Retronix were contacted by the R&D dept of Sandisk in Israel.
Sandisk are a huge memory device company, part of an even bigger organisation, Western Digital.
They saw an opportunity to use us to help them, and this has proven to be correct, as we have now spent 6 months being paid by them to develop systems for them :
Sandisk R&D do a number of things –
They are first in the queue to buy a competitors product. They then take it apart to find out what features it has that they can improve their systems with. As they cant get any info from their competitor, what they do is carefully remove an IC from a competitors PCB, then re-attach it with an interposer PCB between the IC and the original PCB. Then they operate the PCB as normal, but the IC performance can be monitored via connections on the interposer pcb that go to specific test equipment
Previously this was a minefield as they carried out multiple desoldering and resoldering tasks, always scared they would damage the hardware or software, leading to false readings and possibly wasting millions on going town a development route with false information
If they find a field failure that is happening regularly, they will investigate the failed units as above to identify the root cause and feed it back to design
They will mount new chips onto PCBs via interposers to trial them on new products, as part of developing better systems.
In all of the above situations, what is important?
– Cost is always important, but speed of turnaround more important as development is always in a hurry to get to market.
– We meet both of the above comfortably.
– However by far the most important aspect is accuracy of result. As has been said, if an extra reflow cycle or a rogue solder joint causes an issue, it could jeapordise completely the R&D programme.
We have assembled a series of our unique processes into assemblies for them, that do things much safer, cheaper and faster than they could previously, and they love it.
This video explains the old and new way
Below is the traditional method on the left, and the newly developed Retronix method on the right, which has many advantages –
- Original IC is removed using top heat (possibility of damage)
- IC is reballed using reflow cycle
- Grypper socket is soldered to the PCB to achieve height
- Another gripper socket is soldered to the interposer pcb, to allow the reballed BGA to be mechanically connected to avoid too many reflow cycles causing potential damage
- Interposer pcb is then mechanically connected to the first gripper socket
- The IC under test is then mechanically connected to the interposer PCB gripper socket
- Test is then carried out
The disadvantages of this process are :
- Multiple reflow cycles on IC and PCB could cause damage and give incorrect data
- IC is connected via 2 mechanical connections and 2 soldered connections, possible causing a communication issue possibly causing a data transfer issue.
- Original IC is removed using safe heat from underneath (Click Here)
- IC is reballed using laser, avoiding reflow cycle (Click Here)
- Solder spheres are attached to the interposer in a column, achieving the standoff height required (Click Here)
- IC soldered to interposer using low temperature solder, and the interposer PCB connected to the original PCB using the same method
- Test is then carried out
The advantages of this process are :
- No loss of signal due to mechanical connections, only solder connections
- No time delay or cost waiting on matching gripper sockets, solder ball columns instantly created
- No dangerous reflow cycles to cause damage