Almost 30 years old. Old good times without BGAs and modern barely visible components. While some basics are still applicable the modern problems are not covered at all.
Provided you have good eyesight and steady hands, I've mostly found what happens as you get smaller is:
- Heating becomes easier. There's no large sinks to take the heat away. It's also easier to overheat things.
- You need finer tweezers, and don't drop them because if you do the tips will bend.
- The solder's surface tension does more of the work. It feels a lot more like sticking together things with tiny droplets of glue. Having the correct amount of solder in the right place is critical.
- Solder and flux become two separate things you have to care about individually
- It is easier to burn yourself
- learning how to brace your hand against something in a way that gives you very fine control. One reason soldering with an iron can be difficult is because your hand is so far away from the tip, like trying to write with a pen held by the end.
When I started my first job a coworker encouraged me to learn how to solder SMDs and do "microsoldering". Like most people I thought I was going to need high precision and a much steadier hand. Probably like most people that learned I was impressed at how quick I picked it up. I think the hardest thing was learning about part "tombstoning" but that's not that difficult to deal with. I'm not going to say it is easier than soldering through-hole components, but I think for most people the mental barrier is much higher than the actual barrier.
I now highly recommend learning it to anyone doing electronics. It's well worth the (small) time investment and makes things a lot easier, opening lots of doors. Even for a hobbyist you immediately get benefits. Everything becomes more compact, 2 sided boards are much more usable, and, of course, it opens up a lot of repairability (and recycling. Are you really a hobbyist if you aren't desoldering and reclaiming parts?).
> Are you really a hobbyist if you aren't desoldering and reclaiming parts?
Fun memory from who-knows-how-many years ago:
While installing a Playstation mod chip, I accidentally dislodged a nearby surface mount resistor, pulling off one of its metal contacts in the process. (Is that what happens when you overheat them?) I didn't think that was fixable, and since it was Sunday, the local electronics shop was closed. I ended up disassembling an old junk digital camera that hadn't yet been taken to the e-waste recycling drop, and finding inside it a resistor that seemed close enough to maybe work. The transplant was a success, and the Playstation ran great thereafter. Very satisfying.
I recently got rid of a lot of components that I have salvaged and hoarded over the years. If I need a doodad for something I'll just buy it. I'm done storing all this junk I will never use
Agree that SMD hand assembly is easier than it looks, at least down to 0603 imperial. If I can wait the week for boards to arrive, I’ll often skip the breadboard step and go straight to a proto PCB, especially since most parts aren’t available in throughhole without waiting on dev boards anyway.
When you hand someone a board with 0603s on it that you hand-assembled, it seems like magic to people who stop to think about it.
It's worth noting that the longest one there is already much shorter than the classic mid-century unregulated irons, and all of those can be held like a pencil.
> The solder's surface tension does more of the work. It feels a lot more like sticking together things with tiny droplets of glue. Having the correct amount of solder in the right place is critical.
I believe this is why I have an easier time hand-soldering BGA than QF[np]: I can't screw up solder amount/evenness.
What tools do you use for BGA soldering? I’ve seen people (well, dosdude1) using board preheaters and hot air stations, but I could never justify the expense for the amount of board rework I actually do.
I have Chinese hot air station and I am super happy with it. Airflow is nice, temperature constant. Simple manual control. The device has writing 998D on it. This device also has soldering iron attached, but it is very bad, does not have enough power for usual soldering tasks.
Important caveat: The downside to this is you can't inspect it (without an x-ray machine), and if you screw up, you're going to need a new chip (Re-balling does not look approachable/time-efficient)
Reballing isn't too difficult. It's mostly about buying all the special tools and supplies. You need a specific template for each BGA pad layout and jars of the right size solder balls.
Some people also just put solder paste on the PCB with a stencil and reflow that into solder bumps that you then solder the chip to. I find that method suspect but if it works it works
Thank makes sense, thank you! Is there a rule of thumb for how high you set the hot plate? I’d be worried about SMD components on the side in contact with the hotplate.
You're right that that is a concern. I think I set mine for 190C? Don't remember. But as you infer, my goal is to get it slightly below melting temperature so the the other components' solder doesn't melt. (Disaster if they do, then the board gets jostled or you knock a component with tweezers). Then when you hit it with the air, just that component melts.
Also, it seems to not melt if you don't flux the pads prior. Not really sure why.
How are you burning yourself? I've only ever worked with one person who burned himself soldering when working on a SMT PCBs, and it was while desoldering a through-hole connector, when a desoldering station was long past its cleaning interval and it dripped some solder onto a metal ring he was wearing. This was a guy who would lick a soldering iron to see if it was hot and touch the molten solder in the wave solder machine. The Leidenfrost effect goes a long way.
My #1 way is from impatiently touching the board to see if it's cool enough to touch yet. That sounds dumb and it is.
More generally, with iron soldering only the iron and the last couple joints are hot. For SMD, there's more places for the heat to go; sometimes the entire board can be hot. Sometimes, you might need to balance being close enough in to get a good grip on the tiny parts, but far enough to not get burnt. You will feel the heat when SMD soldering - it's not always dangerous but another thing to pay attention to.
I have been impatient, and pulled a board out of the reflow oven while it's hot enough to burn, but it's fiberglass and resin, so it has very low heat transfer, giving enough time to put it down, without getting burnt.
I do a lot of soldering at my day job to bodge boards, tune networks, etc. I burn myself on the time because when I'm working through the microscope I seemingly forget I have hands or lose track of them and bump the iron into them when pulling it away from the work. Not sure why, but it's really easy for me to get into this mode where the view through the scope is the only thing in the world
I don't think that modern boards are really repairable at all beyond component replacement- 4+ layer stackups being the big reason. If there's a way to do anything to those boards besides total replacement I'd be super interested to know.
The techniques here are also way beyond basics I think- like, you look at most guides for repair and it's "idk just solder some bodge wires on there, here's what a good joint should look like"
Andrew Zonenberg posted a Twitter thread a year or two ago where he fixed a missing PCB trace some layers down a PCB, with a stereo microscope, precision mill and very steady hands.
Edit: here's the thread. It's a 6 layer PCB with a short on L5 that needs to be fixed from the L1 side.
If you enjoy that sort of thing, check out this guy's videos. Lots of trace repairs (including below the surface), pad replacements, etc. Quite impressive to see it done.
For boards with a bunch of layers and BGA/LGA packages, that have internal manufacturing errors or damage (e.g from overflexing), repairs can be untenable.
If the parts all have pads on their perimeter, then a jumper wire can replace internal traces. If the pads are underneath the part, and the trace is only internal, than a jumper may not be feasible, unless the damage happens from the surface in, in which case each layer can be jumpered at the damage.
nope, people still repair >8 layers boards like it's nothing. Some even do "chip repairing", literally remove the expoxy of the IC to fix bonding wire, or remove the security key/module on the chip.
There are a surprising number of Indian cell phone repair shops with YouTube channels that do feats of soldering, like repairing torn flat-flex cables… I bow down at their craft. If I ever get a sabbatical I’ll go to India and ask to be an apprentice.
If it's QFN, you're probably best deadbugging it and jumpering all of the pins individually, but first make sure that you need the connection. QFP wireframes all have a center pad, whether or not it's electrically connected to the die or needed for thermal dissipation.
If it's an SO package, e.g. SSOP, TSOP, etc., Desolder the IC, add the jumper wire, bend the pins down enough to account for the thickness of the jumper wire, and resolder the IC.
Either way, make sure you have enough thermal mass connected to it for thermal dissipation. If there's components on the opposite side of the PCB, it's probably not much.
> Destructive static charges are induced on nearby conductors, such as
human skin, and delivered in the form of sparks passing between
conductors, such as when the surface of printed board assembly is
touched by a person having a static charge potential. [..] It is important to note that usually the static damage level for components cannot be felt by humans. (Less than 3,000 volts.)
Less than 3000 volts cannot be felt by humans? Should be 3000 millivolts right? i.e. 3 volts...
It's 3000V, but the energy delivered can be absolutely minute on a human scale, even for the highly sensitive nervous system. But metal oxide layers (what static discharge is often blowing holes in) are not on a human scale, they're atoms thick. And highly insulating, which leads to teravolts/metre field gradients.
For discharge that you can feel and see, the energy is even higher, but damage can be done far below this level.
Considering static discharge is a spark, I assume 3.000 volts is the correct scale. When I wince from a static discharge generally there's visible spark.
SATA ports are often rated for only tens to hundreds of cycles, so they're often made out of fallapartium.
If you're replacing one, it's easiest to cut all of the plastic off first, then desolder the contacts one at a time. Sometimes you can even pull the plastic off, without damaging it. It's usually necessary to preheat the board.
Things like that are often the only through-hole components on a PCB.
I like to melt the original solder one contact or area at a time, add a little bit of fresh flux-containing solder if needed, and as soon as it gets to the consistency of mercury, vacuum it clean out with a good soldersucker.
Almost 30 years old. Old good times without BGAs and modern barely visible components. While some basics are still applicable the modern problems are not covered at all.
Provided you have good eyesight and steady hands, I've mostly found what happens as you get smaller is:
- Heating becomes easier. There's no large sinks to take the heat away. It's also easier to overheat things.
- You need finer tweezers, and don't drop them because if you do the tips will bend.
- The solder's surface tension does more of the work. It feels a lot more like sticking together things with tiny droplets of glue. Having the correct amount of solder in the right place is critical.
- Solder and flux become two separate things you have to care about individually
- It is easier to burn yourself
- learning how to brace your hand against something in a way that gives you very fine control. One reason soldering with an iron can be difficult is because your hand is so far away from the tip, like trying to write with a pen held by the end.
When I started my first job a coworker encouraged me to learn how to solder SMDs and do "microsoldering". Like most people I thought I was going to need high precision and a much steadier hand. Probably like most people that learned I was impressed at how quick I picked it up. I think the hardest thing was learning about part "tombstoning" but that's not that difficult to deal with. I'm not going to say it is easier than soldering through-hole components, but I think for most people the mental barrier is much higher than the actual barrier.
I now highly recommend learning it to anyone doing electronics. It's well worth the (small) time investment and makes things a lot easier, opening lots of doors. Even for a hobbyist you immediately get benefits. Everything becomes more compact, 2 sided boards are much more usable, and, of course, it opens up a lot of repairability (and recycling. Are you really a hobbyist if you aren't desoldering and reclaiming parts?).
> Are you really a hobbyist if you aren't desoldering and reclaiming parts?
Fun memory from who-knows-how-many years ago:
While installing a Playstation mod chip, I accidentally dislodged a nearby surface mount resistor, pulling off one of its metal contacts in the process. (Is that what happens when you overheat them?) I didn't think that was fixable, and since it was Sunday, the local electronics shop was closed. I ended up disassembling an old junk digital camera that hadn't yet been taken to the e-waste recycling drop, and finding inside it a resistor that seemed close enough to maybe work. The transplant was a success, and the Playstation ran great thereafter. Very satisfying.
I recently got rid of a lot of components that I have salvaged and hoarded over the years. If I need a doodad for something I'll just buy it. I'm done storing all this junk I will never use
Agree that SMD hand assembly is easier than it looks, at least down to 0603 imperial. If I can wait the week for boards to arrive, I’ll often skip the breadboard step and go straight to a proto PCB, especially since most parts aren’t available in throughhole without waiting on dev boards anyway.
When you hand someone a board with 0603s on it that you hand-assembled, it seems like magic to people who stop to think about it.
When you're regularly dealing with 01005's, getting to work with an 0603 feels like a luxury.
I tune my resistors by whittling away at the PCB traces you insensitive clod.
One reason soldering with an iron can be difficult is because your hand is so far away from the tip, like trying to write with a pen held by the end.
Newer irons, especially for SMD work, have gotten smaller and the grip-to-tip distance also shrunk; here's a good visual comparison:
https://www.eevblog.com/forum/reviews/grip-to-tip-distance-o...
It's worth noting that the longest one there is already much shorter than the classic mid-century unregulated irons, and all of those can be held like a pencil.
> The solder's surface tension does more of the work. It feels a lot more like sticking together things with tiny droplets of glue. Having the correct amount of solder in the right place is critical.
I believe this is why I have an easier time hand-soldering BGA than QF[np]: I can't screw up solder amount/evenness.
What tools do you use for BGA soldering? I’ve seen people (well, dosdude1) using board preheaters and hot air stations, but I could never justify the expense for the amount of board rework I actually do.
I have Chinese hot air station and I am super happy with it. Airflow is nice, temperature constant. Simple manual control. The device has writing 998D on it. This device also has soldering iron attached, but it is very bad, does not have enough power for usual soldering tasks.
Hot plate, flux pen, hot air gun.
Important caveat: The downside to this is you can't inspect it (without an x-ray machine), and if you screw up, you're going to need a new chip (Re-balling does not look approachable/time-efficient)
Reballing isn't too difficult. It's mostly about buying all the special tools and supplies. You need a specific template for each BGA pad layout and jars of the right size solder balls.
Some people also just put solder paste on the PCB with a stencil and reflow that into solder bumps that you then solder the chip to. I find that method suspect but if it works it works
Thank makes sense, thank you! Is there a rule of thumb for how high you set the hot plate? I’d be worried about SMD components on the side in contact with the hotplate.
You're right that that is a concern. I think I set mine for 190C? Don't remember. But as you infer, my goal is to get it slightly below melting temperature so the the other components' solder doesn't melt. (Disaster if they do, then the board gets jostled or you knock a component with tweezers). Then when you hit it with the air, just that component melts.
Also, it seems to not melt if you don't flux the pads prior. Not really sure why.
[dead]
Provided you have good eyesight and steady hands
I see the flaw in your clever plan...
How are you burning yourself? I've only ever worked with one person who burned himself soldering when working on a SMT PCBs, and it was while desoldering a through-hole connector, when a desoldering station was long past its cleaning interval and it dripped some solder onto a metal ring he was wearing. This was a guy who would lick a soldering iron to see if it was hot and touch the molten solder in the wave solder machine. The Leidenfrost effect goes a long way.
My #1 way is from impatiently touching the board to see if it's cool enough to touch yet. That sounds dumb and it is.
More generally, with iron soldering only the iron and the last couple joints are hot. For SMD, there's more places for the heat to go; sometimes the entire board can be hot. Sometimes, you might need to balance being close enough in to get a good grip on the tiny parts, but far enough to not get burnt. You will feel the heat when SMD soldering - it's not always dangerous but another thing to pay attention to.
I have been impatient, and pulled a board out of the reflow oven while it's hot enough to burn, but it's fiberglass and resin, so it has very low heat transfer, giving enough time to put it down, without getting burnt.
I do a lot of soldering at my day job to bodge boards, tune networks, etc. I burn myself on the time because when I'm working through the microscope I seemingly forget I have hands or lose track of them and bump the iron into them when pulling it away from the work. Not sure why, but it's really easy for me to get into this mode where the view through the scope is the only thing in the world
An easy one is heat transferring through an SMD component to your tweezers, while trying to gently remove it from a heavy ground plane.
I don't think that modern boards are really repairable at all beyond component replacement- 4+ layer stackups being the big reason. If there's a way to do anything to those boards besides total replacement I'd be super interested to know.
The techniques here are also way beyond basics I think- like, you look at most guides for repair and it's "idk just solder some bodge wires on there, here's what a good joint should look like"
Andrew Zonenberg posted a Twitter thread a year or two ago where he fixed a missing PCB trace some layers down a PCB, with a stereo microscope, precision mill and very steady hands.
Edit: here's the thread. It's a 6 layer PCB with a short on L5 that needs to be fixed from the L1 side.
https://xcancel.com/azonenberg/status/1468825231225540611#m
Holy cow! I've been pushing around a TQFP48 tonight and thought I was pretty good.
If you enjoy that sort of thing, check out this guy's videos. Lots of trace repairs (including below the surface), pad replacements, etc. Quite impressive to see it done.
https://www.youtube.com/@northwestrepair
10/10 read
Mostly when things fail it's not a trace, it's components
For boards with a bunch of layers and BGA/LGA packages, that have internal manufacturing errors or damage (e.g from overflexing), repairs can be untenable.
If the parts all have pads on their perimeter, then a jumper wire can replace internal traces. If the pads are underneath the part, and the trace is only internal, than a jumper may not be feasible, unless the damage happens from the surface in, in which case each layer can be jumpered at the damage.
>If the pads are underneath the part, and the trace is only internal, than a jumper may not be feasible.
BGa pad repairing is very common, here is one example: https://www.facebook.com/watch/?v=1077341457703029 (sorry can't find non-facebook link)
nope, people still repair >8 layers boards like it's nothing. Some even do "chip repairing", literally remove the expoxy of the IC to fix bonding wire, or remove the security key/module on the chip.
It's great for working on vintage equipment, stuff that might need (and warrant) that kind of repair. Less so if you run a cell phone repair shop.
There are a surprising number of Indian cell phone repair shops with YouTube channels that do feats of soldering, like repairing torn flat-flex cables… I bow down at their craft. If I ever get a sabbatical I’ll go to India and ask to be an apprentice.
It started getting really annoying earlier with PGA, to be honest.
Missing: how to solder a wire to the thermal pad on the bottom side of an IC. Assume there are components on the back side.
If it's QFN, you're probably best deadbugging it and jumpering all of the pins individually, but first make sure that you need the connection. QFP wireframes all have a center pad, whether or not it's electrically connected to the die or needed for thermal dissipation.
If it's an SO package, e.g. SSOP, TSOP, etc., Desolder the IC, add the jumper wire, bend the pins down enough to account for the thickness of the jumper wire, and resolder the IC.
Either way, make sure you have enough thermal mass connected to it for thermal dissipation. If there's components on the opposite side of the PCB, it's probably not much.
Isn't it just ground? If not, can you drill a corner of the chip? Most of chip material is just plastic.
Not always. Also, drilling the corner of the chip is a bad idea, as you'll likely hit one of the corner pads. There's not enough room.
> Destructive static charges are induced on nearby conductors, such as human skin, and delivered in the form of sparks passing between conductors, such as when the surface of printed board assembly is touched by a person having a static charge potential. [..] It is important to note that usually the static damage level for components cannot be felt by humans. (Less than 3,000 volts.)
Less than 3000 volts cannot be felt by humans? Should be 3000 millivolts right? i.e. 3 volts...
It's 3000V, but the energy delivered can be absolutely minute on a human scale, even for the highly sensitive nervous system. But metal oxide layers (what static discharge is often blowing holes in) are not on a human scale, they're atoms thick. And highly insulating, which leads to teravolts/metre field gradients.
For discharge that you can feel and see, the energy is even higher, but damage can be done far below this level.
Considering static discharge is a spark, I assume 3.000 volts is the correct scale. When I wince from a static discharge generally there's visible spark.
That's not 3 volts.
I have a problem of dismounting brittle SATA ports, maybe I just need more preheating but they are sitting really tight in the board.
SATA ports are often rated for only tens to hundreds of cycles, so they're often made out of fallapartium.
If you're replacing one, it's easiest to cut all of the plastic off first, then desolder the contacts one at a time. Sometimes you can even pull the plastic off, without damaging it. It's usually necessary to preheat the board.
Things like that are often the only through-hole components on a PCB.
I like to melt the original solder one contact or area at a time, add a little bit of fresh flux-containing solder if needed, and as soon as it gets to the consistency of mercury, vacuum it clean out with a good soldersucker.
Watch the PACE videos as well, they’re on YouTube and they also cover how to remove conformal coating and using different tools.
Beautiful illustrations!