Embedded Vanguard Blog: Optimizing deployed system design begins with dialogue
BlogMarch 27, 2014
Jacob Sealander
Curtiss-Wright
Today, defense and aerospace system integrators are under increasing schedule and budget pressure. While funding for new programs is harder to come by, the technical requirements for deployed systems continues to increase as platforms become more digitally connected and sensor rich. One frequently overlooked strategy that can significantly aid a system integrator’s efforts to develop an optimal system solution, especially when confronted with space, weight and power (SWaP) constraints, is to engage COTS hardware vendors early in their design process.
Focused on the application space, a system integrator is typically tasked with assembling a system solution from multiple LRUs, either developed by themselves or by third parties. On the other hand, COTS vendors can provide unmatched expertise in the components and packaging approaches with which those LRUs are built. By giving the hardware vendor the opportunity to consult with the customer about their design goals and constraints early in the design process they are able to bring to bear this unique expertise in embedded electronics.Three recent design challenges provide good examples of how engaging with a COTS vendor at the front-end of the embedded subsystem design process can result in an optimized design.
In the first example a customer had a requirement for a Man-Machine Interface (MMI) processing system using an Intel processor card and a graphics card. The MMI subsystem would be used to display data received from a separate real-time computer that gathered data from a launch system and from radio-sourced information. One of the design goals was to co-locate the MMI and real-time computers in a single 6U chassis because they were sharing data. The platform had significant weight constraints and not a lot of available space. In addition there were stringent power requirements because the space and weight limitations prohibited the use of a fan or other cumbersome cooling methods. Through dialogue the COTS engineers were provided a clear understanding of the constraints that motivated the customer’s initial design choices. They were able to identify an alternative approach that effectively divided the functions of the proposed 6U-based system into two separate 3U systems, and a new location was found for the smaller 3U MMI system. The redesign resulted in improved cabling, eliminated the space problem, and freed up additional space where the earlier 6U chassis was intended to reside, enabling additional functionality to be deployed.
In a second example, a customer had a design for an LRU to be deployed in a high altitude unmanned aircraft in which four processor cards would be stacked vertically in a single chassis in an attempt to reduce the weight of multiple chassis. Weight was by far the paramount concern. In fact, the customer’s sensitivity to weight on the platform was expressed in a formula of how many thousands of dollars per ounce it would cost to fly their platform. Again, because of weight constraints, no fan could be used to cool the cards. To cool the modules using the vertical stacking approach proposed by the customer added a relatively heavy cross-section to the chassis to facilitate conduction cooling. The design’s 12 lb. weight wasn’t ideal but it enabled a module vertical stacking approach that eliminated the need for an additional chassis.
After the COTS engineers understood that weight was the prime constraint that drove the initial system design they were able to propose an alternative two chassis solution that reduced the overall weight by 30%, from the 12 lbs. required for the single vertically stacked chassis approach, to a total of 8 lbs. for two separate populated chassis.
The third example shows how good communication can result in an unexpected solution. The customer was engaging with COTS engineers on an unrelated project when discussion led to a recent frustration that the customer had experienced on a different project. For a graphics intensive application that they were ready to deploy, the customer identified a desired card set from a COTS vendor. The cards worked as expected in the lab during system development, but when it was time to package the system for deployment, successfully cooling the system was found to be quite difficult. The customer was willing to consider an alternative card set proposed by the COTS engineers, but were uncertain that unmodified COTS modules could be properly cooled sufficiently to match their particular application’s requirements.
Given the opportunity to analyze the failed system design and understand the required functionality, the COTS engineers were able to propose an alternative thermal management approach that still enabled the customer to use an unmodified COTS card solution. As a bonus, the alternative cooling approach came with no performance compromises so the system could be run at full performance in worst case environments. Using the new cards, the COTS engineers were able to come up with a better way to spread the heat of the conduction-cooled cards that eliminated a heat bottleneck at the card guide/chassis wall interface, and shortened the heat path. The new approach also involved re-orienting the modules and the use of a thermal interface material to maximize the removal of heat.
Combining a COTS vendor’s expertise with that of the customer can usually result in a system solution that optimizes the original design and eliminates the cost, weight, and schedule penalty that it might otherwise require. COTS module and integrated systems suppliers see many of the same design problems repeatedly. Because they have probably already solved the customer’s particular problem on earlier projects, the customer can then benefit from the multiplier effect.
