Brand New KSGER T12 Glowing Red

[ABSTA]

Hi friends


I just received my KSGER T12 Solder Station.

When i turned on ,it was mentioning error on the screen.

after few seconds The tip get glowing red until i unplug the station .

The tip is okay, i tried it with multimeter on beeper mode.

i hope to find some help with you please

[CaptDon]

Looks like the thermal control device (scr, triac, thermostatic switch) is shorted.
I am not familiar with the make and model. Chinese?? Good Luck or should we
say Good Ruck.

[Manul]

I imagine the tip material to be damaged by such heat. So it will not be a good soldering tip anymore. Also possibly temperature sensor and heating element may have experienced some damage. Would be best to return to seller for these reasons alone if possible. If there is no other way and you need to fix that, look how the current to heating element is controlled. Something is probably shorted, like CaptDon said above.

[ABSTA]

Looks like the thermal control device (scr, triac, thermostatic switch) is shorted.
I am not familiar with the make and model. Chinese?? Good Luck or should we
say Good Ruck.

yes it's a chinese Model  :-[
Yes it seems to be a faulty controller... i will try to find the shorted component ...

So it begins with Testing : Scr,Triac ans the thermostatic switch :-)

thanks to your help ...

if anyone knows this station ?

[Manul]

I do not know this station, but schematic looks simple. Also found another schematic on the net, it is not exactly like what you posted, but basic structure is same. So it seems Q1 (P-ch mosfet TPC8107) is driving heating element and Q2 is driving Q1.

[wraper]

It should be MOSFET, not triac since KSGER uses SMPS.

[ABSTA]

It should be MOSFET, not triac since KSGER uses SMPS.

Hey

I traced some pins of the MOSFET

it seems to be an ao4354

AND All pins are shorted to the + side

there is a sample of my tarces whith multimeter (beeper mode)

[ABSTA]

 

[Manul]

it seems to be an ao4354

?? I see AO4409.

[Manul]

If I would change it, I would choose something more trustful like IRF9321 for example. Also Vishay makes some nice mosfets.

At least good news is that it is p-ch mosfet driven by Q2, so MCU PWM pin should not be damaged. Should be replace and go situation.

[ABSTA]

it seems to be an ao4354

?? I see AO4409.

Yes ,sorry i was wrong its an ao4409 ,

How to make sure it is faulty? ( how i can mesure it with multimeter )

Thank you ...

[Manul]

With what you measured with beeper mode it is pretty much guaranteed to be faulty. Also it agrees with the symptoms of heating element beeing on full power all the time. There is not much else to do, just to find a replacement and perform a quality soldering, while taking some care to not damage new one with static electricity (do not wear synthetic clothes for example).

[floobydust]

I drew a KSGER T12 STM32 controller V2.1S schematic here:
https://www.eevblog.com/forum/reviews/t12-stm32-v2-1s-soldering-station-controller-schematic-etc/msg2463234/#msg2463234
it looks the same as what you have OLED-V3.0 but different designators. I'm not sure what R0 is for.

AO4409 is rated 15A 30V and 20V on the gate, and the power supplies are usually 25V, so not much margin for spikes. I add a diode across the heating element. My controller had TPC8107 which has a gate zener.
The driver Q2 is also a mosfet, so check that it is not shorted D-S as well.
Your power connector has bad solder joints and not sure why it's on the front side. Those JST connectors aren't really rated for 3A so people just hardwire it.

[ABSTA]

I drew a KSGER T12 STM32 controller V2.1S schematic here:
https://www.eevblog.com/forum/reviews/t12-stm32-v2-1s-soldering-station-controller-schematic-etc/msg2463234/#msg2463234
it looks the same as what you have OLED-V3.0 but different designators. I'm not sure what R0 is for.

AO4409 is rated 15A 30V and 20V on the gate, and the power supplies are usually 25V, so not much margin for spikes. I add a diode across the heating element. My controller had TPC8107 which has a gate zener.
The driver Q2 is also a mosfet, so check that it is not shorted D-S as well.
Your power connector has bad solder joints and not sure why it's on the front side. Those JST connectors aren't really rated for 3A so people just hardwire it.

Q2 Isn't short

i tried to measure VOLTAGE between drain and source of the mofset and i don't find Any 24v,20v or 30v
i put it off the circuit , and (maybe i did a mistake) when i turn on the station she seems to works and displays "INVERT"

its definitely a faulty MOSFET

to close the topic , please ,where i can find a website to order it (shipping to France) and what are the other transistors  equivalent to this one

[Mp3]

I drew a KSGER T12 STM32 controller V2.1S schematic here:
https://www.eevblog.com/forum/reviews/t12-stm32-v2-1s-soldering-station-controller-schematic-etc/msg2463234/#msg2463234
it looks the same as what you have OLED-V3.0 but different designators. I'm not sure what R0 is for.

AO4409 is rated 15A 30V and 20V on the gate, and the power supplies are usually 25V, so not much margin for spikes. I add a diode across the heating element. My controller had TPC8107 which has a gate zener.
The driver Q2 is also a mosfet, so check that it is not shorted D-S as well.
Your power connector has bad solder joints and not sure why it's on the front side. Those JST connectors aren't really rated for 3A so people just hardwire it.

Q2 Isn't short

i tried to measure VOLTAGE between drain and source of the mofset and i don't find Any 24v,20v or 30v
i put it off the circuit , and (maybe i did a mistake) when i turn on the station she seems to works and displays "INVERT"

its definitely a faulty MOSFET

to close the topic , please ,where i can find a website to order it (shipping to France) and what are the other transistors  equivalent to this one

You can't return the KSGER to the seller or have them compensate you?

TME in poland seems good, i ordered from them recently.

[BreakingOhmsLaw]

Holy shit, you should really consider to get something more reputable. Thermal runaway protection is no small issue, and the lack thereof seems a design flaw. That said, i'm not sure the likes of Weller and other major brands have runaway protection, but i've yet to read or hear from red-hot tips on quality solder stations.
Point in case is: you will *never* be able to leave this puppy unattended, even for just minutes, even if you get it replaced.
Better spend a couple of bucks more, perhaps on a used brand name unit than to burn down your house. That's just scary. Good luck mate.

[Manul]

to close the topic , please ,where i can find a website to order it (shipping to France) and what are the other transistors  equivalent to this one

You can choose IRF9317PbF, IRF9321PbF they should be ok and even better then original. In Europe I recommend Farnell.

[wraper]

Holy shit, you should really consider to get something more reputable. Thermal runaway protection is no small issue, and the lack thereof seems a design flaw. That said, i'm not sure the likes of Weller and other major brands have runaway protection, but i've yet to read or hear from red-hot tips on quality solder stations.

None of them have. And basically any other station will do the same if pass element fails short. Particularly Hakko 951 will glow red if triac fails as it uses the same cartridge.

[Ian.M]

If a station has a beeper and is MCU controlled, it would be fairly simple to have the firmware do some tests on initial powerup and if it fails, sound a repeating alarm tone and display a message like "System Fault - Disconnect Supply Immediately".  Ideally it would sense element current to determine if the iron is correctly connected and under control, but it would be possible to get a reasonable idea if there is a fault just from the bit temperature sensor behaviour quickly enough to sound the alarm well before the element overheats.

e.g. start by checking the bit is near ambient and there is no temperature rise with the element commanded off, then command a moderate duty cycle and check the temperature rises as expected, then close the control loop and check it can stabilise at lets say 50 deg C above ambient, before releasing it to normal operation.

[TheMG]

You were sold a defective product, why don't you contact the seller and get it exchanged for a new one?

I certainly wouldn't trust it anymore. If anything, the temperature sensing element could have been damaged meaning the temperature calibration could be way off. The heater's reliability would be affected too, it's not designed to ever get that hot.

[floobydust]

Chance it might not be a shorted MOSFET, if the thermocouple is wired backwards the controller will always think it is too cool and keep the heat on. Crappy sofware should never put full heat on for over a minute say. It could also be a wrongly wired wand if the cold-junction compensation thermistor is reading something wrong.

What temperature is it displaying? I would check the temps are not wrong.

[wraper]

Chance it might not be a shorted MOSFET, if the thermocouple is wired backwards the controller will always think it is too cool and keep the heat on. Crappy sofware should never put full heat on for over a minute say. It could also be a wrongly wired wand if the cold-junction compensation thermistor is reading something wrong.

What temperature is it displaying? I would check the temps are not wrong.

Maybe read opening post and some further info?

[ABSTA]

It was definitely a faulty Mosfet. I replaced it by an IRF9321PbF and it's fixed now.

This station is amazing


thank you for your support 

[wf789]

Hi I am having the same issue but the SMD cap @ C20 on the power board has blown for me... Would anyone happen to know the value? I can't find it anywhere online. The board ver is V2.01

Thanks

[floobydust]

Do you mean C20 on the 24VDC power supply? It's an 1uF across the output, not critical, put in anything 1-2.2uF.
Original is 25V running at 25V  so put in a 50V part for less smoke.

[wf789]

Do you mean C20 on the 24VDC power supply? It's an 1uF across the output, not critical, put in anything 1-2.2uF.
Original is 25V running at 25V  so put in a 50V part for less smoke.

Hi Floobydust thanks for the reply. Correct, here is a pic of the blown cap:



Great i'll have a look in my caps box to see if I have one within this range of values, if not I'll order it and hopefully it will be back to normal again!

[wf789]

Do you mean C20 on the 24VDC power supply? It's an 1uF across the output, not critical, put in anything 1-2.2uF.
Original is 25V running at 25V  so put in a 50V part for less smoke.

Hi Floobydust thanks for the reply. Correct, here is a pic of the blown cap:

Great i'll have a look in my caps box to see if I have one within this range of values, if not I'll order it and hopefully it will be back to normal again!

I replaced the cap with a 1uf 50v but the problem still occurs. It shows as ~348 degrees celcius at all times no matter what I set it to (but the iron is cold). If I set it to 350 or higher, the iron will start to heat up, but won't stop and glow orange.

The problem must be else where causing this... I checked the MOSFET as per earlier in the thread, and only the pins highlighted in green are shorted to '+' i suppose that is normal ?

[wraper]

Do you mean C20 on the 24VDC power supply? It's an 1uF across the output, not critical, put in anything 1-2.2uF.
Original is 25V running at 25V  so put in a 50V part for less smoke.

How do you even know it's voltage rating? There are no markings.

I replaced the cap with a 1uf 50v but the problem still occurs. It shows as ~348 degrees celcius at all times no matter what I set it to (but the iron is cold). If I set it to 350 or higher, the iron will start to heat up, but won't stop and glow orange.

IMHO there is something wrong with temperature sensing circuit. Apparently it reads about 350^oC regardless of actual temperature. Thus when set above it, iron heats up uncontrollably not being able to reach the set-point, when set below, there is no heating as temperature readout is above the set-point. You need to check the circuit where temperature reading happens. Does it continue to do so even when cartridge is disconnected?

[wf789]

I replaced the cap with a 1uf 50v but the problem still occurs. It shows as ~348 degrees celcius at all times no matter what I set it to (but the iron is cold). If I set it to 350 or higher, the iron will start to heat up, but won't stop and glow orange.

IMHO there is something wrong with temperature reading. Apparently it reads about 350^oC regardless of actual temperature. Thus when set above it, iron heats up uncontrollably, when set below, there is no heating as temperature readout is above set-point.

I see. It must be something on the temperature control board, or maybe the iron itself? I'm not sure but I'd guess the former. I think that MOSFET which the OP had the problem with is OK for me (see prev post) so I'm not really sure where to look next, I'm only a tinkerer so not that experienced. I hoped it would be a quick fix 

You need to check the circuit where temperature reading happens. Does it continue to do so even when cartridge is disconnected?

What do you mean by the cartridge?

[wraper]

What do you mean by the cartridge?

[wf789]

What do you mean by the cartridge?

Ah, yes. I tried 3 different tips, the problem still occured 

I just tried without. Instead of 350 degrees it says 320 degrees, but when I adjust the temperature, it won't move -- just stays at 320.

[floobydust]

Check (depends on your F/W rev) Setup Menu 04. or Settings Menu 05. "Cold End" is not scrambled or reading a silly temperature. Wand plugged in, it should read room temperature if the CJC thermistor bead in the wand is wired properly, regardless if a tip is plugged in.
Unplugged wand: The controller should display ERROR and the LED will only be lit if a shorted MOSFET.
Since the thermocouple input is floating and picking up noise, the controllers freak out and display ERROR or flicker between run/error mode, this is normal.

Otherwise I would suspect the thermocouple op-amp is not working.
You can unplug the wand and connect a 10k resistor to GND from heater connector pin at the top T12+ or pin 5 on the GX12. With that very low mV input, a low temperature should be displayed and the controller will think it's ok to run and turn on the heater (which is all unplugged).
If you still get that silly high temperature reading, check the SGM8551XN op-amp input (pin 3/LED) is around 0V and the output pin 1 is low under say 250mV.
edit: op-amp output is about 5mV/°C so 350°C is about 1.74V to the A/D. Room temp 25°C should be about 125mV

"OLED-V3.0" schematic is basically the same circuit as V2.1S, with a different 6-pin OLED, no pump control mosfets, different reference designators etc.

[wf789]

Check (depends on your F/W rev) Setup Menu 04. or Settings Menu 05. "Cold End" is not scrambled or reading a silly temperature. Wand plugged in, it should read room temperature if the CJC thermistor bead in the wand is wired properly, regardless if a tip is plugged in.
Unplugged wand: The controller should display ERROR and the LED will only be lit if a shorted MOSFET.
Since the thermocouple input is floating and picking up noise, the controllers freak out and display ERROR or flicker between run/error mode, this is normal.

Otherwise I would suspect the thermocouple op-amp is not working.
You can unplug the wand and connect a 10k resistor to GND from heater connector pin at the top T12+ or pin 5 on the GX12. With that very low mV input, a low temperature should be displayed and the controller will think it's ok to run and turn on the heater (which is all unplugged).
If you still get that silly high temperature reading, check the SGM8551XN op-amp input (pin 3/LED) is around 0V and the output pin 1 is low under say 250mV.

"OLED-V3.0" schematic is basically the same circuit as V2.1S, with a different 6-pin OLED, no pump control mosfets, different reference designators etc.

When the wand is plugged in with no tip, it shows the room temp (24-25 degrees).
When the wand is not plugged in, it does not show "ERROR" but just "325C". The room temp stays at 23 degrees and has a "!" instead of the thermometer symbol next to the degrees.

I will try the rest of what you have said tomorrow. Just to make sure I've got it right (or not), is the heater pin the "+" pin on the 5-pin heater connector?

The pic of the OLED-V3.0 board in my previous posts was just one I took from the main OP's post to highlight the pins that I previously tested with my multimeter to the "+" pin on the heater.

The pic attached below is the board from my solder station. The main board rev is V2.01

Thanks 

[wf789]

Otherwise I would suspect the thermocouple op-amp is not working.
You can unplug the wand and connect a 10k resistor to GND from heater connector pin at the top T12+ or pin 5 on the GX12. With that very low mV input, a low temperature should be displayed and the controller will think it's ok to run and turn on the heater (which is all unplugged).
If you still get that silly high temperature reading, check the SGM8551XN op-amp input (pin 3/LED) is around 0V and the output pin 1 is low under say 250mV.

"OLED-V3.0" schematic is basically the same circuit as V2.1S, with a different 6-pin OLED, no pump control mosfets, different reference designators etc.

I put a 10k resistor between GND and the heater connector pin at the top and the temperature was still high and not changing. (345-350c)
I think the op-amp you are referring to is U3 on my board? I measured it and pin-3 (input) measured 0V but the pin-1 (output) was 1.74V. I assume this op-amp is the problem then? Since you are saying it should be ~250mV.

I looked at your schematic and saw OPA336NA, can I use that as a replacement part? I'm ordering some parts for something else really soon from Mouser so could grab one of those as well.
https://www.mouser.co.uk/ProductDetail/Texas-Instruments/OPA336NA-3K?qs=7nS3%252BbEUL6vwF%252Byt2%2F21Rw%3D%3D

Cheers!

[floobydust]

From those measurements, the thermocouple op-amp U3 is not working properly.
The T12 controller has a few odd requirements for the thermocouple amp.
I'm hesitant to give substitute op-amp numbers because I have not tried them and datasheet shopping can still get it wrong. OPA336NA known to work too. Maybe look at OPA376 if you can't find that.
SGM8551XN has good specs:
rail-rail in and out, 12uV offset, 13pA IB, 0.1msec overload recovery time, 0.84V/usec slew rate and noisy 53nV/sqrt(Hz). It runs off 3.3V with 4k7 to the thermocouple. The op-amp gets saturated while the heater is on, then it has to recover and give the correct TC voltage, which is the trap for phase-reversal as well.
The op-amp input has to float high when the tip is disconnected because somebody saved $0.01 omitting a 1MEG pullup resistor. Or else it will not properly detect a disconnected wand/heater.

I forgot to say, be careful when attaching the 10k to the heater output, if you short it to ground when the mosfet is on, it can damage it.

[dreamcat4]

From those measurements, the thermocouple op-amp U3 is not working properly.
The T12 controller has a few odd requirements for the thermocouple amp.

I appear to have the same PCB here, with same Op-amp on it. This opamp must be replaced, for sure. This Russian guy already has tried...

Information Source:

https://radiokot.ru/forum/viewtopic.php?p=3753771&sid=6d6a79087d7d8bd346c6b744de8b10bc#p3753771

Translated:

User: Kst26

Tue Dec 03, 2019 01:27:36 p.m.

Yes, in it. I also bought a blue Ve2.12S controller with a 61LJV opamp, the difference between the temperature on the display and on the sting was 50-53 degrees. The difference was almost constant in the temperature range of 150-350 degrees and decreased at a temperature close to 450. Moreover, if you try to calibrate the difference decreases to 30-33 degrees and on a cold sting the controller instead of room temperature starts to show 9-13 degrees (although in room 22) .
The op was replaced by the AD8605ARTZ-REEL7, without calibration, the difference became 3-10 degrees, and with calibration + -2 degrees.
All temperature measurements were made with a thermocouple multimeter.

Unfortunately he does not mention which tip / handle. If it's T12 or JBC. Which I guess then require more digging. There are other discussions on the same Russian thread, which is much longer. More little details like this over there. Unfortunately it's a bit difficult to search / find.

[wf789]

Sorry I'm a bit late to reply but thanks for your responses.

@dreamcat4, although mine is 2.01 not 2.1S, the problem sounds a bit similar to mine. Hopefully replacing this chip will work.

@floobydust Well... I've ordered some replacement of the SGM8551XN chip. They will arrive in 1-2 weeks hopefully. I'll replace it and return with the result! & yes I did check when I tested what you previously mentioned and there was no shorts to GND

Cheers

[wf789]

From those measurements, the thermocouple op-amp U3 is not working properly.
The T12 controller has a few odd requirements for the thermocouple amp.
I'm hesitant to give substitute op-amp numbers because I have not tried them and datasheet shopping can still get it wrong. OPA336NA known to work too. Maybe look at OPA376 if you can't find that.
SGM8551XN has good specs:
rail-rail in and out, 12uV offset, 13pA IB, 0.1msec overload recovery time, 0.84V/usec slew rate and noisy 53nV/sqrt(Hz). It runs off 3.3V with 4k7 to the thermocouple. The op-amp gets saturated while the heater is on, then it has to recover and give the correct TC voltage, which is the trap for phase-reversal as well.
The op-amp input has to float high when the tip is disconnected because somebody saved $0.01 omitting a 1MEG pullup resistor. Or else it will not properly detect a disconnected wand/heater.

I forgot to say, be careful when attaching the 10k to the heater output, if you short it to ground when the mosfet is on, it can damage it.

Hey, just want to say that the replacement chip arrived today. I replaced U3 and its now working perfectly!
Thanks for the help. Here is a link for anyone else who may need to replace it if they are experiencing the same problem: https://www.aliexpress.com/item/4000859866686.html?spm=a2g0o.productlist.0.0.1bb324c3qw1YFN&algo_pvid=2e58b573-f1eb-4526-bd83-cd44cdf76914&algo_expid=2e58b573-f1eb-4526-bd83-cd44cdf76914-0&btsid=0b0a3f8115942332811715809ebd8a&ws_ab_test=searchweb0_0,searchweb201602_,searchweb201603_
I will say that the temperature reading "jitters a lot" i.e. moves around very often but only up and down by 10C, whereas my new one doesn't really do that. But it's not a problem, at least I can use this soldering station now.

[dreamcat4]

I will say that the temperature reading "jitters a lot" i.e. moves around very often but only up and down by 10C

[EDITED, corrected]

That might be 3.3v regulator / 3.3v rail, which (generally speaking) is something depended upon by bother the op-amp and the ADC. Which seems to be U2 (a jw5026). And yep... these 'jw5026' are a switching regulator. So indeed.... that may be the reason. Actually I am not sure (myself) if that is the reason. Still gathering information. However that is what other people seem to have remarked elsewhere.

Source: https://github.com/dreamcat4/t12-t245-controllers-docs/blob/master/controllers/stm32-t12-oled/v2.1s-and-Ve2.1S/KSGER%20STM32%20Ver2_1S%20schematic%20rev%201_floobydust.pdf

What we do know for certain is that the 3v3 rail on the *known good quality* 'v2.1' original hardware. Which is not what you have (I know because I have the exact same). The original 3v3 regulator is a different one. Before they started meddling around changing things. You would think a subsequent PCB revision would result in an improved end product, wouldn't you eh? Well not here.

This schematic indicates an ST LD117 (ld117as18tr to be exact).

Source: https://github.com/dreamcat4/t12-t245-controllers-docs/blob/master/controllers/stm32-t12-oled/v2.0-and-v2.1-original/r1/iron.pdf

However photos show instead of that a MaxLinear SPX2954. Which seem to be a similar type of device (LDO with similar characteristics).

It is a different sized package and with more heatsinking around it. Which suggests that it's a liner style VRM. Generates more heat? IDK how much. I includes in the datasheet a schematic, and looks quite simple design. It also advertises 'reject 75db from input supply'. So that might also be a reason why it was originally selected to begin with. IDK.

Your PCB might not be designed to cope with these other type of 3.3v regulator. Might need to be be an 'off board' mod. Still... fairly easy though. I suggest also that maybe these original component are not so easy to source / get ahold of? Maybe an older design?

Well if those cannot be sourced so easily then there may now be other suitable (not switching) 3v3 regulators, with a large enough drop-down from '24v' to '3v3'. That can be used instead of that one.


BTW @floobydust please get me your original .sch file for your schematic. At least 3 known hardware variants have been identified. And they are slightly different from each other. So really it would be nice to edit your schematic to fully document those variants. Especially this is useful for help identifying any smaller differences which might affect compatibility (or not) with the v2.1s branch of the Open Source Firmware (PTDreamer fork).

[wf789]

I will say that the temperature reading "jitters a lot" i.e. moves around very often but only up and down by 10C

[EDITED, corrected]

Thanks for the infos. Well I've been using it for a few hours and the temperature is fine now. It's not jumping around but is stable! I think I will leave it as is, as it's working great.

[keiser1080]

Hi all,
ordered the the t12 solder station from this supplier: https://fr.aliexpress.com/item/4000183089084.html
the build quality semes to be ok but both tips missing in the package.
so i orderer from banggood a set of tips: https://www.banggood.com/Drillpro-10pcs-T12-Soldering-Iron-Tips-Set-for-HAKKO-FX951-FX952-p-1191594.html
before to put the tip in the handel the station was wrorking but with error message in the screen (using my power cord).
After i have insteted the tip as soon as i have inserted the power cord in the wall socket the my home fuse go down.
I have checked inside the station and i diden't find any shortcut on the power suplly board.
So i have used my lab power supply set 24V 1A and i powered the board , the magic smoke come from the diode labeled D2.
C2 seems to be shorted, and M2 (433M2918 https://www.mouser.be/datasheet/2/348/bd4xxm2-c-e-1874137.pdf ?) seems to be also shorted .




could you help to diagnose?
Can i use an alternative diode i have a bunch of schottky 30v diode?
Do i need to check other component surounding the diode?
I think i have found a schema of the controle board, the board look different but the component seems to be the same) The faulty diode is labeled D2  https://github.com/dreamcat4/t12-t245-controllers-docs/blob/master/controllers/stm32-t12-oled/v2.0-and-v2.1-original/r1/iron.pdf

i have the following tools:
lab power supply : https://www.amazon.fr/gp/product/B07SS77N7K/
multimeter fluke 87: https://www.eevblog.com/forum/repair/fluke-87-i-repair-help/

[dnwheeler]

Some general tips for anyone that gets a T12 (do these when you get the iron):

• Check the solder joints at each end of the handle cable (disassemble the handle and DIN connector) - cleanup/fix if necessary
• Open base and look at connections and make sure RTC battery isn't being "squashed" by the top of the case - cleanup/fix anything suspect
• Add a ground wire from mains input jack to case - the case comes "floating" which can be dangerous
• If the RTC battery drains quickly (or is D.O.A.), you probably have a model with a resistor across the battery - look online for instructions for your hardware rev. to remove the resistor
• If you're in the U.S. (or other locations where "up" = on), consider flipping the power switch over

[clanderson]

I just had this identical problem with my new t12 3.1s. Trouble presented itself only a few short weeks after my first use (and initial grounding/handle solder rework), eventually leading to the infamous cherry red tip. Replacing the specified mosfet with a comparable part solved my issue, though this doesn't inspire confidence.

[floobydust]

I would not hot-swap tips, that can short the socket and overload the power mosfet. Which one did you have? I've seen wimp ones SOIC-8 but with no thermal pad.
But really, these are 30V rated parts being run at 25V so any voltage spikes I think damage the part (thus the add a diode recommendation of mine). Or they're from the gutters of guangdong.