Coax faults have a narrow set of causes and a wide set of symptoms. Connectors, water and splitters account for most of them, and cable itself — the expensive thing to replace — accounts for fewer than people expect.
Working in a fixed order, changing one thing at a time, finds the fault faster than anything else.
Quick answer
| Symptom | Most likely cause |
|---|---|
| No signal at all | A disconnected or corroded connector, a dead splitter port, the wrong input selected |
| Sparkles or block artifacts on digital | Marginal loss or reflections. Connectors first, then the loss budget |
| Ghosting or smearing on analog | Reflections — an open splitter port, a faulty connector, an impedance mismatch |
| Works in dry weather, fails when it rains | Water in an outdoor connector |
| Intermittent, changes when the cable moves | A loose connector or a damaged cable |
| Hum, or a bar rolling through the picture | A ground loop from more than one bond to ground |
| One outlet worse than the others | That outlet’s lead and plate, then the splitter feeding it |
Work from the source outward
The order matters because it eliminates whole sections at a time.
Start at the source. If the antenna, dish or camera output is bad, nothing downstream can fix it. Check the source with a known-good lead into a known-good receiver before going further.
Then the head end. The splitter, the amplifier, the grounding block. Substitution is the fastest test: move the feed to a different splitter port, or bypass the splitter entirely and see whether the fault survives.
Then the cable run. Damage is usually visible if you look for it: a staple driven through the cable, a sharp bend at a cable clip, a crushed section where something was set on top of it, or rodent damage in a loft or a duct.
Then the outlet. The wall plate, the patch lead and the connector at the equipment end. These are the parts most likely to have been disturbed.
Then the equipment. Only after everything upstream is eliminated.
Substitution beats measurement
Without instruments, the most efficient diagnostic is swapping in a known-good part.
Three swaps cover most faults. A known-good patch lead between the wall plate and the receiver. A known-good lead between the source and the run. And moving the run to a different splitter port.
If the fault follows the swapped part, that part was the problem. If it stays where it was, that section is clear and you move on.
The mistake to avoid is replacing a long cable run before testing the connectors at each end. Re-terminating two connectors is twenty minutes; replacing a run is a day.
What a multimeter can and cannot tell you
A continuity test proves two things: the center conductor is connected end to end, and there is no short between the center conductor and the shield. Those are worth knowing.
They are also all it proves. Continuity says nothing about the cable’s impedance, its loss, whether the shield is properly bonded along the length rather than only at the ends, or whether a connector will stay connected when the cable is moved.
A cable can pass a continuity test and fail the service it is installed for. What testing can and cannot establish is covered here.
Water, which is the fault that hides
Water in a coaxial joint produces the most misleading fault on the list, because it behaves differently depending on the weather.
The mechanism is capillary action. Once moisture reaches the shield through an imperfect seal, it travels along the braid and foil by capillary action — which means the damage is often well away from the connector that let it in.
The symptom is a service that works in dry weather and fails in wet. Technicians then test on a dry day, find nothing wrong, and leave.
The fix is to re-terminate the affected joints with compression connectors, seal the outdoor joints properly, and arrange the cable so it enters the connector from below with a drip loop. The outdoor installation detail is here.
The faults that look electrical but are not
Two worth knowing, because both send people down the wrong path.
A ground loop produces hum or a slowly rolling bar. It is not a signal-strength problem, and no amount of amplification fixes it. It comes from more than one bond to ground at different potentials, and the fix is to resolve the second bond rather than to remove grounding. The grounding arrangement is covered here.
An amplifier that makes things worse. If a signal is already strong at the receiver, adding amplification can overload the input, which produces distortion that looks like a weak signal. If adding an amplifier made the picture worse, that is the diagnosis.
Scenarios
A camera that worked for a year and now produces sparkles. Re-terminate both ends first. If that fixes it, the fault was a joint degrading in the weather.
A television that works except in heavy rain. Water in an outdoor joint. Re-terminate and seal, and check the drip loop.
A new outlet that performs worse than the existing ones. Compare its patch lead and plate with a working outlet’s before suspecting the cable.
An installation that hums after a second service was added. Two bonds at different points. Find and resolve the second one.
A fault that appears only when the cable is touched. A loose connector or a broken center conductor inside the jacket. Re-terminate, and if it recurs, the cable is damaged somewhere and should be replaced.
FAQ
Why does my coax signal cut out in the rain? Water in an outdoor connector. The seal has failed and moisture is traveling along the shield.
Can a multimeter find a bad coax cable? It can find a broken center conductor or a short. It cannot measure impedance, loss or shield quality.
Why do I have sparkles on digital but the signal meter looks fine? Digital services have a threshold, and the meter may be reading level rather than quality. Reflections and marginal loss cause errors before the level looks alarming.
Should I replace the cable first? No. Test the connectors and the splitter first. They are the more likely faults and far cheaper to fix.
What causes ghosting on an old television? Reflections, usually from an open splitter port or a poorly fitted connector.
Can a bad splitter affect only one outlet? Yes. A splitter with an unterminated spare port can affect the circuit it feeds rather than the whole system.
Related reading on this site
- Coax cable testing: what a meter proves
- Coax connector types
- Coax splitter loss
- Coax grounding and surge protection
- Coax attenuation: why loss rises with frequency