CFW for KSGER/Quicko STM32 Soldering Stations

[DavidAlfa]

No, I bought these:
https://a.aliexpress.com/_uwap9r

[wickated]

tuned to this settings. for me works good enough

[DavidAlfa]

I use default 20mS with 20xpwm, seems to help with the jumpy tips. They will bounce 2 -3 times but then get stable.
I don't think I could do anythign about that. Tested a lot of things during the morning.
Filtering these huge spikes will also filter quick temperature changes, adding a lot of delay between the real and  the filtered temperature, causing continuous oscillations.
Now is when I think on hiding these oscillations like rest of firmwares do. It's just impossible, at least for me.

[wickated]

since its not possible to get really quick response on t12 tips, maybe it would be good to tweak filtering to the max even if it will add like 2 seconds delay ? on jbc i only would like to see 1c estimation, so temp wouldnt jump like 299-300 all time and thats all any1 would need.

[DavidAlfa]

Please don't suggest such things like if I didn't try everything already, the current is the most stable. What you say performs really bad.

[wickated]

even estimation ? i mean cpu counts everythiing as usual but on display applies average func

[FransW]

Hi all,

This is quite a discussion.
Certainly if one hopes for a definite conclusion that moves the soldering station into a reliable competitor situation w.r.t. JBC, Unisolder, Hakko.

The basic set-up is already competitive, certainly for a multi holder solution.
I am not a development specialist. However, needing over 1300 posts seems a bit overdone.

Regards,  Frans

[Tugo]

Now that I am back home I did quite a lot of tests but without oscilloscope.
First of all I checked again Ksger and his overshoot. I didn't test with all tips so there might be differences but I found out it overshoot by about 10C. It is a little bit slow on the tip (on display is fast) regarding the temperature (when starting or when turning it up).
David, what I like on the Ksger and I would like to make a suggestion is:
- changing (setting) a temperature even when you don't have a tip inserted (if possible)
- changing (selecting) a different tip when you don't have a tip inserted (if possible)

Now to my tips that I have been testing (all on 24,7 V and FW from 7th of July)...
K (China) - 1566/2287/3230
K (Hakko) - I don't know what happened but I forgot what was the case with it 
DL32 (Hakko) - 2178/2950/4095 - I set those valus manually because I couldn't get it by calibration and also 400 C was the highest stable.
D12 (Hakko) - 1530/2290/3650 - I set those valus manually because I couldn't get it by calibration and also 380 C was the highest stable.
BL (China) - 1900/2580/3400 -I set those valus manually because I couldn't get it by calibration and also 390 C was the highest stable.
All of the tips have been used before in my Ksger soldering station and was having pretty much the same problem just on the higher temperaures (420 or higher).
What I get are "spikes" when PID ar trying to go the set values so I will have to work around on getting better PID settings.
Manual values was set so that I set temperature to 250C and then I changed values to get to the correct temperature. Then I repeated all the same for 350C. Since I couldn't set to 450C I set it to highest possible and changed values for 450C so that set value was correct. I believe I set everything as it shoould be done but a little deviation is possible.

David, what do you think if we would use lower temperatures? For example 200C, 300C and 400C? This is just a question I don't make you do anything.

I will do some more tips but I don't have a lot more. I also have about 10+ more but are new and they need to be "break in".

[wickated]

did u tested on default 200/20 settings ?

[Tugo]

I forgot to put that info.
It was on 200/20/10x just as you told me I should do it.
I even tested some other settongs but I forgot how they were. I will try to get the worst tip and try to change all of the settings to something that it will work and see if that will improve others tips too. I don't know when I will have time but I will try to do it as soon as I can.
If you have any othet suggestion please let me know 
Thank you, David.

[Tugo]

David, you will hate me 

I had time just to test 1 tip and I used my worst tip. That is original Hakko D12.
Then I set it to 100ms/1ms/1x which was my first sweet settings.
After that I set PID as you described how to set it and I figure out that at the higher temperatures Kd is giving me hard times so I set it to 0 for now.
Kp is 45, Ki is 10 and Imax is 0.20.
My cal values were 1585/2350/3700.
I was also able to do an automatic calibration so 
It gave a little to high values but not so much.

I also found out that after some time when NTC is going higher and higher... from 25 to over 30 I had temperatures also a bit higher... or lower... I forgot 
Do we really need that temperature to compensate the temperature in tip?

Today I probably will not have time but I will test the same PIDs with 200/20/10x if it will work at 450C.

[Tugo]

D12... I hope my last post about this tip I will try next tips now that I know how to do the test.
D12 Hakko tip:
100ms/1ms/2x pwm
PIDs are 45/15/2/0.20
Cal250/350/450
1600/2335/3500

Higher multiplier makes things go up and down a lot and I see that everything else is going great with other settings. David, please give a suggestion how to do any step better like you did for PID if you have 

And another thing for you David: is it possible to ask for another question when you click on Delete inside a tip? I hate when I was  and deleted all my tests in the middle and had to start over.

[tatel]

Well I'm trying to calibrate after installing July, 7th update. A guy would think that, after releasing it on that date, this should have San Fermín blessing

I got half a dozen Quicko-branded tips: three "standard" ones...

D16: https://www.aliexpress.com/item/32834486432.html

B2:   https://www.aliexpress.com/item/32835767458.html

K -This came with the station: https://www.aliexpress.com/item/32994824865.html

And three "High grade" ones BC2, ILS, KU : https://www.aliexpress.com/item/32834661555.html

I don't think there's much difference between the "standard" and "high grade" tips, aside the black coat the "high grade" ones have. Probably they aren't worth the extra expense. Yet they are cheap, so I decided to give them a try.

All of them have been on the "warm up" phase for at least 15 minutes at 350ºC

First stage: done with default values. Temperatures measured with FG100 clone.

Tip	Grade	250	350	450
BC2	High	258	275	291
ILS	High	244	257	273
KU	High	298	314	327
B2	Standard	254	276	294
K	Standard	258	272	293
D16	Standard	265	284	303

I did my best to get good readings. I did put a somewhat big solder blob on the tip, then measured twice, letting the FG100 fall under 50ºC between measurements. After the six measurements, I wicked the solder on FG100's thermocouple and cleaned the tip. So, no oxidized solder working as thermal barrier. Then did the same with the next tip. Got quite good repeatability, always with less than 2-3ºC differences. I think measurements are as good as I'm able to do.

Since FG100 readings are way under the set point at 350 and more at 450, I guessed next step was go to calibration->adjust. Maybe I'm getting it wrong? Anyway, I'm unable to do that adjustement. BC2 tip gets very close to 250ºC with 1050 ADC value; to get it to 350ºC ADC value has to go up to 1600; to get it to 450ºC it should go over ADC 2000 but here I'm getting errors... So I stopped the process here. What am I getting wrong?

I guess San Fermín is quite busy caring for the runners every morning at 08:00, then he has to get drunk as everybody else in Iruña these days. Maybe should I try again after "Pobre de mí" has been sung?  

One last question: Is calibration a whole-system thing? I mean, is it possible to have a calibration for each different tip, or the station should be calibrated so all tips are approximatively OK?

[wickated]

Is calibration a whole-system thing? I mean, is it possible to have a calibration for each different tip, or the station should be calibrated so all tips are approximatively OK?

both ways are possble. u can manually add different tips. preadjustment is made for purpose of avoiding tip burning. u can manually add new tip from tip menu if u have another. if u use different types of handle (for example t12 with shake switch and pace/jbc/anything another with stand switch) u can switch whole system setting profile from settings menu. there are 3 there - t12, 245, 210.

[DavidAlfa]

Tugo, you can select a different tip at any time.
No, I don't plan adding anything else. Almost no memory left. It was a challenge to fit the firmware into 10KB devices.
And no, I'm not breaking my head because the iron drifts a few degrees. I can't control it better in the way it's built.
Start repeating this: $30 Hardware limitations.

Tatel, calibration is per-tip, although new tips copy their initial values from the first tip, or from the selected tip if you use copy function.
But if you run a calibration, it will for the selected tip only. I suggest to calibrate the initial one before adding more tips, so you have a good base to start with. The difference between tips is usually pretty small and doesn't worth the effort.

[tatel]

you can select a different tip at any time.
(...)
It was a challenge to fit the firmware into 10KB devices.
And no, I'm not breaking my head because the iron drifts a few degrees. I can't control it better in the way it's built.
Start repeating this: $30 Hardware limitations.

Tatel, calibration is per-tip, although new tips copy their initial values from the first tip, or from the selected tip if you use copy function.
But if you run a calibration, it will for the selected tip only. I suggest to calibrate the initial one before adding more tips, so you have a good base to start with. The difference between tips is usually pretty small and doesn't worth the effort.

I think I'm missing something. Can I calibrate just a tip differently, then set the station to use that calibration? How can tip type settings be put at work? I mean how to choose which T12 tip i'm going to use. I'm unable to guess it. I know how to change profile. But that's not what I'm looking for.

Of course, I understand hardware limitations

[StephenR0]

It's in the manual.

https://github.com/deividAlfa/stm32_soldering_iron_controller/blob/master/Readme_files/Operation.md

Main screen
Tip selection
Click and rotate clockwise to show the tip selection. Then, long-pressing will enter the selected tip settings.
It will return to normal mode after 5 seconds of inactivity.

There's a lot of subtleties. 

[Tugo]

Here I go...
$30 Hardware limitations.
$30 Hardware limitations.
$30 Hardware limitations.
$30 Hardware limitations.
$30 Hardware limitations.
$30 Hardware limitations.
$30 Hardware limitations.
...
$30 Hardware limitations.
$30 Hardware limitations.
$30 Hardware limitations.
David, I think I got it   
Maybe...
Kinda...
I hope 
---------------------------------

From my point of view I think first thing to set the controller is to know your board. Each board (each version, release,... or manufacturer and even each element on the board is not exactly "the same") is different in a way that for example David can run things on 10x with 200ms/20ms and is stable, but mine is not. I don't know if wickated has the same board as mine or somebody else but he had very tight settings and was better for his board. Mine don't work so good like that.
That is why I believe we should use the worst tip and first try setting up Period, Delay and Multiplier (maybe even set Ki and Kd to 0 so that you don't have problems like I had, I even set Kp to 45) that it would work stable on lowest temperature and also on the highest. When you will have it stable then you tweak PIDs like David explained how to set Kp, Ki and Kd.
Cal250 and Cal350 values are set points for calculate values for temperatures below 350 C and Cal350 and Cal450 for temperatures over 350 C. So you can set these values manually like I did or use calibration process which is great if you have stable temperatures. Even on not stable temperatures you can first use calibration process and then tweak it manually if needed. David did a GREAT JOB! Again, thank you David! 

Tatel: I had exactly the same problems as you and had to figure out how my board could be used with higher set temperatures. First of all not all tips are the same. Find the worst one you can find and try experimenting with settings that will be used in all of the tips (PWM period, PWM delay and PWM multiplier). Try basically everything as I explained above and before doing anything, read a link that StephenR0 gave you to have a basic knowledge what you should and could do.

[wickated]

I don't know if wickated has the same board as mine or somebody else

noone here. i changed:
1) op amp to genuine and reliable AD8605
2) gate disharge resistors to 1k
3) opamp input snabber
4) Ldo
5) PSU to stable and reliable 200w by meanwell
Yes, David is absolutely right. 30$ hardware works only by miracle. and some very fancy-pricy genuine 400$ stations doesnt even allow to change temp without master operator key. because hobby level is nothing compared to business solutions.

[tatel]

Oops, I knew I was missing something. I forgot the push-and-rotate thing since last week

I see half my tips are in 2ºC range, and all of them are in 54ºC range, using default values. So I bet this $30 hardware will serve me well. And I don't think it's working only by miracle, but by good work.

Now I'll have to see why I'm getting that error while calibrating, but it will have to wait some more days. I have very little spare time right now.

Thank you all

[DavidAlfa]

The calibration error is probably because the temperature is very unstable. New tips do that a lot. So when the temperature rises, the spikes trigger no iron detection.
And It will stop calibration,of course.

Doesn't work my miracle, there's lor of work behind. But you lack a lot of accuracy. You must understand how thermocouples work. The NTC compensation is just a cheap way, does something but far from perfect.
Probably there're proper algorithms, I don't know.

I'm just tired of the damn T12 tips. How to drive them correctly to work with all? I don't know.
I've explored other projects, including the Unisolder, it has a lot of nice calculations that I don't understand but I didn't find anything that avoids that problem.
Maybe the best option is to just use proportional gain and live with 3-5°C less.

[tatel]

Well, I see my MAX 6675 can have about 7ºC error over what the K thermocouple gives, that will probably have her own error... So I'm not expecting anything better than 10ºC error here. I'm happy with that. I will be happy even if there is a 20ºC difference. I can see that San Fermín release is successful in getting the tips pretty darn good stable. Quite a lot more that the previous release. I bet it will let me do what I need and that it will be able to get my way forward with what you have given to us so far.

I was getting about 100ºC differences between the point set on my Yihua station and the tip readings from the FG100. So this is much, much better, and you can be proud of your work. It's really useful. Please don't let my cluelessness lower your morale. I'll have plenty of caffeine next time
 

[DavidAlfa]

I consider the accuracy to be pretty good. Yes, I get some drift when the handle heats up and NTC reading changes, but maybe 5-7°C at worst.
Are you soldering or growing bacteria in Petri dishes?
It seems better to show nothing to the user. Comments like "I set 360 but I only get 358". Oh dear.
Now I completely understand why they just put the setpoint. Lots of problems solved!
For the picky people  let me tell that you'll be the only suffering because it's 3°C off .

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.

[wickated]

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

[mastershake]

thats pretty sweet but i doubt many will be able to afford or want to spend that kind of money. i get it id love a station like that but most people 30$ is a great price point and for most people they work plenty good enough. just imo though maybe