You will need a steady hand...

From a recent ABUG event, a demonstration of soldering some relatively conventional but rather fine pitch surface mount devices:
Micro-soldering demonstration (JudgeBeeb)

Here's a nail-biting short video, 15min, hand-soldering a 12-ball WLCSP device, dead-bug style:
Hand soldering a WLCSP package

Here's a long video, with more by way of commentary and recommendations, again dead-bug soldering, this time a 4-ball package:
Beginners' guide to hand soldering BGA and WLCSP for prototypes

(Apparently, as the solder balls on these parts are lead-free, it's better to use lead-free solder, even if one's usual preference is for leaded.)

I've routinely soldered 0.4 mm with a fine tipped iron, but wiring up a flipped BGA with a rough iron, that is just masochism. Like Youtube videos about dropping molten metal into pools, or dropping giant rubber balls from the top of buildings. Fascinating, sure. But in a circus kind of way.

Here's a 49 ball tiny BGA being wired up dead-bug style to tombstone resistors. With 30um wire. I think the idea is to substitute a suitably-programmed FPGA into an existing footprint - and it's a success.

Story here, with videos:
https://threadreaderapp.com/thread/1145 ... 82113.html
"This is either a great idea, or a terrible idea." - @GregDavill

_________________
http://WilsonMinesCo.com/ lots of 6502 resources
The "second front page" is http://wilsonminesco.com/links.html .
What's an additional VIA among friends, anyhow?

Hand soldering to BGA is not as hard as it looked. It is a matter of practicing on few spares before trying on a real part. This is my attempt on a 68EZ328 BGA.
https://hackaday.io/project/36309-melti ... dragonball

Bill

You would not want people to come in and say 'could you do mine', though.

In my experience, having built hundreds of SMT boards with a homemade hotplate - 1mm grid BGA is by far the easiest of them all (shiny side down). Even the big scary (and expensive) Xilinx chips.

Glop on some flux, no paste, visually align the BGA chip to better than .5mm (I silk-screen the chip outline to make it really easy), and stick the board onto the hotplate for x seconds. When the balls melt surface tension pulls the chip into perfect alignment - every time.

It's hard to screw up (even though I managed a few times - overheating melts all the balls into a hot mess, literally; jarring the hot board while removing from hotplate can knock the chips off).

On the other hand, 0.5mm TSOPs gave me much grief during my last project. The pins are exposed and it seems it should be much easier than BGA, but no. I had a lot of trouble adjusting just the right amount of paste - too much and it shorts to adjacent pins; too little and you get unconnected or intermittent connections. Lots of rework...

Moral: don't be afraid of BGAs.

_________________
In theory, there is no difference between theory and practice. In practice, there is. ...Jan van de Snepscheut

_________________
http://WilsonMinesCo.com/ lots of 6502 resources
The "second front page" is http://wilsonminesco.com/links.html .
What's an additional VIA among friends, anyhow?

I started out with an oven, but was never too happy about it. After building a hotplate from an aluminum block almost 10 years ago, I never looked back.

It is wonderful to be able to see the BGA chips sag into place and all solderpaste melt. Looking through the window of my toaster oven was not helpful and relying on the timing device often did not work out so well. For one-off boards you really want to see that happen; for small runs you can time the first reflow and the rest are very consistent as long as you place the boards into the same place on the hotplate.

You can't do double-sided boards on the hotplate, but it was never an issue.

I've had limited success reflowing BGAs using a hot-air tools, but it's a bit stressful, and it's hard not to blow off all the 0402's.

_________________
In theory, there is no difference between theory and practice. In practice, there is. ...Jan van de Snepscheut