EEE Links -- Vol. 1, No. 4 -- October 1995
by Harry C. Shaw
Packaging and Processes Branch
Goddard Space Flight Center
301-286-7293
harry.shaw@gsfc.nasa.gov
The need to reduce the costs and development time for complex, spaceflight reliable MCMs
has been apparent for some time. Code 312 and Pico Systems of Ann
Arbor, MI are working together to address this need through the qualification of the Pico
Systems dielectric-antifuse based programmable silicon substrate. This is
the first NASA evaluation of the technology for spaceflight usage. The enabling element in
these substrates is the antifuse, which consists of a dielectric sandwiched
between two metal electrodes. The application of a suitable programming pulse to the
electrodes causes an irreversible transition from a non-conducting amorphous
state to a metal alloy. Having an elaborate system of conductive traces and interspersed
antifuses permits circuit routing by programming selected antifuses. This
defines a standard configurable substrate. The technology has been proven in both
commercial and military applications. MCMs at clock speeds of 200 MHz have
been successfully fabricated. We have implemented a typical application, in this case, a
dual low-noise preamplifier developed by Jeff DuMonthier of Code 663 for
a Calorimeter Analog Processor circuit on the XDS instrument.
The Code 312 evaluation is accomplishing three objectives:
1.Assessment of the electrical safe area of operation of the programmable substrate. This
is necessary for long-term reliable operation.
2.Establish any functional differences between circuits developed on the programmable
substrate versus development on traditional MCM substrate
technologies. This is necessary to ensure cross-compatibility between technologies.
3.Establish actual cost and schedule metrics and compare those to the previous GSFC
experiences. Objective 1 is being pursued by use of a novel feature of
the programmable substrate which allows "strings" of antifuses to be programmed
in series, creating a customizable load consisting of programmed antifuses
and Al/Cu metallization runs. The internal substrate temperature is extracted from the
precise temperature-resistance coefficient of the Al/Cu metallization.
Objective 2 is being pursued through evaluation of a digital and an analog MCM circuits.
The digital circuit, a dual 32K x 8 SRAM module is being
characterized for reliability, performance, and workmanship. Steady state, high
temperature bias conditions at 125 C and 85 C to accelerate circuit stresses
are being utilized.
Functional issues such as simultaneous switching performance and performance degradation
are being investigated. Both digital and analog circuits have been
fabricated as dual circuits on a single programmable substrate to allow failed structures
to be compared to operational structures.
The Pico Systems technology has proved cost effective. The analog circuit proved to be
approximately 25% of the cost of the same circuit fabricated as a traditional
thick-film hybrid. Deliveries of digital circuits began 3 weeks after the netlist was
delivered from Code 312 to Pico. It is anticipated that the same economies will be
achieved with spaceflight production MCMs.
The internal architecture of the PSCB invites a number of possibilities for exploration.
One of the most intriguing possibilities involves the substrate as a circuit
element. Usage of both programmed and unprogrammed antifuses as lumped elements in
microwave circuits along with sections of signal lines could provide for
some novel circuit implementations. It may prove possible to perform rapid prototyping of
certain classes of microwave hybrid circuits. Future investigations and RF
simulations are planned to explore the limits of the technology for development of RF
microcircuits, stacked-die technologies and mixed signal MCMs.
Mr. Harry Shaw is the Processes and Packaging Branch (Code 312) Engineer for
microelectronics and microwave devices at GSFC.
Home
Last Revised: January 09, 2002
Digital Engineering Institute
Web Grunt: Richard Katz