get there when powering the heater.

Not absolutely required, the amplifier esd diode will clamp the voltage, but this could cause weird behaviour, better to solder a led or a 1n4148 there to clamp the voltage at lower level than VCC.

[ststefanov13]

That's how I did it. My GND is PIN 3.
My measurements 1to3 and 2to3 seam to be okay. The only strange thing is that I measure about 2.9Ohms from 1 to 2. That should be around 2.9 + 1.0 = 3.9 Ohms if I'm not wrong?

Here is what I re-measured on the connector with an inserted (used) C245 cartige a moment ago:
Pin 1 to 2: 2.9 Ohm
Pin 2 to 3: 1.0 Ohm
Pin 1 to 3: 2.9 Ohm
With no cartige inserted I have no connection/resistance between any connector pins.

I measured it multiple times, but will re-measure once I'm back home.

EDIT: I just realized that I probably wired the thermocouple and heater the wrong way around. According to this...

David, the HANDSKIT uses a different arrangement of the GX15-5 connector. See my last picture.
H is the thermocouple, T+ is Plus on the soldering tip, SW is the vibration sensor, GND is the signal ground, PE is the protective ground. With the Handskit, the NTC is soldered to the board.

...I should measure the 1Ohm from PIN 1 (marked H on the board) to 3 (marked with ground symbol on the board).
And 3 Ohm from PIN 2 (marked T+ on the board) to 3.

Another question: Do I need to add a diode here? (Attachment 1)

These measurements are taken at the handle connector disconnected from the post tip station.

pin 1  to 4 (Signal Ground) ? Ohm
pin 2 to 4  ( Signal Ground ) ? Ohm

Signal Ground (pin 4 )  is not connected with  Protective Ground (pin 3) on the board.


1=RED=Theromocouple
2=BLUE=Heater +
3 bridge 4 = GREEN = Protective Ground 
4 bridge 3 = GREEN = Signal Ground

!!!!  Signal Ground is not connected with  Protective Ground  on the board. !!!!

5 =Vibration Switch to back panel for stand

The diode is not necessary for C245 in this place, but it does not prevent having it.

[ststefanov13]

To David
on HANDSKIT station, soldering tip change  in version 1.10.2 doesn't work while working on KSGER ?

I have three HANDSKIT stations, the first with T12, the second with HANDSKIT soldering tweezers and the third with C245.
on those with T12 and C245, changing the soldering tip does not work.
I also have two KSGER v3 stations, the first one with T12 (there, changing the soldering tip works). This makes me think that I should do a test with c245 at KSGER station. 
Second is with HANDSKIT soldering tweezers.

What could be the problem with changing the soldering tip on the HANDSKIT stations.

[ststefanov13]

That's how I did it. My GND is PIN 3.
My measurements 1to3 and 2to3 seam to be okay. The only strange thing is that I measure about 2.9Ohms from 1 to 2. That should be around 2.9 + 1.0 = 3.9 Ohms if I'm not wrong?

Here is what I re-measured on the connector with an inserted (used) C245 cartige a moment ago:
Pin 1 to 2: 2.9 Ohm
Pin 2 to 3: 1.0 Ohm
Pin 1 to 3: 2.9 Ohm
With no cartige inserted I have no connection/resistance between any connector pins.

I measured it multiple times, but will re-measure once I'm back home.

EDIT: I just realized that I probably wired the thermocouple and heater the wrong way around. According to this...

David, the HANDSKIT uses a different arrangement of the GX15-5 connector. See my last picture.
H is the thermocouple, T+ is Plus on the soldering tip, SW is the vibration sensor, GND is the signal ground, PE is the protective ground. With the Handskit, the NTC is soldered to the board.

...I should measure the 1Ohm from PIN 1 (marked H on the board) to 3 (marked with ground symbol on the board).
And 3 Ohm from PIN 2 (marked T+ on the board) to 3.

Another question: Do I need to add a diode here? (Attachment 1)

What is the function of the 10 ohm resistor in the picture on power supply ?

Isn't that 10 ohms you're measuring?

!!!!  Signal Ground is not connected with  Protective Ground  on the board. !!!!

pin 1=RED=Theromocouple
pin 2=BLUE=Heater +
pin 3 bridge 4 = GREEN = Protective Ground
pin 4 bridge 3 = GREEN = Signal Ground

These measurements are taken at the handle connector with a post tip without the handle plugged into the station.

pin 1 to 4 ( Signal Ground ) ? Ohm
pin 2 to 4  ( Signal Ground ) ? Ohm

The field effect transistor that controls the heater is connected to + 24 volts and signal ground and not to the protective ground and its control from stm32 is relative to signal ground. !!!!!!

[yelkvi]

What is the function of the 10 ohm resistor in the picture on power supply ?

This is a revision of my power supply. When using 245 (when the soldering iron tip is cold), the power supply went into protection cyclically.
If you use T12, then this is not observed.
Changed the parameters of the current sensor. On the factory circuit, two 0.43 Ohm resistors are in parallel.
I added another 1.0 ohm resistor. And the power supply stopped going into protection.
Perhaps other instances of Handskit power supplies will not have such a feature.

[DavidAlfa]

That's a hack to let the PSU output more power, but it will instantly blow up if the transformer core reaches saturation.

All PSUs can handle more than their nominal power but they must leave some room for safety margin.

If It works, great, but if it goes boom you know why 

[ststefanov13]

This is not necessary, the HANDSKIT power supply unit is 150 watts, the same AIXUN is used in T3A. just in David's firmware menu for the C245 profile :
1. power 150
2. pwm 10x
3. ADC time 50 or 100ms

[ststefanov13]

What is the function of the 10 ohm resistor in the picture on power supply ?

This is a revision of my power supply. When using 245 (when the soldering iron tip is cold), the power supply went into protection cyclically.
If you use T12, then this is not observed.
Changed the parameters of the current sensor. On the factory circuit, two 0.43 Ohm resistors are in parallel.
I added another 1.0 ohm resistor. And the power supply stopped going into protection.
Perhaps other instances of Handskit power supplies will not have such a feature.

Am I to understand that you have a HANDSKIT station at your place? Does the tip change feature work in version 1.10.2?

[yelkvi]

I have an old version of the firmware. I haven't programmed a new one. No free time. And there is no need for new firmware.

[Rixi]

Does the tip change feature work in version 1.10.2?

What is the “tip change feature” exactly?

[DavidAlfa]

For convenience, it enables the tip selection for 5 seconds after changing a tip, so you can chose the new one by rotating the knob, or simply wait the timeout for it to go away.
Don't go Superman, it will ignore it if the tip is inserted again in less than 3 seconds, assuming it was a a connection failure instead.

Yeah, I have to update the Operation manual, too busy...

[Rixi]

I'm back home and making progress with the C245:

EDIT: I just realized that I probably wired the thermocouple and heater the wrong way around. According to this...

David, the HANDSKIT uses a different arrangement of the GX15-5 connector. See my last picture.
H is the thermocouple, T+ is Plus on the soldering tip, SW is the vibration sensor, GND is the signal ground, PE is the protective ground. With the Handskit, the NTC is soldered to the board.

...I should measure the 1Ohm from PIN 1 (marked H on the board) to 3 (marked with ground symbol on the board).
And 3 Ohm from PIN 2 (marked T+ on the board) to 3.

I re-measured (with a better, more precise multimeter) and confirmed I had the thermocouple and heater the wrong way around. So I swapped them. Attachments 1+2 show how is how it's now wired up.

!!!!  Signal Ground is not connected with  Protective Ground  on the board. !!!!
pin 1=RED=Theromocouple
pin 2=BLUE=Heater +
pin 3 bridge 4 = GREEN = Protective Ground
pin 4 bridge 3 = GREEN = Signal Ground

I bridged PINs 3 and 4 inside the connector (Attachment 3)

Another question: Do I need to add a diode here? (Attachment 1)

[...] better to solder a led or a 1n4148 there to clamp the voltage at lower level than VCC.

Added a diode (Attachment 4)


The good news: The station is now now showing the tip/cartige error any more And it's heating the tip
...but the temperature reading (or somethig else) is still not sufficient:
When switching on the station or putting in a new tip it's heating up strongly for 1-2sec and then reads temperatures up to 800-900 °C. The tip is then hot and you don't want to touch it, but not hot enough to melt solder. After reading the 900°C it stops heating and the temperature slowly drops to the set temperature which it then tries to hold. Still hot, but not melding solder. Various new and used tips all show a similar behaviour.

I do have a (fake) Hakko 191 Tip Thermometer, but can't read the temperatures acurately, because the tip is not melting solder.
However, I tried to run a calibration. It did not worked instantly for the 400°C calibration step: It showed "Wait..." forever and did seam to stabilize the read temperature enough. At some point the calibration seamed to work and showed "Successfull" (or something similar). But the behaviour did not really change (still heating up to ~900°C, tip get's hot but not melding solder).

The T12 handle and profile still work.

Any ideas why the C245 temperature reading/heating is so much off? Or what heating settings are not set the right way?

[DavidAlfa]

Genuine C245 have higher probe output than T12, almost double, so it might be saturating the amplifier output.
Enter calibration settings, click cal250 and keep increasing until it melts solder.
If the limit is reached ( cal400 value) and still not melting, increase cal400 by 500 and try again.
If you get to ~4000 value and still not heating enough, then it's the amplifier saturating.
There's a 249K resistor in the amplifier circuit, solder another one in parallel of 220K-150K to reduce gain and try again.
Be careful, now the output will double, set cal250/cal400 to original values before modifying.

[Rixi]

Perfect! Will try it tonight.

[Rixi]

I finally calibrated one C245-906 tip:
Cal 250°C: 2065
Cal 400°C: 3339

Thank you so much, David!

Is there a process to calibrate/optimize the PID values?

[DavidAlfa]

Was it the gain resistor?
Please document a bit the steps you took to make it work, so everyone with the same board can have a look at it.

The default values are pretty conservative, trying to avoid issues with the 4 million controllers out there.
You can try:
- Increase Pid Ki until having too much overshooting/oscillations.
- Reduce tip filtering to 30-50%.
- C245 sensor output is much cleaner than T12, try lowering ADC delay to 1ms and ADC time to 50ms, it might work nicely.

Do only one step at a time, it's easy to mess things up!

[Rixi]

Please document a bit the steps you took to make it work, so everyone with the same board can have a look at it.

Sure! The correct wiring and connector PIN-out is described in Post #3466.

Was it the gain resistor?

No, I didn't have to put a new resistor.

Beside the correct wiring, the main part was temperature calibration. I first read your manual about this and like you explained here I needed to go into the Calibration-Settings and increasy the values for Cal 250°C and Cal 400°C manually. I did this step by step and ended up with Cal 250°C = 1950 and Cal 400°C = 2950. With these values (already with some lower values) my iron melted solder, so I was able to use my fake Hakko 191 Thermometer to measure the actual tip temperature.
I then started the automated/guided Calibration and finally was able to calibrate the tip precisely (Result: Cal 250°C: 2065 + Cal 400°C: 3339 ).

- Increase Pid Ki until having too much overshooting/oscillations.
- Reduce tip filtering to 30-50%.
- C245 sensor output is much cleaner than T12, try lowering ADC delay to 1ms and ADC time to 50ms, it might work nicely.

Finally, I did not change the PID Ki setting yet, but I did change the tip filtering to 40%, ADC delay to 1ms and ADC times to 50ms.
This made the temperature curve much smoother! 
I still need to learn and understand what all these seetings do and maybe play around with them a bit.

[Murdo]

I apologize for my English, I have to write through a translator.
There was one idea that could compensate for the poor quality of T12 (clones), and unleash the full potential of C245 / C210.
Auto-boost, what I mean is -
1.The power (% of load) threshold is set (for example, 15% for C245, 50% for C210)
2. When this threshold is reached, the temperature rises by 10 degrees immediately.
3.If the load is not reduced, +10 degrees are added every 1 seconds
4. With adjustable upper limit, such as +30 degrees.
5. As soon as the load disappears, the temperature setpoint returns to normal.

[rokko]

For convenience, it enables the tip selection for 5 seconds after changing a tip, so you can chose the new one by rotating the knob, or simply wait the timeout for it to go away.
Don't go Superman, it will ignore it if the tip is inserted again in less than 3 seconds, assuming it was a a connection failure instead.

Yeah, I have to update the Operation manual, too busy...

Really nice function, but any chance to add that 5 seconds delay as a setting in future updates to make it longer? Sometimes I find it too short, when I start selecting changed tip it already timed out and changes temperature instead...

[DavidAlfa]

Easy, change checkScreenTimer(5000) in this line to the time you want, it's in milliseconds:
https://github.com/deividAlfa/stm32_soldering_iron_controller/blob/master/Drivers/graphics/gui/screens/main_screen.c#L558

Compile following the instructions from github, basically it's this: Install CubeIDE, Download the project in .zip, modify that line, run the build script, done!
Shouldn't take more than 10 minutes.