A few generals before, the OP pic had some cute anime girls in a PCB board putting SMD components on their pads. Does anyone have that pic? It was cute.

File: pick and place waifus.jpg (943 KB, 3840x2400)

943 KB JPG

>>2828200
gotcha

1. if i need several different rails for a circuit (+-12 and +24 in my case) should i just use a (switching) regulator for each one?
2. do i need a special inverting regulator to make the -12 rail, or is there some way to configure a regular one to be inverting?

>>2828210
Many thanks.

>>2828201
Just do the normal RX-TX TX-RX GND-GND stuff, but twist each signal wire alongside its respective ground wire. But the strange thing is the ground pins are on 4/5 of the 8P8C, which are twisted together. So I guess it shouldn't matter.

>>2828213
If it's an enhancement-mode NFET, then you'll have issues of the op-amp's output range and the NFET's gate threshold voltage causing a nontrivial dropout voltage. The op-amp is on a 5V supply, if it's an RRIO op-amp then it can go all the way up to its 5V rail, but the 10k/10k voltage divider makes that only 2.5V. The NFET needs at least ~1-5V to turn on, so you can probably only get up to 1.5V as an output voltage, maybe 2V if you have a really low threshold voltage FET.

What kind of voltage range do you want the output to have?

>>2828214
Depends on the use-case. If it's analogue, I'd probably want to make a linear supply with a multi-tapped transformer, assuming it will be mains powered.

If it's battery powered, then yeah probably a switching regulator for each one. For the -12V, you could either use an inverting converter, or an isolated converter that you can just tack below the 0V rail. I'd look at efficiency, size, and EMC ratings before choosing. Some switching regulators have multiple rail outputs and only one feedback path, just by having the different rails share the magnetic flux inside the transformer, but isolated switching topologies are a pain to design.

What's the project?

>>2828220
>the 10k/10k voltage divider makes that only 2.5V
i can take that out, i just threw it in there while troubleshooting the thing because i thought the gate wasnt draining to ground or something.
>What kind of voltage range do you want the output to have?
if by output you mean Vout, then typical will be 10-15, but id like to have room for maybe 18-20 just for robustness. i can bump the op amp to 9 or 12 +rail if need be.
>If it's battery powered, then yeah probably a switching regulator for each one.
yea im grabbing some off the shelf power supply.
>For the -12V, you could either use an inverting converter, or an isolated converter that you can just tack below the 0V rail
>isolated switching topologies are a pain to design.
alright, ill just grab an inverting one
>What's the project?
SPAD driver (you may remember me from previous threads), very sensitive to noise so im putting a lot of consideration in power delivery. ill be using linear regs for anything important.

>>2828224
>typical will be 10-15, but id like to have room for maybe 18-20 just for robustness
Your current topology can never go above the +rail of the op-amp. You can boost it above using low-side transistors and pull-ups to the 24V rail, but those circuits can be a pain to keep stable. Any reason not to run it off the 24V rail? Nothing stopping you from running a choke or capacitance multiplier off the 24V rail to a cap to lower the noise. Also consider using a BJT instead of a MOSFET there, the lower threshold voltage will work in your favour.
>SPAD driver
I have no idea what that is
>very sensitive to noise so im putting a lot of consideration in power delivery
Where noise is an issue, switching supplies may well become a problem. If an AC transformer isn't an option, then at least aim for some linear post-regulation after your noise-sensitive switching rails. A capacitance multiplier is a decent way of doing this, though personally I'd recommend making a TL431 regulator with a pass transistor, set the voltage to your desired level and put the SMPS output 2V higher than that. Conventional linear regulators, especially LDOs, have terrible ripple rejection at SMPS frequencies, they're intended for rectified mains.

File: si_spad_dark.png (114 KB, 2560x1920)

114 KB PNG

>>2828245
>Your current topology can never go above the +rail of the op-amp.
ahh youre right, i forgot thats how MOSFETs worked. i can switch to 24V and a BJT. btw, this particular portion isnt noise critical, a tiny bit of ripple likely wont be a huge issue.
>I have no idea what that is
if you have a P-N junction with incredibly few defects, you can reverse bias it past breakdown without it triggering any avalanche. it you inject a carrier into it, maybe with a photon, the avalanche will begin and the device essentially acts as a photomultiplier. if you choke the current, maybe with a sufficiently large resistor, the avalanche will quickly terminate and the cycle begins again.
pic rel is a Si SPAD. unfortunately when you try making these out of III-Vs they arent very good at terminating themselves, so you have to do it manually with a lot of very fast circuitry.
>Conventional linear regulators, especially LDOs, have terrible ripple rejection at SMPS frequencies, they're intended for rectified mains.
huh, good to know.
>I'd recommend making a TL431 regulator with a pass transistor, set the voltage to your desired level and put the SMPS output 2V higher than that.
ill look into it. thanks fren <3.

>>2828245
>I'd recommend making a TL431 regulator with a pass transistor, set the voltage to your desired level and put the SMPS output 2V higher than that.
actually i have some questions:
would this work for inverted rails?
would this work for voltage differences larger than 2V?, say, 12V down to 5V?

File: Series-regulator-passing-(...).png (7 KB, 250x197)

7 KB PNG

>>2828256
Ah, it stands for "single photon avalange detector". Very cool suff. Can you use an off-the-shelf diode or transistor for this, or do they have too many defects?

>>2828259
>would this work for inverted rails?
I haven't thought about that before, maybe? Pic related is the normal circuit, it sucks current into its cathode until the voltage at its sense pin is equal to 2.5V, so it is intrinsically polarised. I'd mess about in a circuit sim to see if there's a method that puts a PNP on the anode and properly takes feedback from the output.

You could definitely use a zener diode instead of a TL431, that's a simple method for both positive and negative pass regulators, and while the ripple rejection is still damn decent, the regulation isn't nearly as good, in case you want exactly 5.0V. The same applies for using a TL431 with its feedback loop tied to its cathode. One method to get around this, if you have enough dropout to work with, is to use a zener pass regulator followed by a conventional 3-pin linear regulator like a 7805/7905.

>would this work for voltage differences larger than 2V?, say, 12V down to 5V?
Definitely, but the extra power dissipation in the pass transistor might be a problem. If it's a low-current situation, then no problem. You get better ripple rejection as you increase the dropout, but it's diminishing returns, I usually aim for 1-2V. Without a heat-sink, a TO-220 transistor package can handle 1W or so, that's 0.5A at 2V, but only 0.2A at 5V.

File: file.png (1.44 MB, 2565x1608)

1.44 MB PNG

>>2828264
>Can you use an off-the-shelf diode or transistor for this, or do they have too many defects?
i greatly simplified the operation of a SPAD, theres more to it than a simple P-N junction. theres absorption layers, multipliers, gradings, mirrors, and a host of other shit that barely works in the finished device because the stupid fucking chinese grad students cant be bothered to learn how to use an MBE properly.
>in case you want exactly 5.0V
nope, +/- 200mV is fine, just need it to be steady.

ill sim some stuff and see what works. googling around, this stack exchange links claims that some LDOs are decent at filtering ripple (-60dB up to around 9kHz for the LM317), although the switching im looking at is closer to like 2MHz, which might be closer to what most switching regs operate at. i should probably just figure out how to use your pass transistor regulator thing instead of relying on ICs for everything.

>>2828273
>i greatly simplified the operation of a SPAD
I see, good luck then
>+/- 200mV is fine
You can probably get that good with a zener, though ±0.5V is more likely once you include temperature variations.

>the switching im looking at is closer to like 2MHz
I've seen switching down in the 10s of kHz, but usually it's in the 100s of kHz if not the low MHz. You can use a choke or inductor to clamp down on the really high frequencies, the stuff the amplifier in a TL431 or linear regulator is too slow for.
I've always liked the idea of making an inductor with a second shorted winding and using that for series high frequency rejection, though I've never tried it. You shouldn't use any inductive filter if the load on the filter's output can be pulsed, and open-core inductors are prone to picking up EMI, but they're still a good tool for the ripple rejection toolbox.

>>2828282
>the stuff the amplifier in a TL431 or linear regulator is too slow for
wait, so is the TL431 with the pass transistor not gonna work? ill sim it to be sure, but if you think its hopeless at 2MHz then i wont bother (or ill get the switching frequency lower).

>>2828253
Thanks for letting me know that I'm braindead. Anyways, that made the text even worse. Could it be because I'm destroying the strands in the wire while trying to put it in? I have 0 finesse.
>>2828220
>Just do the normal RX-TX TX-RX GND-GND stuff, but twist each signal wire alongside its respective ground wire. But the strange thing is the ground pins are on 4/5 of the 8P8C, which are twisted together. So I guess it shouldn't matter.
Not sure if this changes anything, but the documentation says that pin 4 is not connected.

>>2828283
>wait, so is the TL431 with the pass transistor not gonna work?
The TL431, when used as a pass regulator, is preceded by an RC filter that strips out the high frequencies pretty effectively. Besides the filter, it's possible for high-frequencies to couple through the collector of the BJT, but I don't think that is a major pathway if the base is held at an approximately constant voltage, feedback or no. The only advantage of the TL431 pass regulator compared to a zener pass regulator is the internal error-amplifier's ability to correct for DC offset. A zener pass regulator's DC output will vary with temperature, input voltage, and output current, not that any of those particularly matter if you're sending a few mA into an op-amp circuit.
>ill sim it to be sure
I've tried using the built-in shunt regulators (chips very similar to the TL431) in LTspice and they just don't work at all, so be careful with that.

>>2828309
>the documentation says that pin 4 is not connected.
Well that's even worse. How long is the total cable between the serial port and the RJ45 network switch? If it's less than a few meters it shouldn't matter. If it's greater than that, I'd ensure that the cabling is shielded. And maybe add a ferrite suppression sleeve.

File: 1718557828561710.gif (17 KB, 253x305)

17 KB GIF

I want to thank the person last thread who mentioned doing the spring idea on my floppy disk sleeper mod.

I didnt really understand your advice and I still dont entirely but i think it lead me to some kind of solution of using a torsion spring to push contacts into the position of the sd card.

>>2828214
>or is there some way to configure a regular one to be inverting?
yes. the datasheet should have example circuits for boost, buck, and inverting topologies.

File: 451445885_292694272412948(...).jpg (674 KB, 1180x2481)

674 KB JPG

>>2828380
Thanks. Yeah, i'd have to remove 4 screws each if i actually want to remove both the top and bottom of the syringe pump. Or i can just remove 2 screws and remove the top of the syringe.
Though in practice i plan to just uptake and then expel some alcohol then water a few times to wash. No need to take it all apart.
I did consider some latching mechanism, and i may do that yet, but for now this works.
Vref is 1.8V on the potentiometer of the DRV8825 so current should be 0.9A. It has never gotten particularly hot or even really warm on testruns i have done, but i have not used it for long periods of time yet. I may need to run it for about an hour at most when in actual use. Should a hole in the box be good enough for cooling? Should i add like a small CPU fan or something?

>>2828406
Sorry, Vref is 0.9, current is 1.8A.

>>2828309
>Could it be because I'm destroying the strands in the wire while trying to put it in?
Maybe it's not making a good connection in the terminal. Ferrules are good for those, but for now just use wires long enough to comfortably work with but short enough to minimize external interference. Try to keep it under 5 inches or so.

>>2828367
looking at some TI ICs, the buck-boost/SEPIC chips dont seem to have examples of inverting topologies, but the buck, non-boost ones do. im not confident enough to slap one on a board without an example.

File: file.png (226 KB, 3840x1916)

226 KB PNG

>>2828341
>I've tried using the built-in shunt regulators (chips very similar to the TL431) in LTspice and they just don't work at all, so be careful with that.
i found a model for it. the sim is rejecting 100mV at 2MHz which is nice. pk-pk is 6uV which is like uhhhh 20log(100000/3) which is like uhhh -90dB, cool. i probably fucked up the reference calculations but at least it works. gonna see if i can get it to work with inverted voltages.

>>2828264
>You could definitely use a zener diode instead of a TL431
do you mind sharing a negative topology for this? i cant seem to figure it out.

I wish I had passion like u but all I do is goon and spread fake news on pol

>>2828598
are you that guy i linked from /sci/? behave yourself.

>>2828598
That's just a long winded way of saying you're a filthy kike bastard.
You will never have a constructive passion.

File: file.png (246 KB, 3840x1914)

246 KB PNG

>>2828567
nvm figured it out. about -74dB, almost as good as the TL431. not sure why it took me most of the work day but i never claimed to not be retarded.

>>2828518
68Ω and 100nF is pretty extreme too. You'd normally choose R1 such that:
>hFE * (V_in - V_out) / R1 >> I_out
Or in other words, such that the bias current into the base multiplied by the transistor's current gain is significantly greater than the maximum required output current. I use a factor of 5-10 for this, alongside the absolute minimum hFE value from your datasheet. Using a power transistor like a TIP41 means the hFE is pretty low, it's an option to use a darlington pair if you can handle the extra dropout, or to use a sziklai pair.

As for C1, make it as big as you want. Bigger cap = better ripple rejection. I've seen some people using a MOSFET instead of a BJT, that allows them to use a stupidly slow time constant (1-10MΩ + hundreds of µF) before the zener diode, though the high dropout voltage of the FET and greater Vgs_thres drift is a tradeoff.

>>2828632
>Or in other words, such that the bias current into the base multiplied by the transistor's current gain is significantly greater than the maximum required output current.
got it, ill do some math and size that resistor appropriately. i just used some online calculator and grabbed those values for the time being. i think my max current is like 100mA so hopefully i wont have to use a pair.
>As for C1, make it as big as you want. Bigger cap = better ripple rejection.
sim says that the difference between 100n and 10u is pretty small, less than 10dB if i had to guess, but i might as well use a big one if theres no downside.

>>2828341
>How long is the total cable between the serial port and the RJ45 network switch?
Less than a foot and a half.
>I'd ensure that the cabling is shielded
I'm using a solid-core copper F/UTP 23AWG Cat6A cable. I don't have any shielded ones, but come on this shouldn't be that sensitive... right?
>>2828482
Using < 1 inch solid-core wires now, still garbage output.

Could it be that I just killed the chip in the switch by using a "standard" USB -> RJ45 console cable before?

File: Screenshot 2024-07-24 200009.png (58 KB, 757x547)

58 KB PNG

Should I buy 2nd edition now for $35 CAD or use a PDF and wait til December for the 3rd edition

>>2828649
Pdf both and buy the one you like.

>>2828649
PDF and print out useful pages

>>2828636
100mA means you can get away with a mid-power transistor, probably like a BD139, maybe even a 2N2222A, but calculate the power dissipation first to see if it will get too hot. Without a heat-sink, a TO-220 can handle ~1W, a TO-126 can handle ~0.5W, and a TO-92 can handle ~0.25W.
>i might as well use a big one if theres no downside
Do a bit of work to ensure the impedance at your switching frequency (and its harmonics) will actually be lower. A mid-sized film cap (1µ-10µ) will probably be better than a large electrolytic. Using a monolithic/multilayer ceramic (100n-1µ) is probably most cost effective for low current ranges.

>>2828645
>a "standard" USB -> RJ45 console cable before?
Well normal ethernet uses low voltage signalling, while RS232 is like ±12V, so it's more likely you killed the USB-to-ethernet adapter rather than killing the switch. Do you have a scope or logic analyser? Try wiring both sides for loop-back to ensure they're both working fine.

>>2828671
The USB-to-ethernet adapter works perfectly fine with other switches.
>Do you have a scope or logic analyser?
Nope. What's the cheapest one I could get away with? Not interested in investing too much in this just to find out the motherboard's fried or something.

File: hmm.png (37 KB, 860x603)

37 KB PNG

buy once cry once or work with junk

>>2828679
>What's the cheapest one I could get away with?
Even the cheapest chinky logic analysers are usable, though personally I'd use a nanoDLA instead. Apparently the cheaper ones can overwrite their own firmware or some shit, never had that problem myself but idk. You will need some voltage dropping and clamping circuitry, not a big deal, just a resistor and a pair of diodes for each input.

But a logic analyser alone won't pick up the finer minutia of the signal, if the voltage droops or whatever. I have no idea if your problem is one that needs analogue measurement, I suspect not, but it's impossible to say. If you were to buy a scope, you'd need a digital storage oscilloscope for this kind of work. I have a CRT scope that happens to have a limited digital mode that can do single-shot-captures, but most CRTs don't have this. If you think you'll use it in the future, peruse some youtube reviews for cheaper digital scopes, be they Rigols or Hanteks, or even USB scopes. I'd probably lean towards 4-channel, and electrically isolated from any computer, but you don't exactly need those features here. If you don't think you'd use it much, it is a pretty expensive thing to justify, so just go for the $20 logic analyser. Or have a look for 2nd hand scopes near you.

Not op but I have a two channel Hantek DSO5102P on the way. It looked like the best value available at the time. Did I fuck up?

>>2828719
>DSO5102P
Doesn't look too bad, though personally I'd go for a 4-channel scope instead. If you'd bought a 10-20MHz scope I'd have laughed.

I wonder if there's electrical isolation of some kind on the RS232 or USB outputs?

File: 1690703434115547.jpg (58 KB, 828x840)

58 KB JPG

>>2828707
>buy once cry once or work with junk
depends.
quite often you can find middle-of-the-road tools that work as good or even better than expensive shit.
sometimes, the price is not an indicator of quality or longevity of a tool. sometimes the price is just a brand tax or 'country of manufacture' tax.
be more specific if you want actual advice on what to spend money on.

>>2828719
I have the same scope, which I bought years ago when I started getting interested in (analog) electronics and a DMM wasn't cutting it anymore. It still serves me well for my extremely limited needs. Sometimes I wish I had a third channel, I have yet to feel the need for a fourth channel; firmware, UI, and screen are fine, but far from great. If you're at the start of your learning curve and you don't know any better it's possibly a good pick, depending on how much you pour yourself over it you won't outgrow it in quite some time. Also, it can be modded to 200MHz if you ever may need it. Frankly I'm surprised nothing better for the price came up in these years.

File: file.png (172 KB, 2035x1071)

172 KB PNG

are coupled inductors necessary in dc-dc converters like this? looking at the recommended PCB layout for this chip, theres only the one inductor (the one "in series" with the load), so is having a coupled one optional or something?

File: file.png (369 KB, 2323x1486)

369 KB PNG

>>2828922

File: file.png (801 KB, 1310x818)

801 KB PNG

over-under on me being able to hand-solder 0402s?

Does anyone know of any small cheap linear actuators? I'm having a hell of a time finding any. Either I'm not looking in the right place or they don't exist. (Shit I might just buy two crappy CD drives for their actuators)

I want to use two of them to make a shitty steerable mirror.

Hi, I have this chink MCU (AM8EB153) and I can't find the datasheet.
if anyone could help me find the datasheet that would be appreciated.
if not possible, any ideas of what to do with this SOP14 piece of silicon, that's welcome too.

>>2828936
Why not just steal an adjustable mirror module from a car? Wrecking yards are full of them.
Typical mirror prys off module. 12v on 2 axis.

>>2828946
Did you search on digikey and mouser?

File: stepper actuator.png (306 KB, 572x350)

306 KB PNG

>>2828922
>>2828924
Your schematic is of a SEPIC converter. But that layout diagram is definitely not a SEPIC at all, there's no second inductor nor is there a capacitor between LM and D1, it's just a boost converter. I'd guess the same IC can be used for SEPIC and boost.

>>2828936
There's those cheap little steppers coupled to a screw you find on alibay for like $5 each. Pic related. Though you can also go for a galvo method instead if you want to steer a mirror, they're much faster and probably cheaper too, but less stable when it comes to vibrations.

>>2828998
>I'd guess the same IC can be used for SEPIC and boost.
yea, it can. from what i can tell the C_ac and inductor are the only differences. ill defer to the schematic and use a coupled inductor. theres no spice models for these so im just gonna pray it works.

>>2828982
ofc! its some chinese shit made by Alpha, they don't even sell this on Arrow.
didn't have any luck searching digi and mouser either

>>2828998
>a galvo method instead
Do you have any keywords I can use to search for this?
The fancy mirrors I use at work (Optotune MR-15-30) uses fancy pickup coils to steer the mirror.

I would love to make one, but I figured it might be a bit too difficult for DIY at home. So if there's simpler ways of doing a non-linear-actuator based steered mirror I'd love to hear about it

>>2829000
>theres no spice models for these
Well you can model a transformer by just adding a coupling coefficient like "K1 L1 L2 0.95" in its own spice command. I don't know the difference between a coupled inductor and a transformer.

>>2829005
Search their website, or browse their parts on LCSC.

>>2829029
No keywords. I remember William Osman making a cheap galvo out of a speaker, but I suspect to make an accurate and precise galvo you'd need some of those tiny ruby bearings like what watches use, and a rigid but lightweight frame to mount the mirror and coils/magnets on. If conventional brass bushings don't have too much backlash, it should be pretty doable to diy a larger slower galvo. Most people use galvos for pulsed laser scanning, where it needs to be damn fast, so I can understand why there wouldn't be much interest in lower-speed more diy-friendly designs.

>>2829048
>Well you can model a transformer
sorry, i was referring to the IC

>>2829048
Wew that's about what I expected in terms of effort. It seems like I might just have to drop my project altogether since the optotune mirrors are so damn expensive. I just don't have the MechE prowess to solve the mechanical side of the problem.

File: PXL_20240726_004622936.jpg (292 KB, 1820x2096)

292 KB JPG

>phone charge blade terminals melted at the base
how the hell does this even happen? i don't think the female terminals on the wall outlet even reach where it's melted.

also any tips on how to open this up besides mangling it with a dremel? i tried freezing it and smacking it with a hammer and then squeezing it in a vice. did not work for shit.

File: jeetspeaker.png (40 KB, 158x84)

40 KB PNG

hello sirs! what is calling this 5 pin connections pls??

>>2829065
i did this as a kid...partially plugged in, then drop something across the blades, cool noise and light!

>>2829065
Looks like a coin or something fell on the pins. Do americans not have plastic halfway down their plug pins like the rest of the world?

>>2829096
Shitty picture, but looks like JST-PA. Maybe XA or VH depending on pin pitch.

>>2829048
not even lcsc has anything about the AM8EB153.
I only know it is an ALPHA elec.
part, but I can't find their website so its useless.
I did find some chinese websites that claim to sell said part, and claimed to have a datasheet.
but I couldn't find an actual datasheet for the part; it returned datasheets for other parts, or redirected to random websites.

I think I'll just give up on trying to use the fella, I prefer authentic Atmel components anyway.
but I'm not sure of what to do with it.
maybe practice handsoldering SOP14?

tl;dr: can't find it on any chink website, dunno wat2do wit it

>>2829097
>>2829130
oh wow that's exactly what it looks like, nice catch. my gf's niece borrowed it so this doesn't surprise me.

File: Screenshot_20240726_155426.png (183 KB, 1366x768)

183 KB PNG

>lcsc
how does this interface not make you want to kill yourself
>fucking cement resistor
are there any suppliers that are more beginner friendly?

>>2828929
I give you 30% chance without a miscroscope.
hope you have plenty of spares.

>>2829358
Yeah they also call MOSFETs tubes or triodes. You can browse the non-chink-brand parts on LCSC using Octopart, but for your Brightpowers and GigaDevices you’ll need to use their dogshit site. I wouldn’t be surprised if someone wrote an LCSC webscraper with regex and better parametric search. Do it yourself if you’re autistic enough.

When building circuits in veroboards, it often happens to me that the copper pads unglue from the bottom of the board. What could be the cause for that, I'm soldering the same as I would on any commercial PCB.

>>2829435
iron too hot, heating for too long (i.e. iron too cold), low quality chinkshit. it helps to sand the bare copper pads before heating, to expose the metal and minimize the time needed for the solder to wet.

>>2829435
In addition to what the anon above said, vero-boards are usually made from FR2 paper-phenolic composite (brown or orange), as opposed to FR4 glass-epoxy composite (clear-green or maybe grey). FR2 is much easier to delaminate pads from. I used to have the same problem, but now with a good technique, a decent iron (and non-conical tips) I seldom delaminate FR2. Fast joints are cool joints, practice soldering as fast as possible. I like to have a bit of solder on my tip before starting the joint, that small amount of liquid metal vastly increases the thermal conductivity from the tip to the joint. For through-hole joints I use a 1.2mm screwdriver tip and flux-core solder.

>>2829446
>iron too hot
>iron too cold)
Mine doesn't have temperature regulation. Maybe I should invest into that? Seems quite expensive though...
>>2829450
I've got the dark orange/brownish kind. I do apply a bit of solder at the tip but still have that issue sometimes and I use a conical tip.

>>2829453
wedge is better
>doesn't have temperature regulation
not even a little wheel? it's not perfect, but a half decent 60w iron with a wheel is like $17 on amazon

>>2829364
30%? ill remember this post when my board shows up, hope youre hungry for some humble pie in a couple weeks.

>>2829456
>not even a little wheel?
No nothing.

>>2829453
If you’re quick, an iron that’s a bit too hot shouldn’t be a problem. If it’s above 400C then maybe it’s an issue. But too low a temperature can definitely cause joints to take too long and delaminate boards that way. You might need to wait for it to get up to temp between joints if this is a problem.

A temperature controlled iron is hardly a requirement for this kind of soldering, but they do allow you to work on really small and really big joints thanks to their smaller size and higher power output. Also the common 936, T12, and T245 tip families have great variety, a tip for every occasion.

>>2829456
Those are often triac-based, without measuring temperature. I’d recommend one of the ones with two or three buttons on the handle. Also a Yihua 929D desoldering iron, they’re great if you ever intend on doing desoldering.