KSGER STM32 T12 soldering stations

[redwire]

Picked up a soldering station with H/W Ver 2.1S and S/W 2.10 from KSGER store on Ali. Build quality is very good, price was high.

The handle has a NTC thermistor inside that doesn't seem to do anything.
What do your station do?

Bottom right corner of OLED shows a temperature 40°C all the time. Even with a handle plugged in or not, or shorting the "N" pin to GND.

Menu 04. Cold End Adj. -> NTC-> 40°C {dial in}

The temperature I dial-in is always displayed.

I drew a full schematic of the controller board and 24VDC power supply.
The NTC input is just a pullup resistor going to the MCU input. So the hardware seems fine, it must be crappy firmware?


RTC battery was DOA and a new one also died quickly. Took out the stupid resistor R10 from this video: https://www.youtube.com/watch?v=B-xWhPcr6jk

[sam_sam_sam]

Thanks for sharing it I am going to get one in couple of weeks I am going to remove the switch and put a non conducting cardboard ( under the switch ) that is used battery pack so I do not have that problem

[redwire]

Here's the schematic I drew of the KSGER soldering station 24V PSU for service and repair.
Soldering Power V2.04 and pig with bull horns icon on the silkscreen.
KSGER 96W 24V 5A Electric Power Supply Unit For STM32 STC OLED T12 Digital DIY Soldering Station Controller.

It has some clearance mistakes running +HV trace underneath heatsink and a thin PE ground trace, no connection to the chassis
The TL431 is set to over 25VDC output and it has no bias resistor. The snubber cap is pretty cheap. Could not find datasheets on the CR8642S PWM controller.
Earlier version PSU's like V2.01, V2.02, V2.03 are the same but no pads for a mystery DPAK part or different DC connector.

[stj]

are the heatsinks insulated? the semi's are.

it's a good design - caps well away from the sinks - unlike japanese shit i used to fix.

[redwire]

Semi's are isolated TO-220 parts (fullpack), but the heatsink mounting pin pads have very little clearance to the fills. So the path is PCB trace->heatsink->mounting pads-> nearby pours.
I'll put Kapton tape over the two traces if I can fit that in.

The older version 24VDC PSU (left) is apparently not very good. Primary capacitors are low value 2x22uF, 1A diodes instead of 8A bridge, no 3A fuse but a 1R resistor in heatshrink. The Schottky rectifier runs 112C at steady 3A. Quicko sells them too.

But it has to put out 3A for short periods.

[sam_sam_sam]

I have to wonder how long this switching power supply would last if you ran it 5 or 6 hours a day 5 days a week

[redwire]

It's a good power supply. Just the two issues with 100V snubber cap and heatsink on the HV pcb trace, and it takes a few seconds to power up which is strange. KSGER sells it with their better soldering stations.
Actually you don't need 3A for very long, just bursts.

There are a few other T12 soldering station PSU's out there- this one pictured below is what I consider not very good. KSGER, Quicko, Banggood sell them too. At $12.50 it is $1.50 USD cheaper than the previous one.
It has 1A rectifier diodes, small heatsinks, very small primary caps 22uF 450V each, super thin PE GND trace. Runs hot, high stress on parts and not for heavy duty use.

[sam_sam_sam]

Which of one of photos is a good quality one let see if the one I have is a good one or I should use the medical grade switching power supply that I just bought and change the low voltage power cord on

[redwire]

Left is the lower quality one, right is the better quality one.
Before I was mentioning the right better quality one that had a few weaknesses.

[stj]

Originally Posted by redwire

It's a good power supply. Just the two issues with 100V snubber cap and heatsink on the HV pcb trace, and it takes a few seconds to power up.

can you remove a few mm of the heatsink where it crosses the track??

btw, i dont think it does take a few seconds to power up,
i saw a video - may have been VoltLog, with a meter on the output.
the psu startup was instant but the controller firmware took a few seconds to init the oled display - maybe while it averaged some samples from the thermocouples.

[redwire]

A person could grind off the fin on the heatsink for more clearance. I used Kapton tape and lifted the heatsink up a bit, it does have sharp edges which would cut solder mask.

It really does take several seconds to light up the red LED on the 24V rail. Not as long if you switch it off and then on.
I can't find datasheets for the CR8642S PWM IC. It has 1.5MEG HV DC feed and 22uF 50V cap, so I'll bet it's taking its time charging that to start up. Or a 120VAC version needs a lower resistor value. It's kinda weird. Voltlog is 240VAC so maybe that is covering it.

edit: Changed out the 22uF 50V 105C Chong-X capacitor. It tested fine. No difference in 3 second start-up time. Power-off it quickly goes from 16V to 9V and takes many minutes to discharge to zero. When it still has charge, next start-up is quicker, maybe 1.5 seconds.

[sam_sam_sam]

I have the version the one on the right side of the photo

[redwire]

Originally Posted by sam_sam_sam

I have the version the one on the right side of the photo

That's the better one, it's still pretty good for stuff out of china, with the usual couple of weak points.
I think most important is adding a PE ground wire from the IEC outlet to the chassis.

[sam_sam_sam]

Yes I agree with you about that I am surprised that this was done

[redwire]

Anyone know the T12 handle components? They ship two NTC thermistors and two different motion switches, a ball shake switch and a mercury switch.

The black blob NTC I have only seen on a controller board and the glass MF58 part I used. The 0.1uF cap sometimes across the NTC, other times across the heater, the handle wiring is always different.

Got a replacement T12 controller board V2.1S and the cold-junction temperature is now working. Display shows !24°C if the handle is unplugged, where the old one was just stuck at what was entered at the cal menu.
Impossible to desolder the GX12 connector, I had to trash it to get the board out.

[sam_sam_sam]

That funny because the version I have which is the 2.1 if you do not have the soldering iron hook up it shows error message on it which I prefer and not some temperature

Originally Posted by redwire

Impossible to desolder the GX12 connector, I had to trash it to get the board out.

Are you talking about the connector that the soldering iron connects to the controller board

Because I was able to remove the one on one of mine it was a little bit difficult but not impossible to do I had to more solder on the joints a couple times and then I was able to remove it

But come to think about I also think I had to use a temperature controlled hot air gun to help it along as well

[redwire]

The GX12 has 1.6mm pins and the desoldering tools I have couldn't clear the holes. No tips fit over the pins close spacing. I thought hot air would melt everything around it like the JST power connector and capacitor, buzzer?

The temperature I am talking about is always displayed on the bottom right, even if the handle is unplugged and ERROR is up.
When the handle is unplugged it goes to the default value from menu 4.
I suspect the firmware 2.10 got a sneak update or they programmed the STM32 wrong, as the hardware works fine on the defective controller.

[sam_sam_sam]

Originally Posted by redwire

The GX12 has 1.6mm pins and the desoldering tools I have couldn't clear the holes. No tips fit over the pins close spacing. I thought hot air would melt everything around it like the JST power connector and capacitor, buzzer?

When I remove that connector I use my Desoldiering Gun (ZD-917 ) and a temperature hot air gun and with this setup it still took some doing

But if you have a hot air gun put a very small nose cone on air gun and point it away from the main controller board and go around in a circle until the solder melts and bang on the table to nock most of solder off it you might have to do a few times or when the solder melts pry up on the board until you clear the board

[redwire]

When I pried, the pins slid right out of the GX12. They do not have collars (like in pic) to grab the epoxy, it's a cheap connector from KSGER.

[roman44d]

Originally Posted by redwire

could not find datasheets on the cr8642s pwm controller.

sg6842