[Carsten] messed up. He was soldering an ARM CPU onto a quadcopter board in haste, failed to notice that the soldering iron was turned up to eleven, and pulled some of the traces up off the PCB. In the process of trying to fix that, he broke three pins off of the 100-pin CPU. The situation was going from bad to worse.
Instead of admitting defeat, or maybe reflowing the CPU off of the board, [Carsten] lasered the epoxy case off of the chip down to the lead frame and worked a little magic with some magnet wire. A sweet piece of work, to be sure!
Of course, we’ve covered decapping chips with nasty chemicals a number of times before, but we think that the ease of removing the epoxy case with a widely available laser cutter is a quantum leap in convenience. In [Carsten]’s case, it made it worth attempting a decap instead of board rework. Indeed, we’ve just featured another project where someone lasered off an epoxy blob just for idle curiosity. We expect we’ll be seeing this technique a lot more in the future.
If you’re interested in peeking behind the epoxy curtain, you should also watch this presentation from the 31st Chaos Computer Congress that [Carsten] credits with inspiring him to pick up the laser rather than attempt a risky desolder. The talk covers a range of techniques to get down to the bare metal that are much easier than nitric acid, which frankly gives us the willies.
Thanks [Itay] for the tip!
59 thoughts on “CO2 Laser Decapping To Fix Soldering Mistake”
THIS, this is awesome to a new level.
I have damaged chips this way before, and it meant changing the chip as trying to remove enough material to solder again was unfeasible…I have a CO2 laser at home and this opens up tons of possibilities.
My hat off for this person.
It would be a huge waste to throw away a perfect arm micro just because a few pins broke off.
depends on what value you put on time, unless you are doing it for fun it is limited how much time you can spend trying to fix a $10 part
We don’t all have $10 to spend every time we bork a part. It’s nice to know there are options.
I think that the argument here is that if you have a laser cutter at home, you likely have $10 to toss at a new chip.
Yeah.. because the mentality of “throw it away and buy new” has done so much for the maker/hobby/repair community not to mention our landfills.
If the laser cutter your using belongs to a hackerspace, they probably also won’t allow it, since epoxy gives off fumes which are bad for you and bad for the laser.
If you have a laser cutter yeah, $10 would provably be nothing. BUT… how long will it take to get one if you have none left? That is where I believe this hack shines!
For me it is about the value of my time, but the couple days of precious spare/project time that I have to watch pass by waiting for that inexpensive component to ship is painful. When I can justify it I order spares that either go in the parts bin for later projects or get used to fix mistakes, but sometimes it is better to spend some time on a hack to salvage a part than to just wait idley while an antistat mailer creeps across the continent in my general direction…
I have some something similar with a scalpel. It helps to practice on some scrap boards first.
According to Mr Robot, epoxy residue causes leukemia.
This guy had too much free time on his hands to use this method when a hotair station gets the job done on a $10 part.
He had a laser and a problem that could be solved by a laser. I see zero situations in which a laser is the best possible solution.
NOT the best possible solution. Argh, why is there no edit button?
Because the Over Evillords whant us to tink before posting! B^)
Wow, that’s something I’ve never thought of doing. I’ve always used some extremely thin wire to solder to the bit of metal where the broken pin was attached to. This is…. considerably better.
I also have never thought of doing this…which leads me to wonder at this phrase…”ease of removing the epoxy case with a widely available laser cutter “. Widely available as in “if you have a spare couple of hundred bucks just laying around”. Not everyone has a laser cutter or access to a laser cutter. Therefore, perhaps widely available(anyone can buy one), but not widely accessible.
It’s a cool use of a laser cutter and ingenious thinking in the scenario.
Honestly they’re surprisingly available. My local hackerspace has one, and many hobbyists rent out their personal ones. Its much better than it was just two years ago, that’s for sure.
Could you use this technique to take the top off chips for reverse engineering purposes? (i.e. remove the package and expose the die then use other techniques to read out an embedded ROM from inside the chip or whatever)
There has been a recent (within the past 2 weeks) at least one HaD article about laser de-capping of epoxy.
have had a crack at it couple years ago, it destroys the silicon, in much the same way as using the laser to engrave glass. I have however etched away enough material to allow for a very quick session of nitric acetone heatgun to finish it off.
Is, that nitric, then acetone, then heatgun?
No, altogether. Nitric acid is a hell of an oxidiser. Acetone is the fuel. Those piddling little electric heatguns don’t know what “heat” is.
On the subject of laser cutters, how about a poll of HaD readers who own one to find out which is the current favorite? I really want to buy or build one, but have been overwhelmed by the many choices and mixed reviews. Surely I’m not the only person with this interest.
eBay, 40W, A4 bed, sub-$500. Replace controller if enthused enough.
Upgrade to the $2,000 500x300mm (50-100W) version a couple of years later.
Ideally in 2 years it won’t still be 2 grand.
Probably will. If the $500 model are any guide they maintain the price by dropping the quality even further.
The latest models (small viewing windows) have gone to a really light steel gauge and even done things like omitting the fan in the electronics bay.
The bits that matter (laser & mechanical) are the same.
Omitting that fan is actually a good thing. It’s rather useless mounted so far away from the power supply, so you end up sucking smoke out through the electronics bay. I took my fan out and put a new fan directly on the power supply. Much less odor. Just gotta seal up the rest of the machine now.
Considering the fan is under a few inches from the power supply, your definition of ‘uselessly’ differs somewhat to mine.
I tend to prefer to keep my electronics cool, each to their own I suppose.
I’m looking to start out at the higher end, rather than upgrading later. I already have a 60W 18″x24″ available at our local TechShop, but it’s too small for my needs. Any recommendations for bigger machines?
Next size up are the 1400x900mm (say 5’x3′) ones, usually 100W. Starting around $5-6000 not including import fees.
That’s about the size where people consider DIY; a 100W tube & power supply can be had for $1,500, and $500 for the other bits & pieces like motors, mirrors etc. Maybe $1,000 for nicer stuff. Then whatever for the case.
Thanks, but I know what the parts cost. As I said in my initial question, I’m looking for reviews/recommendations for specific models (including standard DIY builds) from folks who own them. Surely some people in this crowd own (or have access to) larger laser cutters and have opinions about them. Cost effectiveness is what I’m looking for–Epilog lasers (for example) aren’t going to be on my short list.
If you’re looking at the Chinese machines, they’re all the same.
Same glass tubes, same motors, same controllers (AWC608), same mirrors, same lenses, same air pump, same construction.
There are four basic machines, the ‘little’ desktop aka eBay machine, the next one up with the 500-600mm bed, the next group is what you want with the part-sheet bed at around 1200-1400mm wide, and finally the massive full-sheet machines.
I’ve got 3 small ones (3020), one 500x300mm 50W one (5030), and one 1300×900 100W (1390).
No idea who the manufacturer is on any of them, can’t even remember who the sellers were. Don’t care either.
The big one is one of these: http://www.aliexpress.com/store/product/China-popular-cnc-laser-cutting-machine-1300x900mm-Intech-laser-metal-cutting-machine-price/1819740_32341047133.html
Well, not that one as mine is yellow, but exactly the same.
Someone once made a comments that the Chinese lathes (7×12) are a ‘kit of parts’, they kinda work out of the box but need a bit of finishing; same goes for these lasers.
Pick a bed size, pick a tube, pick the cheapest supplier and allow for a week for fixing the weird & wonderful stuff; from putting back the bits that fell off to adjusting everything to adding foam to the doors so they don’t rattle.
You can spend the next year or two ‘researching’ or looking for someone to hold your hand, or learn everything you need to know by spending $500. If you’re comfortable with lasers then buy the one you need. If not get the $500 eBay one, play with it for 6 months and then buy a bigger one.
AGAIN,. Dear HaD editors, research for at least 2 minutes to provide some CONTEXT for the news article you are posting.
This is alternative method (dremel), done in 2011: https://forum.openwrt.org/viewtopic.php?id=37368
and a direct picture link, because attention span:
From the linked story: “Initially I was thinking of using a ‘dreml’ tool to remove the exoxy” Yeah, exoxy [and epoxy] removal sounds like fun.
A bright, powerful laser burning away a dark material, suddenly exposing a shiny reflective surface is something I’ll consider when I have spare eyeballs, and / or safety goggles.
Wait are you using a CO2 laser without safety goggles?
Can safety goggles block 40W or whatever of IR, reflected right into your eye? I thought they were more for the odd millisecond specular flash, the sort of freak thing there’s no predicting. The full power of the laser bouncing off shiny metal, I think would be a challenge. Just dissipating that amount of heat in such a tiny amount of time would be difficult.
What are the limits on laser goggles? Over-relying on a safety feature is a danger itself.
If you’re getting the full 40W in your face, you really doing it wrong.
Here’s what a short unfocused burst at low power does to your finger: http://www.dropbox.com/s/39ozslx52laxlut/Finger.jpg (that’s after the black bits fell off). Note the heat affected zone!
The beam spreads out so the power drops rapidly, the amount hitting your glasses is easily handled by them. There’s a YouTube video “CO2 laser vs regular safety glasses” showing an unfocused beam taking 16 seconds to burn through.
But never mind the glasses, your unshielded face is going to resemble that finger picture.
I have felt the scattered beam on my hands, it’s is noticeable – wear your glasses.
That method is also really awesome. And I think its more impressive.
At first I also wanted to use a dremel, but was to afraid to damage the leadframe.
I also could have just replaced the part, but that was not the goal. Instead I wanted to see if its possible to do that repair. And perhaps inspire other people to try similar stuff when a replacement is not that easy.
A scalpel is much more controlled. In fact, with careful work, only the pin that needs to be connected will be exposed, making soldering much easier.
Personally I think this is a terrible way to do it, a dremel could cause unknown vibratory stresses, possibly to the point of breaking bond wires. that would be terribly unfun to troubleshoot.
you are already in a failure mode scenario before powering on the dremel
You’re nowhere near bond wires – dremel from the edge would be a much quicker and easier way to do it.
The bond wires are encapulated… you will not cause enough deformation localized stress to break a wire or crack a heel.
Laser beats dremell -kind of like rock beats scissors ;-)
This put the meaning of Hardware Hacking to another level!
I need this or a dremel attempt to restore some bad cap electrolyte leak damaged ics (custom jobs, nothing that could be replaced), a few legs are gone but I hope the rot hasn’t gone far into the epoxy itself.
the key to doing this with a dremel is to use a soft wire wheel at full speed. I do it all the time.
As is written by several people, I would prefer hot-gun removing. No damage on PCB at all. If the chip is not so expensive, this method will be far better and handy…?
A proper hot air rework station is a valuable thing to have if you’re thinking about using small pitch or high pincount devices in your designs. I’m lucky to have one at work and makes short work of safely removing big chips without damaging the board or the chip since temperature control is provided. A laser is an interesting approach to this problem though.
How about using the laser not instead of hot air rework but for hot (not air) rework? Piu piu.
>> a widely available laser cutter is a quantum leap in convenience
Is this ironic or just stupid?
A quantum leap is the change of an electron from one quantum state to another within an atom. Therefore it is the smallest possible change in a system. Nothing to care about…
The idiom “quantum leap” refers to something that changes from one state to another with no intermediate steps, suggesting a sudden and dramatic change. I don’t think that the •size• of the leap is usually taken literally.
Or maybe it’s just people who don’t really understand the word “quantum” saying things that sound significant without knowing what they actually mean.
It’s rare for me to break a pin like that, and as others have said, it’s preferable from a reliability point-of-view to completely replace the damaged chip.
However, sometimes there is something *in* that specific chip that you need to get out. For example a damaged ROM or a PAL/GAL device. In cases like these I’ve uses a delicate temporary jig balance on top of the device programmer, just long enough to read the damaged device.
Now that I have one, next time around, I’ll give the laser a try.
Mad props, resilience for the win.
now THIS is what i like to see.
im suprised no one mentioned that they had just used a regular steak-knife,
i used a steak knife to save a 15$ chip that took 3 months to deliver,
i had ONE spare and i was too stubborn to admit partial defeat and use the spare.
instead i decided to sit there and saw at it for a very long time. i eventually uncovered about a half millimeter and soldered it. the solder still works to this day.
the repair (ed)(module) outlasted the prototype it was installed in :)
just remember, humidity can get in there much easier and ruin a chip/resistor/capacitor/ect faster then you can redesign (and build) a prototype, i was part lucky and part smart, i made sure to cover the spot with a thick layer of flux (before, during and after) and covered the flux with glue to keep it there.
it actually kept beer out ;)
Please be kind and respectful to help make the comments section excellent. (Comment Policy)