A telephone route is only as reliable as the product of everything on it: ten repeaters that are each up 99% of the time leave the route up about 90% of the time, while a second independent path raises two 99% paths to 99.99%. The gain exists only if the paths do not share a failure, such as the same pole line, battery or hub. Fault detection then has to find breaks quickly: loop-current supervision, scheduled test calls, and resistance and capacitance tests that locate a fault by distance.
Part of Telecommunications
Redundant paths, single points of failure and fault detection for a telephone network, with the arithmetic of series and parallel availability.
Series Elements Multiply Their Failures
Availability a is the fraction of time an element works. For elements in series, a route is up only when all of them are, so the availabilities multiply. Take a = 0.99, which is 99%, per element as an illustration, not a measurement:
- 10 repeaters: 0.99¹⁰ = 0.904, so the route is up about 90% of the time
- 8 elements (line, 5 repeaters, 2 exchanges): 0.99⁸ = 0.923
Every repeater added to stretch a route (Amplification) therefore costs reliability as well as noise.
A Second Path Helps Only If It Fails Independently
Two paths, each with availability 0.99, are both down only if both fail at once:
- Both down: 0.01 × 0.01 = 0.0001
- Route up: 1 − 0.0001 = 0.9999, or 99.99%
That arithmetic assumes the failures are independent. A second wire on the same pole line is not: one fallen pole, one fire or one storm takes both. Corpus guidance on pole lines is that a carefully built line lasts 30 to 50 years while a careless one may need work within 5 and fail within 10 (Pole Line Building), so the support structure is itself a common cause. Route the second path over different poles or a different valley, power it from a different battery, and terminate it at a different hub. See Topology Design for ring and hierarchy layouts that create a second path.
The Single Points Of Failure Are Known
- The central battery. If it fails, every instrument loses its microphone current. When two or more subscribers on one exchange fail together, check the exchange first, starting with the battery, which should read 48 V ± 2 V under normal load (Troubleshooting).
- Common control. Switching Systems warns that a crossbar exchange’s common control is a single point of failure and must be duplicated.
- The hub. In a star, every inter-village call.
- The operator. A manual exchange fails at night, or when its one trained person is away.
Detecting A Fault Early
- Supervision. An exchange detects subscriber loop current to know a handset is lifted (Battery and Powering). A line that never shows loop current when it should is a suspect.
- Scheduled test calls. The telegraph guidance for repeater stations is a test sequence about twice daily and a battery check daily (Repeater Stations). It is telegraph practice, but the habit transfers.
- Design margin. The same article designs each segment at 30 to 50% of its theoretical maximum length to leave room for faults, weather, corrosion and battery decay.
- Reference parts. Keep a known-good receiver beside any receiver under repair (Electromagnetic Receiver).
- A bridging test set. A high-impedance test set across a 600 Ω line costs less than 1 dB when its impedance is 6,000 Ω or more (Impedance Matching).
Locating A Fault By Arithmetic
Three tests turn a symptom into a distance. The sample figures are the ones the source articles use.
| Fault | Symptom | Test | Worked figure |
|---|---|---|---|
| Open circuit | Line dead | Capacitance: a line open at some distance shows only that fraction of its expected capacitance (Open Circuit Test) | 60% of the expected capacitance on a 50 km line: 0.6 × 50 = 30 km |
| Short | Loop current with the handset on-hook; crossed pairs | Loop resistance to the short (Short Circuit Test) | 400 Ω at 40 Ω/km per conductor: 400 / (2 × 40) = 5 km |
| Ground fault | Noise, intermittent faults | Murray loop test (Ground Fault Test) | Balance at 30% of a 100 km loop: 0.30 × 100 = 30 km |
Good insulation reads far above the floor of 1 MΩ that Wire Selection for Telephony gives for a practical line; Ground Fault Test says hundreds of megohms per kilometre. A reading near the floor is a fault developing.
Repair Speed Is A Reliability Figure
A spliced break restores service within hours instead of days (Cable Splicing). Keep wire, insulators, tape and a splice kit at each exchange, a map marking every splice, and a named person with authority to cut and join. A fault found in minutes and fixed in hours matters more than a route that rarely fails but takes a week to mend.
What You Need
- A log of every fault, its cause and repair time, so availability is measured rather than assumed
- A multimeter and bridge for resistance, insulation and loop tests
- Spares for the parts that fail first: contacts, cords, wipers, batteries
- A second route and a second battery where the critical lines justify them
- A written escalation: who is called when the exchange battery is low
Common Mistakes
- Two paths on one pole line. One storm cuts both.
- A redundant exchange on one battery. The battery is the common cause.
- Believing a calculated availability. The 0.99 here is an illustration; log your own.
- Testing only on failure. Scheduled tests find a developing fault before it is an outage.
Safety
Fault finding on live lines
Any voltage found on a pair is a power cross and an emergency to be cleared before any resistance testing (Short Circuit Test). Aerial work on a pole line during a storm risks lightning induction of thousands of volts (Lightning Protection). A lead-acid exchange battery can deliver very high short-circuit current and gives off hydrogen on charge: fuse every feed and ventilate the room.
What’s Next
- Learn next: Data Communication, where detection of errors moves into the signal itself
- Combine with: Troubleshooting for the symptom-first method