CFW for KSGER/Quicko STM32 Soldering Stations

[Jono434]

Thanks for your reply.  If I am understanding you right, on the handle.. the shake on/off is controlled via SW-200D to pin 1 (in floobydust's schematic).  Is it required to modify the handle and remove SW-200D or will it be ignored when in "Stand Mode"?

Is it as simple as grounding the chassis and then connecting the solder stand to the grounded chassis?  Or does pin 1 need to be involved?

For example, even without a solder stand, if the menu is set to "Stand Mode" would touching the tip against the grounded chassis exterior put it into standby mode?  I mean the "E" part of the tip, not the actual solder tip.

[DavidAlfa]

Your question makes no sense, if you are told to disconnect the shake wire from the handle and wire it externally, then no, it won't be magically ignored if you don't disconnect it and it won't magically work.

Tip earth and gnd are usually joined, but must be checked, won't work otherwise.

Also read WAKE MODE here:
https://github.com/deividAlfa/stm32_soldering_iron_controller/blob/master/Readme_files/Operation.md

Wake mode
How to detect activity. SHAKE or STAND.
SHAKE uses a motion sensor present in T12 handles, shake or hold the handle tip up to wake.
STAND uses the same input, but disconnected from the handle. Must be shorted to gnd when the handle is in the stand.
(Stand mode operation operation: Shorted to gnd = sleep/standby, open = run ).

[Jono434]

I don't have any STM32 stations yet, just trying to plan it out in terms of which handle and stand to buy.  There are so many choices, it's all a bit of a headache.  I'm also coming off an 8 year break in terms of electronics.  My background is terms of soldering is mainly in guitar valve amplifiers.  Around 8 years ago, I learned to do some fun stuff but yeah, my electronics/circuit brain is a bit rusty, I need to learn the fundamentals all over again.

There seems to be a good payoff in the long run for these units, if you can do the right mods and use them the right way.  And they are cheap, so it's easy to experiment with things.  The thing that worries me about the higher end Chinese units like AIXUN T3A is that they are still prone to component failures, no internal fuses, cold solder joints etc.  And the front access to the controller board is bad.  Where as with Quicko or KSGER.. you can easily access the controller board or replace it, if needed.  The CFW seems to open up a lot of options and expand the value of these units, and that is appreciated.  The UI also looks really good and 148 pages of development means that it is well tested.

Sorry, I hadn't read the GitHub Manual, I'll make sure to do that now.  I've mainly just been trying to spec out the parts so far.  I did find some good info earlier in this thread about Stand Mode with others trying to do a similar thing.  There might be more but two good discussions I found were here and here.  Those discussions helped a lot to understand more of what is required and there was also some useful pictures.

Even finding a stand has been hard, for example Wickated made a switch to use a plastic handle with his stand.  But it's been hard in the last few days to find a heavy/non slip stand that makes good connection with smaller T12 handle.. and also is conductive and allows for an output or jack plug.  I like the look of the Eruntop T12-ST91 for example, but there are two versions.. the earlier version fits T12 handle perfectly, the new version doesn't. (attached pic)   Even without using Stand Mode, there is a thread here that says all T12 stands are bad.  With the right combination of some modding or patience, I believe something can work well.  I have been trying to learn everything in the last 7 days, sorry if some of my questions are silly.  I need to learn more, but I think after a few days of having the unit infront of me, it will become more simple.

Is there any difference if using Mini version vs PSU version?  I believe the grounding can be different between the two units.  But for Stand Mode, it probably doesn't make a difference?  Disconnect the connection of Pin 1 in the handle, and run a wire from controller board Pin 1 to the rear of the case (female connector).. connect this wire to the stand, and make sure that it makes connection when the handle is resting in the hand.

[Jono434]

Forgetting about Stand Mode.. I've been trying to work out the best configuration in terms of mini vs full size.  Or green PCB vs blue PCB.

Is there any unit or controller board that is preferred over another?  In terms of the step-down converter on the controller board or other issues.  Or a safe way to know you are getting a real chip vs a fake chip?  I'm considering buying a KSGER V2.1S controller board and handle but using it with a DIY chassis.  With a DC 5.5x2.5mm, so that the same PSU adapter could be used for both this unit and also for Pinecil or TS100. 

The AliExpress link for the controllers is here.. in the feedback, the pics appear to be real STM32 chips but I'm not sure how to tell the difference.

Is one better than the other in terms of reliability or suitability for CFW or are they both good (or bad)?

I have so many questions stands, handles, display sizes with different brand chassis but don't want to overload the thread.  I would also like to remove the potentiometer and replace it with a higher quality one.

[DavidAlfa]

None of these, they're the worst ones, with the linear regulator running off 24V, often overheating or failing.
This one is better made.
The green has a buck regulator, the blue con is not shown.
Ask for pictures before buying!
https://www.aliexpress.com/item/1005005590057638.html

But I would avoid the DIY part and buy the full thing dor just $15 more.

Personal advice for the handle: The cheap, blue plastic one.
The others are just stupid, containing aluminium, a great heat conductor, so the entire handle will be burning hot after 10 minutes.

Mini vs full is the same thing, the only difference is the PSU.

[Maksim Berezhnoj]

 :-/Oдорогой Дэвид. А ты не думал над прошивкой комбо поколдовать?

[SynthKeyWizard]

:-/Oдорогой Дэвид. А ты не думал над прошивкой комбо поколдовать?

А ты не думал переводчиком воспользоваться?

[aress124]

Hello,
The search engine doesn't give me the answer, maybe I'm using the wrong words.
I have a V3.1 R0 board, I saw that in R1 and R2 the 24V to 3.3V converter was changed to a linear stabilizer, is it worth converting your board to a linear stabilizer?
I still have the original firmware but it seems that the tip warm-up and temperature stabilization is unstable, could this be the reason?
I ordered new STM32 and when they arrive, I will upload the firmware from the forum, but I am still wondering about the voltage conversion.

[Jono434]

None of these, they're the worst ones, with the linear regulator running off 24V, often overheating or failing.
This one is better made.
The green has a buck regulator, the blue con is not shown.
Ask for pictures before buying!
https://www.aliexpress.com/item/1005005590057638.html

But I would avoid the DIY part and buy the full thing dor just $15 more.

Personal advice for the handle: The cheap, blue plastic one.
The others are just stupid, containing aluminium, a great heat conductor, so the entire handle will be burning hot after 10 minutes.

Mini vs full is the same thing, the only difference is the PSU.

Thanks, that helps a lot.

[DavidAlfa]

Hello,
I still have the original firmware but it seems that the tip warm-up and temperature stabilization is unstable, could this be the reason?

Post a picture/link of your board before asking. This one?
It's a good board, has a switching regulator convertign 24V to 5V, then a linear regulator lowering 5V to 3.3V.
This power scheme rarely caused any issue, if any.

It could just the usual behavior of new tips:
https://github.com/deividAlfa/stm32_soldering_iron_controller#temperature-unstability

You don't need to replace the stm32, we have all the original firmwares:
https://github.com/deividAlfa/stm32_soldering_iron_controller#backing-up-the-original-firmware

[aress124]

Post a picture/link of your board before asking. This one?
It's a good board, has a switching regulator convertign 24V to 5V, then a linear regulator lowering 5V to 3.3V.
This power scheme rarely caused any issue, if any.

Yes, it's the same board, except I have STM32F103.
Are you sure you're thinking about a good board? Shematic shows that 24V goes to the JW5026 pulse converter and 3.3V comes out of it.

I know the problem of unstable temperature in new tips, but my tip is already heated. Maybe it's the poor quality of the tip.
I used to have an earlier version of the station and it seems to me that the temperature was more stable and warmed up faster.

You don't need to replace the stm32, we have all the original firmwares:
https://github.com/deividAlfa/stm32_soldering_iron_controller#backing-up-the-original-firmware

Ok, thank you for information

[DavidAlfa]

I can't really tell. I don't have a board collection, only my old Quicko.
Yeah it would be better to take 5V off the converter and use a 3.3V LDO.

[aress124]

Ok, thanks for the tips, I will do it and change the firmware, I hope I will see an improvement.
Regards

[SynthKeyWizard]

Here is a simple example of installed dc/DC module directly to a PS board of my Ksger. The PS board already has a third pin under a 24v connector and all wirings are already there. You just need to install a dc/dc and replace a 2-pin connector with a 3-pin one.

[Jono434]

Here is a simple example of installed dc/DC module directly to a PS board of my Ksger. The PS board already has a third pin under a 24v connector and all wirings are already there. You just need to install a dc/dc and replace a 2-pin connector with a 3-pin one.

Wow, really elegant solution there.  I'm not sure how the circuit works though, maybe I'm misunderstanding it.

In the last few days, I've been reading around, checking datasheets and trying to understand how all of this works and what the problem is.  Coming from the PSU, 24VDC is fed to the controller board, where it separates in two directions.  One to feed the TPC8107 MOSFET and the soldering iron itself (24V is desired) and the second is to feed the STM32 chip at 3.3V.  The caps and resistors before the regulator are to reduce ripple and noise.  The regulator steps down the voltage from 24VDC to 3.3VDC.  On some boards linear regulators are used (bad) and sometimes buck converters are used (good), because they use a switching method and are more efficient and generate less heat.

The custom firmware draws more current from the regulator because it uses a larger font.  It makes the regulator get hotter than usual, so the regulator chip shuts down due to overheating.  One solution is to switch to a buck converter, and the other is to add a resistor or heatsink.  Or just turn the brightness down on the display.  The hard thing with this is every board seems to use a different regulator and every board/schematic is different.

In the solution above, is the 3rd pin the 3.3VDC?  So you have 24VDC, 3.3VDC and negative/ground?  And then bypassing the regulator/resistors/capacitors on the controller board, and feeding the STM32 directly with 3.3V?  Or am I completely misunderstanding this?  Would your solution work with all KSGER boards, or only with your controller board?  Every board I see looks different.

[Jono434]

I haven't messaged the AliExpress seller about the controller board yet, I'll do that in the near future.  But I've had bad luck in the last week or two, when it comes to communicating with sellers and getting back good answers to questions.  I think the language barrier is the biggest thing.  And without pics, it's hard to know what you are getting.  Or you could order the same thing twice and get two different things.

I've been looking at the idea of modding one of the other two boards (green or blue), because these are the boards that KSGER official store sells and I'm hoping they would still be available later in the future.  Between the two, I've been favouring the green board, despite how crude the design is, there are some advantages.

With the GX12 connectors, I've had a hard time finding a high quality version.  They all seem to be cheap Chinese ones.  And they can't be desoldered without damaging the connector.  So it seems easier to chassis mount the connector and have a bunch of spare connectors in a draw somewhere.  Also, the KSGER design puts the connector lower than the center line and this makes it incompatible with other front panels.  Chassis mounting the connector would mean that it is floating and could be used with other front panels.  There is an option to cut the blue PCB and do the same thing.  On the blue board, the thermal resistance varies between 46C and 90C depending on the board it's mounted to.  The blue PCB is a more dense board and the pins of the regulator are reversed.  The pin orientation makes it hard to plan a mod in advance because some of the components are on the other side of the board, where the display is.  It's hard to know if there would be access or not.

The green board is crude and basic, but it makes it easier to mod.  At first, I looked at replacing the regulator but didn't have any luck finding anything better than what is in there.  On the datasheet the thermal resistance is 62C.  Other alternatives, were cheaper but the thermal resistance was only 50C.  So it seems better to keep the SPX2954A in there.

I've explored the idea of using a 2W Metal Oxide resistor to lower the input voltage.  I did some simple diagrams at first, but realized the scale was way off.  These boards and all of the components are really small.  I looked at some datasheets and tried to get a better idea of the scale, and tried to draw it closer to real scale.  I attached a diagram with two options, with resistor before or after the cap.  The resistor could be soldered to the input pin of the SPX2954A, or from the empty pad/trace from C2.  I generally like to favour strong mechanical connections, but also.. I'm worried that soldering directly to the regulator could damage it.  So maybe laying the lead wire flat and soldering to the pad would be less risk.

The diagram positioning isn't literal, it's more to make it easier to read with the labels.  For example, the positioning of the cap or resistor could be re-orientated to make it fit better.  The components could float in the air 1cm away from the PCB, or be quite close, depending on internal space available.  I found some super small heatsinks locally.. 4.5x6mm and a 6.5x6.5mm.  The SPX2954A is 6.5mm wide.  It could be an option to further reduce temperatures if the resistor alone isn't enough.  I don't even know how much heat such a small piece of metal would dissipate, but the heatsink could be attached with a mixture of CPU thermal compound and thermal glue.  I've done this before and it works ok.

The mod would have to be right first time because apparently these PCB's are so cheap, they can't handle being desoldered and resoldered much.  The thermal glue would also be permanent, but those boards are pretty cheap, so it's easy to have a spare lying around.

I can't find the capacitance value of C2.  Would it be better measure it and then order the same value, or is there a value that could be suggested that would work well with this 2W 120R resistor?  Would 3W 150R or 200R be better?

My main question is if I am understanding the problem properly, and secondly would this solution work if these components can fit?  And is it better to have the cap after the resistor?  It usually looks that way in schematics.  What uF range would be good for the cap?

[SynthKeyWizard]

In the solution above, is the 3rd pin the 3.3VDC?  So you have 24VDC, 3.3VDC and negative/ground?  And then bypassing the regulator/resistors/capacitors on the controller board, and feeding the STM32 directly with 3.3V?  Or am I completely misunderstanding this?  Would your solution work with all KSGER boards, or only with your controller board?  Every board I see looks different.

The third pin is 5VDC from dc/dc converter board. I have a 24VDC, GND and 5VDC at this connector.
5VDC then fed directly to a linear 3.3V regulator input. In this case it dissipates much less power into heat, and works OK.
 It is easy to calculate:
If you fed 24V directly to a regulator, and stm32 consumes 300mA for example, then we have a voltage difference 24-3,3=20,7v.
20,7v × 0,3A = 6,21 watts of a energy must be dissipated. If you fed only 5 volts, then regulator dissipates only 0,51 watt.

This solution would work everywhere. The only thing that not every board have a comfortable spot to solder a dc/dc subboard to it like in my case. But you can solder it where you want, it will work.

[Jono434]

In the solution above, is the 3rd pin the 3.3VDC?  So you have 24VDC, 3.3VDC and negative/ground?  And then bypassing the regulator/resistors/capacitors on the controller board, and feeding the STM32 directly with 3.3V?  Or am I completely misunderstanding this?  Would your solution work with all KSGER boards, or only with your controller board?  Every board I see looks different.

The third pin is 5VDC from dc/dc converter board. I have a 24VDC, GND and 5VDC at this connector.
5VDC then fed directly to a linear 3.3V regulator input. In this case it dissipates much less power into heat, and works OK.
 It is easy to calculate:
If you fed 24V directly to a regulator, and stm32 consumes 300mA for example, then we have a voltage difference 24-3,3=20,7v.
20,7v × 0,3A = 6,21 watts of a energy must be dissipated. If you fed only 5 volts, then regulator dissipates only 0,51 watt.

This solution would work everywhere. The only thing that not every board have a comfortable spot to solder a dc/dc subboard to it like in my case. But you can solder it where you want, it will work.

Oh right.  That makes a lot of sense.  Dropping from 5V down to 3.3V must be very on the regulator.  Having reread the Github Manual, it also mentions this but at the time of posting, I'd forgotten.  I actually have all that info saved to a text file in case the Github ever disappears.  It probably won't, but sometimes stuff online does that.

Your math is very helpful and easy to understand.  And using 300ma seems like a good general figure.  Does anyone know what the actual current draw is in total?  I haven't had any luck finding specific numbers, but the datasheet for STM32F101C8T6 says 150ma max total draw.  The display is also drawing some current as well though.  Googling small OLED displays, it seems to be around 20-50ma depending on the screen size etc.  So 200ma but with CFW and large font, maybe it's more.

[SynthKeyWizard]

I can measure a consumption a bit later today. I'll give you know.

[Jono434]

I've been exploring the idea of using Meanwell power supply instead of KSGER, QUICKO or any of the others out there.  Reading around, I see random component failures on all of them.  Even Auxun T3A. A PSU with a better design, more reliable etc.  There's many different options in terms of sizes or options.. I should probably make a separate thread in terms of the DIY side of things.

But in terms of the regulator.. I realized that there is a 12V 0.5A fan output for all three sizes that I am looking at.  It says in the datasheet that it is for fan only, but the nature of these solder stations is that they don't stress the PSU very much.  It's brief spikes of high load, rather than constant high load.  For medical power supplies I can understand the warning, but for a hobby soldering station it should probably be fine to use the fan output for a 3.3V STM32/Display?

If using that 12V 0.5A output into the input of the regulator.. it seems like it could solve this problem without any need for PCB mods.  A resistor could also be put in series and covered in a few layers of heatshrink to drop it down to 10V or 8V for example.  The electronics math isn't my strongpoint, I usually just try a few different values to see what works and feel if the resistor is getting hot or not.

Using SynthKeyWizard's math above.. 12V down to 3.3V would be 8.7V X 0.3A =  2.61W.  The right value of resistor in series could drop it to 8V.  4.7V X 0.3A = 1.41W.

Another advantage of this method would be running the 2 pin fan cable to a switch on the front panel that could connect to the controller board and turn the display on or off.  For example.. with a 24V output at the rear of the case, the enclosure could be used as a standalone power supply for Pinecil or TS100 or TS101.  The display could be turned off when using the Pinecil, and turned on when controlling the KSGER controller board.  And the GX12 connector could be mounted on the rear of the enclosure similar to the Aixun T3B.

[Jono434]

I can measure a consumption a bit later today. I'll give you know.

Thanks, that'd be interesting information to know.

[xfint34]

The KSGER blue board (v2.1s) has two linear regulators in series. An L78L08 converts 24V to 8V. An AMS1117 converts 8V to 3.3V. In addition there's a 6.8 ohm resistor in series too.

I have not seen anyone mention fatal problems with that other than the regulators do get a bit hot.
See the thread about it here:
https://www.eevblog.com/forum/reviews/t12-stm32-v2-1s-soldering-station-controller-schematic-etc/

That's the board I have from a full unit with the power supply. I do not use CFW so I can't attest to its performance. I don't have a thermometer to do calibration. I flashed the CFW once before just to see. And then flashed back to the stock firmware.

[DavidAlfa]

A 3W 220 Ohm resistor should work pretty well up to 90mA, which I doubt it consumes, more likely around 50mA.

Code: [Select]

I (mA) LDO ONLY PWR RES (W) PWR LDO (W)
10 0,21 0,02 0,19
20 0,41 0,09 0,33
30 0,62 0,20 0,42
40 0,83 0,35 0,48
50 1,04 0,55 0,49
60 1,24 0,79 0,45
70 1,45 1,08 0,37
80 1,66 1,41 0,25
90 1,86 1,78 0,08

For 50mA, you might increase it a bit more to 330 Ohms, the LDO will only have to get rid of 0.2W.
An excessively high resistor will cause power strain issues (If the current raises too much).
If you have strange behaviour, resetting... lower the resistor in small steps until it works fine.
Like 330->300->270->240->220 Ohms.

The other cheap way is to screw a isolated 7805 (TO220F package) to the aluminum housing, the heat will go into the casing, and the onboard LDO only has to work dropping 5V to 3.3V.

[SynthKeyWizard]

I can measure a consumption a bit later today. I'll give you know.

Thanks, that'd be interesting information to know.

Here is my Ksger consumption measured with Owon B41t+.

[Jono434]

The KSGER blue board (v2.1s) has two linear regulators in series. An L78L08 converts 24V to 8V. An AMS1117 converts 8V to 3.3V. In addition there's a 6.8 ohm resistor in series too.

I have not seen anyone mention fatal problems with that other than the regulators do get a bit hot.
See the thread about it here:
https://www.eevblog.com/forum/reviews/t12-stm32-v2-1s-soldering-station-controller-schematic-etc/

That's the board I have from a full unit with the power supply. I do not use CFW so I can't attest to its performance. I don't have a thermometer to do calibration. I flashed the CFW once before just to see. And then flashed back to the stock firmware.

Learning this stuff in the last week is all a bit of a blur, but I remember seeing a pic of a blown resistor on the power board due to failed MOSFET.  It was a black PCB, the green PCB ones seem ok if you modify the heatsink and add a ground wire.  The main filter cap can be dodgy though and it's hard to find quality replacements in that size.  Also, there's a Aixun T3A video on youtube with a failed diode (it was intermittent and resoldering it helped, but then it failed weeks later).  There's been a few more I've seen mentioned in text but hard to remember off the top of my head.  Generally, the failure rate isn't that high, but for example quality stuff can last 10+ years.  And it seems hard to expect the same from these cheaper units.  It's hard to know for sure because most are only a few years old.

Thanks for that info it helps think about some extra possibilities.  I've messaged the seller from AliExpress that David mentioned, hopefully I get a reply with some pics.  There is a sale on AliExpress at the moment, it'd be nice to order some of this stuff before the sale ends.

I've been reading through some of the thread and all of the talk about regulators makes a lot of sense.  Trying to keep heat on the controller board as low as possible, lower noise, better performance, reliability etc.