Inrush Currents Tamed – Part 2

In part 1 of my blog, we looked at the capabilities that the EDA tools provide in the area of supply network analysis. Now, we look at the different methods of power shut off control in the supply network.

Power Shut Off Control

The various methods used to control power shutoff switches each provide a way to control inrush currents and a way to indicate when a stable supply level has been achieved. Generally speaking, they approach the problem of minimizing peak inrush current by starting with a large power switch resistance (since I = V/R) and then gradually reduce the effective resistance as the difference in voltage across the power switch decreases.

Single power switch with an external timed delay – the simplest control with all switches turned on simultaneously and a simple external timer running to indicate when the transition is complete. This method can produce good results for very small domains that cannot generate enough inrush current to cause sagging and results in small implementation area and quick transition delay.

Multiple power switches with an internal delay line and an external timed delay – multiple groups of switches with each group turned on based on a built-in delay in the switches and a simple external timer running to indicate when the transition is complete. This method can produce good results for all sizes of domains and results in small implementation area and moderate transition delay.

Multiple power switches with an internal delay line and an acknowledge signal – multiple groups of switches with each group turned on based on a built-in delay in the switches and an internally generated acknowledge to indicate when the transition is complete. This method can produce good results for all sizes of domains and results in smaller implementation area (since no timer function is needed) and moderate transition delay.

Multiple power switches with externally configurable timed delays– multiple groups of switches with each group turned on from externally configured delays and a simple external timer running to indicate when the transition is complete. This method can produce good results for all sizes of domains and results in moderate implementation area and better transition delay (since the delays can be more tightly controlled).

Here is an example of a domain controlled by multiple power switches split into 4 groups. Typically, higher resistance power switches are enabled first to minimize the initial inrush current. Then, as this current settles additional switches are enabled at an increasing rate to speed up the transition time and still keep within a reasonable current limit. Finally, larger sets of switches are enabled to complete the transition and provide the final low impedance needed for normal circuit operation. The effect of lowering the impedance can be seen as the rate of the voltage change increases with each step.

Taming the Inrush

Sonics’ solution to taming the inrush currents is to provide a full featured set of control methods in our EPU products. Designers can quickly choose which control techniques to apply in each situation to best solve the problem with the minimal amount of logic.

Typically, the final switch control may not yet be known early in the development cycle, so designers can choose a simple technique that allows the development to proceed. As the development proceeds, they can tweak the software configuration to optimize the solution such that additional delays might be added to account for a larger logic domain or to reduce delays to improve responsiveness. If this is not sufficient, designers can change the control technique in the configuration quickly with little effort and impact to the design.

Check back next week for part 3 of my blog that will look at how inter-domain switching control can be applied to further tame the inrush current.

Leave a Reply

Your email address will not be published. Required fields are marked *