EEVblog® Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: AmeliaBuns on May 29, 2024, 06:42:55 pm
-
Hi! I have been designing a filament drier PCB and it's one of my first projects.
Everything on this board works great when connected via SWD or when I inject 3v on the line (except my clumsy stupid brain wiring I2C to PA6 PA7 instead of PB6 PB7....)
The problem is, when I connect the board to 12v, the power LED flickers, and the 3.3v is at 2.2v peak with 1v of noise with a frequency of 700hz~ which is really strange. After bumping the voltage to 24v, the IC just got busted and yeah.... now it's practically a short. The load on the step-down was only 50ma~.
Does anyone here have an idea of what could be wrong? Thanks! (also yes I know the design is awful, I got really tired of messing with it and I'm a beginner so I figured, as long as it's wired correctly xD)
I also put a very poor photo of what my scope saw, sorry I had to hold the board with my hands and inject the power so I only have this awful photo before my board went permanently kapoom!
EDIT: General To-do for V3 of this board:
- Fix i2c.
- move fuse to be after the capacitor
- move inductor and caps close to IC by rotating.
- Remove Fuse
- Change caps to 0805 and better ratings (Voltage etc) (appearntly ratings on Caps are not that great or detailed, it might not be 22uf at 3.3v dependign on the rating!)
- Change inductor, it's overkill for 1mhz and my current, the value can be changed too.
-
I think that part of your problem is that there is a fairly long path from the output capacitors back to the input ground. I am rather new to switching power supplies, but here is a board I laid out using precisely that regulator. Note that I tried to keep all loops as short as possible, as I have seen others recommend.
Also, what is the purpose of the fuse? In my use, the maximum load is only about 100 mA, but I do believe this chip is capable of protecting itself during an overload. BTW, what value inductor did you use?
-
4.7uH! and tbh I only realized that the IC has built-in protection after adding the fuse!
Also, I doubt a path that's slightly long would cause SUCH a drastic failure and killing 2 AP63203WU ICs
Your design is much better, but My confusion is why the issues are THIS drastic, because Worst case scnario on my design is just poor ripple performance as far as I understand.
-
I would tend to agree with you that that does not look all that extremely bad, but do remember that the chip switches at 1.1 MHz.
While we wait to see if anybody more experienced in switching power supplies chimes in, I will tell you one thing I would've tried in your situation. It is a very simple, very low skill thing to do, but if you remove the damaged regulator and supply 3V3 from a separate power supply, Does the rest of the circuit work as predicted? It should be simple to test, on your design the regulator is completely disconnected by removing the fuse. Also, before you blew it up, did you see what the regulator does with no load?
Regardless, I still consider you lucky. Anytime I destroyed a component, it always seems to be expensive or I don't have another one laying about, oftentimes both. Seriously, don't feel bad, you say you are a beginner. Part of learning is making mistakes. With perseverance and advice from those more experienced than yourself, you will ultimately be successful.
-
It does! although I tested this before burning the step down, I'll try again after i'm home with the IC removed.. And you're right, I forgot about the switching frequency, such a high freqency as far as I understand means you have to be careful.
My fear is that I'll wait a month and waste yet another 5 precious PCBs that I have to throw out, So i'll wait a bit longer, otherwise I'll do what you said. Maybe even design the step down as a seperate PCB to plug/solder into this one (Like an ESP32 module)
And thank you, you're very sweet! I am a bit down but I won't give up! Third time's the charm :)
-
Okay, you need to clarify. It does what? It works fine with an external threefold supply, or the switching supply regulates at 3V3 without a load?
The circuit board situation is something I have a problem with. I am a person has a very hard time throwing out something I may be able to use in the future, and I've got for each of many circuit boards laying around because I only needed one and just can't throw the others out.
Okay, just edited this message to add:
I don't think it is a good idea to design the regulator as a separate module. That adds the cost of another board plus connectors, and every connection is a possible failure point.
-
First of all - remove fuse (or move it after output capacitor). Output capacitors are a part of output circuit and they should be connected directly to inductor (by shortest possible path).
Second - move ground of output capacitors directly to IC (as Doktor suggested). Ground path between them and IC included in high frequency and high current output circle. This ground path will influence not only IC itslef, but all other design.
And last - feedback wire should be routed from output capacitor directly to IC pin. If you can't route it directly try to move it away of input/output current loops.
And the last one - check you inductor. Is it really 4.7uH, what it's rated current and working frequency.
PS. If you intended to redesign PCB you can use ready made power blocks, like Bxx03-1W.
-
Oh sorry.
Before I burned the IC, I did try it with 3.3v coming from my Bench PSU into the SWD pins and the circuit (The stm32 etc) worked flawlessly, I even uploaded my code to it
But I haven't tried to manually "Inject" 3.3v into the rail (with the connectors, or my sharp needles) AFTER I burned it with 24v.
-
Okay, that's what I wanted to know. That seems to indicate that there was no short-circuit on the output, although the regulator should not have burned anyway. Relax and wait an hour or two, I'm sure there will be more advice coming.
I should have asked earlier, did you check very carefully for any solder bridges? That could definitely lead to a failure. Also, how much heat did you apply during soldering and for how long? I personally have pretty good luck with using 250° C, and often apply the heat from the underside of the board (Hot air, not soldering gun). You need to make sure that the solder all melts, but it's not a pizza, do not bake for 20 minutes :-) what I am getting at is that too much heat or too much time cannot only damage semiconductors, but possibly the insulation in an inductor. If you are not sure about this, watch some YouTube videos on soldering surface mount devices.
-
Actually, This tempted to me to try and solder another board before proceeding...
I used a low temp bismuth based solder for the first time and it was a mess. my hot air was set to 250c, and I had to manually rework most of my ICs with my lead-free solder and pinecil.
there was a LOT of solder balls or unmelted solder paste everywhere, but I used IPA and an old toothbrush to thorughlly brush and clean everything, and to be fair, there was no short, Although, if there was there's protection but not all shorts are on power rails!
I'm tempted to do another board by hand even tho it's a lot of work, or with my old leaded solder paste that always worked well and just be careful, I always tend to brush my hair and mess with everything like my PC/keyboard/phone etc when working with these things so I got scared of using leaded products.
-
So you are saying U2 has failed short. Are you sure it has dumped 24V into the output or not? If not, wouldn't it be easier to replace U2 on its own.
Or if you build another board, only populate U2/C14/C17/etc and test that part of the circuit (at say 5V with a limited current supply). Then try the modifications suggested above.
-
- What's the purpose of F2 fuse? Remove it.
- I would add some extra input MLCC capacitance, and a .1uF decoupling cap very close to the U2 Vin and GND pins.
- The feedback trace should be routed away from noisy components and connected to an output cap teminal, not the inductor.
- Join the GND copper fill of the input caps, the output caps, and the IC together. Use several vias to connect it to the bottom layer GND plane.
Also, you don't use low-melt bismuth mix to solder, you use it to desolder. It's brittle.
-
A lot of people use those to solder too appearntly, but I guess it doesn't work great without a stencil etc. My lead free solder wire is also awful I cna't get it to work well at all even at 360c. maybe I need a new soldering iron....
I did only the step down section and no luck, something strange tho....
when I leave the circuit as is, except with the fuses removed and replaced by a jumper, I get 300mv on the output.... when I short BST and SW pins together I get 1.9v?!!!
it's so strange tho, my old circuit works perfectly, and the schematic isn't too different from this (except the boost/sw pins are swapped and mistaken for each other on accident, so I just shorted them together lol)
-
I did some experimenting, And the IC heats to 35c (from 20c-16c ambient) and I can clearly see the switching action happening at the input of the inductor.
The IMG_3561 is with the default config, the top one is with the BST shorted to SW
This all makes absolutely no sense....... And my old design is equally awful and even has a mistake (BST/SW shorted) but it works really great (although with 100mv~ PK-PK of noise if i remember right)
-
A couple of things to note. As someone else said, do not use the bismuth solders. I think the circuit will probably operate just fine if you use them, as they are conductive, but they are very brittle and will lead to problems in the near future. As to how easy they are to solder with, I don't know, and don't plan to find out. I personally have only used the lead-based solder, in both pace form and in wire. Lead is indeed a severe health problem, especially for children whose brains are still developing. A bit less so for adults, but still not a good thing to be exposed to. But if you are doing things correctly, you should not be getting the paste on your hands. Regardless, don't do soldering where you handle food, and always wash your hands when you are done soldering. With those precautions you should not have any health concerns.
As far as soldering irons go, there is a correct temperature, and going above it is not necessarily a good thing. But you don't need a high dollar iron, From what I have read even the Chinese temperature controlled irons are not that bad, and are available quite cheaply. A Hakko is probably even better, and still affordable by most. Personally I really like my Metcal soldering and desoldering equipment, but that is a bit expensive and less you're a really serious hobbyist or have a lot of money.
As far as the SW and BST terminals, you need to understand what those are for. The SW is the output from the internal FETs, it sends the pulses of power out towards the inductor when one FETs is turned on, and acts as a diode when the FET is turned off. It needs to go to the inductor and to one side of the boost capacitor, and nowhere else. The BST terminal gets connected to the SW terminal by 8.1 µF capacitor, that set up works as a charge pump to provide a higher gate drive voltage for the output FET. It should only be connected to SW through a capacitor, not directly.
As far as capacitors go, for input and output I used 1206 (inch) sized components of the same values you chose. The spec sheet for the regulator specifies those values to be acceptable in most cases, it certainly seems it is specifying the rated value of the capacitors, not their effective capacitance operating voltage. That is how I interpret it, and it worked fine in both of my applications. I also use an inductor which was about the same physical size as the one you used, I can give you the part number if you like.
Since you've got the entire circuit bill anyway, at least that's what I understand, there is something else I suggest you do. Before fixing the problem with the regulator, make sure the rest of your circuit works exactly the way you want. Supply with power from an external supply, program it, hook it up, and see how it does. If it does not work correctly, you can also try to repair those defects on your neck circuit board. It would be a shame to get the regulator working and still need to order another board.
One other thing I should probably mention. You say this is going to be a filament driver, I suppose that could be a lightbulb, but I am assuming it is actually a vacuum tube? If so, is this the final output tube, or a low-level amplifier? I am a little concerned that switching noise on a filament may show up on the output. Also, is the tube designed for use with DC on the filament?
-
Bismuth is OK for prototyping. Yes the bond will be weaker and more brittle but your parts won't fall off the board on their own sitting on a bench.
But yes it can take time to figure out what temperature to use, 250C probably too high.
https://www.circuitinsight.com/pdf/new_generation_low_temperature_lead_free_solder_smta.pdf (https://www.circuitinsight.com/pdf/new_generation_low_temperature_lead_free_solder_smta.pdf)
https://en.szfitech.com/newsinfo288.html (https://en.szfitech.com/newsinfo288.html)
https://www.sciencedirect.com/science/article/abs/pii/S0026271421000317 (https://www.sciencedirect.com/science/article/abs/pii/S0026271421000317)
Also, you don't use low-melt bismuth mix to solder, you use it to desolder. It's brittle.
You can use it to solder: https://www.chipquik.com/store/product_info.php?products_id=470008 (https://www.chipquik.com/store/product_info.php?products_id=470008)
-
A couple of things to note. As someone else said, do not use the bismuth solders. I think the circuit will probably operate just fine if you use them, as they are conductive, but they are very brittle and will lead to problems in the near future. As to how easy they are to solder with, I don't know, and don't plan to find out. I personally have only used the lead-based solder, in both pace form and in wire. Lead is indeed a severe health problem, especially for children whose brains are still developing. A bit less so for adults, but still not a good thing to be exposed to. But if you are doing things correctly, you should not be getting the paste on your hands. Regardless, don't do soldering where you handle food, and always wash your hands when you are done soldering. With those precautions you should not have any health concerns.
As far as soldering irons go, there is a correct temperature, and going above it is not necessarily a good thing. But you don't need a high dollar iron, From what I have read even the Chinese temperature controlled irons are not that bad, and are available quite cheaply. A Hakko is probably even better, and still affordable by most. Personally I really like my Metcal soldering and desoldering equipment, but that is a bit expensive and less you're a really serious hobbyist or have a lot of money.
As far as the SW and BST terminals, you need to understand what those are for. The SW is the output from the internal FETs, it sends the pulses of power out towards the inductor when one FETs is turned on, and acts as a diode when the FET is turned off. It needs to go to the inductor and to one side of the boost capacitor, and nowhere else. The BST terminal gets connected to the SW terminal by 8.1 µF capacitor, that set up works as a charge pump to provide a higher gate drive voltage for the output FET. It should only be connected to SW through a capacitor, not directly.
As far as capacitors go, for input and output I used 1206 (inch) sized components of the same values you chose. The spec sheet for the regulator specifies those values to be acceptable in most cases, it certainly seems it is specifying the rated value of the capacitors, not their effective capacitance operating voltage. That is how I interpret it, and it worked fine in both of my applications. I also use an inductor which was about the same physical size as the one you used, I can give you the part number if you like.
Since you've got the entire circuit bill anyway, at least that's what I understand, there is something else I suggest you do. Before fixing the problem with the regulator, make sure the rest of your circuit works exactly the way you want. Supply with power from an external supply, program it, hook it up, and see how it does. If it does not work correctly, you can also try to repair those defects on your neck circuit board. It would be a shame to get the regulator working and still need to order another board.
One other thing I should probably mention. You say this is going to be a filament driver, I suppose that could be a lightbulb, but I am assuming it is actually a vacuum tube? If so, is this the final output tube, or a low-level amplifier? I am a little concerned that switching noise on a filament may show up on the output. Also, is the tube designed for use with DC on the filament?
Sadly, I'm very clumsy, and I work right next to my soldering station as i only have a rented room in a 3 bedroom. I eat on the left side of my work desk and my solder desk is on the right 90 degrees to the other table, so it's hard to keep things absolutely safe, specially as I scroll and touch my computer when.checking schematics, and immidietly go probe something, and then continue soldering.
Yeah metcal is insane, I have a pinecil v2 which is only 30$, but I was eyeing the Aixun TD420 (As the t320 and t3a have voltage leak issues that scare me, I don't want to damage an IC etc)
Oh and yeah I understand how it goes, I'm just confused about what's happening in the circuit... Again my much worse old prior layout (with the pins connected where they shouldn't) was making stable-ish 3v but this can't!
and Oh no, this is a Filament drier, it heats up a small container that I insulated, and opens a valve to let moist air out everynow and then. It's for my 3D Printer. Filaments such as PETG/PA absorb a ton of moisture and need to be dried and stored in a low humdity enviroment.
I'm just very suspicous something else is at play... but thankfully I did a check and I don't need new components for a new PCB (have 3 Step down ICs left!) maybe i'll just bite the bullet and order lead-free regular paste and try again with that. or use my leaded paste for now and be extra cautious.
I use this solder at 340c https://mgchemicals.com/products/discontinued-products/4925-4926-sac305-ra-solder-wire/, sometimes tho on hard ground pins I use the boost function, but i'm sure the tip is actually much colder?
Maybe instead of a new iron, I can just get an MHP 50 hot plate. I DIYed an AC powered hot plate but if i'm being honest, I'm too scared of using it. It's grounded by it's very janky... Don't want to start a fire or like get a shock!
-
I'm still skeptical of the layout being the only issue here, but I redesigned the step down to be like the following. I'll double check the board again to make sure I didn't miss out on anything.
Any final words before I put an order?(and wait a month :\.)
-
Okay, for the type of filament you are talking about, I don't think the noise of a switching regulator will be a problem.
I think your new regulator layout looks much better, but I do have one thing I don't like. I don't like the long loop you made on the feedback at all. I would drop through the board with a via, make eye diagonal trace to write at the top of the output capacitors, come up to the top layer was another via their connect to the capacitors. Other than that, I think it looks pretty reasonable.
In your living condition, I guess you are correct about the use of leaded solder. You should not really be using it near where you eat, that is for sure. I have a separate room where I do my electronic work, and the computer in there is dedicated just to my test bench. It is also equipped with a 43 inch 4K monitor, which is quite nice for schematics and circuit board layout. And when assembling, you can have the schematic on one side and the board on the other, when you highlight a component on one of the two KiCad windows, it also automatically highlights on the other. Very nice when you have components so close together you can't put labels on the PCB.
I think you are heading in the right direction, wait for other comments, think it over for a few days, and then order. I don't know why it is taking you a month to get a board though, for a two layer board it's usually just over a week for me. I order from jlcpcb and use the very cheapest shipping possible, For a small board the total cost is just under 4$ US and the board arrives in about a week.
-
Updated it! I guess I went a bit overboard with the whole "don't go near inductors and swithching components" part xD
-
What is the third scope screenshot, the output of the inductor? Sometimes posting a photo of the soldered circuit itself can help.
I would be very surprised if this regulator did not work with some simple modwires/modifications.
-
Sorry, forgot to give info on the "Screenshots" (Sorry my mac is dumb and can't do FAT32, only exFat! and I have no lan ports near my scope)
yes that's the SW pin behind the inductor. This is a very low quality image of the circuit, sadly my phone even tho pretty recent and modern, couldn't get a decent photo! But i'm ordering a mciroscope this june 1st after months of wanting one! My eyes hurt....
I'm not sure how the PK-PK is 6v tho, the base (Lowest point right?) is near 0v and the top is 3v. so that's 3v PK-PK.
I might actually get a reflow plate (MHP50) instead of the soldering iron too, figured it might be a better investment as I only use my iron for reworking errors anyways 99% of the time. It's just right now that I had to use it (and it annoyed the hell out of me because it'd not melt anything that's a near a copper pour :P)
This is all super strange, I wonder if the boost cap is malfunctionnig in a way, and not providing enough voltage to switch the high side.
-
replaced the inductor with a very small 4.7uH one i had laying around and wired up the LED too... Strange.. now my voltage is just proportionally going up with the voltage. My output is just a very consistent sudden rise and a slow fall. it goes way past 3.3v if the voltage is 7v.
the feedback pin is also receiving 4v like the output at 7v. the frequency is 54hz... IT makes no sense!!!!!!!
EDIT: INTERESTING!
behind the inductor, the switching action is happening, then suddenly shutting down for a few MS then starting again. those large spikes are what I see when I probe the SW pin instead of the 3.3v output, those sharp tall spikes are periods of 1mhz switching (Frequency of the chip)
I'll upload the second pic in another reply due to it's size.
-
Sorry for the spam! But here's the SW pin, it starts switching and then pauses for some time.
connected a 10ohm resistor to the output and everything sorta goes dead. no output :| the chip also heats up to 35c instantly.
Also, there's some switching noise on the feedback pin, very very small amounts at 1.1mhz, but it's otherwise the same as the output 3.3v. tried connecting FB to 3.3v with a wire, no change.
-
That pattern is normal for a lot of buck converters. That's what it looks like when they're in PFM (pulse skipping) mode. Try putting some more load, the pulse should be denser.
the feedback pin is also receiving 4v like the output at 7v.
I don't understand. VFB is 4V while VOUT is 7V?
-
That pattern is normal for a lot of buck converters. That's what it looks like when they're in PFM (pulse skipping) mode. Try putting some more load, the pulse should be denser.
the feedback pin is also receiving 4v like the output at 7v.
I don't understand. VFB is 4V while VOUT is 7V?
No, when VIN is 8v (sorry not 7v, I made a mistake!), the vout is 4.7v RMS (5.1v top, 4.5v base). bringing the voltage higher just increases it.
The feedback pin is nearly the same as the output.
My only guess so far is that my poorly designed PCB layout is causing noise on the FB pin which in turn makes the step down go crazy.
-
Sorry, forgot to give info on the "Screenshots" (Sorry my mac is dumb and can't do FAT32, only exFat! and I have no lan ports near my scope)
https://www.eevblog.com/forum/index.php?action=dlattach;topic=430201.0;attach=2264716;image (https://www.eevblog.com/forum/index.php?action=dlattach;topic=430201.0;attach=2264716;image)
Why is C17 missing in this photo? You tried testing it first with it in place right?
Also why does C14 look white, like a low value capacitor? If its 10uF it should be a darker brown color. But I could be wrong.
As you say microscope should help.
-
Sorry, forgot to give info on the "Screenshots" (Sorry my mac is dumb and can't do FAT32, only exFat! and I have no lan ports near my scope)
https://www.eevblog.com/forum/index.php?action=dlattach;topic=430201.0;attach=2264716;image (https://www.eevblog.com/forum/index.php?action=dlattach;topic=430201.0;attach=2264716;image)
Why is C17 missing in this photo? At least from what I see, hard to tell. As you say microscope should help.
Also why does C14 look white, like a low value capacitor? If its 10uF it should be a darker brown color. But I could be wrong.
It's dark grey! And yes, C14 was missing because I took the picture when testing something, it's back on the board now :)
Speaking of which I tested everything else and I think i'm getting ready to put an order. (found out I forgot a pullup on the single wire port).
I also have another theory which TBH I don't like, that is during my process of transferring my components to an SMD reel, I misprinted the labels and as a result put another value of capacitor on the BST pin. This means i have to wait longer and pay 25CAD of shipping to LCSC for just another set of every cap I own! I'll go with quality ones this time and not any random cheap ones I find (say Samsung instead of no-name ones)
EDIT: Cap indeed reads 92.5nF which is close enough :) It's an xr7 50v rated one appearntly!
-
So no meters available that can measure capacitance?
Quality or noname is not a huge deal as long as they are X5R/X7R and not Y series.
-
I thought my multimeter could only measure in the uF range as it's kind of crappy (AstroAI DM600AR) but it can do nF! yes, it's the correct value.
-
Before committing to ordering a new PCB it might be a good idea to investigate why the current PCB is failing.
-The feedback: the feedback can't pick up noise to the extend to make the regulator go crazy because the trace is low impedance. (directly connected to the buck output capacitors.)
-The layout is not perfect but it should work fine.
Other comments:
-NRST pin requires a pull up resistor (10K will do).
-The SWD interface also requires NRST. I don't see NRST going to to the debug header.
-
Before committing to ordering a new PCB it might be a good idea to investigate why the current PCB is failing.
-The feedback: the feedback can't pick up noise to the extend to make the regulator go crazy because the trace is low impedance. (directly connected to the buck output capacitors.)
-The layout is not perfect but it should work fine.
Agreed. First thing would be to make absolutely sure the populated buck converter was indeed the right chip, the inductor the right value (and properly sized) and that there was no short or open-circuit.
Also, not sure I understood everything the OP did, but shorting the BOOST and SW pins doesn't sound like a great idea.
-
That’s the strange thing, my only other guess is that I damaged a component during soldering….
Me and my friend both spent an entire day looking st the data sheet and could not figure out what was happening… it just behaves the same strange way. Ramping up and down, the frequency changes with the voltage, the vout goes high with VIN.
When I added a 10uF electrolyte cap to the output my power led starting dimming and getting brighter like a heartbeat..
I’ve heard of someone using this IC on a breadboard successfully, so it should work despite a suboptimal layout. I’m tempted to order an overpriced SMD to TH PCB from Amazon to test it further
You can also use the swd without NRST and I thought it has an internal pull up already?
At this point, I’m at a a loss, the only explanation I have is heat damage, or my poor storage (dropped in a 3D printed plastic SMD container.
Capacitors and inductors don’t go bad easily AFAIK so it must be the IC.
-
I have used AP6xxxx switching converters in several projects and never had any problem. They work very well. In particular, I've used the AP64501, it's rock-solid.
Yes your layout should be fine, unless we missed something.
Possibly you had damaged or non-genuine chips. I don't know. From your initial schematic, I think you had placed output capacitors after a fuse, which wasn't such a good idea as it would add some series resistance, but apart from that, I don't know. Actually given the package those converters come in, you could re-prototype just the step-down part on a piece of proto board, just to make sure.
-
I have used AP6xxxx switching converters in several projects and never had any problem. They work very well. In particular, I've used the AP64501, it's rock-solid.
Yes your layout should be fine, unless we missed something.
Possibly you had damaged or non-genuine chips. I don't know. From your initial schematic, I think you had placed output capacitors after a fuse, which wasn't such a good idea as it would add some series resistance, but apart from that, I don't know. Actually given the package those converters come in, you could re-prototype just the step-down part on a piece of proto board, just to make sure.
Guess what, this was already my second attempt, but I made a third PCB in desperate before going to bed (Only the step-down components,) and it works perfectly. It has 10mv pk-pk output ripple under a 10ohm load, stable, even voltage even after drastically changing from 7v to 17v. This is with one 22uF Cap not two!
so now my question is, how on earth did I manage to burn through 3 ICs each time, despite using Low temp solder?! is it my storage method (dumping them in a PLA 3d printer container) ESD from my body (I do put the IC under my finger to double check the alignment before putting it on the board) or perhaps just bad luck?!!! the components are from LCSC.
Were they soldered fine and burned on a specific start condition? I do have to note that my pencil does make my LEDs glow as I use it to solder, but I assume it's such a low voltage that it'd not matter. I did have to rework a pin on my PCB all 3 times, including this one. The only thing I did differently was my 195c (vs 250c) temp, and discharging my body with ground before starting. I also did use another power supply with my pinecil (140w capable)
are these ICs just super heat sensitive?!
This little white LED makes me so happy...
I tried to kill it by sparking the input (suddenly connecting the alligator like I did first instead of connecting, then turning on my PSU) and a few other stuff but it didn't die.
-
> I do have to note that my pencil does make my LEDs glow as I use it to solder, but I assume it's such a low voltage that it'd not matter.
It's matter a lot! You probably burned out your IC with this pencil. Glowing LED mean that tip of your pencil radiate high frequncy voltage or just under high ESD level. Ideally soldering tip should be grounded.
-
Well I guess I have an excuse to treat myself to the auxin T420 now xD
But jokes aside I’ll try to do some research to find something that’s grounded and cheap. I like the modern cartridge style ones a lot
EDIT: Excuse me being a beginner, but I measured 55v AC relative to ground on my pinecil's tip?!!! what... how could that be happening?
I guess I could also start using a power bank instead. or use the DC plugin.
-
Well I guess I have an excuse to treat myself to the auxin T420 now xD
But jokes aside I’ll try to do some research to find something that’s grounded and cheap. I like the modern cartridge style ones a lot
EDIT: Excuse me being a beginner, but I measured 55v AC relative to ground on my pinecil's tip?!!! what... how could that be happening?
I guess I could also start using a power bank instead. or use the DC plugin.
That's not unusual with many of the cheap soldering stations and can cause damage. Note that I've read similar issues with the T420 you're mentioning, so you should definitely look for reviews before buying.
-
One remark on the ground vias, as long as you're not using exotic technologies like copper-filled vias, packing them so dense is not necessary and increases both the electrical and thermal impedance in your ground plane. Not that it makes any difference in this particular case, but it's something to keep in mind for other designs where it will be detrimental.
-
The auxin T420 looks like a good soldering iron to me. The tip is not directly connected to earth but trough a high value resistor.
-
Well I guess I have an excuse to treat myself to the auxin T420 now xD
But jokes aside I’ll try to do some research to find something that’s grounded and cheap. I like the modern cartridge style ones a lot
EDIT: Excuse me being a beginner, but I measured 55v AC relative to ground on my pinecil's tip?!!! what... how could that be happening?
I guess I could also start using a power bank instead. or use the DC plugin.
That's not unusual with many of the cheap soldering stations and can cause damage. Note that I've read similar issues with the T420 you're mentioning, so you should definitely look for reviews before buying.
Oh shoo, I chose that one over T320 purely because it has an isolated power supply. What soldering iron would you recommend? That Iron is pretty much the upper limit of my budget, I was going to grab a miniware mhp50 but this seems more important TBH.
-
One remark on the ground vias, as long as you're not using exotic technologies like copper-filled vias, packing them so dense is not necessary and increases both the electrical and thermal impedance in your ground plane. Not that it makes any difference in this particular case, but it's something to keep in mind for other designs where it will be detrimental.
Oh that's strange... I was expecting the absolute opposite! Why does this happen?
-
The auxin T420 looks like a good soldering iron to me. The tip is not directly connected to earth but trough a high value resistor.
is the resistor a bad or good thing? Why would they use a resistor?
Also I'm open to any recommendations, the upper limit of my budget is the TD420. So I'd like something in the same price range. Maybe I should start a new post in the equipment subforums.
-
EDIT: Excuse me being a beginner, but I measured 55v AC relative to ground on my pinecil's tip?!!! what... how could that be happening?
I guess I could also start using a power bank instead. or use the DC plugin.
You can either use a power bank, DC plug (assuming your DC source is earthed, most are not), or earth the unit as shown in this photo:
https://www.aliexpress.us/item/3256806229994814.html (https://www.aliexpress.us/item/3256806229994814.html)
Shame they don't have a clear explanation of this in the wiki, just a short mention at the bottom: https://wiki.pine64.org/wiki/Pinecil#History_of_Pinecil_Changes (https://wiki.pine64.org/wiki/Pinecil#History_of_Pinecil_Changes)
-
EDIT: Excuse me being a beginner, but I measured 55v AC relative to ground on my pinecil's tip?!!! what... how could that be happening?
I guess I could also start using a power bank instead. or use the DC plugin.
You can either use a power bank, DC plug (assuming your DC source is earthed, most are not), or earth the unit as shown in this photo:
https://www.aliexpress.us/item/3256806229994814.html (https://www.aliexpress.us/item/3256806229994814.html)
Shame they don't have a clear explanation of this in the wiki, just a short mention at the bottom: https://wiki.pine64.org/wiki/Pinecil#History_of_Pinecil_Changes (https://wiki.pine64.org/wiki/Pinecil#History_of_Pinecil_Changes)
yeah, it all sounds even more clunky and annoying than it already is, but 320CAD is a LOT of money to drop on an iron, I think I'll hopefully manage to resist and wait until I come into my senses again.
-
Hakko FX-888DX is ~170CAD and won't zap your chips.
-
Yeah but at that point I might shell 80 extra cad for a nice JBC clone, I can’t stand those old school irons that take ages to heat up or choke on ground planes!
-
To be fair, I think the Aixun station I had seen a review of, and which mentioned a relatively high DC voltage on the tip, was a lower-end model and not this T420D, which claims to be isolated. So possibly this one is clean.
Thing is, while cheaper than a genuine JBC, it's already a sizable amount of money for a clone. You can get a decent Weller soldering station for that price.
I do second the greatness of the JBC tips though. Fantastic stuff.The C210 cartridges heat up to over 300°C in under 3 s. Crazy. So with JBC stations and all decent clones, the station goes into a "sleep" mode as soon as you put the tip back into its stand, and usually the sleep mode regulates the temperature to around 150°C (and finally powers it off after a user-defined delay), so when you pull it out of the stand, it gets back to temperature in like 1 s, while prolonging its lifetime significantly compared to keeping it at > 300°C for hours during a working day.
-
is the resistor a bad or good thing? Why would they use a resistor?
The soldering tip is floating which is good. Why you'll need that:
1. The tip will not cause a short to earth when soldering something onto a powered board powered by an earthed power supply. (That will happen sooner or later in the form of doing just a quick fix. Let's put 10 K here or there. Meanwhile you forgot to switch off the power.)
2. The resistor limits ESD discharge currents to safe levels.
4. The tip is always at earth potential and can not be a source of ESD.
-
is the resistor a bad or good thing? Why would they use a resistor?
The soldering tip is floating which is good. Why you'll need that:
1. The tip will not cause a short to earth when soldering something onto a powered board powered by an earthed power supply. (That will happen sooner or later in the form of doing just a quick fix. Let's put 10 K here or there. Meanwhile you forgot to switch off the power.)
2. The resistor limits ESD discharge currents to safe levels.
4. The tip is always at earth potential and can not be a source of ESD.
Ooh that makes sense :)
I think that I'll just run a single thin ground wire to my pinecil for now, lets hope it won't disconnect mid-soldering!
also I have a weller WESD51, I am tempted to order it fresh tips but that'd be 30$. I think i'll deal with this or a power bank and pinecil until I can make sure that it's a good investment, or until I give up and order it at 3am. I also am tempted to try my luck at DIYing a PCB but gosh, I DIY everything... I think I need a break xD Also the td420 has really nice reviews, but I do get the hate on a product like this tbh, the issue is that.... 800cad fo a genuine one is insane to me! and i'm not a profassional.
Thank you all :)
-
If by DIY a PCB you mean produce the board yourself, I will give you a one word recommendation. Don't.
In my experience the quality of all the Chinese PCBs I had made has been excellent, only time I've had a problem is when I sent them a bad design. And the price is unbelievably low. Think about it, 100mm x 100mm for $2 US, Just under $4 total price in your hand if you choose the lowest-cost shipping option. And that price is for either a 2 layer or a 4 layer board, with holes drilled, slots routed, plated through holes, and solder mask. You'll probably have a hard time getting just the blank PCB material for that price, then you have to deal with cutting it, drilling it, and worst of all etching it yourself. It is only a good thing to do if you are a true masochist.
For the soldering station, I don't think I would spend that much money on a clone. I would personally choose either something cheaper like the Hakko, which is very well known to be a decent quality piece, or perhaps spend a bit more and get a nice used JCB, Metcal, etc.
-
Oh sorry forgot context yet again.
I meant to DIY a soldering station. But I already have so many side projects…
-
Bismuth is OK for prototyping. Yes the bond will be weaker and more brittle but your parts won't fall off the board on their own sitting on a bench.
The problem is that if you use such solder, you will contaminate your components and your soldering iron tip, etc. with it. And that will introduce it even where you don't intend to be just prototyping anymore. All it takes for joints to become brittle are trace amounts of bismuth.
And yes, parts will fall off the board by themselves with this solder after a while in extreme cases - the joints do crack. However, far more often you will get broken joints and flaky connections which are a huge pain in the ... to debug and repair constantly.
There is no reason to use such solder for regular soldering. This is a special solder that is used only for assembling very sensitive specialist components, such as some RF ones or in special alloys for desoldering. All other components can withstand regular lead-free reflow temperatures without issues.
-
I had same problem, but with a similar IC of the same manufacturer, "AP62300TWU-7" (https://www.lcsc.com/product-detail/DC-DC-Converters_Diodes-Incorporated-AP62300TWU-7_C1883519.html (https://www.lcsc.com/product-detail/DC-DC-Converters_Diodes-Incorporated-AP62300TWU-7_C1883519.html)), I burned two of them and finally I found out that it was my mistake because I connected the EN pin, which following the datasheet should have " absolute max." 6V, to VIN which was about 7 - 10V.
I compared the block diagrams of "AP62300TWU-7" and "AP63203WU" and I noticed that they are almost identical (at least for the EN pin) but for the former it's specified "absolute max." 6V and for the latter it's specified "absolute max." 35V for the EN pin. Pretty strange. I think it's worth giving a try with max. 6V on EN pin or even leaving it floating as it has internally a pull-up current source. As work-around I cut the EN pin and the problem was solved.
My custom PCB uses 2 buck converters of this type , one for 5V and one for 3.3V. Both worked fine at 7 – 10 V input and EN connected to VIN , the former with a neopixel strip of 19 LEDs and the second with a uC STM32F401 clocked at 80 MHz. But when I assembled the DFPlayer mini mp3 player to the buck converter from the 3.3V side, that step-down converter was repeatedly damaged until I applied the previously mentioned work-around (this is also pretty strange).
-
I had same problem, but with a similar IC of the same manufacturer, "AP62300TWU-7" (https://www.lcsc.com/product-detail/DC-DC-Converters_Diodes-Incorporated-AP62300TWU-7_C1883519.html (https://www.lcsc.com/product-detail/DC-DC-Converters_Diodes-Incorporated-AP62300TWU-7_C1883519.html)), I burned two of them and finally I found out that it was my mistake because I connected the EN pin, which following the datasheet should have " absolute max." 6V, to VIN which was about 7 - 10V.
I compared the block diagrams of "AP62300TWU-7" and "AP63203WU" and I noticed that they are almost identical (at least for the EN pin) but for the former it's specified "absolute max." 6V and for the latter it's specified "absolute max." 35V for the EN pin. Pretty strange. I think it's worth giving a try with max. 6V on EN pin or even leaving it floating as it has internally a pull-up current source. As work-around I cut the EN pin and the problem was solved.
My custom PCB uses 2 buck converters of this type , one for 5V and one for 3.3V. Both worked fine at 7 – 10 V input and EN connected to VIN , the former with a neopixel strip of 19 LEDs and the second with a uC STM32F401 clocked at 80 MHz. But when I assembled the DFPlayer mini mp3 player to the buck converter from the 3.3V side, that step-down converter was repeatedly damaged until I applied the previously mentioned work-around (this is also pretty strange).
This is old, but where di you find this? I can't find that info anywhere. The datasheet says 35v max
-
THE SOLUTION:
Put a 2.2 ohms resistor right before the Vin.
Reasoning & Background:
I had the same issue, ended up blowing 5-7 ICs before i found out that it is due to power in-rush current. Basically, the circuitry works fine when you supply it lower voltages , but when you supply it with a 24V supply , the capacitor which is empty charges rather quickly drawing in ALOTT of current and the inductor isnt able to stop it down in time , this leads to a quite a large current draw and the IC in question just gives up.
Adding a 2.2ohms resistor throttles it so it charges smoothly , no resonant circuit is created and the current stays in limit of the IC's absolute maximum.
This was my thesis, and adding a 2.2ohms resistor basically fixed this problem instantly for me and after 100s of trial my board is working super fine.