xray'ing BGAs

...snip...

My experience has been that the BGA assembly process is so well understood and easy to get right that universal X-ray inspection is not needed. Most of the assembly houses I've contacted have an X-ray machine, but they only use it when a problem is found and that is usually the board, not the BGA.

I'm not sure I am familiar with the "cool sticky-down stencil things". Are you talking about a way to stencil the solder paste on an already assembled board? I guess that will be a bit of a PITA, but any good assembly house can do that for you. Do you do your own assembly in-house? If so, this might be something to farm out, but I've been told *repeatedly* that BGAs are very easy to get right. Much easier than LGAs or some of the QFNs.

My only concern with most of the BGA's I'd like to use is the small board features required which drives up the board costs and makes prototyping harder.

Rick

and

never

that

The first ones I saw were pretty rough.. the systems seem to have improved quite a bit:

It's these things:

You stick them on the board, squeegee on some paste, poke the BGA down, and reflow. The BGA sort of self-aligns and the sticky thing stays. This would be great for hobby-type use, or for any people who don't have a lot of stencil and placement gear. I'd think that these things and a convection oven would do good BGA work.

We use these for rework only; we have the fancy gear to do the initial stencil/place/reflow stuff.

I guess that will be a bit of a PITA, but any good

Yes, we do most BGA assembly and rework in-house. BGAs are a lot more reliable than fine-pitch leaded parts. And a lot easier to inspect, since you really can't inspect them.

--

John Larkin         Highland Technology, Inc 

jlarkin at highlandtechnology dot com 
http://www.highlandtechnology.com 

Precision electronic instrumentation 
Picosecond-resolution Digital Delay and Pulse generators 
Custom laser drivers and controllers 
Photonics and fiberoptic TTL data links 
VME thermocouple, LVDT, synchro   acquisition and simulation

I'm not a fab guy so I don't know all the details, but my understanding is that the ball is supposed to merge with the paste and squish down a bit. This stencil would alter the way that works to some extent. I have no idea if this would be significant. I know the PCB fab guys have really optimized this over the years to get the optimum repeatability. For rework I expect it is a mox nix and would work ok. Time will tell! If your BGAs fall off after a year in the field you will know they didn't do so good.

Yeah, I guess you can consider them easier to inspect, but then isn't that why you started this post? Your customer wants you to inspect them, eh?

I once saw a device that used fiber optics to slip a camera view under the BGA to visually inspect the balls. You might look for that as a better alternative... although it won't look inside the balls... which is what you want to avoid I believe.

Rick

and

never

that

The X-ray is to verify the thermal profile of the process of flowing the BGA. I can see why they would want it. But its typically done at the begining of the work flow to verify the setup. And probably then an again to verify the process.

They are not comfortable with your rework.

Cheers

the

units

and

never

black

that

This will be a one-time upgrade of 25 boxes. By the time we tune the process, it will be over.

We've shipped them around 400 boxes so far, containing roughly 600 BGAs, with zero problems. I don't know what they are thinking.

--

John Larkin         Highland Technology, Inc 

jlarkin at highlandtechnology dot com 
http://www.highlandtechnology.com 

Precision electronic instrumentation 
Picosecond-resolution Digital Delay and Pulse generators 
Custom laser drivers and controllers 
Photonics and fiberoptic TTL data links 
VME thermocouple, LVDT, synchro   acquisition and simulation

ents to

Altera

lace the

ngs.

ll units

ycled and

entire

d very

eldom

we never

es...

-s...

mb...

rep=...

me black

ven

ests that

They are just covering the asses, but to be fair I can understand say there is some difference in the confidence of a bga soldered on a board in the normal everyday automated fine tuned fashion done millions of times and a rework done by hand to replace a BGA

but I'm sure there are people who do it day in day out with success

-Lasse

and

never

that

I've seen X-ray photo's of a BGA and I could see the balls but I wouldn't dare to speculate on the quality of the solder joints. Reading an X-ray photo is just like looking at an ultrasound of an unborn baby. A trained eye knows where to look for. In short: if your customer doesn't trust you just have the boards reworked by a specialised company.

--
Failure does not prove something is impossible, failure simply 
indicates you are not using the right tools... 
nico@nctdevpuntnl (punt=.) 
--------------------------------------------------------------

We do have an optical gadget that gives us a side view of the chip at about board-surface level. It can reasonably inspect the first three rows of balls, and spot shorts and debris in the lanes pretty much all the way through.

--

John Larkin                  Highland Technology Inc 
www.highlandtechnology.com   jlarkin at highlandtechnology dot com    

Precision electronic instrumentation 
Picosecond-resolution Digital Delay and Pulse generators 
Custom timing and laser controllers 
Photonics and fiberoptic TTL data links 
VME  analog, thermocouple, LVDT, synchro, tachometer 
Multichannel arbitrary waveform generators

the

units

and

never

black

that

or just xray the stuff, send some photos and be done with it.

it's like a long report, nobody reads them anyways, but it's a big deal if it's missing.

Probably the best advice yet. Have the x-ray done & give to them to interpret. But that may open a can of worms and get a 3rd party involved.

John, you might just repeat to your customer what you said here: you've had no problems so far, and provide an "extended warranty" for these such that if they have any issues with the *reworked* boards specifically caused by the BGA soldering you'll cover the costs.

If the customer is being unreasonable, there's always the option of saying "no"...?

Good luck.

Missing x-rays... reminds me of the movie, "The China Syndrome"...

Rick

That would be politically messy. Some sort of compromise must be found. As far as the warrenty goes, they say that process downtime will cost a million dollars a day, or a million an hour, or something like that, so a warranty is in the noise. They should buy lots of spares!

--

John Larkin         Highland Technology, Inc 

jlarkin at highlandtechnology dot com 
http://www.highlandtechnology.com 

Precision electronic instrumentation 
Picosecond-resolution Digital Delay and Pulse generators 
Custom laser drivers and controllers 
Photonics and fiberoptic TTL data links 
VME thermocouple, LVDT, synchro   acquisition and simulation

That's precisely what our large customers did. There was one (3/4-way multiprocessor) computer to do the work, duplicated on the mirrored "side" as a backup, another system (6 or 8-way) running hot with that system, and two more forty miles away, running warm backup in case an asteroid hit the first building. I figured around a hundred fifty million $$ sitting there, just in case.

al if

d.

had

hat if

the

ing

not the same kind money but, remember a place that absolutely had to print many many thousands of pages every night delivering newspapers or something like that

So the had some big monster IBM printer with an expensive service contract, but that only meant if it broke they could call a technician wait an hour and hope he could fix it

One day they just bough bunch of standard printers, much much cheaper and if they all worked it was much faster, as long as they didn't break all of them at the same time they were still fast enough to do the job in time

-Lasse

if

if

Speaking of printers (thread drift on:), I thought the IBM 3800s were pretty amazing beasts and for the early '70s, they certainly were. About five years ago I interviewed at Kodak in Dayton, OH., in their inkjet printer division. These things were incredible. They were as fast as (commercial) laser printer printers, with the programmability, and as cheap as offset and full four color. They drank ink by the 55 gal drum. ;-)

to

Altera

the

units

and

entire

never

black

that

That is tricky. There are always people lurking around who see a spot and then start to raise red flags even though they have very little idea what the fuss they are making is all about. Customer politics...

--
Failure does not prove something is impossible, failure simply 
indicates you are not using the right tools... 
nico@nctdevpuntnl (punt=.) 
--------------------------------------------------------------

I have one of these setting on my desk. If I ever get caught up with work, I'll test it.

Very clever pulsed HV generator package, using mylar strips, a few diodes and little else.

The patents are worth reading.

A friend uses ordinary B&W film. He uses Al, Cu, and Nylon sheets to adjust the energy spectrum. He gets beautiful pics.

The shaped nylon block trick is used in the CT machine I used to work on. It lets the designers balance the beam energy distribution.

CCD based Dental Xray detector might also work.

Steve

On a sunny day (Fri, 21 Dec 2012 11:26:15 -0800 (PST)) it happened Owen Roberts wrote in :

Build your own:

He actually has quite a nice video about how to control xray tubes, reverse engineered some commercial unit, it seems intensity is controlled by controlling heater voltage of those tubes, and measuring anode current. The HV is constant (energy).