CFW for KSGER/Quicko STM32 Soldering Stations

[AndyC]

I got some new thermocouples and had some spare time, so here we go:

Code: [Select]

KSGER 2.1 Green board:
DL32: 1733/2466/3323
D4:   1753/2531/3396
D24:  1797/2567/3546~
K:    1705/2367/3065

KSGER 2.1 Blue board:
DL32: 1667/2382/3155
D4:   1693/2423/3241
D24:  1714/2454/3270
K:    1597/2241/2895

KSGER 3.1 Black "OLED 3.1" board:
DL32: 1975/2668/3428
D4:   1985/2683/3442
D24:  1977/2690/3467
K:    1908/2548/3189

KSGER 3.1 Black unmarked board:
D24:  1727/2426/3181

I made the huge leap from "v1.10-fix" to "210804b". Many great improvements! The grinning screensaver was...yeah

[DavidAlfa]

You found it! 

[StephenR0]

This is with the latest firmware 2021-08-05.  After a full reset, I performed a quick calibration.

Code: [Select]

Model:  Quicko STM32/CKS T12
Tip:    T12-BC2

Calibration > Settings >
                        ADC Value
  Cal. Step 250°C       1050
  Cal. Step 350°C       1500
  Cal. Step 450°C       1900

Calibration > Start
                        Temperature ADC Value
  Cal250                252°C       1193
  Cal350                353°C       1640
  Cal450                443°C       2077

Then I tested how controlled the tips were.  I noticed that at 180°C, the temperature was very well controlled and the temperature graph was quite smooth.  Then as I increased the set point at about 250°C, some variation started to come in.  At 320°C where I do most of my testing, the variation seemed to be where it was going to be.

Here's my BC2 tip, which seems to be my most stable.

" alt="" class="bbc_img" />

Here's my D24.

" alt="" class="bbc_img" />

Here's my KU.

" alt="" class="bbc_img" />

Here's my C4.

" alt="" class="bbc_img" />

And finally here's my ILS, my most wild tip.

" alt="" class="bbc_img" />

Then I started changing some of the PID values.  I discovered that I could decrease the amplitude of the variations by setting Kd to 50.  Here's the result.

" alt="" class="bbc_img" />

The actual temperature variation seems to be within a four degree range.  It may not be as perfect as David's, but I'm sure it's quite usable.

Edit:  I tried to get the images in line in the post, but apparently I didn't get the trick to that.  They are in the correct order, though.

[wickated]

welp, ill definitely place swd port on back of my next stations case. 
found some bugs.
1) tip detection delay - can be only swithced with press+rotate until 900ms and lowered as usual after. also can finally get 1000ms detection now
2) buttons wake settings affect stand mode too - before i can preselect temp whil handle in stand, now it will awake iron from standby state if set in shake wake settings. and to sleep back i have to lift-land iron.

otherwise looks cool. also got new china jbc compatible stand, now stand mode works as intended

[dark_hawk]

welp, ill definitely place swd port on back of my next stations case. 
found some bugs.
1) tip detection delay - can be only swithced with press+rotate until 900ms and lowered as usual after. also can finally get 1000ms detection now
2) buttons wake settings affect stand mode too - before i can preselect temp whil handle in stand, now it will awake iron from standby state if set in shake wake settings. and to sleep back i have to lift-land iron.

otherwise looks cool. also got new china jbc compatible stand, now stand mode works as intended

How does the C245 handle and tips compare to the T12 using the KSGR station and David's firmware?

[Duffy82]

@wickated
Do you have a link to the stand?

[wickated]

How does the C245 handle and tips compare to the T12 using the KSGR station and David's firmware?

overwhelming

[wickated]

@wickated
Do you have a link to the stand?

https://www.aliexpress.com/item/4000270557037.html
dont forget that it require rewiring to work with genuine jbc handles.

[DavidAlfa]

Yes, I rised the detection limit to 1000mS.
Before, the label was too big and would overlap.
Confirmed these bugs, will be fixed

Stephen, your plots look really noisy, also the waveforms are quite strange.
Usually you'll get a sine-like waveform when filtering is too heavy.
But you get a double oscillation in the highs and in the lows.
Really strange.
Anyways, I'll repeat what happens when you blindly use it: It might be doing worse in the original fw and you won't notice!
Time me to plug the scope and see what the hell is going on.

Any other Ksger v2.1 users getting these weird oscillations no matter what?

1 minute of silence for my deceased D52 tip.
I got a hard fault while making the debug menu, the debugger will stop without you noticing and...
Btw, data trace is the most useful feature ever. Had something writing out of bounds, affecting some widget data. Ran the data trace setting the affected variable, and Bang! Variable changed at line 552.
Unitialized data writing to random address!

Really, are these extremely slim tps like the ILS useful gor anything? Maybe for soldering a 0402 pad..

[wickated]

Any other Ksger v2.1 users getting these weird oscillations no matter what?

usually cuz of faulty fake opamp

Really, are these extremely slim tps like the ILS useful gor anything? Maybe for soldering a 0402 pad..

nope. absolutely useless.

[DavidAlfa]

Maybe for soldering a lifted bga pad. I've had to do that few times when worked on that. Scratch the 1st pcb layer until you get to the via and solder an extremely fine wire.
That's the only application coming to my mind. Definitely you're not going anywhere with a D32 tip .

Fixed these little bugs and made new builds.

The widget issues are caused by some small changes I made recently.
For instance, I save the iron delay as time/100. ex. 1=100mS. Same with other settings that go in 100mS steps.
It saves some space, but I forgot to adjust the widget .

Also added a check to the setpoint widget, so it only triggers boost mode when the station is in run mode. So if you're in sleep or standby, clicking won't do anything.
This is the polishing state already!

[dark_hawk]

@wickated
Do you have a link to the stand?

https://www.aliexpress.com/item/4000270557037.html
dont forget that it require rewiring to work with genuine jbc handles.

How big a PSU board do need to drive those tips?
I think the one in the KSGR stations is 24v5A. So an upgraded PSU will be needed for those tips I think.

[ricktendo]

Please, can you add Tetris next?

[DavidAlfa]

Snake better?

[Tugo]

Please, can you add Tetris next?

Snake better?

On which drugs are you two? 

/joke off 

I have no oscillations but I still don't know if I have Ksger v2.0 or v2.1. It say v2.1s but I have linear voltage regulator on the board and not a switch one.
Bought it here: https://www.aliexpress.com/item/32844179319.html
It is normal version.

[wickated]

Bought it here: https://www.aliexpress.com/item/32844179319.html
It is normal version.

and you should have bought GX version like i did 

[DavidAlfa]

2.0/2.1 whatever. There only 3 real versions.
If the fw works nicely with 2.0 but
reads noise with a 2.1, it's either a design problem or a defective controller.

[Tugo]

Wickated: I already owned the same version for a couple of years and didn't have any problems so I bought the same. I would buy the same if I had to 

David: what if some pcbs don't like the same settings like other pcbs? Keep in mind about your 200/20 and my 100/1. Wickated is using T245 so he is out 

[Tugo]

achao, everything you need to know is here: https://github.com/deividAlfa/stm32_soldering_iron_controller

[DavidAlfa]

I don't know... I'm really tired about that, even I could say I don't care anymore.
The fw is there, it's time for some Ksger owner to play with it, check the signals, voltage ripple, noise sources, try replacing the op amp...

[wickated]

Wickated: I already owned the same version for a couple of years and didn't have any problems so I bought the same.

gx is at least equipped with 8551 opamp instead of noname fake sheet.

[wickated]

The fw is there, it's time for some Ksger owner to play with it, check the signals, voltage ripple, noise sources, try replacing the op amp...

sorry i dont have affordale quality oscil yet, but u see, i dont see trouble in reverse rotating encoder, i can just swap wires without complains. i dont complain for strange t12 spikes anymore, cuz i can just swap to 245 and be happy with amazing firmware and u see, i dont have any strange behaviour with 245 tips which can be seen with just bare eyes and ears (my psu make enough acoustic noise). u did enough work when tried to enable dma for i2c display module but it was found that factory screen module use different display controllers that dont support any changes (i guess yours had ssd1306 or smth), but u gave a try. thats pretty enough.
im using ksger board cuz i dont like pcb design of qicko one and also that they dont sell controllers alone, only full kits
p.s. have to upgrade to newer build once again, got some bugs with sleep stand state. i know u dont have stand, so i dont complain there are less options for stand users than shaker users
p.p.s rechecked, all seems ok (got bug that rising handle went from slp to stby only and not work mode), now fixed and i can set temp while handle in stand. also thanks for up to 2s tip detection time, i have now +1second to swap tip with bare hands 

p.p.s. to everyone - all that ali stuff is just diy, piece of pcb that require lot of work to make it work good. chinese fw just draw digits like if its all ok, but u have to change schematics and components to make it working for real. so for now i only have display, quartz, stm and some passive smd left from factory state, but thats still ksger 2.1.

[DavidAlfa]

Please, just tell what's currently wrong, listing old bugs is simply confusing.
I think you didn't understood correctly.
There're a lot of Ksger models, very similar ones.
If model A works, but model B, using the same op-amp circuitry and same fw settings, reads noise, it's something wrong with the circuit itself, and I can't do anything about it.
In that cases the owner will need to find the source of the problem. Or live with it, because It will usually work just fine.
It's not that I'm not willing to help. But I can't do anything.
The firmware is doing correctly. All the timings are spot on.
Then the problem must be from the outside. Power supply, op amp, tip, handle connection... Even a badly soldered capacitor on board, which has already happened.


I'm adding a fine filter tuning. Instead using rough 0...8 factors, now it can be tuned from 0 to 99%.
Calling the filter coefficient "k": New average = (k*previous average) + ((100-k)*new reading).
This means: As hig the value, the less impact the new reading has on the average. 100% can't be set, the average would never update - Yep, red glowing things would happen.

[Tugo]

Wickated: I already owned the same version for a couple of years and didn't have any problems so I bought the same.

gx is at least equipped with 8551 opamp instead of noname fake sheet.

Don't be negative about this.
Normal version also has 8551 on it. At least I have it.
On my old (the same on the outside but with STM32F101 and S06AC opamp) it is an opamp that I can tell you it has the same noise as 8551 (I tested both). I could be wrong but if you try to order 8551 you usually get S06IA or S06AC.
When I thought my opamp was wrong I replaced it with S06IA and it is working without any problems (no oscillations).

David: I would suggest using PID as about 45-0-0 so that it is stable enough and see how period, delay, filter, (pwm multiplier)... maybe something else but it was just those 3 (4) things really had an impact on all tips. You know that I had a huge problem before you completely rewritten some things and from that time I never complained. I believe before that time I was pretty much the only one with complains. After that I just tried to help you back with testing and how I did my testing so that others could see what should be done to see how things could be better. But others used default values and whoever have problems didn't change that. Stop repairing what should be ok and try to chill for a bit with some other stuff. After you will have your head away from the project you could also think and see something better. Have a 
And again THANK YOU VERY MUCH FOR A GREAT FW!!! 

[DavidAlfa]

45-0-0 is pure crap. The temperature will alway be lower than the setpoint, also the responsiveness will be close to nothing.
Updated code with new filter settings. From now the builds will be less frequent, so I recommend to install stm32 Cube ide and set up the project, it only takes 3 minutes, you have the video showing how to do it.