PCB based system brings increased flexibility to satellite test systems
4 mins read
The proliferation of artificial satellites is quite staggering; almost 7000 satellites have been deployed to date, around half are still in orbit and something like 1000 are still active. It's no exaggeration to say that modern life wouldn't be the same without satellite technology and so it's not surprising to learn it's a growing industry.
Operating sophisticated electronic equipment in any harsh environment carries huge demands in terms of quality and reliability; satellite technology remains a specialist area for these reasons. Whilst repair missions aren't unheard of, a satellite's viability is influenced directly by its ability to operate without fault when in service — which can be anywhere up to 20 years, or longer.
Preparing a satellite for operation in space, therefore, offers no room for error, so the test and verification of a satellite prior to lift off must be thorough, exhaustive and – wherever possible – conclusive.
The challenge
Because testing is a critical part of the production cycle, factoring in the resources required is essential. The problems arise, however, when those resources begin to dominate the timescale and budget. This is increasingly the case with satellite technology; as complexity increases, so do the demands on the equipment needed to verify those systems.
The problem is compounded by the fact that, typically, every satellite is different from the power, communications and payload points of view. This means the test equipment is also bespoke.
The challenge faced by companies developing bespoke test systems, such as Siemens Convergence Creators Aerospace (CVC), is not only to achieve a solution within the budget and timescale, but also one which can test increasingly complex systems. The satellite industry is understandably conservative; once a piece of equipment is flight qualified, it won't be changed without good reason. However, the number of units now being integrated into a satellite is increasing rapidly, while redundancy merely imposes greater pressure on the test process.
The established method for implementing these bespoke test solutions is to use multiple 19in racks for each major (sub) system – the RF communications, power supply system and payload would typically have more than one dedicated rack and there may be several subsystems for each of these aspects. Furthermore, it isn't uncommon for the subsystems to be distributed geographically, meaning multiple test racks need to be constructed.
All these challenges may be surmountable, but the compounding factor is increased complexity; the number of physical connections that needs to be made between the unit under test and the test equipment continues to rise, as does the number of connections needed within the test racks. Physically, manufacturers are running out of space to implement the necessary level of interconnection using conventional cabling.
The solution
Following a period of discussion, Siemens CVC took a strategic decision to explore an alternative solution to using wire harnesses for distributing signals around test racks. The objective was not just to address the physical limitations of wire harnesses, but also to address satellite testing diversity; in effect, creating a platform that could potentially make it simpler to develop bespoke test solutions.
With demand from the European Space Agency for a satellite test solution, Siemens CVC took the decision to meet that need using PCBs as the main interconnection method. As the system would also be carrying significant levels of power, it was necessary to provide sufficient air flow for cooling purposes. For this reason, the Vienna based design team decided to construct a three sided solution using two innovative 'sideplanes', each of which would be used to route 2500 signal traces and 300 power traces (carrying up to 12A). The sideplanes would be connected using a smaller backplane and flexible PCBs. This arrangement would maintain a level of configurability whilst maximising air flow through the rack.
Because the sideplane PCBs were more than 1m long, the initial challenge was to find a supplier capable of manufacturing them; three companies were identified, with ViaSystems (based in the US, but fabricating in China) selected.
The next challenge was to find software tools capable of handling a design of this size. Modelling the system was addressed using Siemens' Solid Edge suite, whilst PCB design was undertaken using Altium Designer.
The team was confident that Altium Designer could cope with the complexity, but it presented a further challenge in the number of layers that could be supported. Initial assessments indicated the design would need 48 layers; as Altium Designer can accommodate 32 active layers and 16 negative layers, it could be achieved. However, after talking with the PCB manufacturer, it became apparent that in order to remain commercially viable the team had to reduce the number of layers to 'just' 34. Although this would further increase the design complexity, it was deemed to be an acceptable trade off in order to meet the commercial targets.
The next challenge was to route thousands of signals and hundreds of power rails whilst maintaining signal integrity. As the number of active layers had been reduced, fewer shielding layers were available, so the team had to route all signals manually in pairs using adjacent layers. Power traces, meanwhile, had to be routed perpendicularly in order to provide a level of electrical shielding.
Routing a PCB of this size at this signal density would be challenging under any conditions but, in order to accommodate both signals and power traces in the available space, it was obvious the team would have to avoid the use of vias – essentially, all traces had to be routed without changing layers.
Another key element in the success of the project was in sourcing connectors capable of offering the density needed; without them, the project would have undoubtedly failed. The connector chosen offers 18 x 10 connections and the project uses a significant number of them.
The three rigid PCB solution, with bridging using flexible PCBs, has the potential to offer significant reuse – meeting up to 80% of the test requirements for satellite systems. While predominantly passive, there is room on the backplane PCB for up to 300 resistors, allowing it to play more of a role in the overall test process. As a result, the platform offers greater flexibility, shorter development time and higher reliability than a traditional approach using conventional wire harnesses.
While PCBs have been used for interconnect in test racks in previous projects, the requirements of today's systems have now reached a point where the only viable option is to replace, where possible, wire harnesses with PCB interconnect. Even while the project in question still uses a significant amount of wiring internally, it would not have been possible without the use of PCBs.
Significantly, the PCBs were right first time and, boosted by its confidence in the technology, the team has recently completed a flexible 3D PCB design.
The PCB approach has been so well received that it has changed the design culture within Siemens CVC and the opportunities offered by flexible PCBs has encouraged other Siemens departments to consider the approach.
Alfred Fuchs is electrical ground support equipment product line manager with Siemens Convergence Creators Aerospace.