CFW for KSGER/Quicko STM32 Soldering Stations

[DavidAlfa]

Wickated, I see your point now.
I've already explained a few times how the oled dimming should work. That was a bug, not the way it was supposed to be.
I missunderstood you. We aren't masters at English language, so that things will happen from time to time .

Ex. I said: "It turns off after sleep".
You answered:"I suggest you turn it off after sleep" instead "You said it turns off but it's not doing so"
In that moment my thoughts were "wtf, did he even read what I wrote? It's doing so (Or should)" *Lots of confusion*

I changed the oled  dimmer. Now it behaves in a different way, doesn't care about the stand or shake modes, uses only the iron states.

- If dimmer is only enabled for sleep mode, run modes will keep the oled at full brightness all the time.
- If dimmer is enabled for all modes, the screen will only wake up if it was off and the station is not in sleep mode (It woke up).
  But if screen was only dimmed (oled power in sleep is enabled), it will stay dim, because that's what "Dim always" should do.
- In any case, the encoder activity will wake up the screen.

I also added wake filtering (For shake input mode).
If enabled, wake changes faster than 100ms or slower than 500ms will be ignored, so the station won't wake up by a small noise or vibration.

Made changes to the labels. If you're fine with them, time for translations!

.IRON_Boost_Add = " Increase",
.IRON_Error_Timeout = "Err time",
.IRON_Error_Resume_Mode = " Resume",
.SYSTEM_Oled_Dim_inSleep = " In sleep",                 (Maybe "Отдыхает" ?)
.SYSTEM_Shake_Filtering = " Filter",
.SYSTEM_Stand_Mode = " In stand",


The layout looks like;
Boost          [10...120s]
  Increase   [10...200ºC]
Dimmer      [OFF] [SLP] [ALL]
  Delay       [5...250s]
  In sleep    [OFF] [ON]
Err time      [0.1...25.0s]
  Resume    [SLP] [RUN] [LAST]
Wake mode [SHAKE][STAND]
  In stand    [SLP] [STBY]
  Filter         [OFF] [ON]

I know the timeouts are a bit weird, but they are supposed to be fast, ex. 10-30s for the oled timeout (To prevent burning), and 0,5-2s time for clearing errors (When removing/inserting tips, handles...), not meant for high values, although it allows to do so.

[wickated]

- In sleep: What to do after the screen was dimmed in sleep mode and temperature is under 100ºC.  Leave on or turn off.
                It's never turned off in any other modes, and never turns off unless the tip is below 100ºC.

[DavidAlfa]

It seems someone didn't read the last message

Regarding the mesage from the Radiokot user, saying that the temperature actually goes down when manually heating the tip.
Seems caused by the mains ac noise.
The reason it goes up and down is because you're injecting 50Hz noise to the ADC, sampling at 5Hz, but they aren't perfectly synced, so the mains wave will drift over time.
So in the end you'll have a sinewave!
My house earth isn't great, touching the iron tip while having the feet away from the floor doesn't cause the issue.


I'm waiting for the translations, new builds will go when they're ready

[AndyC]

I'm waiting for the translations, new builds will go when they're ready

Swedish:

Code: [Select]

      .IRON_Error_Timeout = "Feltid",
      .IRON_Error_Resume_Mode = " Återst.",

      .SYSTEM_Shake_Filtering = " Filter",

      .errMode =     { "SOV", "KÖR", "FRG" },

I had a quick look at the scrolling text, could be quite useful for the translations...

[DavidAlfa]

Thanks!
Yeah... As I already spent the time on it, I left the code somewhere, could be useful in the future.
The current method has no limits, you can have a very short label and a huge option, the other way around, or overlap.
But sliding text has boundaries, dividing the screen in 2 parts...It's complicated.
There other possible layouts, like expanding the option, but I prefer to see every option with its value.
Testing the sliding text is as simple as enabling the option and using longer strings than the available drawing space.

Wickated, c'mon! Just 4 words! I could use Google but I'm scared, I might end writing some bad things about someone's mother, you never know with these translations

What about translating "boost" as "турбо", and " Increase " as  "Добавить"?
Would it make more sense?

[wickated]

Wickated, c'mon! Just 4 words! I could use Google but I'm scared, I might end writing some bad things about someone's mother, you never know with these translations

What about translating "boost" as "турбо", and " Increase " as  "Добавить"?
Would it make more sense?

sry i was very tired and forgot to post message.
boost is just new layout of boost?

Code: [Select]

.IRON_Boost_Add = " Догрев"
.IRON_Error_Timeout = "Вр.Ошибки"
.IRON_Error_Resume_Mode = " Возврат К"
.SYSTEM_Oled_Dim_inSleep = " РежЭкрана"     
.SYSTEM_Shake_Filtering = " Фильтр"
.SYSTEM_Stand_Mode = " Подставка"
.errMode =     { "ВЫКЛ", "ПУСК", "ПРЕД" }
i also forgot to explain photo. screen shuts down when tip cools down, but after wake triggers it doesnt light up. as i remember at first implementations it was supposed to be ultimate antiburn feature and screen goes off instead of dimming every time. later hot protection was added and screen turns off only after tip cools down. but dont turn on.

p.s. if text lables are short enough (11chars), they wont scroll?
so i mean we can enable scroll but leave short names as is?

[DavidAlfa]

I've said already 2 times it was a bug, not how it was supossed to be.
Affected only stand modes because it was using shake input as source. All that is already fixed.

Text scrolling is a test feature, disabled by default in the code.
With it, the actual text size is even smaller, around 8 chars, because you have to split the screen in 2, one for the value and other for the label scrolling.

New builds! All should be fine now.

The settings structure is different, so everything will be wiped... note down your calibrations!

The fw has changed a lot since the last video... Will have to make a new one.
My camera hates oled screens, specially the refresh rate!

[DavidAlfa]

Nobody complaining? Everything working correctly? Really??   

[ygi]

You miss the complains already? I just bit the bullet and updated from a 2 weeks old build, thus had to spend two hours recalibrating all my tips (my thermometer's so fiddly it takes 10min for each tip).
All I can say for now is that default settings for calibration target points are too high for my unit. Both 250° and 400°C are off by a large margin (80~100°C too high). I lowered them to 1100 and 2000 so it lands within the 50°C delta.

Hardware is Ksger 1.5 with T12 tips. For reference, my values are:

Zero 41
BC2 1092 2338
BCM2 1097 2022
C1 1209 1645
D52 (forgot it, I'll do it later)
ILS 1157 2517
JL02 1184 2350
JS02 1107 1906
KR 1041 1833
KU 1092 2135

[DavidAlfa]

Well, that's easy to fix. I can't find these values by myself 
You have a lot of variation between tips, specially with the C1.
That's crazy, are they new?

[ygi]

I don't use C1 often but it's not new, it's got at least 4-5 hours of burn in. My tips are of 3 or 4 different chinese "brands".

[DavidAlfa]

That makes sense.
I usually get all from the same seller, the calibrations are so close that I rarely calibrate more than one tip, doesn't worth the hassle for 1-5ºC error!
Being your tips so different, backup the flash before flashing the firmware, so you can quickly restore it if the new version fails for any reason, keeping all your settings.

[wickated]

flashed 9-13. on first boot after changing languange once screen clears and new strings doesnt load

[DavidAlfa]

Whoops, I introduced a bug due the new drawing detection mechanism.
I also spotted a little bug in the screen power, it wouldn't turn off if there was an active error, staying dim.
Fixed, updated the builds (Same settings structure, nothing will be erased).

[wickated]

so for late testing, i still dont have external port for swd on backplate.

about newest post on radiokot. so u say precal. values 250 and 400 are linked to each other i.e. 400 is an offset over 250? i thought that both 250 and 400 are set independently above 0. and only rule is "250 < 400"

[DavidAlfa]

Nope, no offsets.

In older releases you had to roughly adjust the ADC value, but when I studied the calibration code, I made it right.
You should have noticed that now the default calibration adjustment is a lot more accurate when calibrating tips, that's because of that.

The system uses linear interpolation, but applying temperature correction.
So when you adjust CAL 250 or CAL 400 values, that's not the value applied in the ADC, but the calibrated value for 0ºC ambient temperature.

So for example, if you have these values:

- Zero offset:    300
- Cal 250:        1400
- Cal 400:        2300
- NTC reading:  38ºC

And you ask the system for 400ºC, it won't use the 2300, what's really happening at low level is this:

- You want 400ºC at the tip.
- But remember, thermocouples works on a temperature gradient.
- As the ambient temperature is  38ºC, you really want to read 400-38=362ºC.
- You have 1400 for 250ºC (at 0°C), 2300 for 400°C(at 0°C), but you need the ADC value for 362°C.
- The system checks what range to use: 0-250 or 250-400, Which are Cal_zero-Cal_250 and Cal_250-Cal_400.
- In this case, 362ºC is between 250 and 400. The calculation would be: (362-250)*(2300-1400)/(400-250)+1400 = 2072.
- This is the real value for the ADC when you're calibrating for 400ºC and your NTC reads 38ºC.
- And this value is changing all the time based on the NTC readings.

So, as you see, if you change CAL_250, it will affect the 400ºC calibration.

When cal 250 hits the limit (Can't be set higher than cal 400 value), you need to increase 400ºC, always watching to not overheat things, repeat cal 250 calibration, then run the 400 calibration to use the updated Cal 250 value.
This is likely to happen only with JBC tips because I set really low values, don't want to take any risks, they're not $3 tips, but $40.
You'll certainly prefer spending some extra time at calibrating than burning your expensive tip.

[wickated]

Thanks for total explanation!
this also persuades me to actually run calibration eventually, not only writing numbers from paper every fw update. even if it still work enough accurate to never bother about.

[DavidAlfa]

These values are temperature-compensated, so it's completely valid to manually restore them.
The differences you may get between calibrations are usually caused by the NTC or the handle connections, causing small errors.
Normally these errors are very small, you might get 2400, 2420, 2385... That's not an issue.

I do that too, I got my calibration values of 1377 and 2216, and I restore them after each build in the tip settings menu.

The only calibration I need to do is the zero cal value the first time after resetting the settings

[wickated]

i also have stupid question - do i still need quartz generator onboard ? 

[DavidAlfa]

No


After a lot of research,I found the type D thermocouple to look very similar, my measurements also showed that slight curve.
Type C is almost the same, but it's more straight.

[ygi]

You could put a clock somewhere at the bottom of the system screen just to give it a use.

[DavidAlfa]

Nope, I've been clear about that.

Found this great PDF with all polynomials(mv to C, C to mv)
https://tempsens.com/pub/media/pdf/Product_Guide.pdf

Will have a look. Luckly I'll be able to use a single vale calibration.

[wickated]

You could put a clock somewhere at the bottom of the system screen just to give it a use.

i cant make any explanation for a 150+watt clock

is that type of tc for t12 or jbc? since i guess not both at the time

[Tugo]

These values are temperature-compensated, so it's completely valid to manually restore them.
The differences you may get between calibrations are usually caused by the NTC or the handle connections, causing small errors.
Normally these errors are very small, you might get 2400, 2420, 2385... That's not an issue.

I do that too, I got my calibration values of 1377 and 2216, and I restore them after each build in the tip settings menu.

The only calibration I need to do is the zero cal value the first time after resetting the settings

I didn't have time to do any tests at all but I read this forum 
When I will have more time I will tell you my new values.
So if I understand you correctly... if real temperature is "23 C" and NTC is showing that at the start before soldering iron wasn't used at all, does that mean I will have a "15 C" (for this case) lower value than wanted if NTC is getting higher values over time since handle is getting warmer (lets say to "38 C") but ambient room temperature will still be the same "23 C"? If that is correct, wouldn't be better to have NTC somewhere where there is more like real ambient room temperature and would not change over time if outside temperature wouldn't change?

[DavidAlfa]

Nope, you need to measure the temperature in the cold side of the probe, which are the T12 contacts.
If the contacts are at 40ºC, the sensor will output the equivalent voltage for 360ºC when the tip is actually at 400ºC.
That's how it works, and why you need the NTC there.
If you want to use the internal temperature, simply disable the NTC in the menu.