Intermittent C-Bus faults are the ones that drive technicians up the wall. Everything works on Monday, half the lights go funny on Wednesday, and by the time you’ve driven back out it’s behaving itself again. Nine times out of ten when a customer rings us with “it works, then it doesn’t”, the cause is one of two things: the network has been pushed past its hard limits, or there’s a marginal cable fault that’s only misbehaving under load or temperature. This article walks through how our team chases both down on advanced jobs.
Before you start swapping units, get the fundamentals straight in your head. A single C-Bus network segment has firm physical limits, and the moment you brush up against one, comms go unstable in exactly this maddening, intermittent way. The diagram below lays out the three numbers you can’t break.
The three hard limits of a single segment
Every C-Bus network segment on the Lighting application has to live inside three boundaries. Go past any one of them and you don’t get a clean failure — you get flaky, on-again-off-again behaviour that looks like a hundred different problems.
- Cable length — around 1000 m total per segment. That’s the sum of every run in the segment, not the longest single run. On big two-storey homes with a detached garage or studio it’s easier to hit than people think.
- Unit load — around 100 unit loads per segment. Each device draws a published unit load (most input and output units are 1 or 2). Add them up; once you’re near 100 the network clock and burden start struggling.
- Enabled power supply — no more than 2 A per segment. You can physically wire more 5500PS capacity in, but only 2 A of it should be enabled (clock/burden) on a single segment. More than that and the bus voltage and signalling go out of spec.
The trap is that an install can sit happily at 95% of a limit for months. Then someone adds a couple of eDLT switches, a hot day expands a connector, or a dimmer ramps a big load, and suddenly the segment tips over the edge intermittently. If the network was designed close to the line, treat the limits as your first suspect.
Split the load with a Network Bridge
If the count says you’re over (or uncomfortably close to) any limit, the fix is not to nurse one overloaded segment — it’s to split the installation across multiple networks joined by a Network Bridge (5500NB). The bridge connects two segments while keeping each one’s clock, burden and unit-load budget independent, and passes the Group Addresses you tell it to between them.
On larger Melbourne homes we’ll routinely run the ground floor on one network, the upper floor and outbuildings on another, and bridge them. Each segment then comfortably sits inside its 1000 m / 100 unit / 2 A envelope. If you’re planning this from scratch, our notes on designing a healthy C-Bus network cover how we carve up segments before a single cable goes in.
Hunting cable faults
If the limits check out, you’re looking for a physical fault on the pink Cat5. These are classic intermittent-drop-out causes:
- Damaged cable — a staple nicked through the sheath, a screw run through it, or rodent damage in a ceiling. The conductors touch occasionally and pull the bus down.
- A short between the bus pair — often at a poorly stripped connector, sometimes from moisture bridging two pins.
- Water ingress — outdoor switch plates, garden lighting junctions and roof-space runs near a leak. Water changes resistance with temperature and humidity, so the fault comes and goes.
- A loose connector — a conductor not fully seated under a screw or in a plug. Vibration or thermal movement makes and breaks the connection.
All of these show up the same way: units intermittently disappear, and the C-Bus voltage sags or wanders. A healthy unburdened bus sits in the high 20s to around 36 V DC; the burdened, clocking bus runs lower. What you care about is stability and the reading at the far ends of the network, not just at the switchboard where it always looks fine.
The diagnostic sequence we follow
- Run a Toolkit scan over time. Connect via your PCI/CNI and scan the network — then scan again, and again, over a period. A unit that’s present on one scan and gone on the next is your smoking gun. Note which units drop and whether they cluster on one physical run. Our general C-Bus diagnostics guide has more on reading scan results.
- Measure bus voltage at the extremities. Take a multimeter to the furthest switch plates and output units on each segment and read C-Bus voltage between the bus conductors. A reading that’s significantly lower at the ends, or that visibly jumps around, points to a cable or burden problem out there.
- Disconnect and divide. If a segment is unstable, split it at a convenient junction and re-scan each half. Whichever half clears tells you which run holds the fault. Keep halving until you’ve isolated the offending leg.
- Inspect the suspect run. Check every connector on that leg for fully seated, correctly stripped conductors. Look for physical damage, and look hard at anything outdoors or in damp areas for water ingress. Re-terminate anything that looks marginal.
- Re-scan and confirm over time. Don’t call it fixed after one clean scan. Leave it scanning, or come back, and confirm the units stay present and the far-end voltage holds steady.
The one fault that destroys units: mains crossed with C-Bus
This deserves its own section because it’s the most dangerous fault we ever find. If 230 V mains has been accidentally crossed onto the pink C-Bus cable — a wiring mistake in a switch box, a cable run through the wrong terminal, or pink cable mistaken for an extra-low-voltage control wire and connected to active — it will destroy C-Bus units and is a serious safety hazard.
How we verify: with the C-Bus power supply isolated and the network disconnected from the PCI, we check for any voltage between the C-Bus conductors and earth/mains references. There should be none. Any AC reading on the pink pair is a crossed-mains fault until proven otherwise. We then trace the offending termination, correct it, and replace any units that have been cooked. Toolkit usually confirms the damage for you — fried units simply won’t answer the scan.
When a whole network comes and goes: check the bridge
If the symptom isn’t a handful of units but an entire network that randomly appears and disappears, suspect the Network Bridge before the cable. A misconfigured or failing 5500NB makes everything on the far side of it look like it’s dropping out, because Group Address messages stop crossing.
- Confirm the bridge’s configuration in Toolkit — which applications and Group Addresses it’s set to pass, and that both sides are addressed on the correct networks.
- Check the bridge has stable power and a healthy bus on both segments — a bridge sitting on an over-limit or under-voltage segment will behave erratically.
- Watch for a bridge that’s intermittently rebooting. If one whole network blinks out and comes back together, that’s a bridge or a power/clock problem on its segment, not fifty separate cable faults.
Putting it together
Work it in this order and you’ll catch the vast majority of intermittent faults: confirm the segment is inside its 1000 m / 100 unit / 2 A limits, scan over time to find units that vanish, measure voltage at the far ends, rule out crossed mains as a safety priority, then divide-and-conquer to isolate the bad cable run — and if a whole network is flapping, go straight to the bridge. If you’re tightening up an existing system, our programming and project housekeeping notes help keep the Toolkit project matching reality so the next diagnosis is faster.
That’s how we chase these down. Intermittent faults reward patience and a methodical sequence — and the discipline of never assuming the install was wired inside the limits just because it once worked.
If you’ve got a C-Bus system anywhere around Melbourne that drops out, behaves differently on hot days, or has a network that randomly disappears, our team genuinely enjoys solving these. Get in touch via our contact page and we’ll bring the gear to scan it properly. — Adam and the DUKE team.
Frequently asked questions
What are the hard limits for a single C-Bus network segment?
Each C-Bus segment should stay within around 1000 m of total cable, around 100 unit loads, and no more than 2 A of enabled power supply. Exceeding any one of these causes unstable, intermittent communications rather than a clean failure.
Why does my whole C-Bus network randomly disappear and come back?
That pattern usually points to the Network Bridge (5500NB), not the cable. A misconfigured, under-powered or intermittently rebooting bridge stops Group Address messages crossing, so everything on the far side appears to drop out together. Check the bridge’s configuration and the bus health on both segments.
How do I catch C-Bus units that only disappear sometimes?
Run a C-Bus Toolkit scan repeatedly over a period of time. A unit that’s present on one scan and gone on the next reveals an intermittent fault. Note whether the dropping units cluster on one physical cable run, then measure bus voltage at the far ends of that run.
Can mains wiring crossed with the C-Bus cable cause damage?
Yes — and it’s a serious safety fault. If 230 V mains touches the pink C-Bus cable it destroys units and creates a shock hazard. Isolate the affected circuits at the switchboard immediately and have a licensed electrician verify and rectify it. The pink cable itself is low-voltage SELV and must never be crossed with mains.
What should the C-Bus voltage read on a healthy network?
An unburdened bus sits in the high 20s up to around 36 V DC; the burdened, clocking network runs lower. What matters most is stability and the reading at the far ends of the network — a sagging or wandering far-end voltage points to a cable fault or burden problem.