Global Sources
EE Times-Asia
Stay in touch with EE Times Asia
EE Times-Asia > RF/Microwave

E-band cost, reliability concerns in MMIC packaging

Posted: 15 Jul 2015 ?? ?Print Version ?Bookmark and Share

Keywords:RF systems? semiconductor packaging? E band? millimeter-wave? microstrip?

Millimeter-wave RF systems have great potential for automotive applications, if they can be made cost effective. Traditional semiconductor packaging approaches are either far too expensive or suffer from signal integrity issues at the frequencies involved. But new techniques are becoming available that can address these problems.

Whether it is highway cruising or neighbourhood trip, automobile travel would be far safer, especially at night and in bad weather, if all new cars were equipped with long-range, radar-based collision-avoidance systems. Shorter-range collision avoidance systems now in the market, typically operating at 24GHz, can activate braking and tighten seat belts before an imminent front or rear crash. But long-range systems can warn drivers well in advance that they are closing on slower vehicles or obstructions not yet in their line of sight, and can dynamically adjust the speed of cruise controls to avoid the need for sudden braking.

Such long-range systems operate in the millimeter-wave domain, specifically between 76GHz and 77GHz in the E band, to provide better object resolution and extended reach compared to the collision-avoidance radar systems that operate at 24GHz. The millimeter-wave and the short-range radar systems complement each other rather than compete, but high production costs have limited the market for long-range systems mostly to luxury cars. Engineering long-range collision-avoidance systems that would be easy to manufacture and therefore inexpensive enough for use in compact, economy carsthe automobiles that are the most vulnerable in a crashis an important responsibility.

For their part, semiconductor companies have not been remiss, with cost-saving CMOS E-band transceiver chips recently announced as near-future alternatives to SiGe devices already in production. But although economy at the die level would help, the real concern is not semiconductor processes but interconnects. How do you maintain the integrity of millimeter-wave signals going between a chip and its package, between the package and a PCB, across a PCB, and through a connector at the board edge, using inexpensive, high-yield technologies compatible with existing volume manufacturing and assembly practices? Approaches common to military and aerospace radar systems, for which cost is no object, are unfeasible. Automotive safety systems must be vastly less expensive to produce, yet must also be absolutely failsafe.

Chip packaging
Packages for millimeter-wave devices typically are constructed on ceramic substrates that route controlled-impedance microstrip or coplanar waveguide interconnects up to the mounted chip, which connect to the die by wire bonds, or flip-chip bumps. But even the shortest wire bond that can be produced acts as an inductor at such frequencies, as do bumps. Neither are impedance-controlled structures. This means that a matching network is necessary within the package to cancel the effect of each bond wire's or bump's stray inductance and thereby maintain signal integrity everywhere along the path from each die pad outward within the package. Unfortunately, the matching networks, though needed to prevent signal reflections, tend to limit the bandwidth the package can serve. Furthermore, such high-frequency packages often have a glass-sealed metal lid for hermeticity and are much more expensive than SMT plastic packages, which have not been practical for millimeter-wave applications.

Yet QFN cavity-type plastic packages are in use with monolithic microwave ICs (MMICs) that operate at 24GHz. The packages have standard JEDEC outlines and footprints compatible with pick-and-place machines and surface-mount assembly processes. The main limitation for using those packages at the higher, millimeter-wave frequencies is the bond wire. Bond wires, as noted, are not impedance controlled.

Figure 1: This X-ray photograph of a coaxial bond wire reveals the conductor, the deposited ground sheath, and the conformal insulator surrounding the wire.

1???2???3?Next Page?Last Page

Article Comments - E-band cost, reliability concerns in...
*? You can enter [0] more charecters.
*Verify code:


Visit Asia Webinars to learn about the latest in technology and get practical design tips.

Back to Top