We are using STM32F765IIK6 176 pin 10x10 0.6mm UFBGA, die rev "Z". Whilst processing these through SMT solder reflow (using an 8 zone forced air convection oven) we had a number of these devices failing, due to the substrate of the BGA package dishing causing solder shorts & opens under the device.
We have followed the J standard pre bake & storage process for these devices (MSL 3) and have profiled the board and device, so as not to exceed the upper reflow profile, as stated at 260°C on the Materials Declaration Form for the device. The actual device peak temperature during soldering was 255°C with a ramp rate of less than 3°C / Second.
We are using a Halogen Free Ultra Life Solder Paste SAC 305 (Sn96.5/Ag3.0/Cu0.5).
We have tried different date code STM32F765IIK6 devices through the process and some are more suspectable to failing than others. All have been pre baked and procesed in the same way.
We found that lowering the peak reflow temperature reduced the dishing affect and the number of failures, but this created problems with other devices on the board not correctly soldering (i.e. the NiPdAu plated parts).
Please can you ;
i: Clarify the reflow profile limits for this device and package substrate - i.e. is the Material Declaration Form parameters correct.
ii: Whether you have other reported failures for this device / package type - and if so any corrective action to resolve issue.
Thanks
Martin
