Via in pad and voids - doing it the Jack Sparrow way

[eTobey]

Some of you may remember the scene of "Pirates of the Caribbean - At World's End" where they turned the ship uside down. When thinking about reflowing of  parts that have vias in the pads, i thought maybe this could work there too!

As a lot of gases coming out of the solder paste when reflowing and creating bubbles, i thought wouldnt they just vent out of the vias if you flip over the board when reflowing? But would the parts (not considering heavy parts) stay on the board then? What do you think.

This channel has some interesting videos:

[Whales]

Testing this would be interesting (you'd glue the parts down like they do for 2-sided SMD reflow), but I forsee a lot of annoying variables:

 - Bubbles are large and seldom, which is the opposite of what you want for consistency.
 - Pad size & via size & component thermal mass
 - Solder alloy consistency and flux chemistry across the entire bubbling range of temperatures and time

[T3sl4co1l]

Gas is trapped when it has nowhere to go, i.e., one-side-tented vias.  You tent both sides, it's fine, it has somewhere to go.  That's why you use untented, or fully tented (preferably plugged and capped) vias, for via-in-pad.

Buoyancy probably doesn't make much difference, as surface tension dominates on these scales.  Maybe it would help some, maybe it would reduce but not eliminate the rate of voiding, for example.  Dunno, would have to see it tested.  Meanwhile, parts simply bubbling/popping off the inverted board (or get shaken off during handling), is probably the elephant in the room.

Tim

[jmelson]

I reflow double-sided boards all the time.  Some of these boards even have a TO220-5 regulator on the back side.  I stuff the back side first, then reflow, then stuff the front side and reflow again.  I have NEVER had a part fall off the back side during the 2nd reflow.  I have had parts get moved if I am in a hurry and move the board before it has cooled.  The rack can bump those parts.  Oh, and I do not glue the back-side parts!
Jon

[T3sl4co1l]

Yeah, once soldered, surface tension usually does a good job.  To be clear, I'm imagining inversion at the paste stage... paste isn't very strong at all.  With glue it would be okay (is glue and paste ever used simultaneously..? I don't know, but it seems doubtful, or complicated at least..).

One could also ask whether re-reflow allows voids to work themselves out (and, there's probably some studies out there to this effect?), but it's probably also an imperfect approach, and one with significant disadvantages (you can only reflow a board so many times before risk of damage starts going way up; usually components are rated for 5 cycles or so).

Tim

[eTobey]

There are some studies! Just look at the youtube channel above. They use some vacuum, but you can find other channels to. I found 2 papers, that concluded that tented vias would not really help. Also overall it looks to be that there is nothing that would improve the voiding really, except vacuum. They also did only very few of those test, so it is actually not much of a study id say.

I had the idea, that thin bands of soldermask could help to vent out the voids.