Power Struggle — Getting the Most Out of Your PSU through Enhanced Design Solutions
It is widely accepted throughout the industry that whenever a system design project is embarked upon the power supply unit (PSU) needed to run everything will not have its parameters finalized until the later stages of the design and development process. In some cases changes will be made throughout, in others this aspect simply won’t even be looked at until the last minute. The result is that often there are performance and space constraint issues that need to be addressed.
It is widely accepted throughout the industry that whenever a system design project is embarked upon the power supply unit (PSU) needed to run everything will not have its parameters finalized until the later stages of the design and development process. In some cases changes will be made throughout, in others this aspect simply won’t even be looked at until the last minute. The result is that often there are performance and space constraint issues that need to be addressed.
In such designs, it will of course be highly desirable for the PSU’s conversion efficiency levels under normal operating conditions to be as high as possible, so that energy consumption is minimised. Also, as excessive space may well have already been used elsewhere in the design, it is likely that the PSU dimensions will need to be curtailed as much as possible. By keeping its switching speed high, the size of the magnetics utilised in the PSU can be reduced. The conversion efficiency has a role to play here too – as the higher this is, the lower the level of unwanted heat generated by the supply will be. This in turn this will allow the amount of heat sinking around the PSU to be lowered and thus lessen the overall size of the implementation.
The specification of standard off-the-shelf PSU products will bring drawbacks in terms of the performance that can be achieved, as well as the form factor into which it can be enclosed. Firstly the voltages will not be optimised for the application in question and poorer conversion efficiencies will result. Secondly, the overall space allocated to the supply will be greater due this lack of optimisation – as an extensive thermal management mechanism, plus cumbersome magnetics and inclusion of other discrete components, will take up valuable space. Finally, the larger quantity of heat being generated, because of the poor efficiency levels realized, will also raise questions about the operational longevity of the system.
The design and development of a fully customised solution, despite the fact that they can satisfy the performance benchmarks needed, may prove to be a gamble however, with both engineering and financial risks involved. If badly judged then major penalties can result in terms of windows of opportunity being missed or end products’ unit costs being too high. It is incredibly difficult to accurately forecast what the shipments of a new product will be prior to its introduction, but this factor will have a very large influence on the development of the power system. Engineering managers cannot afford to be daring here as the costs associated with these decisions will prove crucial to the end products’ commercial viability.
If volume shipments are absolutely assured, then opting for a full custom approach is clearly valid. Conversely if they aren’t then the revenue generated through product shipments might not be adequate to justify the money that has been fronted up. The increased complexity of a fully custom design also adds an extra element of risk to the project. This could mean that the introduction of the end product could be pushed back while reworking of the power design is undertaken. In the meantime, a rival’s product may be released and steal a large chunk of market share.
In contrast, taking a semi-custom approach allows the realisation of solutions that can serve performance-demanding, low volume markets, as well as testing markets where the opportunity for larger volume, full custom design could be a possibility in the future. This trend mirrors what has taken place in the semiconductor industry, where in many industries application specific chips have in recent years been superseded by more flexible FPGA solutions which do not require heavy NRE costs.
Through the utilisation of ready-made high density building blocks, based on innovative, industry-proven brick technology, commercially aware power solutions providers will be able to meet the performance requirements set by their clients without the need for a heavy initial outlay of money or long winded development. Since each of these blocks already has the necessary approvals (such as UL or CSA, and CE for Europe), the end product’s approval process will be easier and less costly to complete than would be the case for a full custom – where approval would basically need to be done from the very beginning. Working with a company that has an engineering team experienced in dealing with the mechanical and thermal management issues relating to PSU development will allow a considerable amount of space to be saved, as well as augmenting overall system reliability.
A semi-custom approach enables the creation of highly-optimised power solutions for applications where off-the-shelf power solutions simply don’t have the necessary specs, but where the volumes involved may not be enough to warrant the commitment needed for a fully customised power supply. It can markedly cut the upfront capital investment needed and reduce the length of the development period that would be involved if a full custom option was taken. Furthermore, it means that the project will not be left exposed to any technical problems arising that could impact on its success.
In conclusion, the nature of PSU implementation needs to change dramatically if the engineering community is going to be in a position to meet the exacting pressures now being placed upon it. End product manufacturers are increasingly reluctant to pledge their limited financial resources onto expensive development activities. It is clear that there are now many technical and economic arguments which are making semi-custom power solutions more attractive to the electronics industry. A broader cross section of applications are being shown to be far better suited to this approach. Through it markets can be addressed where there is still some uncertainty on how much business is likely to be generated, but a migration path for embarking on a full custom design is still available once serious volume demand has been secured.
