manual BGA placement

We are increasingly using chip-scale parts, like the EPC GaN BGA fets. Manufacturing handles these fine in production, running panelized boards through giant machines. But I want to do onesies of engineering prototypes. Our old prototyping techniques won't work with tiny chip-scale parts. We can do layouts and get cheap fast-turn boards, but we don't want to panelize and program the production machines.

I want some optical assistance for placing these parts manually onto single boards fast.

This video is vague about placement, and crazy about verifying placement during the process.

formatting link

They don't use paste, which is good. We can use sticky flux to hold the parts in place until we heat the board and melt the balls onto the pads.

I'm thinking about something like two USB microscopes added to a manual p-p machine, at 90 degrees, peeking at low angle at the sides of the part as it's lowered onto the board. Or something like that. Any ideas?

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin
Loading thread data ...

torsdag den 7. marts 2019 kl. 23.21.42 UTC+1 skrev John Larkin:

formatting link

Reply to
Lasse Langwadt Christensen

Well, it might be way more than two USB microscopes but I'm using an A.P.E. Sniper II Vision rework station that I got from Ebay for $850. Here is their Sniper III station that is almost the same as older Sniper II:

formatting link

It's reflow part on the right is very good but it is what's on the left that makes BGA (and others) rework/assembly very easy and pleasant.

There is a vacuum part holder (vacuum tweezer? something like regular vacuum pens with rubber cone) that holds your part pads down, towards the board and the board underneath on the table. Between them goes beam splitter with camera looking to it. What you see on the monitor connected to the camera is both PCB and part pads overlayed over each other. You move the board to visually align the pads, step on the pedal switch and see the magic -- that beam splitter gets pulled out, vacuum pen with the part drops to the board releasing the part when it touches that board, then everything retracts to original position and you are ready for the next part...

Those are pretty expensive and it was pure luck to get mine for $850 from one of the vultures selling busted business' remains at the peak of dotcom bust but it is the idea :)

--
****************************************************************** 
*  KSI@home    KOI8 Net  < >  The impossible we do immediately.  * 
 Click to see the full signature
Reply to
Sergey Kubushyn

How do you plan to heat the solder? An assembly house tried to mount a BGA for me and found the flux was not of any real use once the board was heate d. But then he was heating with a hot air desoldering tool.

If you can't heat with IR I'd say give it up.

As for alignment, I'd say use the silkscreen outline. You don't say what a ccuracy you need, but if the silk screen isn't good enough, you should use fiducial marks by the part in the copper layer. For machine vision they us e dots, but for your eyeball placement you might want crosses on diagonal c orners.

Rick C.

Reply to
gnuarm.deletethisbit

I'd use three or four, so you can detect angular misalignment more easily, I've done optical inspection of large BGAs, and more angles really helps.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

The part I'm using now only has four balls, so I figured that 2 cams might work. My mfg folks are giving me a hard time on this, but if I harass them enough maybe they will invent a way to fast-turn my test boards. There are tons of Youtube videos of people doing this sort of thing at home.

The big battle is going to be to get them to try it without solder paste. Good grief, I'm the President and everybody should instantly do exactly what I say. That doesn't work nearly as well as it's supposed to.

I don't suppose that donut pads would help align the balls, sort of snap into place.

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
John Larkin

USB microscopes seem slow (lots of latency), (but I've only tried cheap ones)

Possbly a closed-circuit video setup camera with the right lens would work better. Perhaps a reversing camera with new optics.

Perhaps prisms could be used to alter the view angle of vertical cameras - like your BGA inspector does.

--
  When I tried casting out nines I made a hash of it.
Reply to
Jasen Betts

Huh, well you can encourage that sort of behavior if you want. That wasn't my impression of you.

Four balls? I know nothing of BGA's but doesn't' surface tension suck it into the right place or not. Some simple electrical test might be easier? (how many units?)

George H. That doesn't work nearly as well as it's supposed

Reply to
George Herold

This is why I like leaded parts. QFPs are great! QFNs are ok too as long as they don't have the hidden row behind the front row. Then they are really land grid arrays (LGAs).

There's a new MCU out that is only 6 uA/MHz! Unfortunately it only comes in balled packages, BGA and Chip Scale. I might have to learn to design BGA type boards.

Rick C.

Reply to
gnuarm.deletethisbit

Yes, surface tension tends to align the chip on the board, however with only four balls, the force applied isn't all that great. BGAs are such easier to use than one would expect. In a previous job, we had zero fallout on BGAs but QFNs drove manufacturing nuts. Again, the more balls the merrier. Serious companies still use XRay inspection on anything with solder pads or balls to make sure their process is good.

Reply to
krw

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.