I was working on a project where I need to solder a BGA (0.8mm pitch). I have previously hand soldered SMD components like QFP, QFN but never a BGA.

I was reading about BGA routing and why it's a bad idea to place vias under BGA pad (solder gets sucked into pad via capillary action).

That got me thinking, while this is a problem for a mass produced reflow design, could this be leveraged for easier BGA prototyping (without having to use reflow/stencils).

Basically place a PTH on every BGA pad, smaller than the pad. Now the BGA can be placed and centered on the pads (the balls will sit on the holes).

Now flip the board around (while maintaining BGA location), and feed in solder from the back side into every hole (agree this is time consuming for 400 holes, but not a deal breaker).

I have previously soldered the QFN thermal pad with this technique. I tried to look up a lot, but didn't find anyone using this. And there's the universal wisdom that PTH under BGA pad is a bad idea.

What am I missing here?

Thanks,
Sagar

 

Are you sure the solder will come out of the holes to solder to the ball?

 

Capillary action should work both ways right? If it can suck solder from BGA side to bottom side, it should suck solder from bottom side to BGA side as well? My experience with solder (especially the leaded variety) is that as long as the PTH is not corroded or something, it should be wetted throughout by the solder.

Another variant of the said technique I thought of just now:

1) Fill the holes with solder paste
2) Place BGA with balls resting on the holes
3) Flip the PCB (while maintaining the BGA in place) and blast the underside with hot air.
4) Profit??

 

Sounds like an idea worth trying. Schmartboard (http://schmartboard.com/schmartboar...-5mm-pitch-65mm-pitch-2-x-2-grid-202-0016-01/ ) does that for the center contact on QFN and similar. I have used its breakouts in the past. One additional thing it does is to route out/groove where the "leads" go. That makes alignment of some packages easier.

 

Yes, I have built some PCBs myself with a hole under the QFN thermal pad, and it seems to work quite well.

I myself have never seen something like this for BGA though. My only concern is how reliably the solder will wet the balls with 300+ balls, or if there may be some cold joints.

I guess the only way to figure out is try it, but the BGA I was working on was a ~300 pin ARM, and the last thing I want is to route the whole thing and end up with an unworkable board. No pain no gain I guess

 

Agree, whether something like that works has to be determined empirically. One possible glitch might be whether your PCB manufacturer will object to so many small vias.

 

I wouldn't think that is a problem in itself, as the typical way to route a BGA is to place a via in the space between four pads. So you end up the same number of vias as pads and any manufacturer who does BGA typically supports these many pads.

 

Or, you could use this product I saw years ago... http://www.solder.net/products/stencilquik/
.