Hi there, I have here a nintendo switch console with blod and already tried reflowing, without much success, so I decided to remove the APU, reball it and then put it back on.
I reballed the APU and was making a last check for the pads on the Mainboard, just to discover a ripped pad. It is the pad to the most top left, when the mainboard lays as it does inside the switch. So, most left line of pads and the most top one on this line. Problem is, I cannot see any trace going to this pad. So I cannot just restore the pad with a pad-strip… So the question to you guys: Is this pad even “doing” anything, besides holding the APU Chip? Or do I need to fix this pad and if so, how can I do it? Where is the trace of it?
Thank you guys for the help
PS: How can I upload any pictures here? I cant use the picture button above and I cant embed them via link… Are there no pics in this forum?
Well, on your picture, it would be the pad on the bottom right.
I managed to find a tiny point somewhere in the middle of the pad, I assume a via and when checking in ohms mode of my multimeter, it has 0-1 ohms to ground, so I guess it is GND then? Can I leave it unrepaired then?
(How many more posts do I need to make before I can post pictures?)
The website you posted it awesome by the way, I love it.
I never repaired this kind of damage before, only attacked padstripes to traces. How should I do the repair with a via under my padstrip? Or should I just take the GND from some other point and just lead it, where it belongs under the APU?
Here a pic, you might see the golden glimmer of the via: imgur (dot) com/a/kOJsEJK
You have a couple of options. You could tin the via nub and then use some fine wire and solder one end and do the twirl to turn it into a “pad” or alternatively you could tin the via and use the circular pad which fits from one of those pad strip sets and place it over the via and while holding it with some tweezers use your soldering iron on top which will join it to the via… Ideally the circular pad you select will be ever so slightly undersized because that way you can use some UV curable solder mask around the edges to offer further strength.
Paractice on some dud boards first until you’ve got the process down
Yeah definately via / ground plane which should be going to that pad afaict
All right, thank you guys so much for all the help. My padstrips are about the same size as the original ones and usually I would apply some solder mark with my tweezers tip under the pad and cure it into place. As the soldering point is under it however, I might just use a pad strip wit a longer leg connecting to it and just solder it somewhere near to a ground point, right? It does not have to be the via underneath, where I solder it to…
Or i just do this twist thingie with some copper wire.
The UV solder mask only cures when the light comes into contact with it, so if the UV curable mask is underneath something (non transparent) then it won’t cure, I’d imagine if you felt like it was cured it’s probably more down to the UV cure which has spilled over the edges. You can test this if you’d like on a larger scale, grab some aluminium tape or copper tape and cut a large circle and apply the solder mask to the underside and then place it on some random board with a bare spot and attempt to cure, what you’ll find is after peeling the foil tape up is that the solder mask will still be wet and uncured
In your particular case, the solder will keep the pad in place and solder mask around the edge will provide additional strength and help it stay in pace during your eventual SoC reflow.
One advantage the wire coil method has, is, you can solder mask the entire area on top afterwards, because the wire is circular in it’s diamter and because the wire will inevatably have gaps between each coil wrap around, the solder mask will have a lot af surgace area and a good grab and after curing you can gently scrape (or abrade) the surface of the cured mask with a blad to re-reveal the copper (though this method is somewhat time consuming)
All righty. I will repair it with some wire then. No big deal, just time consuming, as you said.
Just one more thing, I reballed something this large for the first time and used a stencil for this job, but it was also my first time reballing with a stencil. I needed to do it multiple times and never got it “perfect”, neither with the little solder balls of 0,35mm width, nor with solder paste, which was a lot worse. In the end, I made it work for almost every pad and had about 50-60 left, where I put the 0,35mm balls manually, one by one and soldering it afterwards.
Problem is, there might be some balls, which are teeny-tiny bigger then the rest. Will this definitely cause trouble? Or are such small differences not that important?
Well you can use the circular pad but just understand that the only point of “anchoring” with solder mask is around the edge, which is perfectly fine and also understand that applying solder mask underneath the pad is somehwhat pointless as it won’t cure So if using the circular pad, just ensure you’ve got some amount of solder mask around the edge and your good, don’t worry if you over apply as you can scrape it off following curing.
You can’t reliably use solder paste on a chip this big, I don’t care what anybody says it’s just not possible.
You have to use preformed balls and the correct jig and the correct stencil, if your doing so then you will have no issues with 100% coverage and ball size all being the same. (you can search the forum regarding the jig and stencil which i’m pretty sure I’ve provided links to)
Trouble with the direct heat stencils and paste is as you say, inconsistent ball size, warped stencil (as a result of heating while the stencil is in place as opposed to when the stencil is removed on proper jig etc) missing balls, likely damage to the SoC as a result of the repated attempts to reball due to the solder paste falures mentioned above.
I treid the paste only once and saw the mess and just put it away to never be used again in such a case XD
The stencil is from aliexpress, it warps under heat extremely. I secured the stencil with capton tape, put on the balls and heated the chip from underneath, so the stencil gets warm the last. I can post a picture later and show the balls under the APU, if you give your ok, I will try to solder it back on. I really want to save this switch, as it is an unpatched one from 2017.
Yeah not saying it’s impossible (i’ve done it) but the sun, moons and stars all have to align for it to work… just not repeatable
With the correct jig and stencil you don’t reflow the balls with the stencil in place, you put a thin layer of flux on the chip, seat the upper portion of the jig with the stencil above, swirl the balls around, remove the upper portion of the jig that has the stencil attached and then reflow the balls (no stencil in sight) that way you avoid all these issues and it’s repeatable
In regards to the image, your better of showing me an image of the die in the center of the wafer, if the resin around the edge is still a blueish or greyish tone, then your good, if on the other hand it’s brown then unfortunately the SoC has been cooked and it’s highly likely it’s dead
You are making me very sad man, the color is kind of amber. I worked with 390 °C, but I am certain that the color was like this to begin with.
Lets say the chip is cooked, will any other switch APU fit onto this switch? If yes, it the switch then patched? I could not find out which parts got “pathed” on the v2 switches.
If you mean does the board determine whether it’s patched or unpatched or not, then no, this is defined by the SoC iteself, so if you put a patched SoC on the board then it’s patched and if you put an unpatched SoC on the board then it’s unpatched.
Yeah, I mean it’s worth a go considering you’ve already reballed it. Though often times (and rather sad) it’s not the SoC which was ever at fault and causing the blue screen, more often it’s fuel gauge or ram (or Ram joints) … seen a lot of boards like this
As I say, you can use a patched SoC on your board (they are interchangeable) but obviously you won’t be able to use the exploits you’d wanna use like on an unpatched SoC.
Just a side note, the reason for the issues with the resin getting cooked around the die is, I think of these chips a bit like underfilled chips (those you might see on apple devices) and I presume the die is just soldered onto the wafer, i’d imagine there is ball squeezout going on under the die and/or physical damage due to the nature of the hard resin being used (internal fractures due to the heat and hard bonding)
All a guess mind as I’ve never been able to cleanly separate the resin which nintendo uses and remove the die for confirmation, it’s so hard and brittle and almost has ceramic qualities but from experience the browning usually confirms death… often times you can measure the resistance across the caps for the boot CPU rail and secondary CPU rail - often times they are quite a bit lower or dead short which is typical in these cases, and if you find that’s the case on yours then it’s not worth the effort of putting the SoC back on the board as this is confirmation of death
This switch is in pretty bad condition, i mean the case and front screen (touch I guess) are pretty beaten up. The joy cons need new analog stick caps, they can be “undocked” without pushing the button. All in all the only reason I wanted to fix this one was the fact that it was unpatched. If this fails, I will just sell it as parts or so, I am done with it. I lost way too much time on this one.
