Application Note – 17
Redundant Systems – Fully Redundant, No Single Point Failure, Multi Channel
Introduction
This paper describes a redundant system architecture that provides a fully redundant, no single point failure, multi channel output system. If required, additional components can also be added to this system to make it a “seamless” switch over, avoiding any phase discontinuity (or phase jumps), as described in another paper, however in a multichannel configuration this can become highly complex.
Review Of MTBF Basics
The MTBF for a system can be calculated by combining the individual MTBF’s of all the system components. Failure rate F is inversely proportional to MTBF. The way the system components are combined depends upon the configuration.
1. In a normal (non-redundant) system, the individual components can be considered as being in series.
Calculation of the system MTBF is as follows
Component A, failure rate Fa = 1/1000 = 0.001 Component B, failure rate Fb = 1/2000 = 0.0005
Overall failure rate Ft = Fa +Fb = 0.001 + 0.0005 = 0.0015
MTBF = 1/Ft = 667 hrs
2. In a redundant solution, components can be considered as being in parallel (i.e both need to fail for the system to fail)
Using the same MTBFs for component A and B:
Component A, failure rate Fa = 1/1000 = 0.001 Component B, failure rate Fb = 1/2000 = 0.0005
As they are in paralell, the overall failure rate is calculated as the product of the individual failure rates:
Overall failure rate Ft = Fa x Fb = 0.001 x 0.0005 = 0.0000005
MTBF = 1/Ft = 2,000,000 hrs or over 200 years, much greater !
Multi Channel Application
A number of applications need output drive to multiple channels e.g. in a WiMax system there is an array of three antenna systems spaced at 120 degrees to each other to give full (circular) ground coverage. Each of the antenna systems needs a precision reference for its drive system, and each of these references should be fully redundant.
Many multi channel redundancy implementations use redundant sources, which after switching are then distributed to multi channels. This system is fairly robust because distribution amplifiers and systems have very high MTBFs as the designs typically use low risk components such as resistors, low value capacitors, and operational amplifiers.
Having said this, this type of architecture does NOT meet the criteria of NO SINGLE POINT FAILURE, as invariably if the output distribution does fail it would cause a system failure, therefore overall system MTBF can never be better than the MTBF of the distribution amplifier.
The alternative architecture described here overcomes this “Achilles heel” by generating multiple channel outputs from each source (primary and backup) and then combining them through a “failsafe” switching unit.
The output channel switching is done by mechanically latching mechanical relays, and therefore even if a relay fails, it continues to pass through the last selected signal, thus does not impede system operation.
Considering the system schematic below ;
Both source 1 and source 2 have 4 x 10MHz outputs in order to drive the four auto switches AS1 through AS4. The outputs from the auto switches then provide the four fully redundant outputs. This system may exhibit momentary signal loss during switching as the relays are “break before make”, however in many instances this is not an issue as the down-stream equipment itself has an internal oscillator that is phase locking to the reference, and this acts as a “flywheel” during the switching time (typically approximately 1ms)
If multi channel seamless switching is required, this is attainable and is addressed in another paper.
The above system is a solid and relatively low cost solution to providing full redundancy. This scheme is implemented in the ptf 1207A Auto Switch /Distribution unit, that can house up to 8 redundant output channels in a single 19 inch rack mounting, 2U high unit.
Precise Time and Frequency, Inc. has the most comprehensive selection of redundancy and distribution instruments and modules for time and frequency applications in the world, and is continually providing custom solutions to meet system performance and budget requirements. To discuss this or alternative methods of redundancy, please contact Precise Time and Frequency, Inc.