Any opinions on the AiXun T3A?

[c0d3z3r0]

Do they use PWM? If yes, what frequency?

It looks like they do: https://github.com/sparkybg/UniSolder-5.2/blob/main/software/front/US_Firmware.X/PID.c#L340. Their PID controller is adjusting a duty cycle in the end. I don't know what their frequency though, It's not clear from their github.

Oh well, yes, actually any sort of regularly turning power on/off is in fact "PWM". What I mean is that Aixun uses a PWM signal that is again turned on/off regularly, so low-freq PWM on a high-freq PWM signal.

EDIT2: Checking the output of the Unisolder with an oscilloscope for comparison could help... anyone?

EDIT3: Hmm... had a quick look, I might been wrong... maybe their PID algorithm is just weird or they use bad or non-optimal parameters. I wasn't able to full reverse-engineer it, yet

[drksy]

maybe their PID algorithm is just weird or they use bad or non-optimal parameters. I wasn't able to full reverse-engineer it, yet

Interesting, hope you can figure it out. The unisolder has historically been convoluted, confusing, and unnecessarily complicated.

I’ve asked a related question on Electronics Stack Exhange here https://electronics.stackexchange.com/questions/682699/pwm-circuit-that-switches-mosfet-with-soft-edges/  to get some help in calming the harsh power circuit down.

I’ve got one answer so far with some very interesting insight. He actually produced a spectrogram of the audio in my YouTube video demonstrating the buzzing phenomenon, take a look:



Knowing that the Aixun switches at 166Hz, it appears that the buzzing frequency is curiously the 9th harmonic of this fundamental at around 1.5kHz. Something somewhere in this circuit is resonating! And this resonance worsens when the handle shell touches the wire cover.

While softening the switching edges filters these harmonics. I’m more interested in the source of this resonance at this peculiar frequency. Something about just the 9th harmonic that is getting amplified in some loop. I don’t hear anything in any other multiple of 166Hz.

Also he suggests to put a behemoth of a low pass LC filter before the heating element with the cutoff frequency of 40Hz. I don’t really have the inductor he is suggesting to us. There is probably better ways to filter it.

[dreamcat4]

how did you generate the spectrogram? looks cool 

[T3sl4co1l]

Oh, you are in this thread, lol. I didn't look.

Spectrogram is just an Audacity option (or whatever variant of -acity is politically acceptable these days; I have an old version for what little audio work I do so I haven't paid attention to that stuff).

Tim

[drksy]

Oh, you are in this thread, lol. I didn't look.

Spectrogram is just an Audacity option (or whatever variant of -acity is politically acceptable these days; I have an old version for what little audio work I do so I haven't paid attention to that stuff).

Tim

Hi Tim! Didn't realize you are on this forum either, though your icon looked familiar

[drksy]

So after some more experiments tonight, I finally managed to eliminate the buzzing. Here's the circuit I built, essentially moving the high power path out of the circuit, and also providing a separate, isolated supply to run the high current loop.



I didn't have any high power PMOS at hand, so I used one of my spare IRF3710 n-channel MOSFETs as a high-side switch by adding an extra, isolated 8V supply to drive the gate. A boost converter could also be used here.



Running this, it was absolutely silent. I didn't hear any buzzing whatsoever from the board or the any of the supplies.

Next, I removed the 24V bench supply and used Aixun's own 24V supply:



And what do you know, everything started buzzing again! What's going on here!! It seems when the high power path shares the same supply as the analog control circuit, the thing just starts buzzing.

It was already getting late, so I ended it here. My next try is going to be supplying the 2.8V for the digital section straight from an isolated bench supply and see if that does anything.

[dreamcat4]

And here i was just assuming that it depended on the grounding... (for example if you touch the tip to a grounded objects whilst soldering, or if the aixun base station is plugged into a poor grounding, or if the 1mohm static grounding resistor explodes due to static shocks, things like this).

[T3sl4co1l]

Aha, so it's not the element buzzing, it's the supply.

On the upside, the proposed solution still works, if just moved over to the input to the switch (i.e. an LC filter between supply and switch); of course, it's no more economical there, either.

That suggests the noise is probably core magnetostriction from the supply throttling up and down regularly (and suddenly); it might even be going idle inbetween pulses.  Replacing the supply with a quieter one is the most expedient route.  Digging into the SMPS yourself would be... more challenging.

I suppose I'm still only assuming it's SMPS, as I haven't seen complete up-close internal photos of this thing.

Tim

[dreamcat4]

i think you have to ultimately realize that at least 1 component somewhere in the box is making a sound. and whether thats from a magnetic coil, or its a microphonic peizo from an mlcc or something else. then that component either isn't locked down properly enough, or it's basically gone faulty / bad and may be a subject for possible replacement. since it's making a noise after all...

but the reason that it is resonating is because there is a resonance occuring at some specific frequency, due to a resonance occurring or being driven in the circuit.

so if you think about somethingg like a power brick (or smps) if it's not grounded but isolated / floating. then this is why i am saying maybe there is some grounding issues. whereby that floating is not tied to anything, thus permitting such a resonance to occur between multiple parts of the circuit. to generate that.

but what do i know eh? but what i am really saying is look at your active circuitry and whether powered internal or external. and whether the tip is getting grounded or is left floating when touching the target metal object (the thing that you are actively soldering). that is when the output is being driven to pwm and put the heat / power into the tip.

and then maybe something can fall out here and makes a better sense from it. to inform what you might try to measure next. or identify how to disable buzzing more reliably. and under which specific conditions.

unfortunately i cannot help you here on my unit because... it never buzzes. it's totally fine under all of the circumstances (because 0 components is faulty, nothing is due for replacement for this matters). however if the station did buzzes, i just simply would not care to matter enough. because it's too minor matters to really bother fixing it. i would probably just choose to live with the buzzing issue. (but so long as it was not too high pitched or nasty sounds, there is different types of buzz).

[T3sl4co1l]

The fact that no one else has reported noise, does sound suspicious.

Maybe they changed the design, maybe you got a counterfeit, but maybe more likely it's just a magnetic core or capacitor or something has come loose.

Note that capacitors and magnetic cores (by piezoelectricity for ceramic caps, and electro/magnetostriction otherwise) make noise even when fully anchored; these effects result in strain of the bulk material itself.  It's usually a small effect, and in particular I would expect voltage changes to be small (at least, if the supply is operating nominally, i.e. compensation is adequate) so that capacitors can be ruled out, and that leaves cores, hence my proceeding reply.  But again, this would apply to any unit, and a loose core able to rattle, or move due to attractive forces, or wires in the windings by repulsive forces, would be a viable explanation for additional noise.

Not necessarily "faulty" components, but, different in some way, seems likely.

Tim

[dreamcat4]

and pretty much (quite often) if the component is not sufficiently fixated with enough of glue or potting compound. then any degrees of freedom in vibration will mechanically couple to the substrate that the component is mechanically mounted to or touching. which often is the pcb, which can have resonany modes, in different locations on the board, and its mounting holes. and so on. such that a resonance can additively be efficient (or less) to convert such a small and tiny piezo energy vibrations into a more audible sound.

and this is specially easier when lower energies are requied to audibly hear that frequency, such as in the higher frquencies. however you can get lower pitched hitting or buzzing also. and electro magnetic sources can create much more energies for lower frequency vibrations.

so there is some sense to understanding those things. both in the debugging. but also for pcb designs, or to redesign to attempt to try to reduce such issues. (and for various ways to mitigate, there are different strategies, or multiple strategies in different ways).

[c0d3z3r0]

And here i was just assuming that it depended on the grounding... (for example if you touch the tip to a grounded objects whilst soldering, or if the aixun base station is plugged into a poor grounding, or if the 1mohm static grounding resistor explodes due to static shocks, things like this).

What grounding resistor? There is none in the T3A.

Aha, so it's not the element buzzing, it's the supply.

...

That suggests the noise is probably core magnetostriction from the supply throttling up and down regularly (and suddenly); it might even be going idle inbetween pulses.  Replacing the supply with a quieter one is the most expedient route.  Digging into the SMPS yourself would be... more challenging.

i think you have to ultimately realize that at least 1 component somewhere in the box is making a sound. and whether thats from a magnetic coil, or its a microphonic peizo from an mlcc or something else. then that component either isn't locked down properly enough, or it's basically gone faulty / bad and may be a subject for possible replacement. since it's making a noise after all...

The noise not only comes from the supply, but also from the handle. The supply might be the root cause, though.

I suppose I'm still only assuming it's SMPS, as I haven't seen complete up-close internal photos of this thing.

Yup, it is.

The fact that no one else has reported noise, does sound suspicious.

Multiple people reported that issue

[c0d3z3r0]

@drsky does it also buzz when supplying the 24V from your bench supply but without your circuit?

[c0d3z3r0]

Oh, look what I found... doesn't this supply look incredibly similar?

https://ae01.alicdn.com/kf/Hb3d3ba1e50064c2a89720def7f234ec76/T12-L-ten-Station-LED-Digital-Temp-Einstellen-Reparatur-L-tkolben-Ausr-stung-Tragbare-Home-Elektronik.jpg
https://de.aliexpress.com/item/1005005125820177.html

So, Aixun has copied the Handskit T12 supply, which is probably based on WX-DC2416.

https://oshwlab.com/FrankCA/wx-dc2416-psu

... which seems to have other problems as well \o/ https://www.diyaudio.com/community/threads/smps-killed-my-radio-reception.373045/page-2

[dreamcat4]

i did not look too closely but afaikr somebody else said at the beginning examinations that the t3a original psu was based on the kesger smps. however they might have made some slight modifications or allowances to enable the higher burst power withough shutting itself down due to ocp/opp.

so the implication is that they are both clones branched or copied from slightly earlier t12 smps design. but like you say, it is indeed seems based on that other design before to begin with (remembering that even before kesger it was like some no-name project of this 1 chinese guy, the original 't12' stations).

[c0d3z3r0]

Well, not sure about the KSGER, because it's using a different pwm controller and it has a different layout. While they all are very similiar, the Aixun and HANDSKIT share controller, count of schottky diodes, layout and even component ids/names.

Anyway, I wonder how the new combined T3A/T3B supply differs

[dreamcat4]

ok. but perhaps there might be more than 1 version of ksger psu over the years? at least they have had revisions for other part of the station (for the controller pcb, although i know we are not talking about that)

[c0d3z3r0]

Oh, interesting... I just had a look at my SMPS... I got version 2.0, the one in the picture above is v1.3. Let's see if I can spot any differences (probably difficult using only the picture :/)

EDIT: Yeah, I see two small differences. See R15 and 0 ohms resistor below.

Schematics available:
WX-DC2416: https://oshwlab.com/FrankCA/wx-dc2416-psu
There also is a schematic for a CR6842-based SMPS, which is used by the Aixun SMPS. It is very similiar to WX-DC2412. https://oshwlab.com/kisly.va/cr6842-power-supply

Some random differences and details regarding WX-DC2412/16, Aixun versions:
- Different PWM Controller (probably doesn't matter much)
- Aixun is lacking Y caps, while WX has both Y caps and X cap
- Aixun moved the NTC from behind the bridge rectifier (DC) before the Fuse (AC) - I have no idea what is right or better or if it matters...
- High resistance path for MOSFET gate uses 100 ohms, while WX uses 27 - no idea... I read that too low values would increase ringing and voltage spikes
- Aixun has a 0 ohms resistor between MOSFET gate and high/low resistance circuit, while WX has 5.6 ohms. v1.3 has no resistor there but a simple trace, hmmmm
- As already noted, Aixun has a single shottky diode. However the new T3A/T3B combined version has two
- v1.3 is lacking R15 (or is it R14?), for v2.0 R15=4.7k. Connected between pins 1, 2 of the opto coupler. This is not present in any public schematics

- Thermal protection is not being used
- switching frequency is 72.5kHz.

I had another look a the current limit. This is the circuit: (https://www.eevblog.com/forum/reviews/any-opinions-on-the-aixun-t3a/msg4421320/#msg4421320)

                            MOSFET Source
                                       |
                            680     |               R10, R11 (0.43 each) -> R_s = 0.215
SENSE <----------===----+------+---===-----+--- GND
                                                +---===-----+

The current limit threshold voltage is 0.76V (max. 0.9V). So the peak current limit should be 3.5A. EDIT: Oops, this is the current on the primary side. What does that mean for sec?

Well, I measured 17A (at ~0V), so something is wrong here. The 11A fuse in my BM789 didn't burn as well o.O because the voltage dropped to 0V? Probably this is a very bad way of measuring the max current...
I then measured the max current with an Aixun tip connected. 7.5A (Voltage dropped from 24.2V to 22.5V).
That sounds more like the peak current of a WX-DC2416 (6-9 A depending on exact model), but still I would expect a maximum of 3.5A.

One more thing: the SMPS is completely quiet even on full load. I still believe the noise has to do with the heater PWM frequency...

[floobydust]

I don't own a AiXun station, the hardware appears to be a copy of the T12 clone but with stupid changes and penny saving.

I see many H/W mistakes in the T3A - the V1.2 PSU has terrible values for the TL431 circuit, it could be the "whistle" is due to the PSU going unstable as I would expect. A 510Ω resistor for the opto LED off a 24V rail? OUCH. No R14 top keep the TL431 alive? All values are not good there, as if they have some problems.
Check if your V2.0 PSU addresses any of this.

[c0d3z3r0]

@floobydust thanks for your comment! Check out what I found out in my post above

[dreamcat4]

Hello... seems I have a v2.0 one here too. (and not any buzzing or other issues).

510Ω resistor for the opto LED off a 24V rail?

As per c0d3z3r0 comment I also measured 4.7k there (total, between the 2 opto pins on low voltage side). But you know... he seems to have really scrutinized it now. Well done!

 

[drksy]

i cannot help you here on my unit because... it never buzzes. it's totally fine under all of the circumstances

Thanks. So that puts to rest the theory that all Aixuns buzz.

Aha, so it's not the element buzzing, it's the supply.

I'm still not convinced that it's the power supply alone causing this.

@drsky does it also buzz when supplying the 24V from your bench supply but without your circuit?

Yes, it does. I went back to the lab today for further hacking. In fact, I was wrong in my previous assumption. The circuit that I previously said does not buzz, with external power switch and bench supply, did in fact make a buzzing sound from the handle . When the cartridge was connected mid-air, there was no noise, but when I used the handle, the handle was buzzing. Also, providing an isolated supply for the digital section did nothing.

This setup, powered by the bench supply, with everything soldered in, was buzzing from the handle, and faintly from the station.



Furthermore, I discovered some weird effect. For example I tried installing a 47nF across R6:



With this cap installed, suddenly I witnessed the weird bug that Tony shared not too long ago. The screen would turn off, complaining about "no tool installed" and then quickly coming back alive a second later showing a readout of the temperature, and then go off again half a second after. This is getting more and more confusing as I go down this rabbit hole. Why would this happen?

My understanding of the way Aixun determines if a hand-piece is installed is by 1-checking the ID pin, and 2-reading out the thermocouple voltage. So putting this 47n cap is likely throwing off the thermocouple readout, but how though?

And here i was just assuming that it depended on the grounding... (for example if you touch the tip to a grounded objects whilst soldering, or if the aixun base station is plugged into a poor grounding, or if the 1mohm static grounding resistor explodes due to static shocks, things like this).

Maybe it is a grounding issue. Could it be a bad solder joint? Maybe the power header that I've trusted so far is just a bad one. 11PM rolled around and I had to head back. More tests remain to be done.

[dreamcat4]

My understanding of the way Aixun determines if a hand-piece is installed is by 1-checking the ID pin, and 2-reading out the thermocouple voltage. So putting this 47n cap is likely throwing off the thermocouple readout, but how though?

Well, I believe I was told that the id pin is meant for T12 handle identification purpose. Because in my unit, it's T245 handle. And in that handle's wiring the white wire is bridged to the green wire (-). For which green color is the GND wire. So presumably when ID=GND then the MCU on the controller PCB (the software). Checks that GPIO line and determines that 0v = T245 handle. And something else (whether floating or otherwise some specific value). Would then pertain to a T12 tip.

Here I labelled the colors inside the T245 handle.




If you look carefully here the ID printed on the PCB referrs to the through hole header pin. Which then leads to the bottom pin on the GX12-5 connector.



And here is the handle insides (for the Aixun T3a - T245 type hand piece):





Rotating it in my hand:

And here i was just assuming that it depended on the grounding... (for example if you touch the tip to a grounded objects whilst soldering, or if the aixun base station is plugged into a poor grounding, or if the 1mohm static grounding resistor explodes due to static shocks, things like this).

Maybe it is a grounding issue. Could it be a bad solder joint? Maybe the power header that I've trusted so far is just a bad one. 11PM rolled around and I had to head back. More tests remain to be done.

Well I am not any sort of expert on mains grounding - far from it! But my understanding is that a poor quality (or even completely "gone", or intermittent) installation ground can occur within the room level. (or as you say on the socket strips etc). Or it can occur at the building level, or it can be that the stake driven into the soil, with respect to being tied also elsewhere locally in the incoming feed or the wider neighbourhood. But generally speaking: it's more likely in certain places or locations than other locations. So the level of suspicion eh, maybe there are professional electicians tools to measure ground connectivity. In a certified fashion. I do think it's not my business really to comment for how to go about determining such things.

But lets say your wall ground is fine. Ok then, but does the SMPS inside the Aixun carry that to the low voltage side? Well it does on my Version 2.0 Aixun SMPS (which quite funny, it has "T12" written also on the PCB of the SMPS). Anyhow the trace path of the ground on the V2.0 PCB you can see on the bottom layer clearly. It snakes around in an S shape. And then comes out at the "-" DC terminal. The path looks solid here. But perhaps if the jumper power wire itself is a bit dodgy or marginal when you press it? Or perhaps you have different version of PSU.

As for the handle buzzing. I simply do not understand that. However is this T245 handle you are talking about? --> the logical step might be to request new handle then. Because some of these handle can be known issues. And if you are sent new handle then you could also (later on) destructively dissassemble the old buzzing handle to look inside of it.

However if your handle is something else (for example like T12). Then that itself could maybe have different set of possible handle issues. But that is not my area of familiarity for these T3a model. I do not own such handle to comment about those ones.

[c0d3z3r0]

My understanding of the way Aixun determines if a hand-piece is installed is by 1-checking the ID pin, and 2-reading out the thermocouple voltage. So putting this 47n cap is likely throwing off the thermocouple readout, but how though?

Well, I believe I was told that the id pin is meant for T12 handle identification purpose. Because in my unit, it's T245 handle. And in that handle's wiring the white wire is bridged to the green wire (-). For which green color is the GND wire. So presumably when ID=GND then the MCU on the controller PCB (the software). Checks that GPIO line and determines that 0v = T245 handle. And something else (whether floating or otherwise some specific value). Would then pertain to a T12 tip.

Yes the ID pin is used to detect what tip (actually what handle) is connected. It's not only for the T12 but for all tip types.

However, there are two messages: 1) no handle 2) no tool (don't remember the correct messages)
1 is checked via the ID ping, 2 (the tip) is checked by measuring the tip temperature

Well I am not any sort of expert on mains grounding - far from it! But my understanding is that a poor quality (or even completely "gone", or intermittent) installation ground can occur within the room level. (or as you say on the socket strips etc). Or it can occur at the building level, or it can be that the stake driven into the soil, with respect to being tied also elsewhere locally in the incoming feed or the wider neighbourhood. But generally speaking: it's more likely in certain places or locations than other locations. So the level of suspicion eh, maybe there are professional electicians tools to measure ground connectivity. In a certified fashion. I do think it's not my business really to comment for how to go about determining such things.

Regarding grounding issues at room / building level I am very confident, that I have no such issues here.

As for the handle buzzing. I simply do not understand that. However is this T245 handle you are talking about? --> the logical step might be to request new handle then. Because some of these handle can be known issues. And if you are sent new handle then you could also (later on) destructively dissassemble the old buzzing handle to look inside of it.

My handle and others seem to buzz, too. What should be bad there? It's just cables and a connector. You don't have to destroy it, you can easily (well, depending on how much glue they used for your handle, for me it was easy) pull off the upper part (maybe have to twist it a bit).

[c0d3z3r0]

i cannot help you here on my unit because... it never buzzes. it's totally fine under all of the circumstances

Thanks. So that puts to rest the theory that all Aixuns buzz.

Aha, so it's not the element buzzing, it's the supply.

I'm still not convinced that it's the power supply alone causing this.

Me neither. Since both the SMPS and handle are buzzing I really think it's the pwm swichting of the controller...

@drsky does it also buzz when supplying the 24V from your bench supply but without your circuit?

Yes, it does. I went back to the lab today for further hacking. In fact, I was wrong in my previous assumption. The circuit that I previously said does not buzz, with external power switch and bench supply, did in fact make a buzzing sound from the handle . When the cartridge was connected mid-air, there was no noise, but when I used the handle, the handle was buzzing. Also, providing an isolated supply for the digital section did nothing.

Oh well, it depends very much on the temperature set (how much it needs to heat). Set it to 380 or 400 degrees and you'll hear that buzzing each 1-2 seconds.