Atmega8 is a legacy chip, it was used in many projects years ago, then a next generation Atmega88 appeared (along with others from the family) adding more features and improving existing peripherals. Guess someone in China had a batch of then-cheap Atmega8 chips, and they did what they often do: renamed them as a "better" Atmega88. There was a point when new AVR microcontrollers were expensive and scarce, maybe this is the reason?
If pins are so oxidized that they refuse solder, clean them off the board. Many of those solder connections of the fake chips still looked not connected well even after your last attempt. As for wires, do yourself a favor and get a spool of wire wrap wire. That is great for all wire connections that must be thin. They only need to be tinned before using, because they are not meant to be soldered.
If these are harvested and re-branded chips, the harvesting process may have subjected the leads to far more heat than is normal. This could cause damage to the silicon (hence the bad chip you found) and/or so heavily oxidize the solder on the leads that they become extremely difficult to solder. This is the reason I only buy from Digikey or other reputable sources.
I guess a lot of chips on aliexpress are recycled/scavenged from PCBs from e-waste, then they clean them and put new markings to make look like new. The problem is they put the same markings on all chips don't care what exact was before...
Those first chips looked like they were made in 2010. They have had plenty of time to corrode, especially if they have not been stored properly. You can actually see the difference compared to the new ones. The old ones are grey and ugly, while the new ones are nice and shiny. The bad soldering was not your fault.
this! the oxidisation on the legs prevented the solder from sticking... cleans up with some flux and a hot (dry) iron run over the bottom side of the legs
Came here to say this myself I'm 7m into the vid and i can see the pins are corroded and a gentle sweep of the iron on top of the pin to break through the layer of crud and let the flux do it's magic was all that's needed
Learned this lesson a few years ago. Now i use 800 grit sandpaper every time i have to solder a new SMD chip to scrape the top layer and expose clean metal.
4:45 the first mistake is tinning more than one pad. That way the IC can never be flush with the rest of the pads. Only one pad must be tinned, allowing for the IC to be held in place. Solder only one pin while pushing the IC into position, then check the alignment. If the alignment is correct, solder the pin diagonally opposite to secure it in place. Then go for the others.
yeah, we all love success stories and clean soldering but personally i admire most creators who share their failures for us all to learn. Great job and fantastic final product Noel!
@@blakecasimir also good to see that he makes mistakes just like I do. Whenever I mess up soldering I always think "why can't I solder like Noel Llopis? he always does a perfect job first time!" - now I don't feel so bad! 😊
You forgot the basic number 1 rule of soldering.... heat the pin AND pad at the same time while touching the solder to the pin.... when the pin is hot enough the solder wil melt and flow properly to the pin and pad and the joint will be smootth and shiny.
Surprising how few people here recognise the root cause of the difficulty you had getting solder to bond with the IC legs, but it's very simple actually. You could see those off-colour, dull looking legs from across the room. Loads of corrosion and/or some sort of dirty build-up. Fresh tip: don't bother trying to solder onto corroded / dirty metal. Solder needs to actually contact the metal surface, and the metal surface of course needs to be at a similar temperature to the melting point of your solder at the point of contact. Careful mechanical abrasion is usually what you need! Like, a small square of medium / fine grit sandpaper held with tweezers. Or just the blade of an X-acto knife. Then scrub with a bit of IPA, dry it off, and get to work. Sometimes only the scrub is needed, with an organic solvent (IPA, acetone, MEK) if you just need to displace some tacky or oily dirt. There is the option of using an active flux. These are usually acidic (often phosphoric acid!), and typically not noticeably aggressive at room temperature. But, with the heat of the iron the acid gets angry and can eat through most corrosion. Some active fluxes are *very* active and will easily eat through Fe2O3 (red rust), and your soldering iron tip. Note, however, that you must clean active flux residue off VERY thoroughly. And I do mean VERY thoroughly. Any residue hiding, say, under an SMD component, will destroy whatever it's touching (starting with exposed contacts and copper traces) over the course of the next few months / years. It will catch you out. Not recommended for this kind of work.
that... and it is very possible his flux is just hot-garbage. (as much trouble as he had with soldering the pin-headers on the board, the solder definitely isn't "flowing".)
The important part of soldering is to heat up both pieces of metal to be joined at the same time. Applying the tip of the iron just to the pad isn't heating up the pin, therefore you won't get a good joint.
I'm an electronic technician who specialises in rebuilding and restoring older machines, like Neumann matrices engravers for vinyl mastering, as an example. I understand the frustration that comes with that kind of work. I spent a couple of days searching for intermittent trouble in a machine I just finished to rebuild. Since the client and I had already invested 10k in time and parts, I wasn't allowed to give up, I found the trouble, but man, do it get frustrating sometimes, lol 😅
I worked on and maintained a VMS80 for a few years. Can be very nervewracking to troubleshoot, you dont want to blow out a cutterhead because you goofed a repair on the amp ;)
Hey Noel! So, I'm about 7 minutes in. You first mistake with respect to soldering the chip is you DID NOT heat the pins themselves. Solder sticks to hot metal, it really doesn't like to stick to cool surfaces. A bigger tip DEFINITELY would help here, your general technique is great though. You probably could use hot air here. I'll update via comments my notes to you.
Yes, usb ports in windows ARE NOT all the same as far as devices go. It will allocate differently, in fact windows even puts the usb port it's on as part of the device path internally. this has tripped me up a few times. This is why people have issues with OBS, because windows does not treat a device the same across multiple usb ports.
With these small pins usually it’s fine to not touch them. The problem is the tip, I have a similar one and never use it as it does not work well and/or I have to crank the heat with it.
The problem is NOT the tip. Before soldering, ALL PINS must be cleaned (mechanically) and tinned separately. PS, the microcontroller does not look genuine.
Oh boy. I went through 6 pairs of Darlingtons and 4 rare Schottky diodes and finally blew a loudspeaker from replacing power transistors with Chinesium ones. Then I bought UK sourced ones from a reputable supplier, a replacement voice coil from the USA and finally got an old Fender guitar amplifier singing. Like you, I wondered why would they bother to rebadge scrap at such low prices? Then I remembered how Sir Clive Sinclair started in business...repackaging scrap transistors and reselling them as 'special' ...
One of the most braindead things in Windows USB drivers is that they are port specific. It is to me completely absurd that the device gets treated differently in a different port. I mean, it is good to have that as a possibility, but not as a default! This has been the way since at least Windows XP (probably even in older versions), and you can e.g. see Windows installing new drivers whenever you plug a device to a port where it wasn't before. This is particularly problematic when the default drivers don't work out of the box... I think Microsoft doesn't really know what the "U" in USB stands for.
Man, this led ya on a merry chase, no? I am no expert on SMD soldering, but I have done a bit of it in my time. If you don't have some, get some rework flux and low melt solder from a reliable vendor (I like Chemtronics, but that just me). It can make a surprising difference, especially when using hot air. Of course it didn't help that your microcontroller chips were heavily oxidized from the recovery process. That discoloration on the legs was from heat, and it really made things difficult for ya in terms of getting the solder to stick. You figured it out, though. Those chips are cheap enough that getting new stock from Mouser was a genius move, especially since the first ones you got were rebranded junk not actually suitable for your purpose. Time, patience, and perseverance encompasses all things. Your eventual success here is a prime example of that. Thanks for sharing your trials along the way so we can all learn from them! That kind of gold one doesn't find every day.
Lots of good info in the comments on the specifics here, but as a guide for other who may be hobbyist solderers(?) from someone who does a lot of nano and micro rework by hand as a job: I use tips way bigger and iron temperatures way hotter than you would assume for any given task. Plus, not all tools and materials (station, tips, solder, flux etc) are born equal. I have helped beginners who's starter soldering kits make things additionally hard for them and me! One tip when using a monocular scope/screen is that pins can look OK from the top down and a decent nudge test is the only way to check, and absolutely soak the pins in the first place.
true, but the device isn't reporting the correct ID, e.g. it is reporting as a simple RS232 device and not as a JTAG device, the act of associating the driver with the device on the port is why the strange behavior occurs. The reason for this is simple, it stops other devices from being used with the "new driver" when they aren't that type of device. Really what should happen is the device should appear as "unknown" and a driver would be required to be loaded to get the device to be seen at all. Then, when the device is moved to a different port, Windows would then use the correct driver for it. I'm forever getting hassled by people at work on Windows who have reached COM255 from plugging and unplugging RS232 dongles 🤣
16:50 Why don't you crimp some new Dupont male pins? I would still solder the 1.27" pitch header, and solder to the pins on the other side, bent them up and down, not sideways. Shorting power should not destroy your chip, perhaps your programmer could get damaged.
It really looked like there was some kind of oxidation on the pins of that ATmicro. That would make soldering nearly impossible. I'd try some ultra-fine sandpaper on the pin's contact surfaces.
27:37 that's why when i am soldering any SMD components like QFNs i gently push on them to make them flush with the board; unless legs were bent before usually it suffices for good contact. And yeah, apply that much heat during soldering to actually heat pad AND the leg to avoid dry joints.
I know everyone is a know it all with soldering, but maybe something that will help is using a KU knife tip. I very rarely use anything else and works great for through hole and SMD components. Those tiny cone tips are absolutely useless, and as long as there is flux aplenty, I find bigger tips are fine for even tiny components.
This counterfeit chip business is crazy. Wish someone would take an in-depth look into how much is actually made form these low-cost chip rebrandings. I faced that a long time ago when I bought power transistors for my Fender Passport sound system..
I'm not sure what's going on with your solder or solder technique, but for sure those solder beads aren't acting right on that microcontroller. The solder should be "wicking" under and even up onto the chip legs. Like if you were using a reflow station, you don't even *touch* the legs. You just tin the pads, apply hot air to the legs, and the solder wicks up and connects everything. I'd probably take that chip to some sandpaper or emery board to eliminate the possibility that it's got oxide on the underside of the legs. I'd also probably check my tin/lead ratio on the solder. If it's all tin, you just have to deal with the fact that it's not going to wick as well I guess. But definitely if I was experiencing what you were experiencing when you were trying to mount the microcontroller, I'd have stopped and solved that before moving on. Something's not right. Solder's too cold, bad lead/tin mixture, oxidized legs, not enough heat being applied to both the legs and the pads, **SOMETHING**. The way that solder was acting, you're almost guaranteed to have a cold joint.
Yeah, my first thought as well. The pins on the microcontroller looked oxidized and the solder didn't flow as it should under and around the legs. As you say, cleaning the legs before soldering probably would have made things way better.
Great to have you back. I felt your pain. My tip to check SMD solder joints: hold the board at a shallow angle and look through a magnifier. If the joint is good you’ll see a shiny ramp of solder from the pad to the pin. If you see anything else (ie a step) the joint is bad. Since discovering that I’ve not had a duff joint.
There are so many cool projects I'd like to do, but I don't. When I was young, I had energy and was sharp. I also had confidence and better fine motor control. As I get older, I find I'm losing those qualities. I end up getting really frustrated and if I'm lucky, will get the job done with a lot of blood, sweat and tears. If I'm unlucky, I'll put it aside and never come back to it.
This is a great lesson in problem-solving. It also warns us to be careful about the chips we buy. Most counterfeits that I have seen have blurry lettering and incorrect logos. Try to find images of genuine chips online and compare. Another great takeaway from this video is that there is now a new source of SID chips for the Commodore community. Hats off to whoever created it.
Congrats! Genuine distributors should the first place you look, buying chips from online marketplaces is a risk you take when you've run out of other options. Great video as usual, I love your perseverance and glad you won in the end.
I think it is important to see also videos like this. Often I do not remember to read properly the messages. I am quite varied about the fake chips that are around and it is difficult to find a reliable source that suits for hobbyist.
Some chips are harder to solder than others. What stands out to me is that you only heat the pad and not the pin. For the best results, you need to heat both the pad and the IC pin with the soldering iron. If you don't touch the IC pin, it will be at a lower temperature and won't pull solder onto it. Additionally, if it's an older IC that wasn't stored properly, there may be some oxidation on the pins, so it's essential to burn that off with flux. Therefore, heating the IC pin is crucial for successful soldering.
21:27 None of those legs looks like they are solder properly. I wonder if the legs are made of some material that won't like your solder tin. Could be a low budget chip with contaminated legs? You really need to clean your solder tip.
I feel slightly better about my own soldering skills now. Not good, but if you can struggle then I understand my own struggles a lot better. That microscope is looking like a game changer...
Fake components are a real problem. I bought some NE5532 op-amps; they are pretty cheap even from "proper" suppliers, and the entire batch were fake. They were op-amps but not NE5532. It took me ages to pin down what was going on. Don't buy components from any of the "cheap" sources.
Hi Noel !! This happens too much here in Brasil !! Super low quality of the leads of the chip !! Some Chinese chips must have the side down of "legs" *_gently sanded_* and ressoldered to get a good quality of assembly !! Or some drops of _*deoxit_*... Cause of our tropical weather, all stock chips tends to little oxidization !!
I'd use some solder paste instead of a wire for that chip. Has included flux and a lower melting temp (lower temp difference = easier for it to stick), much easier to work with small pads. You can also place it before you add the chip, so that it is much less likely to not make contact
Fantastic video! I love how you explore the difficulties and mistakes during the Nano SwinSID build, especially dealing with soldering issues and counterfeit chips, it must not have been easy. It's really true that even simple projects can have unexpected challenges. Looking forward to more content like this. :)
27:29 not to sound harsh, but you barely soldered them in place. Bottom pins look untouched. Untill the top of the pin is covered in lead, don't consider it a good contact.
With SMD try a K tip for your soldering iron. Don't tin any pads before placing the part. Solder paste can help. Just don't give up. I started with AVR stuff and a few years later I'm now working with 400 pin BGA parts.
Really interesting video. I also get worried about causing a short circuit when I need to take a measurement and the two pins are very close to each other. And let's not even talk about when the plastic melts while trying to solder a connector 😅
As someone who has been doing SMD soldering for over a decade, you pretty much nailed it that the tip was all wrong. Chisel tip, load and drag. You got the amount of flux right at least. Not sure what temp you were soldering at. Looked a bit cool, especially for that tip. Bit painful to watch stuff like this. Edit: Some medium active flux would have helped as well.
Not really. It was obvious the real fault was not applying enough heat and not applying that heat to the pad AND leg simultaneously. Pretty much everyone who commented got it, seemingly except you with over a decade of experience.... Hmmmm....
@@g4z-kb7ct Not sure what your issue is. Not enough heat due to the crappy pointy soldering tip. It's the same net result. The sharp little tips never transfer enough heat. Bob you should know better then to attack people.
As for why they would fake cheap chips with more expensive ones might just be an observation bias. You compared the prices the chips have now but you should compare the prices they had when they were made (according to one commenter here the chips seem to be 10 years old). The 88 series looks to be more recent than the 8 models. It's possile that when they were introduced that they were rare and more expensive than the older and cheaper ones they were supposed to replace.
You've got quite some suggestions so maybe mine is redundant but: 1. yes, you need the biggest tip you can use for soldering. Maybe that chisel one is too big though! :) 2. Heat up the pad and the leg for 1 second, THEN apply solder. 3. Solder wire, flux, solder tip quality they all make a huge difference. I use Loctite solder wire, Kingbo flux (Amtech is just way too expensive) and I believe you have a Hakko? Those are genuine Hakko tips? Fake tips will deliver... fake results :) 4. Apply more solder on the first pad. Only one pad or the whole chip will be elevated. Once the first leg has fallen into the solder - and the whole IC has touched the PCB - then you can solder the second leg, then all the others. Good luck! :)
I have a few nano's rattling around in a drawer somewhere. Now the SwinSID Ultimate is a different story. I love that one. But getting ahold of one these days is a bit tough. My favorite has to be the FPGASID. The psudo-stereo option is fantastic. Voices 1&2 on left speaker, Voices 2&3 on the right speaker makes old games seem line new.
you should be using paste and a hot plate to solder the microcontroller. as for the tiny programming connector, the best option is to make a pcb programming adapter with the proper connectors
It's actually faster to re-design the whole PCB and put a proper AVR header on it... that's what I did. I've been re-designing lots of projects to optimise design, traces and part cost for over 20 years.
Jeez, this video is like watching a train wreck in slow motion - seems like you know/understand everything, yet managed to stumble each step on the way, TQFP is the easiest smd packages to solder, not buying chips from reputable distributors, not using standard ISP connector, not checking the ID, like HOW? I've seen bad days, but this is quite something else.
@@lis6502 must be *very* infrequently - I can solder that TQFP blindfolded, upside-down and drunk at the same time, I've also taught bunch of ppl to solder - once past the 'Aha!' moment, they never forget.
@@VEC7ORlt it's about one's personal abilities. I am able in the same intoxication/ exhaustion condition to troubleshoot not working JBoss or solve network access issue. My friend is able to drive a car or bulldozer( haven't tested but i am sure he'll manage) completely drunk. My point is: i don't know what Noel's innate talent but surely it isn't soldering ;)
@@lis6502 in the end he eventually tried everything - lifting the pins, testing for continuity, maybe it still was a big fat chunk of bad luck on top of another.
thats why IC-s are kept in moisture controlled enviorment. That chip was probably siting in amazon rainforrest for a while. Even the flux could not get rid of oxide. Soldering with a axe, wont help also.
Awesome video. I'm a bit of a novice, but I've been trying to build custom devices with the Raspberry Pi. You channel has helped so much in learning systematic troubleshooting. This one also helped in the don't give up part of building a project. Keep up with the great content.
As I am very comfortable with ATmegas and their programming, your odyssey almost made me cry out loudly. I am not through the video yet and wonder if/when you will discover that you are trying to talk to fast to the ATmega88. And I have seen much better solder skills from you in the past, haven't I? When handsoldering TSOP packages you really need to apply heat to the individual pins - drag-soldering or pin-by-pin. The message about the signature you got from avrdude indicates that your ATmega88 actually is an older ATmega8 - could it be a relabeled fake? Also I can really recommend the GUI avrdudess for avrdude.
Have discovered something... The specs of the atmega 88 is 20 mhz... Swinsid is operating above... I wonder if this makes problems to some atmega . Try programming before installing the crystal. ;) This is why the eprom is not used... Just the flash...
It doesn't require anything more than a simple sanding with some 400 grit sandpaper. I do a lot of arcade repairs (for roughly 30 years) and I have to do that a lot with some boards that are 40 years old.
Thank you for sharing this video, Noel. I have been in a similar situation not long ago, but ended up giving up and concluding a lack of skill. Your video showed me it is not just me. I should give this another go…
I smoked an ESP32 chip the same exact way by stupidly shorting one of the i2c lines to 12v power from a shunt. The slightest little slip a buzz and a spark and that’s it, all that work gone out the window.
I think, maybe the μCs are a bit old and the pins were oxidized. That could explain the soldering problems. What makes SMD soldering fun is a hot plate. I have been hand soldering SMD for probably 30 years, but the hot plate was really a step forward. Much better than hot air. And for cable making, crimping is much nicer than soldering. The IWISS SN-025 tool is great for Dupont connectors (although all crimp tools seem to be recommended for Dupont, the previous tool is the only I know). The SN-28B is good for many other crimp terminals.
Lol, Noel, I have so much resoect for you, and I admire you trying to going into a field that is so unfamiliar to you, but some of the mistakes you make, makes me wish I could have been your hardware-mentor, because those mistakes are so elementary,,,, from soldering AND LEARN TO USE AN OSCILLOSCOPE TO FIND FAULTS…. And then I look at my pathetic attemps to write software, and still I wonder why the software engineers hated my code, lol…. Pushing things beyond our safe spaces is hard, and I admire you for it, especially putting it all on RUclips.👍👍👍👍
I had a similar issue trying to build a Greaseweasel and trying to find legitimate STM32F103C8T6. They weren't rebadged, but the sellers would list the chip but send a different model of STM32.
I had to program an ATMega32 yesterday. I used the AVR mkII. Had to take the chip out and put it in the ZIF socket to get it to program. The ICSP header didn't work. I believe I needed to provide external power to the chip. Had to plug in 5V to the ZIF socket.
I should start to make RUclips videos too. ALL of my electronics projects go wrong. Buying chips from Aliexpress is one of the biggest contributors to that. I totally felt your pain when I saw the soldering challenges. I spent all day Saturday doubting my soldering skills due to dodgy CH340 chips from Aliexpress. When I was a kid I used to kill many ICs trying to desolder them so I've always been wary about how long to keep the iron on pins. That's the whole reason I've been too chicken to solder a Spartan 3 FPGA onto a PCB. Drag soldering scares me - always worried a bridge will form underneath the IC somewhere where it can't be seen. The other mistake I make is trying to design a PCB, getting it manufactured, waiting 2-3 weeks for it to arrive and then discovering there's a mistake or I haven't added something. Just did that - didn't add the ISP pins to program my AtMega2560.. I'm glad to see that other people experience this pain. I keep thinking it's just me and that maybe electronics and I were not to be :-)
You got yourself some expired ICs. These get rejected by electronics manufacturers because they have built up oxide on the legs/pads and get sent off to Huaqiangbei markets. You need to recondition them first by retinning the working surface of the legs by any method that works for you. Device signature doesn't match... are those fake chips as well? Huh. Ah yes, yes they are. Excellent.
I don't know if this has any bearing on the legitimacy of the chips, but looking at the video (around 31 minutes in) the Atmel MCU you replace the 88PA is clearly and sharply labelled ATMEGA8A, if this is one from the batch sold as 88PA that would definitely point to something wrong.
Just because something is open-source does not mean it's a good design or the released info is good. In fact quite the opposite, most projects are poorly designed and pooly implimented. I've been grabbing various open-source projects for roughtly 20 years and making them much better with proper PCB routing and better part selection. For example one SD2IEC version used a sdcard slot that cost $10 lol! Another project used an amp chip that cost $15! Of course I've also been reversing some not-open-source projects too hehe! As for SID replacement's I recently reversed something that is said to be open-source but actually isn't. If you want the schematic and gerbers you can find my contact info on my about page. You know me under a different name and we have spoken in the past. This is something you definitely want ;-)
Its very good you released this video. Its very hard to learn without seeing mistakes or even knowing what one does wrong. - Flux is either bad or not enough heat to activate and clear oxide from those crusty old Atmegas. - no amount of bending legs will help with oxidized pins. - Soldering Iron you are using is ancient Tip on heater ~70 year old Hakko (clone?) design, good for thru hole and fixing headphone cables. Not the best for SMD. Integrated cartridge (Hakko T12/JBC C245) clones are really cheap and miles better. - Tips. @14:10 hardly anybody with SMD experience will ever tell you "this tip is too large" :) First tip is not good for anything, no idea why they even manufacture those needle cone tips, perhaps as a joke to newbies? or for carving wood? Second tip much better in thermal mass department, but bad at transferring solder where you need it - you kept adding solder to the top of the tip with nothing going to actual pads/pins. Optimally you want Beveled tip with small solder reservoir. - @27:20 "no way visually" while its Clear as Day those pins are not soldered :o - @27:42 "are they wiggling at all .. those look fine" while you can clearly see in the video at least the first one is in fact wiggling :o - @gnif already commented below about touching both pad and the pin when soldering, otherwise you are counting on thermal transfer thru parts and flux which might not be sufficient. Hotair would work much better than soldering iron in this case heating everything uniformly, and I think @30:48 you finally had all pins properly connected with hotair :) Irrelevant in light of remarked atmega8 tho :-)
Most of this seems like assembler error.... Soldering technique and not being familiar with tooling Thanks for sharing your learning process and experience for everyone to learn from!
For SMD I *never* pre-tin the pads. Even if you remove it, it leaves a slight "bump" on the pads. Which makes alignment harder. The pads need to be the same height and FLAT. As for counterfeit chips, I always buy modern chips from Mouser or Digikey (or maybe Jameco). I only buy vintage chips off eBay and I try to make sure they are legit. Just not worth the effort and time on questionable chips when a modern one from Mouser isn't much more expensive.
[15:50] DO NOT use these cables for soldering! They are usually made from ALUMINIUM (often even painted), not copper! Thus, they cannot be soldered at all.
Like your video. I always, always check my connections with a multimeter to check for proper contact AND shorts of supply rails and all other traces. When I do apply power , I do so with a bench power supply and limit the current to 50 mA. If anything has been overlooked the current will be very small and most components will survive that much current. Plugging the board without testing to an USB to FTDI adapter that can supply upwards of 1 amp is not smart. Stuff will release magic smoke if something is wrong. Same goes with repairs, always use current limiting and know what voltage is needed for supply and digital/analog interfaces.
just buy good soldering iron. quick 202d + 303k soldering tip on it. u can solder several legs of ic (even 5-6 in row). with good flux u can easily rid of excessive solder. for example kss-800 flux can handle even bga cpu soldering process or u can use "martin" flux but its expencive, and just less foggy than kss-800. dont use thin tips, especially with soldering irons , that use ceramic heater in it. so even oxidated ic legs are not problem. but ive faced problems with china headers 2.54 (as in video) they werent oxidated, but i couldnt use them with regular flux. i dont use active fluxes because they are damaging soldering tips(nickel covering). anyway i am happy to see ur video again. i wish it come more often)
I am doing a Vicky-Twenty build (going on 2 years now) and I am having all kinds of issues. All logic chips test good, all the MOS chips work (I have a CR VIC-20 I test them in and all work when swapped out), so I know your feelings on this
neither your soldering tip nor the PCB nor too much flux was the main issue for your soldering problems in this project. the first atmega looked REALLY crusty, oxidized to oblivion. at that point, even really acidic flux won't help that anymore. apart from that, don't ever try to solder those cheap jumper wire dupont cables, they tend to be really corroded as well making for horrible joints. a cable recommendation: 28awg silicone fine strand tinned copper wire. about the one non-connected pin 17 on the second set of boards: you can actually make those out visually, if there's no solder just next to the pin on the half of the pad, they're generally bad. also, if the solder "balls up" on the half of the pad that isn't under the pin, and goes down towards the pad, it generally is bad as well. last thing I noticed, using a 32 Mhz oscillator can't be correct, the atmega88 is only made for up to 16 Mhz crystals, not 32 M, I'm really surprised that that worked...
I ran into a similar problem recently. I was making a replacement for the keyboard controller in a Mac Plus keyboard. I initially prototyped with 40 pin Mega32 (needed a lot of I/O to match the matrix along with everything else) and same problem - the IDs didn't match. I got that batch from Amazon. They kind of worked anyway, but I got more from Digikey and they were perfect. Finished up with a surface mount version in a 40 pin DIP form factor. I saw you used hot air to remove the chips, but have you tried it for soldering them on? The Mega32s I had to use were much finer pitch that those 88s. I just laid down a thin line of solder paste across the pads, stuck the chip on, and hit it with the hot air. Once the solder melts it's attracted to the bare pads and legs and it just kind of "snaps" into place. Way easier than trying to do it with an iron.
You need a hot air station and a better soldering iron like something from JBC. Ditch the hakko, or like me, keep the old hakko in a box under a table in your garage where it can collect dust. $450USD is a lot for an iron but with quality tools, soldering goes from being a PITA to being easy. Surface mount stuff is generally best done with heat plates and hot air. You can get a quick 861DW (cheap) hot air station and a nice machined hot plate from amazon also for cheap. That would turn a solder job like this into a 10 second easy task. With proper tools, I've found I prefer surface mount soldering over through hole.
Thank you for making this video. I learned a lot through it. It was great to see your process and understand how you went through figuring out each of the various different issues and overcoming the problems.
Never say never but I would be very much amazed if you blew that first Atmega up by accidentally shorting any pin out on that chip. They are very hardy chips. Ive shorted Atmega 328 and 2560 out many times in the past either pin to pin or pin to a power rail and never once destroyed one or lost an I/O pin from doing so. A quick short to ground on a high level pin would be extremely unlikely to blow the chip. Of course if the short persisted for a length of time maybe then but a quick slip I would really doubt it. Perhaps might have died from excess soldering heat more than anything? If anything I could imagine the magic smoke coming from that first USB programmer perhaps depending on how the 5V is running through the board and if its switched etc via a transistor? If the supply power was to short that could very well destroy something on the programmer then? Awesome work, Noel. Always great to watch your video's.
I used to watch Louis Rossmann to learn about microsoldering (and some salt - now it's all salt - so look at early videos) as I wanted to repair lots of broken things I had around me, PS4 and XBox 360 HDMI ports, USB-C port on a phone etc. So I bought the equipment that he used (only in function, not exactly the same) and started learning to solder with hot air needing very little iron work. I got very good at it. When you took the dead IC off you used hot air, do the same for putting it on, it saves a lot of time. Just put some solder onto the board, either paste or wire with the iron, then plenty of flux, heat the board, then place the IC onto the board and keep swirling the hot air around the legs, eventually the IC will magically float into place, gently tap it with tweezers and remove the heat. Job done. Regarding the fake parts, the wrong signature is real red flag - it means you don't have what is marked on the IC so that immediately shouts fake. It is also likely that the process the Chinese component sand-and-etch brigade use affects the plating on the component legs leaving a scale that needs a lot of heat and flux to remove. The connector looks like 1.27mm (0.050" - 50 thou/50 mil) I guess if you were programming in any volume, you'd make a jig to push spring loaded pins onto those pads from below.
These finer Pitch programming throughholes would have really bugged me. I guess i would have bought some 6pole jst connectors which would fit on the board perfectly. Programming an IC without solid connections, that is a real pain....
the pads on the chip may have been oxidized just a layer... that was the problem on the soldering, should be easier... sometimes something abrasive to remove that layer helps. flux and put some solder in there...
4:25 You fail at soldering. Even though I was only a technician, whenever the lecturer wanted to teach students how to solder, they'd call me down to demonstrate and talk them through how to do it correctly. Rule 1, of soldering: Apply heat to the pin AND the pad. You're just applying heat to the pad. Rule 2, don't apply solder to the part you're heating, apply it where they meet, or the other side if you're breaking rule 1. You need to actually press your soldering iron tip down against the top of the chip's leg. Applying solder to the pad while you're heating the pad is definitely the wrong way to solder. 39:29 No, I could tell when you placed them side by side, the new ones look a lot "Cleaner", less oxidised, so it should be easier to solder. The old (Fake?) ones looked like the pins needed cleaning.
6:57 Yeah lol. Years ago I re-designed my own swinsid board and made the design much better including getting rid of those silly small-piched holes and 43:28, yes there is room. I added the proper 2x3 AVR programming header :-) If you want my gerbers or the full source Eagle PCB files let me know.
If a driver is working in one USB port and not in another USB port then the driver is a little faulty, stop searching, end clumsing around. Get better drivers. The decades of Windows programming are over. It is 2024 and close the era of Windows, just like most users did. Often there are many USB thingies to do something serial with something else. Pick another and stop soul searching for mistakes YOU did not make.
Yeah, small needle tips are pretty much useless untill you overheat them. Very small thermal mass, the needle shape doesn't help at all to push some heat into the board, and for some reason they ship with every soldering station. Meanwhile something in the shape of T12-KU or BCF3 does all these small jobs in one pass and doesn't lose temperature or break that often.
helps to tin the legs on the chip too... i think that was alot of the problem.. the chip legs looked like they had a layer of oxide causing the poor solder joints....at least run the iron over the bottom side with some flux to clean off any oxidisation on the legs before soldering to the board also you could have used any other arduino board with an ISCP port to program that micro... you dont need special programmers (the usb one you eventually purchaced is the same chip as on an arduino nano...look up Arduino as ISCP programmer) one last tip.... it always helps to investigate ALL the errors listed when programming fails... starting at the top most error!... the device id not matching is a huge clue the chip is not genuine! Thanks Noel... missed regular vids from you... but it get it..life happens!
So you are tinning the pads but then not tinning the bottoms of the legs of your chip? Watch some of these Indian guys with these giant irons they use easily sticking tiny things together in these very scary skeleton circuits and they are lifting the entire circuit up by literally any component. Tin both items you are putting together. Soldering header pins. Oh man. Plastic melts...simply pause for a few seconds between pins. But even better don't solder header pins. You can get kits with JST connectors. The male receptacle is simply a plastic piece that will slip right over your pins. All the wires in the kit already have ends on them, you just push them in. No crimping required. Just get the right dot pitch. I think xh and ph are the most common sizes I use. PH should be Arduino size. No crimping at all. No one wants to do that crap . Those little Dupont wires are fine for prototyping but not reliable on stuff you are plugging and unplugging, moving around. Under $20 on Amazon for the pre crimped JST kit. All the wires are separate, ends on both ends already, and all the popular colors. Now those big leads on your meter are clumsy. I went ahead and made myself something more precise and easier to control. If you have any plastic laying around, the ones that come bundled with something and you probably have stuffed in a junk drawer. Get em out. Grab a couple of sewing needles(they are sharp and also very strong). You're good at melting plastic, you know what to do next, so do that. Get some super floppy silicone wires and alligator clips or sacrifice an existing test lead with clips I'm sure you have and make yourself a pointy little set of probes. Probably blunt the tip if the needle. I have second set of these with pins on the end if the leads so I can stuff them in a TC1 component tester. Much much easier to test SMD resistors and caps and the thing most meters don't do, inductors. Tweezers leave a hand free so you can press the stupid button that is way harder to push in than it should be.
Unfrikkin' believable.. why would they go to great lengths of rebadging ICs. Definitely good to see you got it working in the end. I have had similar issues with USB ports but in Linux. The port can stop working after resuming from hibernation and the issue is fixed by rebooting.
5:40 i never tought that watching someone's soldering job might be my guilty pleasure, but yet here we are ;d. As mentioned by others: skill issues ;) That's why i never will be grateful enough for local flea markets where you can buy virtually any electronic junk for a dime and train your skills, testing if 400 degs C really helps to make job easy ;)
The Avrdude correctly reported that it probably is an m8 chip (ATMega8). The Avrdude has a pretty comprehensive list of device IDs.
Atmega8 is a legacy chip, it was used in many projects years ago, then a next generation Atmega88 appeared (along with others from the family) adding more features and improving existing peripherals. Guess someone in China had a batch of then-cheap Atmega8 chips, and they did what they often do: renamed them as a "better" Atmega88. There was a point when new AVR microcontrollers were expensive and scarce, maybe this is the reason?
If pins are so oxidized that they refuse solder, clean them off the board. Many of those solder connections of the fake chips still looked not connected well even after your last attempt. As for wires, do yourself a favor and get a spool of wire wrap wire. That is great for all wire connections that must be thin. They only need to be tinned before using, because they are not meant to be soldered.
If these are harvested and re-branded chips, the harvesting process may have subjected the leads to far more heat than is normal. This could cause damage to the silicon (hence the bad chip you found) and/or so heavily oxidize the solder on the leads that they become extremely difficult to solder. This is the reason I only buy from Digikey or other reputable sources.
Or factory rejects/defective chips that got rebranded.
.... Or just random parts with right number of pins.
I guess a lot of chips on aliexpress are recycled/scavenged from PCBs from e-waste, then they clean them and put new markings to make look like new. The problem is they put the same markings on all chips don't care what exact was before...
Those first chips looked like they were made in 2010. They have had plenty of time to corrode, especially if they have not been stored properly. You can actually see the difference compared to the new ones. The old ones are grey and ugly, while the new ones are nice and shiny. The bad soldering was not your fault.
this! the oxidisation on the legs prevented the solder from sticking... cleans up with some flux and a hot (dry) iron run over the bottom side of the legs
@@WacKEDmaN Try a melamine sponge (Mr Clean Magic Eraser, if you're in the US). It'll scrub the oxidation off like nobody's business.
@@AtheistWren good flux usually does the trick quick and easy with tools on hand....Kingbo flux ftw!
Came here to say this myself I'm 7m into the vid and i can see the pins are corroded and a gentle sweep of the iron on top of the pin to break through the layer of crud and let the flux do it's magic was all that's needed
Learned this lesson a few years ago. Now i use 800 grit sandpaper every time i have to solder a new SMD chip to scrape the top layer and expose clean metal.
4:45 the first mistake is tinning more than one pad. That way the IC can never be flush with the rest of the pads. Only one pad must be tinned, allowing for the IC to be held in place.
Solder only one pin while pushing the IC into position, then check the alignment. If the alignment is correct, solder the pin diagonally opposite to secure it in place. Then go for the others.
yeah, we all love success stories and clean soldering but personally i admire most creators who share their failures for us all to learn. Great job and fantastic final product Noel!
Yup. Exactly right. Steel sharpens steel.
Welcome back, Noel. I was a bit worried about you after so much time. Thanks for another great video!
Indeed. It's always a welcome surprise to see a new video from Noel.
@@blakecasimir also good to see that he makes mistakes just like I do. Whenever I mess up soldering I always think "why can't I solder like Noel Llopis? he always does a perfect job first time!" - now I don't feel so bad! 😊
You forgot the basic number 1 rule of soldering.... heat the pin AND pad at the same time while touching the solder to the pin.... when the pin is hot enough the solder wil melt and flow properly to the pin and pad and the joint will be smootth and shiny.
This right here I was going OMG put the iron on the pin as well.
Surprising how few people here recognise the root cause of the difficulty you had getting solder to bond with the IC legs, but it's very simple actually.
You could see those off-colour, dull looking legs from across the room. Loads of corrosion and/or some sort of dirty build-up. Fresh tip: don't bother trying to solder onto corroded / dirty metal. Solder needs to actually contact the metal surface, and the metal surface of course needs to be at a similar temperature to the melting point of your solder at the point of contact.
Careful mechanical abrasion is usually what you need! Like, a small square of medium / fine grit sandpaper held with tweezers. Or just the blade of an X-acto knife. Then scrub with a bit of IPA, dry it off, and get to work. Sometimes only the scrub is needed, with an organic solvent (IPA, acetone, MEK) if you just need to displace some tacky or oily dirt.
There is the option of using an active flux. These are usually acidic (often phosphoric acid!), and typically not noticeably aggressive at room temperature. But, with the heat of the iron the acid gets angry and can eat through most corrosion. Some active fluxes are *very* active and will easily eat through Fe2O3 (red rust), and your soldering iron tip.
Note, however, that you must clean active flux residue off VERY thoroughly. And I do mean VERY thoroughly. Any residue hiding, say, under an SMD component, will destroy whatever it's touching (starting with exposed contacts and copper traces) over the course of the next few months / years. It will catch you out. Not recommended for this kind of work.
Des isch dees!!
NEVER use acid flux on electronics. Never.
that... and it is very possible his flux is just hot-garbage. (as much trouble as he had with soldering the pin-headers on the board, the solder definitely isn't "flowing".)
6:50 oxidization on the chip legs maybe? it was sticking to the pads fine. mayeb the chips were old and had been exposed to the elements?
The important part of soldering is to heat up both pieces of metal to be joined at the same time.
Applying the tip of the iron just to the pad isn't heating up the pin, therefore you won't get a good joint.
I'm an electronic technician who specialises in rebuilding and restoring older machines, like Neumann matrices engravers for vinyl mastering, as an example. I understand the frustration that comes with that kind of work. I spent a couple of days searching for intermittent trouble in a machine I just finished to rebuild. Since the client and I had already invested 10k in time and parts, I wasn't allowed to give up, I found the trouble, but man, do it get frustrating sometimes, lol 😅
I worked on and maintained a VMS80 for a few years.
Can be very nervewracking to troubleshoot, you dont want to blow out a cutterhead because you goofed a repair on the amp ;)
Hey Noel! So, I'm about 7 minutes in. You first mistake with respect to soldering the chip is you DID NOT heat the pins themselves. Solder sticks to hot metal, it really doesn't like to stick to cool surfaces. A bigger tip DEFINITELY would help here, your general technique is great though. You probably could use hot air here. I'll update via comments my notes to you.
rip poor little micro. i've been there before :(
Yes, usb ports in windows ARE NOT all the same as far as devices go. It will allocate differently, in fact windows even puts the usb port it's on as part of the device path internally. this has tripped me up a few times. This is why people have issues with OBS, because windows does not treat a device the same across multiple usb ports.
oh BBB! Saw him at VCF East. We're usually on Dave's sunday livestream together.
With these small pins usually it’s fine to not touch them. The problem is the tip, I have a similar one and never use it as it does not work well and/or I have to crank the heat with it.
The problem is NOT the tip. Before soldering, ALL PINS must be cleaned (mechanically) and tinned separately.
PS, the microcontroller does not look genuine.
Oh boy. I went through 6 pairs of Darlingtons and 4 rare Schottky diodes and finally blew a loudspeaker from replacing power transistors with Chinesium ones.
Then I bought UK sourced ones from a reputable supplier, a replacement voice coil from the USA and finally got an old Fender guitar amplifier singing.
Like you, I wondered why would they bother to rebadge scrap at such low prices?
Then I remembered how Sir Clive Sinclair started in business...repackaging scrap transistors and reselling them as 'special' ...
One of the most braindead things in Windows USB drivers is that they are port specific. It is to me completely absurd that the device gets treated differently in a different port. I mean, it is good to have that as a possibility, but not as a default! This has been the way since at least Windows XP (probably even in older versions), and you can e.g. see Windows installing new drivers whenever you plug a device to a port where it wasn't before. This is particularly problematic when the default drivers don't work out of the box... I think Microsoft doesn't really know what the "U" in USB stands for.
Man, this led ya on a merry chase, no? I am no expert on SMD soldering, but I have done a bit of it in my time. If you don't have some, get some rework flux and low melt solder from a reliable vendor (I like Chemtronics, but that just me). It can make a surprising difference, especially when using hot air. Of course it didn't help that your microcontroller chips were heavily oxidized from the recovery process. That discoloration on the legs was from heat, and it really made things difficult for ya in terms of getting the solder to stick. You figured it out, though. Those chips are cheap enough that getting new stock from Mouser was a genius move, especially since the first ones you got were rebranded junk not actually suitable for your purpose.
Time, patience, and perseverance encompasses all things. Your eventual success here is a prime example of that. Thanks for sharing your trials along the way so we can all learn from them! That kind of gold one doesn't find every day.
Lots of good info in the comments on the specifics here, but as a guide for other who may be hobbyist solderers(?) from someone who does a lot of nano and micro rework by hand as a job: I use tips way bigger and iron temperatures way hotter than you would assume for any given task. Plus, not all tools and materials (station, tips, solder, flux etc) are born equal. I have helped beginners who's starter soldering kits make things additionally hard for them and me! One tip when using a monocular scope/screen is that pins can look OK from the top down and a decent nudge test is the only way to check, and absolutely soak the pins in the first place.
You should have a Linux machine (or dual boot) for this kind of thing, it's much easier than dealing with Windows drivers and idiosyncrasies.
true, but the device isn't reporting the correct ID, e.g. it is reporting as a simple RS232 device and not as a JTAG device, the act of associating the driver with the device on the port is why the strange behavior occurs. The reason for this is simple, it stops other devices from being used with the "new driver" when they aren't that type of device. Really what should happen is the device should appear as "unknown" and a driver would be required to be loaded to get the device to be seen at all. Then, when the device is moved to a different port, Windows would then use the correct driver for it. I'm forever getting hassled by people at work on Windows who have reached COM255 from plugging and unplugging RS232 dongles 🤣
16:50 Why don't you crimp some new Dupont male pins? I would still solder the 1.27" pitch header, and solder to the pins on the other side, bent them up and down, not sideways. Shorting power should not destroy your chip, perhaps your programmer could get damaged.
29:37 oxidation, one that is removed with flux and good amount of heat.
19:25 One trick I've heard of to make pin header "blocks" is the superglue the individual parts together.
It really looked like there was some kind of oxidation on the pins of that ATmicro. That would make soldering nearly impossible.
I'd try some ultra-fine sandpaper on the pin's contact surfaces.
27:37 that's why when i am soldering any SMD components like QFNs i gently push on them to make them flush with the board; unless legs were bent before usually it suffices for good contact. And yeah, apply that much heat during soldering to actually heat pad AND the leg to avoid dry joints.
I know everyone is a know it all with soldering, but maybe something that will help is using a KU knife tip. I very rarely use anything else and works great for through hole and SMD components. Those tiny cone tips are absolutely useless, and as long as there is flux aplenty, I find bigger tips are fine for even tiny components.
This counterfeit chip business is crazy. Wish someone would take an in-depth look into how much is actually made form these low-cost chip rebrandings. I faced that a long time ago when I bought power transistors for my Fender Passport sound system..
I'm not sure what's going on with your solder or solder technique, but for sure those solder beads aren't acting right on that microcontroller. The solder should be "wicking" under and even up onto the chip legs. Like if you were using a reflow station, you don't even *touch* the legs. You just tin the pads, apply hot air to the legs, and the solder wicks up and connects everything. I'd probably take that chip to some sandpaper or emery board to eliminate the possibility that it's got oxide on the underside of the legs. I'd also probably check my tin/lead ratio on the solder. If it's all tin, you just have to deal with the fact that it's not going to wick as well I guess. But definitely if I was experiencing what you were experiencing when you were trying to mount the microcontroller, I'd have stopped and solved that before moving on. Something's not right. Solder's too cold, bad lead/tin mixture, oxidized legs, not enough heat being applied to both the legs and the pads, **SOMETHING**. The way that solder was acting, you're almost guaranteed to have a cold joint.
Yeah, my first thought as well. The pins on the microcontroller looked oxidized and the solder didn't flow as it should under and around the legs. As you say, cleaning the legs before soldering probably would have made things way better.
Great to have you back. I felt your pain.
My tip to check SMD solder joints: hold the board at a shallow angle and look through a magnifier. If the joint is good you’ll see a shiny ramp of solder from the pad to the pin. If you see anything else (ie a step) the joint is bad. Since discovering that I’ve not had a duff joint.
We all stink at soldering until one day we just don't. It's super weird.
Same as welding, which is basically "big boi" or "macro" soldering.
There are so many cool projects I'd like to do, but I don't. When I was young, I had energy and was sharp. I also had confidence and better fine motor control. As I get older, I find I'm losing those qualities. I end up getting really frustrated and if I'm lucky, will get the job done with a lot of blood, sweat and tears. If I'm unlucky, I'll put it aside and never come back to it.
This is a great lesson in problem-solving. It also warns us to be careful about the chips we buy. Most counterfeits that I have seen have blurry lettering and incorrect logos. Try to find images of genuine chips online and compare.
Another great takeaway from this video is that there is now a new source of SID chips for the Commodore community. Hats off to whoever created it.
Congrats! Genuine distributors should the first place you look, buying chips from online marketplaces is a risk you take when you've run out of other options. Great video as usual, I love your perseverance and glad you won in the end.
I think it is important to see also videos like this. Often I do not remember to read properly the messages. I am quite varied about the fake chips that are around and it is difficult to find a reliable source that suits for hobbyist.
Some chips are harder to solder than others. What stands out to me is that you only heat the pad and not the pin. For the best results, you need to heat both the pad and the IC pin with the soldering iron. If you don't touch the IC pin, it will be at a lower temperature and won't pull solder onto it. Additionally, if it's an older IC that wasn't stored properly, there may be some oxidation on the pins, so it's essential to burn that off with flux. Therefore, heating the IC pin is crucial for successful soldering.
21:27 None of those legs looks like they are solder properly. I wonder if the legs are made of some material that won't like your solder tin. Could be a low budget chip with contaminated legs?
You really need to clean your solder tip.
We would say "wenns bärt, dann bärts" which means "When things go wrong, they really go wrong all day long"
Chinese sources will fake everything, even if it is dirt cheap. Somehow this seems to be a lucrative business still.
I feel slightly better about my own soldering skills now. Not good, but if you can struggle then I understand my own struggles a lot better. That microscope is looking like a game changer...
yeah, a microscope is a necessity for good quality work, especially when dealing with more pins and denser pin pitches.
@@davidhunt240 I'm mid 50s. The version of my eyesight is simply not compatible with good soldering without upgrades...
Have you thought about drug store diopter glasses?
@@raul0ca good idea... I'll see what I can find that allows me to focus closely
@@donaldklopper There's also dental magnifiers but they get pricey fast
hey noel , thanks for the channel its filled to bursting with tips and tricks so an amateur like me can tackle our problems
Fake components are a real problem. I bought some NE5532 op-amps; they are pretty cheap even from "proper" suppliers, and the entire batch were fake. They were op-amps but not NE5532. It took me ages to pin down what was going on. Don't buy components from any of the "cheap" sources.
Hi Noel !! This happens too much here in Brasil !! Super low quality of the leads of the chip !! Some Chinese chips must have the side down of "legs" *_gently sanded_* and ressoldered to get a good quality of assembly !! Or some drops of _*deoxit_*...
Cause of our tropical weather, all stock chips tends to little oxidization !!
All the soldering skill in the world won't make up for a fake chip. Shows just how prevalent the damn things are.
ALL of the retro youtubers I follow had to deal with fake chips
You made super hard work of that. The bad soldering early on should have been a sign but you just powered through it. Plus the device ID...
I'd use some solder paste instead of a wire for that chip. Has included flux and a lower melting temp (lower temp difference = easier for it to stick), much easier to work with small pads. You can also place it before you add the chip, so that it is much less likely to not make contact
I was thinking of solder paste!
Fantastic video! I love how you explore the difficulties and mistakes during the Nano SwinSID build, especially dealing with soldering issues and counterfeit chips, it must not have been easy. It's really true that even simple projects can have unexpected challenges. Looking forward to more content like this. :)
27:29 not to sound harsh, but you barely soldered them in place. Bottom pins look untouched. Untill the top of the pin is covered in lead, don't consider it a good contact.
yeah the soldering is so terrible i'm considering unsubscribing ;-)
I am sick of the rebadged chips from china.
With SMD try a K tip for your soldering iron. Don't tin any pads before placing the part. Solder paste can help. Just don't give up. I started with AVR stuff and a few years later I'm now working with 400 pin BGA parts.
Really interesting video. I also get worried about causing a short circuit when I need to take a measurement and the two pins are very close to each other. And let's not even talk about when the plastic melts while trying to solder a connector 😅
As someone who has been doing SMD soldering for over a decade, you pretty much nailed it that the tip was all wrong. Chisel tip, load and drag. You got the amount of flux right at least. Not sure what temp you were soldering at. Looked a bit cool, especially for that tip. Bit painful to watch stuff like this. Edit: Some medium active flux would have helped as well.
Not really. It was obvious the real fault was not applying enough heat and not applying that heat to the pad AND leg simultaneously. Pretty much everyone who commented got it, seemingly except you with over a decade of experience.... Hmmmm....
@@g4z-kb7ct Not sure what your issue is. Not enough heat due to the crappy pointy soldering tip. It's the same net result. The sharp little tips never transfer enough heat. Bob you should know better then to attack people.
@@jeffm2787 not attacking just correcting false and bad information so others don't make the same mistake....
@@jeffm2787 my name isn't bob lol
As for why they would fake cheap chips with more expensive ones might just be an observation bias. You compared the prices the chips have now but you should compare the prices they had when they were made (according to one commenter here the chips seem to be 10 years old). The 88 series looks to be more recent than the 8 models. It's possile that when they were introduced that they were rare and more expensive than the older and cheaper ones they were supposed to replace.
You've got quite some suggestions so maybe mine is redundant but:
1. yes, you need the biggest tip you can use for soldering. Maybe that chisel one is too big though! :)
2. Heat up the pad and the leg for 1 second, THEN apply solder.
3. Solder wire, flux, solder tip quality they all make a huge difference. I use Loctite solder wire, Kingbo flux (Amtech is just way too expensive) and I believe you have a Hakko? Those are genuine Hakko tips? Fake tips will deliver... fake results :)
4. Apply more solder on the first pad. Only one pad or the whole chip will be elevated. Once the first leg has fallen into the solder - and the whole IC has touched the PCB - then you can solder the second leg, then all the others.
Good luck! :)
Yeah soldering the chip on top of that mountain of solder was a huge mistake ;-)
I have a few nano's rattling around in a drawer somewhere. Now the SwinSID Ultimate is a different story. I love that one. But getting ahold of one these days is a bit tough. My favorite has to be the FPGASID. The psudo-stereo option is fantastic. Voices 1&2 on left speaker, Voices 2&3 on the right speaker makes old games seem line new.
you should be using paste and a hot plate to solder the microcontroller.
as for the tiny programming connector, the best option is to make a pcb programming adapter with the proper connectors
It's actually faster to re-design the whole PCB and put a proper AVR header on it... that's what I did. I've been re-designing lots of projects to optimise design, traces and part cost for over 20 years.
@@g4z-kb7ct ah, maybe... good idea
Jeez, this video is like watching a train wreck in slow motion - seems like you know/understand everything, yet managed to stumble each step on the way, TQFP is the easiest smd packages to solder, not buying chips from reputable distributors, not using standard ISP connector, not checking the ID, like HOW? I've seen bad days, but this is quite something else.
the answer is: doing this too infrequently. you have to repeat activity not only to learn but also not to forget. very human thing i guess ;)
@@lis6502 must be *very* infrequently - I can solder that TQFP blindfolded, upside-down and drunk at the same time, I've also taught bunch of ppl to solder - once past the 'Aha!' moment, they never forget.
@@VEC7ORlt it's about one's personal abilities. I am able in the same intoxication/ exhaustion condition to troubleshoot not working JBoss or solve network access issue. My friend is able to drive a car or bulldozer( haven't tested but i am sure he'll manage) completely drunk. My point is: i don't know what Noel's innate talent but surely it isn't soldering ;)
@@lis6502 in the end he eventually tried everything - lifting the pins, testing for continuity, maybe it still was a big fat chunk of bad luck on top of another.
thats why IC-s are kept in moisture controlled enviorment. That chip was probably siting in amazon rainforrest for a while. Even the flux could not get rid of oxide.
Soldering with a axe, wont help also.
Awesome video. I'm a bit of a novice, but I've been trying to build custom devices with the Raspberry Pi. You channel has helped so much in learning systematic troubleshooting. This one also helped in the don't give up part of building a project. Keep up with the great content.
As I am very comfortable with ATmegas and their programming, your odyssey almost made me cry out loudly. I am not through the video yet and wonder if/when you will discover that you are trying to talk to fast to the ATmega88. And I have seen much better solder skills from you in the past, haven't I? When handsoldering TSOP packages you really need to apply heat to the individual pins - drag-soldering or pin-by-pin.
The message about the signature you got from avrdude indicates that your ATmega88 actually is an older ATmega8 - could it be a relabeled fake? Also I can really recommend the GUI avrdudess for avrdude.
Have discovered something...
The specs of the atmega 88 is 20 mhz... Swinsid is operating above... I wonder if this makes problems to some atmega . Try programming before installing the crystal. ;)
This is why the eprom is not used... Just the flash...
You can cover oxidized contacts 100% with solder and still have no connection. Such cases require a lot of attention.
It doesn't require anything more than a simple sanding with some 400 grit sandpaper. I do a lot of arcade repairs (for roughly 30 years) and I have to do that a lot with some boards that are 40 years old.
Thank you for sharing this video, Noel. I have been in a similar situation not long ago, but ended up giving up and concluding a lack of skill. Your video showed me it is not just me. I should give this another go…
I smoked an ESP32 chip the same exact way by stupidly shorting one of the i2c lines to 12v power from a shunt. The slightest little slip a buzz and a spark and that’s it, all that work gone out the window.
I think, maybe the μCs are a bit old and the pins were oxidized. That could explain the soldering problems. What makes SMD soldering fun is a hot plate. I have been hand soldering SMD for probably 30 years, but the hot plate was really a step forward. Much better than hot air.
And for cable making, crimping is much nicer than soldering. The IWISS SN-025 tool is great for Dupont connectors (although all crimp tools seem to be recommended for Dupont, the previous tool is the only I know). The SN-28B is good for many other crimp terminals.
Lol, Noel, I have so much resoect for you, and I admire you trying to going into a field that is so unfamiliar to you, but some of the mistakes you make, makes me wish I could have been your hardware-mentor, because those mistakes are so elementary,,,, from soldering AND LEARN TO USE AN OSCILLOSCOPE TO FIND FAULTS…. And then I look at my pathetic attemps to write software, and still I wonder why the software engineers hated my code, lol…. Pushing things beyond our safe spaces is hard, and I admire you for it, especially putting it all on RUclips.👍👍👍👍
I had a similar issue trying to build a Greaseweasel and trying to find legitimate STM32F103C8T6. They weren't rebadged, but the sellers would list the chip but send a different model of STM32.
[23:55] Wrong device signature confirms that this MCU is fake...
PS, AVRDUDE tells you that this MCU is M8 (ATmega8/8A), not M88!
I had to program an ATMega32 yesterday. I used the AVR mkII. Had to take the chip out and put it in the ZIF socket to get it to program. The ICSP header didn't work. I believe I needed to provide external power to the chip. Had to plug in 5V to the ZIF socket.
I should start to make RUclips videos too. ALL of my electronics projects go wrong. Buying chips from Aliexpress is one of the biggest contributors to that. I totally felt your pain when I saw the soldering challenges. I spent all day Saturday doubting my soldering skills due to dodgy CH340 chips from Aliexpress. When I was a kid I used to kill many ICs trying to desolder them so I've always been wary about how long to keep the iron on pins. That's the whole reason I've been too chicken to solder a Spartan 3 FPGA onto a PCB. Drag soldering scares me - always worried a bridge will form underneath the IC somewhere where it can't be seen.
The other mistake I make is trying to design a PCB, getting it manufactured, waiting 2-3 weeks for it to arrive and then discovering there's a mistake or I haven't added something. Just did that - didn't add the ISP pins to program my AtMega2560.. I'm glad to see that other people experience this pain. I keep thinking it's just me and that maybe electronics and I were not to be :-)
You got yourself some expired ICs. These get rejected by electronics manufacturers because they have built up oxide on the legs/pads and get sent off to Huaqiangbei markets. You need to recondition them first by retinning the working surface of the legs by any method that works for you.
Device signature doesn't match... are those fake chips as well? Huh.
Ah yes, yes they are. Excellent.
I don't know if this has any bearing on the legitimacy of the chips, but looking at the video (around 31 minutes in) the Atmel MCU you replace the 88PA is clearly and sharply labelled ATMEGA8A, if this is one from the batch sold as 88PA that would definitely point to something wrong.
Just because something is open-source does not mean it's a good design or the released info is good. In fact quite the opposite, most projects are poorly designed and pooly implimented. I've been grabbing various open-source projects for roughtly 20 years and making them much better with proper PCB routing and better part selection. For example one SD2IEC version used a sdcard slot that cost $10 lol! Another project used an amp chip that cost $15! Of course I've also been reversing some not-open-source projects too hehe! As for SID replacement's I recently reversed something that is said to be open-source but actually isn't. If you want the schematic and gerbers you can find my contact info on my about page. You know me under a different name and we have spoken in the past. This is something you definitely want ;-)
Its very good you released this video. Its very hard to learn without seeing mistakes or even knowing what one does wrong.
- Flux is either bad or not enough heat to activate and clear oxide from those crusty old Atmegas.
- no amount of bending legs will help with oxidized pins.
- Soldering Iron you are using is ancient Tip on heater ~70 year old Hakko (clone?) design, good for thru hole and fixing headphone cables. Not the best for SMD. Integrated cartridge (Hakko T12/JBC C245) clones are really cheap and miles better.
- Tips. @14:10 hardly anybody with SMD experience will ever tell you "this tip is too large" :)
First tip is not good for anything, no idea why they even manufacture those needle cone tips, perhaps as a joke to newbies? or for carving wood?
Second tip much better in thermal mass department, but bad at transferring solder where you need it - you kept adding solder to the top of the tip with nothing going to actual pads/pins.
Optimally you want Beveled tip with small solder reservoir.
- @27:20 "no way visually" while its Clear as Day those pins are not soldered :o
- @27:42 "are they wiggling at all .. those look fine" while you can clearly see in the video at least the first one is in fact wiggling :o
- @gnif already commented below about touching both pad and the pin when soldering, otherwise you are counting on thermal transfer thru parts and flux which might not be sufficient.
Hotair would work much better than soldering iron in this case heating everything uniformly, and I think @30:48 you finally had all pins properly connected with hotair :) Irrelevant in light of remarked atmega8 tho :-)
Most of this seems like assembler error....
Soldering technique and not being familiar with tooling
Thanks for sharing your learning process and experience for everyone to learn from!
For SMD I *never* pre-tin the pads. Even if you remove it, it leaves a slight "bump" on the pads. Which makes alignment harder. The pads need to be the same height and FLAT.
As for counterfeit chips, I always buy modern chips from Mouser or Digikey (or maybe Jameco). I only buy vintage chips off eBay and I try to make sure they are legit. Just not worth the effort and time on questionable chips when a modern one from Mouser isn't much more expensive.
[15:50] DO NOT use these cables for soldering! They are usually made from ALUMINIUM (often even painted), not copper! Thus, they cannot be soldered at all.
Now I'm through the video, and you've figured out the fake chip name and its real ID. is there way to program the 8A chips?
Like your video.
I always, always check my connections with a multimeter to check for proper contact AND shorts of supply rails and all other traces. When I do apply power , I do so with a bench power supply and limit the current to 50 mA. If anything has been overlooked the current will be very small and most components will survive that much current.
Plugging the board without testing to an USB to FTDI adapter that can supply upwards of 1 amp is not smart. Stuff will release magic smoke if something is wrong.
Same goes with repairs, always use current limiting and know what voltage is needed for supply and digital/analog interfaces.
just buy good soldering iron. quick 202d + 303k soldering tip on it.
u can solder several legs of ic (even 5-6 in row). with good flux u can easily rid of excessive solder. for example
kss-800 flux can handle even bga cpu soldering process
or u can use "martin" flux but its expencive, and just less foggy than kss-800.
dont use thin tips, especially with soldering irons , that use ceramic heater in it.
so even oxidated ic legs are not problem.
but ive faced problems with china headers 2.54 (as in video) they werent oxidated, but i couldnt use them with regular flux. i dont use active fluxes because they are damaging soldering tips(nickel covering).
anyway i am happy to see ur video again. i wish it come more often)
I am doing a Vicky-Twenty build (going on 2 years now) and I am having all kinds of issues. All logic chips test good, all the MOS chips work (I have a CR VIC-20 I test them in and all work when swapped out), so I know your feelings on this
neither your soldering tip nor the PCB nor too much flux was the main issue for your soldering problems in this project. the first atmega looked REALLY crusty, oxidized to oblivion. at that point, even really acidic flux won't help that anymore. apart from that, don't ever try to solder those cheap jumper wire dupont cables, they tend to be really corroded as well making for horrible joints. a cable recommendation: 28awg silicone fine strand tinned copper wire.
about the one non-connected pin 17 on the second set of boards: you can actually make those out visually, if there's no solder just next to the pin on the half of the pad, they're generally bad. also, if the solder "balls up" on the half of the pad that isn't under the pin, and goes down towards the pad, it generally is bad as well.
last thing I noticed, using a 32 Mhz oscillator can't be correct, the atmega88 is only made for up to 16 Mhz crystals, not 32 M, I'm really surprised that that worked...
I ran into a similar problem recently. I was making a replacement for the keyboard controller in a Mac Plus keyboard. I initially prototyped with 40 pin Mega32 (needed a lot of I/O to match the matrix along with everything else) and same problem - the IDs didn't match. I got that batch from Amazon. They kind of worked anyway, but I got more from Digikey and they were perfect. Finished up with a surface mount version in a 40 pin DIP form factor.
I saw you used hot air to remove the chips, but have you tried it for soldering them on? The Mega32s I had to use were much finer pitch that those 88s. I just laid down a thin line of solder paste across the pads, stuck the chip on, and hit it with the hot air. Once the solder melts it's attracted to the bare pads and legs and it just kind of "snaps" into place. Way easier than trying to do it with an iron.
You need a hot air station and a better soldering iron like something from JBC. Ditch the hakko, or like me, keep the old hakko in a box under a table in your garage where it can collect dust. $450USD is a lot for an iron but with quality tools, soldering goes from being a PITA to being easy. Surface mount stuff is generally best done with heat plates and hot air. You can get a quick 861DW (cheap) hot air station and a nice machined hot plate from amazon also for cheap. That would turn a solder job like this into a 10 second easy task. With proper tools, I've found I prefer surface mount soldering over through hole.
Thank you for making this video. I learned a lot through it. It was great to see your process and understand how you went through figuring out each of the various different issues and overcoming the problems.
Never say never but I would be very much amazed if you blew that first Atmega up by accidentally shorting any pin out on that chip. They are very hardy chips. Ive shorted Atmega 328 and 2560 out many times in the past either pin to pin or pin to a power rail and never once destroyed one or lost an I/O pin from doing so. A quick short to ground on a high level pin would be extremely unlikely to blow the chip. Of course if the short persisted for a length of time maybe then but a quick slip I would really doubt it. Perhaps might have died from excess soldering heat more than anything?
If anything I could imagine the magic smoke coming from that first USB programmer perhaps depending on how the 5V is running through the board and if its switched etc via a transistor? If the supply power was to short that could very well destroy something on the programmer then?
Awesome work, Noel. Always great to watch your video's.
I used to watch Louis Rossmann to learn about microsoldering (and some salt - now it's all salt - so look at early videos) as I wanted to repair lots of broken things I had around me, PS4 and XBox 360 HDMI ports, USB-C port on a phone etc. So I bought the equipment that he used (only in function, not exactly the same) and started learning to solder with hot air needing very little iron work. I got very good at it. When you took the dead IC off you used hot air, do the same for putting it on, it saves a lot of time. Just put some solder onto the board, either paste or wire with the iron, then plenty of flux, heat the board, then place the IC onto the board and keep swirling the hot air around the legs, eventually the IC will magically float into place, gently tap it with tweezers and remove the heat. Job done. Regarding the fake parts, the wrong signature is real red flag - it means you don't have what is marked on the IC so that immediately shouts fake. It is also likely that the process the Chinese component sand-and-etch brigade use affects the plating on the component legs leaving a scale that needs a lot of heat and flux to remove. The connector looks like 1.27mm (0.050" - 50 thou/50 mil) I guess if you were programming in any volume, you'd make a jig to push spring loaded pins onto those pads from below.
These finer Pitch programming throughholes would have really bugged me. I guess i would have bought some 6pole jst connectors which would fit on the board perfectly. Programming an IC without solid connections, that is a real pain....
the pads on the chip may have been oxidized just a layer... that was the problem on the soldering, should be easier... sometimes something abrasive to remove that layer helps.
flux and put some solder in there...
What a nightmare! When you compared the two microcontrollers at 39:00 you can clearly see the difference in the pins. New vs old and crappy looking.
4:25 You fail at soldering. Even though I was only a technician, whenever the lecturer wanted to teach students how to solder, they'd call me down to demonstrate and talk them through how to do it correctly.
Rule 1, of soldering: Apply heat to the pin AND the pad. You're just applying heat to the pad.
Rule 2, don't apply solder to the part you're heating, apply it where they meet, or the other side if you're breaking rule 1.
You need to actually press your soldering iron tip down against the top of the chip's leg. Applying solder to the pad while you're heating the pad is definitely the wrong way to solder.
39:29 No, I could tell when you placed them side by side, the new ones look a lot "Cleaner", less oxidised, so it should be easier to solder.
The old (Fake?) ones looked like the pins needed cleaning.
6:57 Yeah lol. Years ago I re-designed my own swinsid board and made the design much better including getting rid of those silly small-piched holes and 43:28, yes there is room. I added the proper 2x3 AVR programming header :-)
If you want my gerbers or the full source Eagle PCB files let me know.
If a driver is working in one USB port and not in another USB port then the driver is a little faulty, stop searching, end clumsing around. Get better drivers. The decades of Windows programming are over. It is 2024 and close the era of Windows, just like most users did. Often there are many USB thingies to do something serial with something else. Pick another and stop soul searching for mistakes YOU did not make.
Great video! Loved watching your troubleshooting process, and learned how easy ICSP is to use!
Yeah, small needle tips are pretty much useless untill you overheat them. Very small thermal mass, the needle shape doesn't help at all to push some heat into the board, and for some reason they ship with every soldering station. Meanwhile something in the shape of T12-KU or BCF3 does all these small jobs in one pass and doesn't lose temperature or break that often.
helps to tin the legs on the chip too... i think that was alot of the problem.. the chip legs looked like they had a layer of oxide causing the poor solder joints....at least run the iron over the bottom side with some flux to clean off any oxidisation on the legs before soldering to the board
also you could have used any other arduino board with an ISCP port to program that micro... you dont need special programmers (the usb one you eventually purchaced is the same chip as on an arduino nano...look up Arduino as ISCP programmer)
one last tip.... it always helps to investigate ALL the errors listed when programming fails... starting at the top most error!... the device id not matching is a huge clue the chip is not genuine!
Thanks Noel... missed regular vids from you... but it get it..life happens!
Mouser is the way to go. Always! 🙂
Thanks for sharing your experience. Your channel is one of the best and it's criminally underrated.
So you are tinning the pads but then not tinning the bottoms of the legs of your chip? Watch some of these Indian guys with these giant irons they use easily sticking tiny things together in these very scary skeleton circuits and they are lifting the entire circuit up by literally any component. Tin both items you are putting together.
Soldering header pins. Oh man. Plastic melts...simply pause for a few seconds between pins. But even better don't solder header pins. You can get kits with JST connectors. The male receptacle is simply a plastic piece that will slip right over your pins. All the wires in the kit already have ends on them, you just push them in. No crimping required. Just get the right dot pitch. I think xh and ph are the most common sizes I use. PH should be Arduino size. No crimping at all. No one wants to do that crap . Those little Dupont wires are fine for prototyping but not reliable on stuff you are plugging and unplugging, moving around. Under $20 on Amazon for the pre crimped JST kit. All the wires are separate, ends on both ends already, and all the popular colors.
Now those big leads on your meter are clumsy. I went ahead and made myself something more precise and easier to control. If you have any plastic laying around, the ones that come bundled with something and you probably have stuffed in a junk drawer. Get em out. Grab a couple of sewing needles(they are sharp and also very strong). You're good at melting plastic, you know what to do next, so do that. Get some super floppy silicone wires and alligator clips or sacrifice an existing test lead with clips I'm sure you have and make yourself a pointy little set of probes. Probably blunt the tip if the needle. I have second set of these with pins on the end if the leads so I can stuff them in a TC1 component tester. Much much easier to test SMD resistors and caps and the thing most meters don't do, inductors. Tweezers leave a hand free so you can press the stupid button that is way harder to push in than it should be.
Unfrikkin' believable.. why would they go to great lengths of rebadging ICs. Definitely good to see you got it working in the end. I have had similar issues with USB ports but in Linux. The port can stop working after resuming from hibernation and the issue is fixed by rebooting.
Commodore made C64's until 1992 I believe. So it would be early 90's maybe 93 or 94 (I would imagine they made a few spares) before the sid shut down.
5:40 i never tought that watching someone's soldering job might be my guilty pleasure, but yet here we are ;d.
As mentioned by others: skill issues ;) That's why i never will be grateful enough for local flea markets where you can buy virtually any electronic junk for a dime and train your skills, testing if 400 degs C really helps to make job easy ;)