Printed Circuit Board Assembly & PCB Design SMT Electronics Assembly Manufacturing Forum

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.

672 PBGA on PC Card .015 thick pcb

Brad Arnold


672 PBGA on PC Card .015 thick pcb | 8 October, 2001

We are trying to produce a PC Card bus assembly with a 672 pin BGA on one side. These pcb's are .015 thick. I believe the ball pitch is .5mm. The pcb's are plated with white tin. The pcb's are in a 1 X 3 array. We had 2 plates with appropriate cutouts made out of .060 AL. We use the 2 plates to sandwich the pcb panel to keep it flat.

Using a 4 or 5 mil stencil produced bridging.

We tried mounting the BGA using sticky flux only on the tin pads. We tried "tinning" the pads with 63/37 then mounting the BGA with sticky flux. These methods produced about a 60% yield.

Anyone have any ideas as to how the yield could be increased?

Thanks for the input.


reply »


672 PBGA on PC Card .015 thick pcb | 9 October, 2001

That�s a 672 BALL BGA, bud!!! It�s a BIG Mutha!!! [I know. I know. For some reason suppliers call balls "pins".]

Let�s focus on: * Clarifying the units on your PCB and aluminum plate thickness measurements. * Describing your use of the aluminum plates through your process [ie, printing, place, reflow, etc.]. * Explaining how do you apply and control the application of your sticky flux. * Checking that your 672 BGA pitch is 0.5mm. [Never heard of one like that. All of ours are 1.27mm.] * Discussing the aperture size & shape that you used with your 4 or 5 mil stencil. So, was it 4 mil or 5 mil? * Telling us about shape and coating thickness of the BGA pads on your board.

reply »



672 PBGA on PC Card .015 thick pcb | 11 October, 2001

OK, OK. 672 ball BGA.

The pcbs are PC Card bus size and are .015 thick. There are 3 on a panel with about 1/2" between and rails.

The aluminum plates are about 3" by 7". There are 3 cutouts which allow for the parts from either side to protrude during palcement and reflow. The panels are printed first then placed into the AL bottom plate. An AL top plate is placed on top and 6 nuts clamp it tightly to the bottom plate thus holding the pcb flat. The plates/pcb are then placed on conveyor for placement and reflow.

We print paste for everyting but the BGA. We use a brush to apply the sticky flux. Therefore I can't say that the flux application is controlled very well.

The BGA is actually 1 mm pitch, sorry. Balls are .024D and the pads on the pcb are .0157.

We tried one stencil, 5 mil thick with .016 aps. After consulting with the stencil house yesterday, we are ordering a 4 mil stencil with 3 patterns, 1 with .008 aps, 1 with .010 aps and the third with .012 aps. We'll have that in a couple of days.

The pads are .016, round. The plating is white tin.

Further FYI, we have a BTU VIP 70 with 5 zones. Due to the expense of the BGA (500.00 ea.) the customer was reluctant to allow us to do a destructive profile. Now, they are giving us a couple of 3 pcb panels and 3 parts. once we get the stencil we will drill holes in the pb and bring 2 thermocouples up under each BGA on each PCB.

We use a CircuitMaster Designs profiler with 6 thermocouples.

Typical profiles we run have temp ramp rates of 1.25 to 2.00 C/second.

Now, what are your thoughts?

Thanks in advance for your consideration.

reply »


672 PBGA on PC Card .015 thick pcb | 11 October, 2001

Bruce I�m with yall. Designers and fabricators say �pins�.

You say: * BGA pins are 1mm [0.039�] pitch. * BGA pads are 0.4mm [0.0157�] diameter. * BGA pins are 0.6 mm [0.024�] side-to-side. * Board thickness is ~0.4 mm [0.015�]

First, as you can tell from earlier comments, we have not done devices both this BIG and with this pitch. Further, we do flex and routine PCMCIA, but have not done your intermediate 15 thou boards.

I�m surprised there's bridging in your 5 thou / 1:1 ap stencil.

Consider: * Changing your fluxing process. We spoke about painting flux a rough process, in another thread on SMTnet within the past two weeks. [] Although, I doubt this is the source of your problem. * Slowing down the ramp on your spike. There�s only 0.4mm [0.0157�] spacing between balls. This leaves little room for thermal expansion, tolerance, and what not. * Making sure your time over liquidous is at least 60 sec. * Changing your bare board solderability protection to ENIG. Immersion tin is fairly uncontrollable. And may not be a good choice for the flatness and uniformity requirements of your device. * Reviewing your fixturing method. Could the huge differences in CTE of the aluminum, open "windows" in the aluminum, and the bare board and firm securing of the board to the fixture be causing board warping during reflow? We fixture our flex and PCMCIA with unclad FR-4.

reply »

High Precision SMT Fluid Dispensers

On-board Dispensing of Dots & Lines