Darryl McKenney, Vice President of Engineering Services at Mercury Systems, talks to CIE about the challenges associated with the effective handling of thermal management. As he explains the challenge is how to do more with less and how to deal with reduced power requirements, in more extreme environments, in less time, with less weight, and at less risk.
CIE: What are the latest trends in thermal management in terms of the materials used, as well as the product and system developments that you are seeing in the market?
Darryl McKenney: The latest trend in thermal management addresses the challenge of how to do more with less: how to deal with reduced power requirements, in more extreme environments, in less time, with less weight, and at less risk. Cost, Size, Weight and Power [C-SWaP] analysis, performed at a subsystem level, is driving the trade-offs between module capabilities and chassis capabilities for subsystem optimisation. In order to support these trade-offs, it is important to always start subsystem development work from a baseline of products that comply to open standards, this significantly reduces cost and lead-time. If a solution cannot be made using only standards-based products, trade-off studies can help identify the areas where the solution may be tweaked and new materials introduced that can reduce weight, improve cooling capacity and rigidity, or address other specific system requirements. Creating solutions using standards-based products that include technology insertion options for potential upgrades delivers the optimal solution for the programme, while reducing development time and risk.
CIE: Air-cooled, conduction-cooled, air flow-by, or jet-spray cooling? What are the benefits/drawbacks of each and what technology is Mercury specifically focusing on?
DM: In order to meet the wide range of our customers’ needs, we offer standard products with an agnostic approach to cooling. We can package the same PCBA module in air-cooled, conduction-cooled, or Air Flow-By configurations. This flexible design approach to our module cooling allows us to maximise our product development velocity and allow the modules to be leveraged across different subsystem mission environments. Since there are pros and cons with each type of cooling method, it is important to have the flexibility to offer our customers a solution that meets their specific needs.
Air-cooling provides easy access to module debug connectors, front panel I/O and mezzanine modules. This combination simplifies system development and configurability while the system is in its greatest state of flux and requirements are not all identified. A major drawback is that air-cooled modules are not typically designed to be deployed in rugged environments.
Conduction-cooling has been the preferred method of cooling for deployed systems for many years. The modules are designed to handle the rugged shock and vibration levels, while the systems seal the modules away from harmful elements. A major challenge with conduction cooling is that it is heavier than air-cooled and thermally challenged with higher power modules.
Air Flow-By – a new cooling technology designed by Mercury – delivers the best of both worlds. It provides the efficient point source cooling of an air-cooled module with the rugged deployment capabilities of conduction-cooling. From a C-SWaP perspective, we believe Air Flow-By is superior to alternative solutions as we execute risk and reward analysis at a sub-system level.
CIE: How is thermal technology changing? What impact are new materials, e.g. graphene, having on thermal management techniques?
DM: Today”s thermal technologies are designed to enable the use of the latest high performance components in modules in deployed environments. Since the mission profiles for each system are diverse, it is important to have a design that is flexible enough to meet varied program needs and not break the calendar or the bank. Every available technology has different implications for C-SWaP impact. This makes it essential to optimise the solution to address a specific problem and create new leverage points for technology insertions.
· Aluminum is still the dominant solution as it is moderately conductive, inexpensive, and widely available.
· As a rule, copper provides two times the conductivity of aluminium, at three times the weight and four times the cost. But copper is still widely available and an effective solution for increasing thermal conductivity when there are no concerns about paying a thermal and cost penalty.
· Heat pipes are highly conductive and have about the same density of aluminium. However, they have a finite energy carrying capability and once they become saturated they are no longer highly conductive. Heat pipes are widely available, but there are potential long term reliability concerns and geometry limitations.
· Vapor chambers have a higher thermal capacity than traditional heat pipes, but can suffer from “pillowing” at higher temperatures, as well as orientation and acceleration effects (Pillowing is the enlargement of the heat chamber and distorts the cooling cavity).
· Pyrolytic graphite is light weight and highly conductive in two dimensions, but not very conductive in the third dimension. The material is very brittle and requires an outer shell to be used successfully in a wide range of environments. The materials for this are expensive and the additional processing steps add to material lead-times. These are used where weight is a primary variable and can override the cost impact.
CIE: In military and aerospace applications there are no shortage of challenges for thermal engineers. What are the key problems you have to address?
DM: Heat is the primary enemy of module reliability. A major advantage of our newer cooling solutions is double digit processor temperature reductions. This has an amazing positive impact to reliability. We are seeing reductions of greater than 10° C on our high performance processors and over 20°C on extreme performance GPUs.
The urgency to quickly develop solutions and maximise design reuse drives leveragability requirements being imposed on programmes. These factors reduce the likelihood of seeing multiple design requests for similar but separate sea-based, land-based or aerospace platforms. Today, one set of requirements has to satisfy multiple deployed scenarios: light weight and low power for aerospace systems, rugged and abrasion-resistant for land-based systems, and highly corrosion resistant for sea-based systems. At first glance, this seems to triple the design complexity for mechanical and thermal engineers. Mercury recently developed Air Flow-By which embraces the light weight and point source cooling of air-cooled designs and the rugged, sealed, corrosion resistant capabilities of conduction cooling and can efficiently cool today’s high powered electronics in multiple rugged deployed applications. Our current Air Flow-By technology was developed to cool about 200+ watts in one 6U form factor slot at 1” of pitch. As we continue to push the high power processor envelope we are moving toward the utilisation of liquid cooling in our modules as well. Next year, we hope to introduce newer module cooling solutions that will allow us to cool 300 Watts per 6U slot in a 1” pitch.
CIE: How do you overcome packaging constraints?
DM: The first step is to properly define the design goals of the product or subsystem, what changed from previously successful products and how to bridge the gap between what has worked in the past and the needs for the current product. It all comes down to identifying and attacking the primary variables, tracking the secondary variables and maintaining compliance to the tertiary variables. For example, if a processor’s power consumption is increasing from 25 watts to 45 watts, is this a heatsink cooling capacity problem, or a thermal spreading problem, or both? While they sound similar and can affect each other, there are different techniques for improving each one with different impacts to the solution as a whole. It’s basic: identify the primary variables, bring it back to the “math”, create a solution, and move onto the next challenge.
CIE: How do you test and measure thermal management? How are modelling techniques changing and what impact are they having on the design process?
DM: Different companies approach the test and measurement of their products’ thermal capabilities in different ways. Mercury believes in the utilisation of CFD Analysis (Computational Fluid Dynamics) before we build anything. Our modelling approach has a direct correlation and allows us to test products with an expectation of passing qualifications the first time. This design for reliability approach allows us to make reductions in the product costs and improvements in reliability during the design process. We also use application-specific software to test the products to extreme environmental conditions including testing at all four corner case conditions (high temperature, low temperature, high voltage, and low voltage) electrically.
Thermal modelling, signal integrity and component placement are “negotiated” very early in the design process. This allows for technical risk reduction and tradeoffs in electrical, mechanical, and reliability analysis to be in the design process to optimise the overall product design, while minimising rework as the design cycle matures.
CIE: Looking to the future how do you expect thermal management to develop?
DM: The future of electronic component cooling is very hot! As subsystem developers employ new methods, you will see new advances in the mass transfer of thermal energy. Higher efficiency heat pipes, new materials, hybrid exchangers and component miniaturisation will all have significant impacts as we head into higher power “thermally managed opportunities.”
Finally, in the near future, there will be a definite need for a scalable “Swiss Army Knife” solution for today’s demanding systems; in other words, one solution for use in multiple platforms. The future of thermal management techniques must address not only how to cool more power, but how to cool more, while taking up less space, adding less weight, and functioning successfully across a wider range of environments.