Investigacion De Causas Raiz

Main POP
Dakota.
15/09/2011

Background
Item: RIMPOP-39388-006 Pareto First Level:

Background

Radio Connect
The main cause for send units to repair is the Radio Connect Fail in BLT, its means that the unit will not start. (Note: This fail doesn’t appear at Oracle because the unit don’t power up and don’t send any BSN. The trend of this fail is reported by manufacture.) Due to thisfailure mode is related mainly with the minimum system (U902 & U801) we review on x-ray to discard some SMT issues. It was not possible found anything abnormal.

At X-Ray nothing was found

The Shield was remove and found that the U902 is lifted

NOTE: We review 10 units and found 1 with this condition

TML Investigation

Hypothesis to investigate
To validate what condition canproduce the lifted or raised component we have the next hypothesis: Ho1= On SMT line, the flux dispenser not wet all the balls of the BGA so the effect of
raised or lifted after the reflow oven.

Ha1= The flux dispenser wet all the balls of the BGA. Ho2= The BGA balls are not uniform and produce that the flux dispenser not wet
correctly.

Ha2 = The BGA balls are under specification. Ho3 = Theshield placed upper the PoP (U801) & BGA (U901) is misalignment and
move the components.

Ha3 = The Shield is placed correctly. Ho4 = AOI coverage include the PoP (U801) & BGA (U901) before to place the shield
in order to confirm the correct position.

Ha4 = AOI coverage don’t include the PoP & BGA. Ho5 = Exists material on the line to produce ESD to the components inducing a
Functional FailureHa5 = Not exists material to produce ESD to the components

Ho1= On SMT line, the flux dispenser not wet all the balls of the BGA so the effect of raised after the reflow oven.
In order to validate this hypothesis, we validate the process and found:

Flux dispenser Panasonic

Flux dispenser Fuji

Exit have two different dispensers, the Fuji is more constant than Panasonic, thisbecause the mechanics that have each machine. Another evaluation we did was the way to measures the height of the flux

The method to measure is not reliable

Ho1= On SMT line, the flux dispenser not wet all the balls of the BGA so the effect of raised after the reflow oven.
We validate the wet flux on the component

On Panasonic we found one ball without enough flux

On Fuji don’t found anyball without flux

o Panasonic dispenser is not as steady as that of Fuji.

This failure is related with the PoP

Hypothesis Accepted

Verification on Critical Area of Hypothesis Ho1
In order to reproduce the problem we did:
A) Take the same flux used on the line B) Wet only a half of the component C) Place the component and review the effect.

Print of the component , only half wetAir Vac Machine used to place the component

Video of the process evaluation

Verification on Critical Area
Picture of the result after test:
No Flux Wet side

Conclusion: With this evaluation we can confirm if the component doesn’t have the enough flux is possible to have a raised or lifted component. Hypothesis Accepted

Proposal to Hypothesis Ho1
The proposal A is generate afixture to take the measure in order to eliminated the human factor. This fixture must be fixed measurement tool, so that reliable measure is taken. The idea is have two poles down when to take the height of the flux.
Owner: Raul Baltazar
Measurement tool

Flux Dispenser
The proposal B is increase the flux quantity in the dispenser from 0.175 microns to 0.200 microns in order to wet all the ballscorrectly.
Owner: Raul Baltazar

The proposal C modify the control of the flux to assure that never have flux with time expired on the line or clean process to assure never have flux on the dispenser in the lunch time.
Owner: Raul Baltazar

The proposal D validate the flux dispenser calibration on the Panasonic machines
Owner: Raul Baltazar

Ho2= The BGA balls are not uniform and...