CFW for KSGER/Quicko STM32 Soldering Stations

[FransW]

The only thing that bugs me is the unstable T12 readings.
I wish I could wipe them off.
From my point of view, that would finish the firmware.

you know, i finished it via dumping all t12 tips i had. 
parallel TC connection instead of series one solves some problems.

main joke is not about just showing set temp close to real tip temp, but in excluding over/undershoots. which results in solid solder joints.

and then u can finally start working on own ac solder station with zero crossing detection, universal pinout, multichanell , big rainbow screen and coffee machine.

like this

Can you spend some words on the technical specs, like a functional block and the expected results?
The pictures are very clear, so I understand what you try to achieve, no mistake here.
The control station is compact and very useful as depicted.

Thanks in advance,

Frans

[wickated]

changed stand switch from 2 wire switch to 1 wire through 1k res, noticed that station goes standby but dont sleep. with 0,1k res works mostly ok , with 0k res also dont sleep. ( stand switch line looks this : (3.3v > 4,7kR)+(stm pin > 4,7kR) > |/ tip gnd < 2,5R heater < 1MR < 3.3v )

[DavidAlfa]

Sorry, but make a damn proper schematic or drawing

The handle mode will ignore any changes in the wake input unless it's steady for at least 500mS.
Where's your electronics knowledge? Have you seen the voltage drop in the gnd wires when drawing 10 amps?
Probably you're injecting a lot of switching noise, and the wake pin is receiving constant changes.

Try using a diode like this [3.3v][4K7][--|>|--][Handle gnd].
But it' probably won't work either.

[wickated]

diode didnt work. 100R did.
i really dont bother, just chatting. 2 wire mechanical stand switch is also good if spend some time on 3d printing proper holder

proper drawing ? yeha im boss of paint.exe

[DavidAlfa]

Just use a reed relay!

[wickated]

Just use a reed relay!

NOOOO. i spent 100$ on jbc handle not to install reed relays in it. but ill try smth better - optical sensor switch.

[DavidAlfa]

Maybe I could check the wake input when the PWM is low, after the ADC conversion, to get a clean reading.
Try this:

[wickated]

Maybe I could check the wake input when the PWM is low, after the ADC conversion, to get a clean reading.
Try this:

i dont get what you actually changed but now it works solid as rock even on 10ms/0,1/10x. no spikes at all. i also left almost default pid this time

p.s.  in older fw, if i touch tip end on stations self ground wire, temp reading got mad due to ground drain, now not. maybe i missed new runaway detection in latest fw.

p.p.s. made some test vid like SteveyG did

[DavidAlfa]

You need more Imax!

[wickated]

You need more Imax!

thanks, now all stable! up  to ~100w output on 10mm shovel tip and overshoots only 5-10c
40/40/2/0.4

i have a question, maybe its stupid a bit but what about AC driving, but using current schematics with 1 p fet pwm ? i can just go buy 2*12/24v donut transformer anywhere but no 9/18v ones

[Tugo]

David, I finally managed to do a lot of test with different tips.
All was done manually so that I could get the same temperatures after a while going up or down and at the end of the tip.

Tip type (C/H... China/Hakko)    Cal250 / Cal350 / Cal 450
D12 (H)   1570 / 2285  / 3500
K (C)       1600 / 2400 / 3500
K (H)       1645 / 2455 / 3880
BL (C)     1900 / 2625 / 3750
DL32 (H) 2200  / 3050 / 4095 *
D24 (H)   1850 / 2575 / 3450
ILS (H)    1660 / 2475 / 4095 *
J02 (C)    1840 / 2555 / 3380
BCF2 (H)  1870 / 2650 / 4095 *
BCM2 (C) 1640 / 2500 / 4095 *
JL02 (H)   1890 / 2760 / 4095 *
DL52 (C)  1760 / 2400 / 3220

Those with a * are not able to get 450 C but I was able to set to about 400 or 410. At those set temperatures are really lower tip values.
Those with * I don't use very much (except DL32 which should be stable at higher values but I cannot set higher Cal450 to be able to use it - I should lower a gain in opamp but my test will then be of no good for you). Any suggestion?

EDIT: FW version 7.7.2021

[DavidAlfa]

Sure, anything can be made... You will have to learn programming

[wickated]

BCF2 (H)  1870 / 2650 / 4095 *
BCM2 (C) 1640 / 2500 / 4095 *
JL02 (H)   1890 / 2760 / 4095 *

you should check opamp gain. it should have 250x multiplyer  . 4000+ adc is too much

[DavidAlfa]

I doubt is the amp. It won't vary the gain between tips.
Likely these tips have little use and need more burning.
I might revert to the old 200/300/400 calibration.
But some solder alloys won't melt at 200°C...
So maybe just use different calibration values like 250,325, 400.
I could add more calibration steps but I don't think it worths it. More hassle when calibrating... The actual accuracy is pretty good in all range.

[wickated]

I doubt is the amp. It won't vary the gain between tips.

i mean gain set by voltage divider. 4000 for t12 is way too much, even on mine ad8605 opamp 450c t12 was near 2500 adc value. also this explains why he get "no iron" and so on cuz opamp output exceeds adc ipnut voltage.

250-350-400 woulld be ok since noone really uses 450 but 350 ftw

and what about ac sine power supply instead of dc voltage? with high pwm frequency will it work?

[Tugo]

I understand you guys but I cannot do anything about it.
Well, I could, but then I cannot do anymore tests for David or he will have to do some calculations 
DL32 is one of my oldest tip and you can be assured it is burned enough 
For the past 2-3 years I usually work with 270 and I almost never go over 330, but there are times I also need 380 (which is veeeeeeeery rare).
Before that I used higher temperatures and this tip was also used at higher temperatures and quite a lot so it is well burned.

I will check both resistors to calculate exact gain. Maybe that 1k resistor is lower than what it should be since it has more impact in gain.
Other than that it doesn't matter if I set it to 4095 and use just temperatures to about 380-400 since I will not go over for example "4080".

David, leave FW as it is, it is perfectly fine. If I will decide to do anything about it I will change it for me 

[DavidAlfa]

The fact you're soldering fine at only 270°C suggests a wrong temp measurement... It's too low!

[wickated]

The fact you're soldering fine at only 270°C suggests a wrong temp measurement... It's too low!

maybe he`s soldering wires with leaded solder? like in good old times 

p.s. bought donut transformer and bridge, ll try pulse 24v (fully rectified ll be 30v, thats too much), hope 1khz pwm would be enough.

[DavidAlfa]

Not a good idea unless the voltage drops a lot or your plan is to ruin the expensive tips.
30V would out more than 300W over them.
The best would be unfiltered rectified, and synced 100Hz pwm ending at zero crossing.
The stm32 would need additional circuitry and code modifications. That's close to how jbc does it (they use ac+triac).
I also use leaded. But by no way it flows like it does at 330.
With original jbc I mostly soldered at 320.

[Tugo]

Temperatures are just fine 
I measured it with fake FG 100 and with a Uni-t DMM with temp probe. Temp was off by a degree or spot on, but if you measure it on different tip spots you get very different temps.
I always calibrate station to set the temperature at the spot of the tip where I use the tip on the pcb. That is usually at the "far end" of the tip where it is all narrow and small. If I would have calibrate it on any other spot it would be even lower. Maybe I am doing it wrong (please tell me why), but that is how I calibrate it.
And yes, I use leaded solder, I don't use or even have any lead free solder. If I would have one that would have been another story.

David, why is 270 C too low for leaded solder?
When I see that there is lot of ground or higher mass to solder I also put +50C to use there (or even higher).

[wickated]

30V would out more than 300W over them.

jbc uses 24v ac transformers. so they have 30v too. i meant usage of only bridge w/o caps, so it rectified unfiltered i guess and 100hz. high frequency pwm should be ok, but as i see, stock circutry is unable to work stable over 1khz.
my plan is too get rid of awful acustic noise i get from pulse PSU and get 12+24v for later c210 hadle+tips expansion.

p.s. i calculated a bit, read some info about mosfet drivers and other stuff and decided to wait and build DC supply but with donut transformer, so i returned mine to shop and ll order 2*9v 10a from manufacturer. 2 month or so waiting time. 

[Tugo]

Resistors measured on the pcb are 0.999k and 247.3k.
I guess about 250 gain is just fine.

[wickated]

Resistors measured on the pcb are 0.999k and 247.3k.
I guess about 250 gain is just fine.

and input ones? should it be 1k-5k
try change 247k to 200k.

[Tugo]

wickated: what?
Those 2 values are 2 resistors that define gain.

you should check opamp gain. it should have 250x multiplyer  . 4000+ adc is too much

It is about 250x.
I agree with you that 4000+ is too much.

I also told in my message which was in general for David: "I should lower a gain in opamp but my test will then be of no good for you."
So I only have 2 options: to leave it as it is and I will use it as it is or I will change the gain and I cannot help David with my other tips. Hmmm... I just thought of something: Maybe I can put a switch on the side to lower the gain for some of the tips and use it just when needed  This could be for example to put a 1 M resistor in parallel with a switch 

[mastershake]

be careful with some of these boards the ones that showed up today both had pieces of solder bridging legs on the stm32 chips and some stray solder pieces all over the board. one had 6 bridges on the stm32 chip i had to pick off. they were not directly soldered together but there were pieces of extra solder bridging them they came right off with a small plastic pick. not that all will be this way but this one was its also a better stm32 then most of the others have had on the with the stm32f103rbt6 my wife should be done with all the tips i gave her in the next day or two again sorry for the delay. the eye went downhill some more and i had to go back in again friday and today