Dual AC Input Power Supply Redundancy Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing information handling systems face inefficiencies and high costs due to the requirement for excessive power units in fully AC redundant power supplies, which are oversized for DC redundancy and wasteful in non-fault conditions, especially in large server systems where power requirements exceed standard connector limits.
Innovation Solution
A dual AC input power supply with each power unit featuring two AC front ends and a single DC back end, optimizing AC and DC sections to provide cost-effective redundancy by using only N+1 power units for both AC and DC redundancy, with each AC front end capable of 2000 watts and each DC back end capable of 3000 watts, allowing for efficient power distribution and utilization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If N+N power units are used for full AC redundancy, then AC reliability is improved, but device complexity and cost increase significantly
Solution Approach 1:
The power supply system is segmented into two functional sections within each power unit: an AC front end and a DC back end. This segmentation allows the AC section to handle AC input redundancy while the DC section handles DC output distribution, enabling N+1 power units to provide both AC and DC redundancy instead of requiring N+N units. Each power unit's AC front end can independently process AC input from different sources, while the DC back end distributes power to multiple rails.
Solution Approach 2:
Each power unit is designed with multi-functionality to serve dual purposes: the AC front end can accept AC input from either of two different AC sources and convert it to DC, while the DC back end can distribute power to multiple DC rails. This universal design allows any power unit to potentially take over for another unit in case of failure, providing system-wide redundancy with fewer units than traditional dedicated redundant architectures.
2Reliability
If N+N power units are configured for AC redundancy, then AC power reliability is improved, but power capacity utilization deteriorates as half the power remains unused in normal conditions
Solution Approach 1:
The power supply system implements dynamic load sharing and automatic failover mechanisms where power units can dynamically adjust their operation based on system conditions. In normal operation, N power units actively supply power while 1 unit stands by as hot spare. Upon detection of a failure (AC input loss or unit malfunction), the system dynamically reconfigures to activate the standby unit, ensuring continuous power supply without requiring permanent allocation of excessive power capacity.
Solution Approach 2:
The system changes operational parameters based on system state: during normal operation, power units operate at optimized efficiency points; during fault conditions, parameters such as load distribution, voltage regulation, and current allocation are dynamically adjusted to maintain power delivery while accommodating the reduced number of active units. This parameter adaptation allows the system to maintain reliability without permanently provisioning for maximum fault scenario capacity.
3Adaptability or versatility
If standard IEC connector limit of 2500 watts is used per connector, then adapter compatibility is maintained, but power delivery capability deteriorates for large systems requiring over 5000 watts
Solution Approach 1:
The DC back end of each power unit is segmented into multiple independent DC output sections, each capable of delivering substantial power through standard IEC connectors. For example, a single power unit can provide multiple DC rails (such as +12V, +5V, +3.3V, and other specialized rails), with each rail capable of delivering high current through its own IEC connector. This segmentation allows the system to aggregate power across multiple connectors and rails to achieve total system power delivery exceeding 5000 watts while maintaining compatibility with standard connectors at each interface point.
Data Source
AI summary
A cost optimized redundant power supply consists of a plurality of power units (PUs) wherein each PU has two AC front ends. Each of the two AC front ends receive power from separate AC power sources, and each produce an isolated DC output. The isolated DC outputs from the two AC front ends are coupled to an input of one DC back end, e.g., DC-to-DC converter. Optimally, each of the AC front ends may have a power capacity of TRP/(N+1) and the DC back end may have a power capacity of TRP/N, where TRP=Total Redundant Power or the maximum system power required for operation and N is the number of PUs of the power supply.