|
Downtime in today's on-demand environment is not only unwelcome—it is costly. If applications are not consistently available, business suffers. If a web site is not live when a customer needs it they are likely to look elsewhere in the time it takes to click a mouse. Downtime can cost between thousands to multiple millions of dollars per hour. Damage can extend well beyond financial into key areas of customer loyalty, market competitiveness and regulatory compliance.
Businesses need to keep applications up and running in the event of planned or unplanned system disruptions. "High availability" is an attribute of a system that provides service at agreed-upon levels and masks unplanned outages from end users. "Redundancy" allows work to shift from failing components to those that prevent the end-user service from being interrupted. Failed components may be removed and replaced while the system is still active, so service may continue.
Hot swappable N + 1 redundant power is the preferred approach for systems requiring high reliability, availability, and serviceability. Hot Swappable N+1 redundant AC-DC and DC-DC converters using transformer based power topologies have been available for many years to provide board level bulk power in Server, Telco and Netcom systems. In these converters the transformer provides isolation of the input voltage from the output and ORing diodes provide the hot swap functionality.
Transformer-based topologies are no longer a viable solution for directly powering digital ICs due to decreasing digital ICs operating voltage, related increases in current, and the proliferation of the number of required power rails. Power loss, size, and cost are all unacceptably large compared to a non-isolated PoL (point-of-load) solution.
Until now, POLs have been unable to provide N + 1 redundant power with the required input to output isolation. This has forced system OEMs to implement redundant power at the board or rack level rather than for every output rail. Today's solution requires an entire board or rack to be replaced if the PoL fails along with the loss of any customer data being processed by the board or rack at the time of failure.
Redundant PoL
To address this problem, International Rectifier developed an N+1 redundant PoL power system. As shown in Figure 1 this solution implements a synchronous buck converter combined with input FET(s) that provide hot-swap and E-fuse functions and ORing FET(s) to ensure complete system protection against failures such as high or low side FET short circuit.
Average current mode control is used to implement active droop sharing between converters. This eliminates the need for any control wires connecting between two or more current sharing PoLs and prevents a potential single-point-failure mode from disabling the system. An analog current monitor signal is provided and allows the system microcontroller to monitor the current being delivered by each PoL. A digital interface is also included to transmit data back to the host system indicating status of the PoL's output voltage, input FETs, and ORing FETs.
The topology is fully scalable to support output currents above 300A by adding the required number of phases. The IR3510 provides overall system control and interfaces with any number of IR3088A Phase ICs, each of which drives and monitors a single phase Sync Buck power channel.
Figure 1 " N+1 Redundant PoL Block Diagram
|
