Advanced power management: The next step in DSP board design

Filling an information gap when it comes to power use

15Rob explains the need for advanced power management in DSP boards to help avoid costly overdesign.

Embedded engines for military applications are typically limited by the amount of power they are able to dissipate or by how much power the chassis is able to absorb. Their on-board processors, often two or four on a single board, and often multiple boards in a single chassis and deployed in a space, weight and power constrained environment, can test the limits of a system’s ability to manage power – both the power required to drive the and the power dissipated by them. In the world of airborne , having a good handle on power can be critical for helping to keep system weight down (Figure 1). It’s no exaggeration to say that half of the design challenge today is about power. This challenge can be eased by a new design approach that adds capabilities, and temperature monitoring, to DSP engines, enabling the measurement of power dissipation on each card.

21
Figure 1: ’s CHAMP-AV6, a quad processor is an example of a multiprocessor engine designed for demanding military applications.
(Click graphic to zoom by 1.9x)


The problem of managing high power electronics on signal processing cards has become increasingly critical. In the past, say 15 years ago, a customer might ask, “what's the maximum power of the card?” and a board vendor would reply with a single “maximum” power number, essentially a “worst case scenario” number. The system integrator would then drop this value into their spreadsheet to drive the design of their enclosure, power supply, cooling requirements, and the like. In the past, when the maximum power was lower, such as 12 W, a variance in one direction or the other, was not that crucial. But today, with the power dissipation of DSP cards frequently measuring in triple digits of watts it has become necessary for system integrators to become much more knowledgeable and precise about how much power is being used and when. The risks of not having accurate measurement include systems that don’t work or, more commonly, the costly, in weight, size and dollars, overdesign of . Exacerbating this problem is the wide variance of power that a DSP card can require, which can be caused both by what the card is tasked with doing and the operational temperature of the system.

Because the actual power usage of a card may not be known until it is in the development stage, it becomes difficult to provide customers with the most useful information that they need for their design (Figure 2). Because there is no “power meter” on a card today, measuring power on an individual card can pose a great challenge. It’s not possible today when a card is plugged into a backplane to easily provide the answer to the question “how much power is it using?” Customers may have the ability to measure their entire system’s power profile, but to isolate the power that a given card is using is near impossible. Even testing the operation of a single card is complicated because the customer has to buy an instrumented backplane and create a special test configuration of card and/or software.

22
Figure 2: A representative thermal plot showing the power dissipation on a multiprocessor DSP card.
(Click graphic to zoom by 1.9x)


Knowing the actual power usage of a card can provide significant benefits to a system designer. It is very common that DSP cards are used in a less demanding way than that which they are tested for. The vendor, though, is obliged to test that card for power consumption using an approach that represents the 95th percentile of power of potential usage. Today’s “maximum power” number, for example 120W, may represent a highly unlikely, but not impossible, usage scenario. Vendors provide system designers with a power specification that represents a demanding use of the card that the large majority of users may never achieve. In fact, some subset of customers will use a given card in significantly less demanding ways. When a customer buys a specific card they have no way of knowing if they will be one of those 20 percent or so who use much less power unless they have the ability to measure the usage, which as we’ve discussed is very difficult today.

When the only number that the system designer has available is the “worst case” number, the consequence can be costly system level overdesign. For example, when you add 10 W to the board you have added 10 W to the power supply and unnecessarily higher capacity heat exchangers and fans, which can significantly burden the final system. Every additional watt in a piece of embedded electronics has a cost, incurring bigger, heavier systems to deliver it and then dissipate it.

If the board can inform the system designer how much power it is using it becomes much easier for the designer to know precisely how much power their application is using. Onboard card advanced is the missing element. The advantages are multifold because it will provide vendors with the data that enables them to know if they have additional “headroom” when designing cards that will require more power. And vendors will have the benefit of knowing that the customer will more fully understand their own power requirements. With power management, customer is empowered to ascertain actual usage numbers in lab environment and design system to match that reality. Today, if a card is used in a less demanding fashion than anticipated, the system designer won’t be able to take that fact into account until after the system is designed, when it is too late to be useful information.

In the near future, to address this issue, new DSP cards products will have greater capabilities to measure their own power. The power monitor will give users much more understanding about what is happening in their system. Precise understanding of power and temperature is required to determine what the overall system will need for cooling and powering. This will help put an end to design for worst-case scenario and to costly overdesign.

Robert Hoyecki is Director of Advanced Multi-Computing at Curtiss-Wright Controls Embedded Computing. Rob has 15 years of experience in embedded computing with a focus on signal process products. He has held numerous leadership positions such as application engineering manager and product marketing manager. Rob earned a Bachelor of Science degree in Electrical Engineering Technology from Rochester Institute of Technology.

Rob can be reached at info@cwcembedded.com.