Browse Prior Art Database

Fault-Tolerant Power System

IP.com Disclosure Number: IPCOM000123092D
Original Publication Date: 1998-May-01
Included in the Prior Art Database: 2005-Apr-04
Document File: 2 page(s) / 87K

Publishing Venue

IBM

Related People

Beer, R: AUTHOR [+3]

Abstract

Many fault-tolerant computer systems and peripherals include an "N+1" fault-tolerant power system. Conversion of AC mains input to low-voltage DC is performed by a number of power-supply units (PSUs). The number of units required to power the whole system is "N". To ensure fault-tolerance the system is implemented with "N+1" PSUs. If any one PSU fails, or is removed for maintenance purposes, the remaining "N" PSUs can supply enough power to keep the system running.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Fault-Tolerant Power System

   Many fault-tolerant computer systems and peripherals
include an "N+1" fault-tolerant power system.  Conversion of AC mains
input to low-voltage DC is performed by a number of power-supply
units (PSUs). The number of units required to power the whole system
is "N".  To ensure fault-tolerance the system is implemented with
"N+1" PSUs.  If any one PSU fails, or is removed for maintenance
purposes, the remaining "N" PSUs can supply enough power to keep the
system running.

   A popular configuration for fault-tolerant products is
"N=2" where the system contains a total of 3 PSUs.  Each of the PSUs
has its own separate AC mains input connector.

   Some users prefer to provide all of their fault-tolerant
machines with two separate sources of AC mains.  This protects them
against the failure of one of the two mains supplies.  However, it is
not possible to connect a dual AC mains system to the 3 PSUs and
still maintain full fault-tolerance.

   The solution described here connects the central PSU via a
dual-input automatic power switch.  The wiring is then as shown in
Fig. 1.

   The dual-input automatic power switch will normally drive
PSU #2 from mains supply A.  If mains supply A fails then a switch is
operated so that PSW #2 is driven from mains supply B instead.  This
leaves PSUs #2 and #3 still powered, satisfying the rule that at
least "N" PSUs be powered to keep the system running.

   In the case where mains supply B fails, no action is
taken.  This leaves PSUs #1 and #2 still powered.  Therefore, either
mains supply can fail, but operation of the system is protected.

   The automatic switch can be a conventional relay or a
solid-state switch implemented in Thyristors or Triacs.  Whatever
implementation is selected, it must be able to perform the
switchover quickly enough to avoid interruption of the out...