SPRAA99C March 2008 – May 2021 AM3351 , AM3352 , AM3354 , AM3356 , AM3357 , AM3358 , AM3359 , AM4372 , AM4376 , AM4377 , AM4378 , AM4379 , OMAPL138B-EP , TMUX646
Reliability is one of the first questions designers ask about any new packaging technology. They want to know how well the package will survive handling and assembly operation, and how long it will last on the board. The elements of package reliability and system reliability, while related, focus on different material properties and characteristics and are tested by different methods.
Package reliability focuses on materials of construction, thermal flows, material adherence/delamination issues, resistance to high temperatures, moisture resistance and ball/stitch bond reliability. Thorough engineering of the package is performed to prevent delamination caused by the interaction of the substrate material and the mold compound.
TI subjects each nFBGA to rigorous qualification testing before the package is released to production. These tests are summarized in Table 3-1. All samples used in these tests are preconditioned according to guidelines of the Joint Electronic Device Committee (JEDEC) A113 at various levels. Typical data is presented in Table 3-2. The nFBGA packages have proven robust and reliable.
| Test Environments | Conditions | Read Points |
|---|---|---|
| HAST | 85RH/85°C | 600 hrs., 1000 hrs. |
| Temp. Cycle | -55/125°C | 500 cycles, 750 cycles, 1000 cycles |
| Thermal Shock | -55/125°C | 200 cycles, 500 cycles, 750 cycles, 1000 cycles |
| HTOL | 125°C, Op. voltage | 500 hrs., 600 hrs., 1000 hrs. |
| HTOL(1) | 140°C, Op. voltage | 500 hrs. |
| HTOL(1) | 140°C, Op. voltage | 500 hrs. |
| Bake(1) | 150°C, 170°C | 600 hrs., 1000 hrs., 420 hrs. |
| HAST(1) | 170°C | 96 hrs. |
| Package Types | ||||
|---|---|---|---|---|
| Leads | 113ZVD | 289ZVL | 289ZWE | |
| Body (mm) | 8x8 | 12x12 | 13x13 | |
| Die (mm) | 4.2 × 4.2 | 5.8 × 5.7 | 10.6 × 8.2 (Die 1) 8.1 × 7.7 (Die 2) | |
| Level | 3 | 3 | 3 | |
| Test Environment | ||||
| T/C, -55/125°C | (500 cycles) (1000 cycles) |
0/78 0/78 |
0/83 0/83 |
0/246 0/245 |
| T/S, -55/125°C | (500 cycles) (1000 cycles) |
0/77 0/77 |
0/77 0/77 |
|
| HAST, 85°C/85%RH |
(600 cycles) (1000 cycles) |
0/78 0/78 |
0/78 0/78 |
0/77 0/77 |
| 150°C Storage | (600 cycles) (1000 cycles) |
0/45 0/45 |
0/77 0/77 |
0/77 0/77 |
| HTOL | (1000 hrs.) | 0/120 | ||
Board-level reliability (BLR) issues generally focus on the complex interaction of various materials under the influence of heat generated by the operation of electronic devices. Not only is there a complex thermal situation caused by multiple heat sources, but there are cyclical strains due to expansion mismatches, warping and transient conditions, non-linear material properties, and solder fatigue behavior influenced by geometry, metallurgy, stress relaxation phenomenon, and cycle conditions. In addition to material issues, board and package design can influence reliability. Thermal management from a system level is critical for optimum reliability, and thermal cycling tests are generally used to predict behavior and reliability. Many of these are used in conjunction with solder fatigue life models using a modified Coffin-Manson strain range-fatigue life plots (number of cycles to failure has an inverse exponential relationship with the thermal cycle temperature range).
In addition to device/package testing, board-level reliability testing has been extensively performed on the nFBGA packages. Various types of daisy-chained packages were assembled to special boards, with electrical measurements made in the initial state and then at intervals after temperature cycles were run. Table 3-3 shows a summary of a wide range of board-level reliability.
| Conditions (With Solder Paste) | Failures/Sample Size | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Requirements | Extended Range | |||||||||||
| Package | TI Mfg Site | Body | Pitch | Die | Temp. | 500 | 800 | 1000 | 1500 | 2000 | 2500 | 3000 |
| Test Site | (mm) | (mm) | (mm) | Cycle (°C) | (Cycles) | (Cycles) | ||||||
| ZVD 113 balls |
TI Hiji | 8×8 | 0.65 | 5x5 | -40/125 | 0/36 | 0/36 | 0/36 | 0/36 | 0/36 | 0/36 | 0/36 |
| ZVD 113 balls |
TI Hiji | 8×8 | 0.65 | 4.2x4.2 | -40/125 | 0/48 | 0/48 | 0/48 | 0/48 | 0/48 | 0/48 | 4/43 |
| ZVD 289 balls |
TI Hiji | 12×12 | 0.50 | 9x9 | -40/125 | 0/95 | 0/95 | 0/95 | 0/95 | 0/95 | 0/94 | 0/94 |
| ZVD 289 balls |
TI Hiji | 12×12 | 0.50 | 6.5x6.5 | -40/125 | 0/36 | 0/36 | 0/36 | 0/36 | 0/36 | 0/36 | 1/36 |
| ZWA 11 Balls |
TI Phi | 2x 1.4 | 0.4 | 1x0.6 | -40/85 | 0/33 | 0/33 | 0/33 | 0/33 | 0/33 | 0/33 | 0/33 |
Table 3-4 summarizes conclusions from the testing. Two important conclusions are that the PCB pad size needs to match the via size, and that solder paste is needed for attachment to give optimal reliability.
| Condition | Improved BLR → | ||
|---|---|---|---|
| Die size | Larger | → | Smaller |
| Die edge | Over balls | → | Within ball matrix |
| Ball count | Smaller | → | Larger |
| Ball size | Smaller | → | Larger |
| PCB pad size | Over/undersized | → | Matches package via (for NSMD ~90% of via) |
| Solder paste | None or insufficient | → | Thickness 0.10 nom. (type matches reflow) |