Maximising the value in modern power system designs

Rob Hill, Powerstax


The issue of the power source is nearly always left until the last minute

Click image to enlarge

Figure 1: CAD drawing & photo of completed semi-custom design for a high-rel radar implementation within a space constrained enclosure

In what can be an increasingly cutthroat business environment, the engineers of original equipment manufacturers (OEMs) are under huge pressure to deliver products that will differentiate themselves from those offered by rival operations. Often they need to do this within tight deadlines, so that windows of opportunity are not missed. Unfortunately, despite the fact that it is vital to the system design, and the success of the product therefore hinges upon it, the issue of the power source is nearly always left until the last minute. In order to augment the effectiveness of the product, the conversion efficiency rates of the power source must be raised. Switching speeds also need to be increased to minimise the size of the magnetics so that the space into which the source will fit can be squeezed as much as possible. The utilisation of standard off-the-shelf power supplies is unlikely to be particularly attractive in such circumstances due to their generic nature. Though these items are easily accessible and reasonably priced, they will bring performance penalties, as voltages will not be optimised, resulting in lower degrees of conversion efficiency for a given application. This will not only increase the overall power budget, but will also mean there is a greater quantity of heat to be dissipated - potentially shortening the operational lifespan of the product. Though the implementation of a fully customised power solution will provide OEM engineering teams with exactly what they need performance-wise for their new products, and do so at the optimum price points for guaranteed high unit volume projects, this is rarely the situation that we see in the real world. Normally, any prediction on the actual quantities to expect will not be defined that well - in fact it is more likely that it will be nothing more than the result of optimistic guesswork conjured up by the marketing department. Though the commercial element of an OEM can, to some extent, afford to be bold about the predictions it makes, the engineering department cannot take that gamble. They need to base their judgment on something that is far more tangible, as key decisions about how the design and manufacture is undertaken will be made on this - with the costs associated with these decisions proving crucial to the end products commercial viability. If volume shipments are assured, then opting for a full custom approach is clearly valid. For example, when producing consumer devices like flat panel TVs, the volumes are expected to be high from the start, so every dimension of the development and manufacturing is geared towards highly optimised custom designed component parts that will augment system performance and lower the end product's unit cost - normally through large upfront investment. However, in most cases OEMs really need to test the market, before committing to this - as the heavy non-recurring engineering (NRE) costs initially involved will only be justified if the volume targets are met. Otherwise the revenue generated from sales will simply not be enough to cover the capital that has been ploughed into the project. As well as the cost considerations there are other drawbacks that should be taken into account before embarking on creation of a full custom power system. Often these systems will rely on unproven technology. This heightens the risk involved and can lead to delays in the end product's release plus unforeseen costs being accrued, as reworking of the design is done. The Middle Way Not wishing to be bound to potentially crippling NRE costs and keen to avoid being left exposed to the possibility that redesign work may need to be carried out, OEMs are now looking for another way of getting the power source they want. A way that combines the plus points of the custom and off-the-shelf approaches previously detailed. Following a semi-custom strategy presents engineers with a method by which to gain marked performance advantages over standard power sources, while simultaneously mitigating both the upfront costs and the risks that are characterised by full custom power designs. Though they have to accept that there may be some downstream costs or reduction in profit margins if large quantities are actually achieved, this is far less likely to impact on their business than the full custom alternative would if volumes failed to materialise. Furthermore, following strategy still leaves a migration route open once the market demand necessary to support a full custom design has been proven to exist. Because the power system is generally, as already pointed out, the last thing that gets attention in the product development process, the time pressures here are at their most pronounced. By basing the power system design on the use of ready-made, high density building blocks through which a semi-custom solution can be created time can be saved. Each block will already have all the necessary approvals (CE, UL, etc.), so the approval process for the end product will be quicker and cheaper to carry out than a full custom, where approval would effectively need to be done from scratch and often proves to be a long drawn out affair. (See Figure 1.) As everything else within the design will have already been taken care of, the space that is left for it may well be difficult to utilise. This is why engaging with a company which is highly experienced in both the mechanical and thermal management aspects of power system design will pay dividends. In conclusion, working with a power supply vendor on a full custom solution that the desired performance targets and low units costs can be reached is not as easy as it first appears. In many cases the time and initial expense involved will outweigh the actual benefits, especially if there is any doubt about the volumes that are likely to be called for. With cash-strapped OEMs looking to avoid heavy capital investment, semi-custom solutions are far better suited to addressing markets where there is still some uncertainty on how much business is likely to be generated. By utilising a modular approach, based on innovative, industry-proven brick technology and power engineering expertise, certain power source suppliers can now enable performance levels approaching those associated with custom designs to be realised without the hazards of large upfront expense, a lengthy development period or technical problems arising. Time-to-market can be shortened and the dent in the OEM's financial resources lowered considerably. Powerstax