CFW for KSGER/Quicko STM32 Soldering Stations

[ricktendo]

delete cartridge configuration and calibrate

Do you mean tip configuration?

Thanks

[deaxx]

delete cartridge configuration and calibrate

Do you mean tip configuration?

Thanks

yes

[joeyjoejoe]

Very cool project.

Do JBC handles plug and play with the existing connector on this board? Or it needs a different connector to be put in the unit as well?

[N9ZN-Extra]

RICKTENDO, a Great discussion here, thank you for this post! :-)

I briefly peeked at this post and did not find, or missed the discussion, of the type of wire used between the thermocouple and the junction at which the temperature is converted to true temperature. My limited knowledge of thermocouple interfacing seems to recall a specific reason for wiring to be of a certain type, dependent upon the alloy used at the hot junction, to maintain potential accuracy. If this is true the cable between the thermocouple and the sensing device very likely must be changed for reliable accuracy. It is doubtful the Chinese have taken this to the table after weighing the cost of cable redesign and manufacturing if it was considered at all.

The ultimate design would be to place the cold junction inside the handle of the iron removing the need for special cable construction. The closer the distance between the junctions accuracy is improved.

Please elucidate or point us to something missed? (Any person with knowledge in this area is encouraged to comment)

Note: This level of accuracy in temperature will be required in very few situations but it is reassuring to know it is on point when required. Too many tradeoffs quickly degrade any product on a logarithmic scale and that practice is too often the definition of made in China. In fairness, I ask what the driving force is behind these decisions? Could it be our constant desire of more for less?

[dreamcat4]

Hey there. I am considering between either this project and another one. Or perhaps make both. However I do have a few questions. Maybe Some of you can chip in and answer these? Because the other project is all in Russian so it's harder for my to find similar answers over there! Many thanks.

1) What is the performance like, because it's difficult to see on the video. The original JCB stations go from cold to about 300C in only 3-4 seconds. However other solutions (like the unisolder) take considerably longer time than that.

2) Is the heating done high current DC pwm based. Or through an ac / zero crossing triac / half waves?

3) Does the heater have to be turned off, in order to measure the temperature from the thermocouple? How often is this done?

4) [edited] Is there no possible software fix, to make a safety mechanism to detect and prevent the tip from glowing red hot on sleep? Or is that something else? For example because of different wiring configuration between the T210 and T245 handles.

Ah... I see maybe that is already fixed now in the latest v1.1.3 release. "This release fixes an issue which resulted on the power to the iron being inadvertently be set to 100% with some calibration values." Good to know! 

5) If i get a revision 3.0 board, will the STM32 controller on it break / die quickly? Why would that happen?

Sorry if these questions seem a bit stupid or obvious. I really haven't had enough time to research all this in depth. Just trying to figure out which way to go first.

[CapnBry]

I have the KSGER soldering iron with the PCB that looks like this and I flashed it with the Oct 31, 2018 iron.bin from iron_hakko_big_screen.zip. However, now all the iron does is run the piezo beeper at insane volume and never starts up.

• Set readout protection disabled
• Program and verify said sectors were write protected
• Went into Option Bytes and turned off all sector write protection
• Program and verify completed
• BEEEEEEEEEEP (device just beeps). The beep isn't 100% continuous, it sort of ticks at a regular frequency, but mostly it just squeals loudly
• Performed chip erase and reset device. Sweet sweet silence
• Program and verify completed
• BEEEEEEEEEEP

Am I doing something wrong or is the hardware different and therefore now I've broken my soldering iron until I reverse engineer the schematic and hope it is close enough to run the custom firmware?

EDIT: Aww dangit. This is an STM32F102C8 so no wonder it doesn't work. Guess I just made myself a brick!
EDIT2: I pulled an STM32F103C8 off a blue pill and plopped it into the KSGER V2.1S PCB. It flashed ok with the iron.bin but I still get just the continuous beep, although it is a little quieter because I sorta melted the piezo while using the hot air gun. I also burst the battery from the heat, but removed it afterward. That's not needed to run though, right? Any step I've missed here?

[Hemi345]

What was wrong with the firmware that came with the station that made you want to upgrade?  I'm just curious as I'm interested in buying one of these KSGER stations.

[Hydron]

What was wrong with the firmware that came with the station that made you want to upgrade?  I'm just curious as I'm interested in buying one of these KSGER stations.

Lets you run JBC irons/tips as well as T12/clones.

As for the previous poster, no you can't expect firmware for one board to work correctly on a different one, and yes you will need to do some work reverse engineering the PCB and modifying the code!

[CapnBry]

What was wrong with the firmware that came with the station that made you want to upgrade?  I'm just curious as I'm interested in buying one of these KSGER stations.

The original firmware goes absolutely bonkers trying to control the temperature with the K shaped tip, even after calibration. The temperature tolerance for it is all over the place; setting it to 320C you'll see up to 360C and down below 300C. With all tips, it also doesn't seem to hold temperature very well. I set it to 300C with the BC2 tip and can hold it on a GND-plane-connected joint and it still says 300C and you can watch the power sort of creep up to as much as 20% but the solder still isn't melting. With I and J type fine points, soldering SMD is hit or miss as well with the temperature and how long I have to keep the tip on the part. I was hoping to get more accurate temperatures / less temperature drop / faster temperature recovery when soldering.

I also figured with the ubiquity of these designs and that the F101 and F103 processors have the same pinout, they would have the same things connected to the same pins and the only difference between then would be the firmware so the CFW might work. I looked at the code and it seems to halt and trigger the alarm if any of the HAL fails to initialize properly so I'm not even sure it is getting past init(), which should work given the MCU is the same as with the CFW. Not sure what crystal my board has though.

[Hemi345]

Hmm, that's discouraging.  A soldering iron is one thing I don't want to troubleshoot and tinker with, just use it for it's intended purpose (I have too many other things like that already ). I've read and watched videos about fluctuating temps with the aftermarket T12 tips.  I wonder if you just got a bad one and a genuine Hakko tip would have solved the issue?

[dreamcat4]

Hey there. Somebody can comment on the heat up time (from cold) wth the T245 handle ?

[trampas]

My unit is also the STM32F101xxx series.

It will take me a bit to determine what is different.

[CapnBry]

My unit is also the STM32F101xxx series.

Wow that's nuts because the board layout is completely different than mine, despite mine also being "Ver 2.1S". Mine has the external pump switch header populated, and the battery is soldered along the bottom edge, with the piezeo closer to the other edge. I know there's also another, third, Ver 2.1S that has the header for the connector to the iron but also has the ability to solder the DIN / aviation plug directly to the board as well.

As far as how fast it heats a T245 I can't say, but the T12s take under 10 seconds to heat up so I can't imagine they'd be very different. It actually took longer for my unit to turn on once I flip the switch than it did to heat up. There's like an 8 second delay on startup between the power supply coming on (<2s) and the iron initializing with a black screen (~6s) or waiting in a bootloader or something.

[dreamcat4]

EDIT2: I pulled an STM32F103C8 off a blue pill and plopped it into the KSGER V2.1S PCB. It flashed ok with the iron.bin but I still get just the continuous beep. Any step I've missed here?

Hmm. I am reading that specific variants seems to differ by the last letters denoting the various different sizes of the various different ram types...

https://www.st.com/en/microcontrollers-microprocessors/stm32f103.html?querycriteria=productId=LN1565

His one was this:



STM32F103C8T6

Which is 64k flash and 20k ram

Where I *believe* that the T6 at the end merely means the 'normal' (not special) operating temperature of -40C to +85C... so 'yes', all C8 part are supposed to have the same amount of memory in them, which should mean they are compatible and flash each other.

So either you did not flash it properly somehow (which I myself don't know the proceedure yet, but it seems pretty darn complicated / jumping through hoops until the cows come home).

Or perhaps the new chip you replaced somehow sustained ESD damage at some point during it's handling. I have seen this myself on other 3.3v rated MCU chip (in the teensy). Where just through regular handling, and ESD discharge killed one of the GPIO pins on the MCU. Whilst the rest of the chip continues to function normally and without problem.

Hope that helps.

[dreamcat4]

Heres a pic of my new v3.0 controller from KSGER. Which arrived today. The table @ https://www.st.com/en/microcontrollers-microprocessors/stm32f103.html?querycriteria=productId=LN1565 says it has enough flash space on it.... so it aught to be compatible i guess



and typical (as per usual), we can see this board arrived with some flux residue left on it. No biggie really. Should come off

[Hemi345]

Heres a pic of my new v3.0 controller from KSGER. Which arrived today. The table @ https://www.st.com/en/microcontrollers-microprocessors/stm32f103.html?querycriteria=productId=LN1565 says it has enough flash space on it.... so it aught to be compatible i guess



and typical (as per usual), we can see this board arrived with some flux residue left on it. No biggie really. Should come off

Your v3.0 controller looks identical to the one that is in my v2.0 KSGER except the battery is soldered to the board rather than having a header. But same ST chip.  Mine seems to work well so far with the KSGER firmware, just a weird issue with boost mode if I try to enable it too soon after coming out of standby.  Mine has some flux around one of the SOT23-3 components like they had to do a little rework on it.

[CapnBry]

So either you did not flash it properly somehow (which I myself don't know the proceedure yet, but it seems pretty darn complicated / jumping through hoops until the cows come home).

Or perhaps the new chip you replaced somehow sustained ESD damage at some point during it's handling. I have seen this myself on other 3.3v rated MCU chip (in the teensy). Where just through regular handling, and ESD discharge killed one of the GPIO pins on the MCU. Whilst the rest of the chip continues to function normally and without problem.

It is very probable that I could have damaged the chip when desoldering it to move it over, but given that the STM32 ST-LINK utility flashes the chip and then can be verified as many times as I want, it is flashed fine unless there are other fuses or flags that need to be set in order for it to run properly.

It just plays the buzzer on startup though, which is a function of the source code which just turns on the buzzer if HAL init fails on startup. If any of HAL_ADCEx_Calibration_Start(&hadc2), HAL_ADCEx_Calibration_Start(&hadc1), or HAL_ADC_Start(&hadc2) fails it just turns the buzzer on. Looking deeper in the code it doesn't halt unless the last one fails. I probably should have just uploaded the code to the blue pill first to see if worked before the surgery.

The board you got there looks like the "2019 STM32 V2.1S" version and mine is the "2018 STM32 V2.1S", which is funny because the 2018 version listing on aliexpress shows the 2019 PCB and warns "STM32 V2.1S controller is changed to this model, before order please be careful". Hey here's an idea guys, why not change the version number when you made large changes?

I'd be interested in hearing if you have any luck with the custom firmware which could mean that all I need to do is just get a replacement 2019 control PCB.

[esonec]

Dear friends,

Is that possible to find the firmware for KSGER 2 in 1 soldering station? I broke the STM32 chip and I want to replace it.
Need the firmware to flash onto the STM32.

banggood.com/KSGER-STM32-OLED-T12-Temperature-2-in-1-Hot-Air-Dryer-Digital-Rework-Soldering-Station-Solder-Iron-p-1407450.html

[Hemi345]

I don't think U5 is up to the task for powering the controller.  Mine continues to sporadically reset when trying to enable boost and then started resetting when entering sleep mode.  I opened up the case to see what it would take to add chassis grounding and the crashing stopped.  That got me thinking that maybe it was a overheating problem.  Touching various parts while it was running revealed U5 was super hot.  If I put my finger on it for more than a second, the station would reset. 

The device is a 250mA 3.3V regulator in SOT89 footprint and heatsinking on the PCB is very poor. Looking that size up on Mouser/Digikey reveals there's not much available to support more current handling in that form factor and with 24V+ input voltage. I might try to just replace it if it's defective, but I have an idea of using a bigger/better regulator on a stand alone PCB and use flying leads to connect it.

[dreamcat4]

while it was running revealed U5 was super hot.  If I put my finger on it for more than a second, the station would reset. The device is a 250mA 3.3V regulator in SOT89 footprint and heatsinking on the PCB is very poor.

When i look at my earlier photo (above) of U5... it seems possible for me to cable tie a heatsink onto the back of the pcb there. And have it stabilized by the 2 other ICs beneath it. With the cable tie avoiding th mounting holes, and passing underneath the gap where the LCD screen is mounted on the other side of the PCB.

This would mean that the heatsink can fully cover the U5 package, however it would be positioned at the very corner of the heatsink. Due to it's awkward location right next to the LCD header along the top edge. Which sticks up a lot. WHich isn't most ideal. However if i trim down those LCD header pins, then the heatsink could be repositioned better. But still not ideal if it's overhanging a lot.

As somebody who only bought the controller PCB, and not the external casing / housing for it.... then there are more options for actually removing the LCD and the buzzer from the PCB. To put them on little extension wires / cables.

But first it would make more sense to veryify / make certain that adding such a size of heatsink actually can solve the issue at all. Before thinking about which of those options (if any) is most reasonable way of getting it mounted properly.

Not going to look at this for some weeks however. But in the meantime maybe you can experiment yourself with a thermal pad and a little heatsink. Mine over here is 20mm x 20mm square with 49 pins / extrusions (cross cut). And that is really the largest one I have that's going to fit without doing any of the other mentioned modifications.

[CapnBry]

That's disappointing news about the newer revision of the circuit board. Comparing your U5 to my U2, my U2 is much larger (SOT-223-3?) and also has a thermal pad in the copper with vias to the back side. Hope you're able to work it out and get a stable controller.

[Hemi345]

Yeah, this isn't what I really wanted to deal with.  Looks like I can use spare holes in the board to get 24V and Gnd over to a little PCB I made that breaks out a larger, more capable Vreg.

But with the cover left off, this station actually does a great job. I have the K, ILS, BC2, D16, and DL32 tips.  The D16 is my favorite, it's awesome for SMD. Works great for SOT23 type ICs and 0805 passives and still big enough for through-hole headers.  I think the ILS is pretty worthless because the tip is so fine.  The K is great for desoldering and removing solder bridges but needs flux to work well.  The DL36 makes quick work of through hole and large copper pours. Not much time yet with the BC2, but looks like it'll work well to drag solder. 

Like others have noted, the temps bounce around when the tips are new, but setting medium heat (250C) and letting them burn in for the first 5-10 minutes, they calm down and have been pretty accurate once calibrated.  Being able to use a corner of my silicon mat to yank out one tip to swap in another is excellent.  I really like the blue handle, it's comfortable to use and fits great in a Pace iron stand a buddy gave me.   

[ealex]

I've got the same "3.0" controller board, with the random reset issue out of the box.
I've tried burning the FW and ended up with a dead unit - it looks like it's freezing somewhere in a loop.
MCU is STM32F103C8U6
Info : device id = 0x20036410


I think it's time to start tracing the circuit. I'll have to buzz the traces out, there's some risk of breaking stuff if I try to separate the display board, etc.
I've quite different from the schematic at "https://www.ptdreamer.com/wp-content/uploads/2018/02/iron_schematic.png" - there is an extra 24C08 eeprom and a SOT23-5 marked "A36" - it might be an OPA336 amplifier.
It matches the connection and the output pin reacts to me touching the input connector.

(edit reason - added device ID)

[Hemi345]

@dreamcat4,  do you have a photo of the backside of the controller PCB that faces the OLED display but taken down at an angle?  I'd like to see how big the traces are coming off the empty header labeled "C D Gnd VCC" above the STM chip.  If Gnd and VCC are reasonably substantial like the connections for Gnd and VCC are on the OLED header, then I'll use those to patch 3.3V back in from my little Vreg PCB. 

[dreamcat4]

I think i understand your question correctly. Unfortunately i did not think to take such a photo. However what i can describe is that both C and D traces are clearly only on the same back side as the STM32. Which I will call the "MCU side". C & D seem to be data lines connecting to pins on the MCU. SO those are already visible in my previous existing photo. So the only 2 that connect on the other side (the display side) are GND and Vcc.

However if you were to de-solder and remove the OLED display / mini board on the display side. Then you could get fine access to scrape & bodge a rather large solder blob for the GND VIA. To make GND perfectly fine. For the Vcc via, the trace connecting to it is the same width as another trace that can already be seen in my previous photo. Which is the short trace and fat directly above the middle of U5 regulator. So that is also reasonably 'a lot wider' than most of the other tiny signal traces. Hope that helps.