ICT Symposium - "Microwave Technology"
Arundel, 28th September 2005.
Forty delegates attended the ICT Evening Symposium in Arundel on 28 th September 2005. The theme was Microwave Technology, and four papers were presented, one from an OEM, one from an EMS provider and two from laminate suppliers.
Alec Groves of BAE Systems described some of the many factors to be considered by a supplier of military radar systems in designing and manufacturing microwave circuit boards. The objective being to gather and manage battlespace information, a practical illustration of the precision of state-of-the-art radar was its ability “to track a beer can at a distance of 200 kilometres”.
Microwave engineers viewed copper circuit features primarily as components rather than as conductors, and Groves explained the various functions of some typical design elements. Materials were required which combined consistent dielectric properties with very low loss characteristics at GigaHertz frequencies. A typical material appropriate to BAE’s microwave designs was a glass micro-fibre reinforced PTFE with dielectric constant 2.33 and loss tangent of .0012 at 10GHz, excellent electrical characteristics being achieved at the expense of relatively poor mechanical properties. The tolerances of the manufacturing process were intrinsic to the design – small variations could have major impact on performance. As high-speed digital switching rates were approaching microwave frequencies, the digital design world was drawing increasingly upon microwave design expertise.
Groves described the design process sequence: establishing the variables, simulating performance, building a test circuit, testing it, then going back and tweaking it. Accuracy of track-width measurement was critical, but open to interpretation depending on side-wall profile, and surface finish had a significant effect on how signals were transmitted. Similarly, the geometry of the interface between circuit features and discrete components was critical – component leads could act as antennae – and one technique for minimising lead effects was to mount the components in cavities cut into the board surface. Even the volume of solder joints had significant influence, and these joints had to be sufficiently robust to sustain in some examples a 40-year working life.
Value engineering in microwave design balanced the cost of finding the answer with the value of the eventual answer, with the emphasis on establishing what was good enough whilst avoiding adding cost by over-specification.
David Woodley of Technograph Microcircuits gave an insight into “thinking inside the box”, with a fascinating overview of the specialist contract manufacturing services provided by Technograph. From a huge range of designs, he selected as his example a hermetic microelectronic assembly operating at 10 gigabits per second, driving a modulator for an optical telecommunications application, capable of handling 150,000 simultaneous telephone calls. This tiny unit represented a range of technologies including thin film and thick film hybrids, a copper-tungsten base and an aluminium package with hermetic feed-throughs, gallium arsenide MMICs, piggy-back surface mount assembly, some soldered joints and many wire bonds. The component value of each unit was over £300, and the design had been productionised on a very short lead-time to a peak of thousands of units per week. Significant cost reduction had been achieved without compromising high frequency performance. In Woodley’s words, “some of the precision assembly involved in these products makes brain surgery look like gardening”.
Bob Nichols of Neltec emphasised that it was the designer’s responsibility to consider price/performance value rather than direct material cost in choosing appropriate substrates for radio-frequency applications, although the laminate supplier was always willing to offer assistance in achieving system specification and cost-effective performance targets.
Nichols described a variety of different high frequency printed circuit types and constructions and the range of available materials options. Recent market research had indicated that the principal applications for low-loss materials were base-station antennae, base-station amplifiers and LNBs (Low Noise Block-downconvertors, the units which collect satellite TV signals from the parabolic reflector and convert them to a lower frequency). Substantial growth was predicted for the use of PTFE materials in microwave circuits, particularly in base-station antennae.
But PTFE was not the only option for high frequency applications and, for example, a controlled-dielectric epoxy resin system was available with dielectric constant of 3.8 and loss tangent of .007 at 10GHz, which easily satisfied the requirements of lower-power circuits operating in the 1.5 to 2.5 GHz range, illustrating the importance of specifying the material to the required frequency range rather than over-specifying and paying an unnecessary cost premium.
Developments in flame retardant technology had overcome the limitations of earlier attempts to produce low-loss bromine-free materials in which the bromine alternatives had downgraded dielectric properties. Bromine-free laminates were now available, compatible with lead-free assembly conditions, with dielectric constant 4.3.
Jim Francy of Taconic introduced a new low-loss laser-ablatable microwave substrate material which had evolved from a Eureka-funded project to develop materials and packaging solutions for millimetre-wave technology. The strategic significance of this material was that it opened-up an opportunity to the PCB industry which would otherwise have been lost to ceramic alternatives. The new material was a pre-bonded metal-back PTFE-ceramic laminate, where the metal back could be copper, brass, or CTE-matched copper-molybdenum or aluminium-silicon. A PTFE-ceramic bond-ply was also available for building multilayered structures. The isotropic dielectric could be CO2 laser-ablated, to form micro-via interconnects, thermal vias, or cavities into which components could be mounted to minimise lead-out lengths and to enable good thermal transfer to the metal back. Thin copper foil was available, to allow close etching tolerances to be achieved.
Francy explained the process route used for fabricating a 3-layer stripline demonstrator with 100 micron dielectrics and 150 micron laser-vias from layers 1-2, 2-3 and 1-3. New applications were progressively being developed and some thin-film designs had already been transferred to the new material.
Thanking speakers and delegates for their support, and acknowledging the generous sponsorship of Artetch Circuits, ICT Technical Director Bill Wilkie declared the symposium a great success, not least that it had encouraged several new members to join the Institute.
Pete Starkey
ICT Council
September 2005