Search by Part #:    ADVANCED 
SEARCH

Home

AC/DC

DC/DC

Other

Applications

Manufacturers

Home  |  About Us  |  Resources  |  Terms  |  Contact Us
Search by Part #:    ADVANCED 
SEARCH
Today's electronics must power  many of our most vital systems. Service interruptions are not easily tolerated in the financial, transportation, telecom, military or healthcare sectors, just to name a few. Redundant or Fault-Tolerant designs provide assurance that power is always available to these critical systems.   

When applied properly, power supplies can be very reliable components. However, as with any physical component, there's a limited and somewhat calculable  lifetime of service. Mean between failures or MBTF helps us measure the reliability of the power supply. Regardless of the reliability and careful design considerations, a failure is inevitable. At this point the system is down or off-line due to the lack of available power. By implementing a parallel redundant scheme we can increase the overall system reliability by increasing the availability of power. Some common methods of load share for parallel redundancy are shown below.
 
Parallel Redundancy

1.        2N: N represents the number of power supplies required to run the system load. As the name implies, 2N is twice the amount of necessary supplies. It's a fully redundant system.
Each equal supply or set of supplies is configured in parallel to the other. In the event of a failure there now exists a fully redundant system capable of supplying the load without interruption. See Fig. 1


Figure 1. 2N Parallel Redundancy
 
2.        N + 1:  N +1 is a system that employs the N amount of power supplies required to run the system load + 1 additional supply of equal wattage. The supplies are configured in parallel to supply the system load. In this method, one power supply failure requires that the additional supply only provide 1/N of the total load. The system load is now maintained without interruption by the remaining functional supplies. See Fig 2.In some mission critical applications there may be a contingency for two power supply failures or N + 2.

Figure 2. N+1 Parallel Redundancy
 
3.        2 (N + 1): This is a 2N configuration of  two complete N + 1 systems in parallel. Each N + 1 system can fully power the system load if necessary, making it a fully redundant system. See Fig 3.
 
Figure 3. 2N+1 Parallel Redundancy

Topics for Parallel Redundancy: 
    
Oring Diodes:  In redundant systems, the Oring diodes provide isolation from the other supplies. In the event of a failure like a short, the diode becomes back-biased and won't allow current from the other supplies to flow into the defective supply. The OR function effectively serves as an automatic disconnect.
Some power supplies have built-in oring diodes. If the diodes are placed externally, consideration should be given to the power dissipation of the diode. Heat may need to be removed, especially in larger applications.

Fault Detection/Hot Swap: Failed units can be identified through power good or power fail signals or other methods of  failure detection.  Once the failure has been detected it can be removed and replaced at a time when the system can tolerate a brief shutdown. If some cases the system can never be turned off and the failures must be replaced  with the system running. This is called a "hot-swap". Always be sure that the power supply has hot swap circuitry before trying this procedure.

Active /Passive Load Sharing: Load sharing can be passive or active. A passive load share is one that allows for uneven loading of power supplies. The highest output voltage will take most of the load current.
This situation may lead to reliability issues and premature failures. A better approach is to employ active or forced load sharing. By the use of dedicated circuitry, active load share forces parallel supplies to contribute to the load evenly, regardless of the number of units.

Power Boost: Power boosting is inherent in parallel redundacy. Boosting power over several supplies allows for a failed output to be compensated by other functioning supplies. N , the necessary amount of power supplies to provide the system load, can be any amount of supplies that can be paralleled successfully (consult with manufacturer).
Power boosting over several supplies allows for the use of smaller, less expensive supplies. It also decreases the amount of load required of each supply, thereby increasing MTBF hours and overall system reliability.



This month's Watts News is pleased to feature Integrated Power Designs (IPD). Integrated Power Designs has many designs that employ load share. Please see the information on the IPD NXT-400M a medically approved 400 watt power supply with load sharing for parallel redundancy and power boost. 
 

NXT-400M Web Page