CFW for KSGER/Quicko STM32 Soldering Stations

[Bozog]

Well, I left you guys with a job. To propose nice fonts for the gui.
I already compiled and tried that program and it's very easy to implement new fonts!

Hi David, how are you connecting to the swclk pin to program the board, it doesn't seem to go anywhere on the quicko t12?
Cheers

[DavidAlfa]

Oh for sure it does!
RST not needed if set software reset method in the programmer options. If you can't connect, the debug interface might be disabled in software, then use RST.

[Bozog]

Thanks for that, dead easy then, I'll have a go later see if I can connect to it.

[Il_Marco]

I agree with DavidAlfa that you should keep it all as simple as possible, at least to start with. Maybe just some simple hw mod that are worth to make. JBC hardware jumper in ksger board just disconnect (if cut) + from JBC TC because in JBC T245 has it's own TC output. Anyway this jumper doesn't help in any way ksger hardware or software to detect if you are using T12 or T245 handle. It's just useful tu move this jumper to GX12 handle connector so to have it right wired based on the handle you are using.
I would find it useful to always power in the unit in sleeping mode, to avoid having it hot due to power/software failures. If it restart for any reason, it should keep cold. User just start it maybe rotating the encoder fast right (and put it in sleep or standby rotating fast left). Maybe powering on in sleeping mode make it possible detecting which handle or no handke at all is connected reading TC adc input? I don't know if some differences could be detected between T12 and T245 handles (leave T210 apart for now). Maybe SW could try to detect it and ask for confirmation for right detection at power on? Then the user can simply not confirm and select T12 option. Anyway it would be good to keep different T12 and T245 TIPS to fast select from T12 and T245 handles. I know that this is simple to say but hard work to do... Thank you DavidAlfa for the good work you have already done and what you are still doing. I still have not received GX12 connectors to try my T245 on my ksger unit (that has some JBC tips in its menu). As far as the JBC jumper way for ksger unit to use T245 handles, we have to live with the fact that common pole (for TC and HEATER) in theese units is GND (connected to the TIP). So, as long as it works (I'll try it ASAP), it would be no problem heating JBC TIP through the TC.

[dreamcat4]

Well what about the NTC thermistor inside the T12 handle? Which is between pins 2 and 4 on the 5 pin aviation connector.

It's reading several kilo ohm, at room temperature. Between 8.5k and 9.5k. If we wire the T245 handle so that it is either an open or a short between those pins. To give a reading that is otherwise impossible.

Am I right to think there is no such NTC resistor inside the T245 handle? According to these information gathered:

https://github.com/dreamcat4/t12-t245-controllers-docs/tree/master/tips-and-handles/t245-c245

Then If we can figure out these wiring scheme:

https://user-images.githubusercontent.com/749470/99142834-a33c4b00-260d-11eb-9dc2-b2753d5301a6.jpg

Such that it does not conflict, to prevent detection (later on) of T210. Since they do a different type of short, between different pair of pins on the JBC HiRose connector. To maintain future compatibility.

So where does the NTC thermistor goes into the Micro? This is shown in floobydust schematic, as PA7 / Pin17, marked "NTC"

https://www.eevblog.com/forum/reviews/t12-stm32-v2-1s-soldering-station-controller-schematic-etc/msg2463234/#msg2463234

So that's an ADC input, which seems balanced (the 9k) with 2x 4k7 resistors. And then the NTC is like a bias? The NTC seems "NTC B=3950 10K@25 degrees Celsius".

Anyhow I am not sure the swing of "possible" adc values, given the bias.  For when the NTC is working. Because it's resistance can vary a lot. However maybe if we were to short across the NTC, or have an open circuit. Or a closed circuit. It would seem possible to measure that as an invalid value in the code.

Unfortunately I am not 100% clear on the wiring, of the ground etc. Because these different diagrams are giving a conflicting information. Really annoying. Sorry. But pin 2 of the aviation connector seems to be going to the NTC, looking at my blue PCB here.

Heck if we just didn't connect it to anything at all, that could select T245. If only shorting it could then also detect T210. At the other end of the scale... however the way JBC wire it, they bridge a pin of the C210 only to the heater negative -, rather than to GND case. Which is a bit frustrating.

Also it is desirable to have a safe failure mode, of the T12 NTC (pin 2), in case the wire breaks inside the handle. This would then mean put the T245 as short, instead of open circuit. Because it would not be possible to distinguish between open circuit and a broken T12 handle wiring.

 

[dreamcat4]

The main problem is... if a short to ground could instead also happen on a faulty handle. For the pin 2 NTC. If both short or open circuit can happen by such accident inside these chinese poor quality handles.

 

[dreamcat4]

Hmm, well it seems possible to detect both open and closed circuit on the NTC. Because I have just tested here on the original FW, and either way, the invalid ADC input value is noticed as out of range, and the "room temp" on the main screen then prints as text: "! 14°C"

Perhaps a sensible compromise would be to think in an extendable fashion. ATM we have 2 settings for the tip choice:

* T12
* T245

However there is no NTC present with T245 handle. So you don't need to display warning message at boot time. Unless you also detect the invalid NTC value too during boot time. Which would then suggest *maybe* the JBC tip. Or *maybe* bad wiring of the T12 handle. Then ask the user to positively manually select between T245 and JBC. So then display the warning message.

Otherwise it should be safe assume (most cases), if you get a valid NTC data it must be because a T12 handle is already connected to the station. Before the user powered on.

But in future, the planning of 1 additional option, is also possible. So then a 3rd option can later appears:

* T12
* T245
* JBC Auto

With "JBC Auto" being the auto sensing option. In this case, the user must manually choose to enable this option, by going into the menu settings after boot. However if somebody chooses this 3rd option, then they are no longer protected from faulty T12 handle wiring, giving a faulty NTC reading. This is then a good idea to display a 2nd warning screen, after the user selects "JBC Auto". To warn them of that.

An open circuit on NTC (--> 3v3 NTC adc max) should set station to --> T245 handle. However in 3rd planning stages. An additional mod board, hardware mod is installed. Which is to say for future expansion. Then a closed circuit on NTC (--> 0v NTC adc min). Would set the station to be T210.

I think that is the best way to map future expansion, to select multiple tips. Simply by physically plugging them in. The auto detect feature is useful when you frequently want to change handles.

For example, if you have cheap T12 tips for some shapes. Then more expensive JBC tips for other shape. Also the T245 has larger size, for large work. And T210 is smaller, better for fine micro soldering.

I think that is really the only way I can see ATM. To keep firmware compatibility, and make an extensible system. For both modding, and not modding, and new hardware.

Otherwise if you don't use the NTC value alone. Then it can be supplemented with additional extra GPIO digital IO pins. But that as you say original boards, which don't break out the pads for not-used pins. So then they are left to micro soldering some tiny fly wire (mod wire) and glueing it down with epoxy. Onto the legs of the MCU. Which is not always so easy, for the initial capabilities of beginner / novice user.

But anyhow, it's just something you can start doing right now! Check NTC value is valid at boot time. If it's valid it cannot be any JBC tips. ---> must be T12. --> No need to display the warning screen at boot up.

Similarly, after boot time: If T12 handle is the selected option. Then constantly be checking that the NTC is a valid value (not ADC max, or ADC min). If that ever happens, then stop all power to the heater. And display an error message.

[Il_Marco]

I was planning to add a NTC in my T245 GX12 connector to use it as a cold point just as in T12 But I can do without it for the sake of autodetection
In ksger unit I can select in the menu what is the cold point I want to use, NTC (in T12 handle) or CPU (mandatory for JBC I think, maybe automatically selected for T245).

[floobydust]

The NTC is for cold-junction compensation, but there are so many junctions - the socket, cable, GX connectors, PCB.

I wonder if it really helps with total accuracy. There is a temp sensor in the MCU, some boards have the NTC on the PCB, almost all (KSGER) builds are in the handle. The menus have an ability to dial in a temp and use that instead. I would limit CJC in firmware, your handle should not be beyond +50C or -30C. If outside that window just use 25C as a default. I don't know where JBC is doing CJC 

Could auto detect perhaps looking at the thermocouple A/D value when the heater is on? Because the JBC TC is separate it should not peg like a T12. But you'd likely end up looking at the cable resistance due to the high heater on currents. Hmm.

[DavidAlfa]

And add that lots of T12 handles don't have the NTC installed. Like mine...
We can't rely on that. A cut or short wire can have nasty consequences.

I agree with DavidAlfa that you should keep it all as simple as possible, at least to start with. Maybe just some simple hw mod that are worth to make. JBC hardware jumper in ksger board just disconnect (if cut) + from JBC TC because in JBC T245 has it's own TC output. Anyway this jumper doesn't help in any way ksger hardware or software to detect if you are using T12 or T245 handle. It's just useful tu move this jumper to GX12 handle connector so to have it right wired based on the handle you are using.
I would find it useful to always power in the unit in sleeping mode, to avoid having it hot due to power/software failures. If it restart for any reason, it should keep cold. User just start it maybe rotating the encoder fast right (and put it in sleep or standby rotating fast left). Maybe powering on in sleeping mode make it possible detecting which handle or no handke at all is connected reading TC adc input? I don't know if some differences could be detected between T12 and T245 handles (leave T210 apart for now). Maybe SW could try to detect it and ask for confirmation for right detection at power on? Then the user can simply not confirm and select T12 option. Anyway it would be good to keep different T12 and T245 TIPS to fast select from T12 and T245 handles. I know that this is simple to say but hard work to do... Thank you DavidAlfa for the good work you have already done and what you are still doing. I still have not received GX12 connectors to try my T245 on my ksger unit (that has some JBC tips in its menu). As far as the JBC jumper way for ksger unit to use T245 handles, we have to live with the fact that common pole (for TC and HEATER) in theese units is GND (connected to the TIP). So, as long as it works (I'll try it ASAP), it would be no problem heating JBC TIP through the TC.

Sure, that was in my mind, also disabling the buzzer (In fact, I removed it on mine the first day I got it. What an annoying thing  )


Do people really switch between JBC and T12? Seems a bit strange to me. After using JBC tips..nothing beats it, at least not T12 tips.
I would understand the different JBC sizes, as you migh use a T210 for micro soldering, and the T245 for regular stuff.

[cosmin1]

I think i'm going crazy! 
Replaced the STM32F101 with new one and still can't access or program. I get the same error.
The F103 i can access instantly.

[DavidAlfa]

Check the voltages in the Vcc pins, also in the reset pin!

[cosmin1]

Connected to St-link:
Vcc pins all equal ~ 3,31v.
RST pin ~ 3,29v.

Connected to 24v power supply:
Vcc pins all equal ~ 3,26v.
RST pin ~ 3,25v.

Same on functional board.

[dreamcat4]

Could auto detect perhaps looking at the thermocouple A/D value when the heater is on? Because the JBC TC is separate it should not peg like a T12. But you'd likely end up looking at the cable resistance due to the high heater on currents. Hmm.

Hey! That is... a pretty good idea floobydust. Perhaps we could probe this with a 1% PWM on. For a very brief period? And then compare the averaged thermocouple adc value, over a brief time window. On, off.

TO see how much it jumps.

[DavidAlfa]

The latest JBC invention! Detachable tips!
The thin one works nice...
So the danger is not with C245 tips, but with 210s...
TC coefficient:
C245 ~40uV / °C
T12  ~21uV / °C
T210 ~10uV / °C

That explains why I had to set It crazy high to start melting the solder.
So we have a trouble here.
No T12 station will work correctly with the C245.
Op-amp gain is ~250.
Max ADC vin 3300mV.
3300/250=13.2mV Max amp input
13200/40=330°C Max temperature.
Can be fixed easily though.

[floobydust]

The T12-clone controller boards are having many issues with the thermocouple amplifier. Mainly due to using too cheap op-amp, and the 3.3V buck-converter makes too much noise to expect a gain of 250 to be clean.
It's like they're shooting themselves in the foot by using the smallest capacitors, cheapest op-amp, no H/W filtering- yet expecting accuracy to a few degrees.

I saw recent T12 V4.4 controller boards are using a 24-8V buck, then a 3.3V LDO to power things. I'll have to check if the TC amp runs off 8V then you can accomodate the JBC TC and just need a second A/D channel with reduced scale.

[cosmin1]

Check the voltages in the Vcc pins, also in the reset pin!

Works on WinXP. But nothing happens when i open one of the .hex files available for these boards...
Am i missing something?

[DavidAlfa]

My idea was pretty simple. As the input has a 4k7 resistor, we could connect another resistor after it and switch to ground by software when c245 is selected.
No amp problems on mine, not perfect but the noise is totally acceptable (2-3 bits of noise)

Works on WinXP. But nothing happens when i open one of the .hex files available for these boards...
Am i missing something?

After opening the file you should see it in the Binary File tab.

[DavidAlfa]

After opening the file you should see it in the Binary File tab.

[cosmin1]

Solved it. The HEX file must be renamed. Long name file is not loaded.
Now i will see how it fits to my board...

[DavidAlfa]

I'll try to port the fw for your board soon...

There's something werid going on.
The measurement is very stable so I don't think it's wrongly wired. I re-checked it yesterday and it was fine.

The adc doesn' match any tip TC. The T12  was spot-on with 21uV/ºC.
My soldering alloy is Sn97Cu3, melts at 220-230ºC. The C245 starts slightly melting the solder at 1250 ADC.
That gives aprox 4.3mV at the amp input.  So 4300uV/230= 18uV/ºC? Almost the half?

This was my approach to connect the c245. As the GX12 connector is soldered with the front fascia in between, I can't access the back of the board without removing it.
I wanted to avoid removing it, the board gets damaged easily.
This was the simplest approach
- Cut the heater to amp trace
- Connect the handle TC wire to the handle NTC pin
- Remove active components that belong to handle NTC circuit (fet and amp)
- Run a wire between the handle NTC input and the TC amp input

The only additional connected components are a 100K resistor and a 100nF cap,  that should't make any difference.
(D1 acts like a 0.6V zener, it's reversed in the schematic)
Original


C245 mod (Heater input trace cutted)

[cosmin1]

I'll try to port the fw for your board soon...

Great!!! 
Friend floobydust and i finished the schematic of my board, we will post it on Github or here.

[DavidAlfa]

That's something important to make the fw work... Post it when you can!

[cosmin1]

One problem solved.
Tested all fw available from russians and one perfectly matches my board (T12_HW20_Encryption_OLED13) and is very bright with unmodified display. Just needed to generate the RG1/RG2 keys and that's it.

But... now other problem apeared.
When i tested all those firmwares, i came across a version (T12_HW20S_Encryption-SSD1306) where the display was less bright.
Then came the idea: what would be if i rewrite the board with F103 with same fw that matches the first board? Just to have the screen more bright.
So... i put back the 510k Iref resistor and flashed the F103. Surprise: dead! And no other available fw matches this MCU.
I am thinking to replace the F103 with F101 and flash it with the working fw.
Sounds crazy?

Anyway, i will be more than happy to use your fw David, if you will port for this board.

PS: this fw works for black board with 101 MCU (but NOT with F103)

[dreamcat4]

Intersting cosmin. But did you also take partial dumps and compare before flashing? I had put up some script for helping that on github.

If not, no worries. Just would be nice to see / compare the partial dumps. If you had taken them.