If your broadband is blinking red, that light is a status code, not a random glitch. It means your router or modem has lost its link to the network. Sometimes the fault is a loose cable in your front room.
Sometimes it is a cut fibre three streets away.
There is a real number behind this. Per ITU-T G.984, the standard governing GPON fibre, most optical network terminals expect a receive signal between -8 dBm and -27 dBm. Drop outside that window and the ONT declares loss of signal.
Your router goes red.
Quick Answer

A blinking red broadband light means your router has lost its internet link. Check for an area outage first. Power-cycle your modem or ONT, then reseat every cable.
If the light stays red, the fault is on the line. Call your provider and ask for a line test.
Broadband Light Blinking Red: What It's Actually Telling You
The lights are a status language, not a decoration
An LED on a router reports one thing: whether a specific link is up. The broadband light watches the connection between your equipment and the provider's network. It does not watch Wi-Fi.
It does not watch your laptop.
That is why the light can be red while your phone still shows five bars of Wi-Fi. The wireless side is fine. The link behind it is not.
Manufacturer documentation from Netgear, TP-Link, Asus, Arris, and Sagemcom follows the same convention. Solid green means the link is established. Blinking green usually means traffic or a handshake in progress.
Red means the link is down, or the device cannot authenticate.
Some gateways split that into two separate lights. One covers the physical connection. The other covers the login to your account.
Why just rebooting is the wrong first move
Rebooting is not useless. It is just badly ordered.
If a fibre cable has been cut two streets away, rebooting changes nothing. You have burned three minutes and learned zero. Worse, you have wiped the error log a technician would have used.
Our research across ISP troubleshooting documentation shows the same recommended order every time. Check for an outage first. Then check your own cabling.
Only then reboot.
The reason is simple. Outages and physical faults are the two most common causes of a red light. Rebooting only fixes the third cause, a stalled session.
That third cause is real, though. A router that has lost its DHCP lease or dropped its PPPoE session can sit in a red state until it retries. A clean power cycle forces that retry immediately.
What you can fix yourself vs. what needs a technician
Many red-light cases get resolved at home. The rest need someone with a line tester and access to the street cabinet.
| Symptom | Likely cause | Can you fix it? | | Red light after a power cut | Stalled session or failed sync | Yes, power-cycle | | Red light on one device only | Loose coax, RJ11, or fibre connector | Yes, reseat it | | Red light on ONT LOS or PON | No light arriving on the fibre | No, call provider | | Red light plus crackling phone line | Copper line fault | No, call provider | | Red light on a new self-install | Provisioning or activation | Usually provider-side | | Red light with neighbours also down | Area outage | No, just wait |
If the light is red on the ONT itself, stop there. That box converts light into data. You cannot restore a missing optical signal from your living room.
Openreach, Verizon, and CenturyLink all treat the ONT as provider-owned equipment.
If the light is red only on the router and the ONT shows green, the fault sits between those two boxes. That is often an Ethernet cable you can reseat in ten seconds.
Which Light Is Blinking Red? Start Here
Every brand names the same light differently. Here is how the labels map across the major providers.
| Brand or device | Label used | What it monitors | | BT, EE, Plusnet | Broadband | Link to the exchange or ONT | | Virgin Media, Xfinity | Internet or Online | Link to the CMTS | | Sky, TalkTalk, Vodafone | Broadband or Internet | Link to the provider network | | Standalone routers | WAN | The provider-facing port | | Fibre ONTs | PON and LOS | Optical signal from the network | | DSL modems | DSL | Sync with the DSLAM |
Broadband, Internet, WAN, Online: same job, different labels
These four words describe the same thing on different hardware. Once you know that, the label stops mattering.
WAN stands for Wide Area Network. It is the port and the link that face your provider, not your devices. On a typical home router, the WAN port is the one that runs to the modem or the ONT.
Everything else on the box, the LAN ports and the Wi-Fi radios, faces inward. Those are irrelevant to a red broadband light. So if your router says WAN and your neighbour's says Internet, you are both looking at the same indicator.
LOS and PON lights: the fibre-specific ones that mean no light arriving
LOS stands for Loss of Signal. PON stands for Passive Optical Network. A red or blinking red LOS light is the most serious reading on any home device.
It means the ONT is receiving no usable light from the provider's network. Common causes include a cut drop fibre, a dirty or unseated SC/APC connector, a failed passive splitter, or a powered-down port at the street cabinet.
There is one absolute rule here. Never look into the end of a fibre cable. Fibre lasers operate at wavelengths your eye cannot see, so the damage happens before you feel anything.
Do not unscrew the fibre connector to inspect it either. A single speck of dust on a fibre ferrule can add several dB of loss. That is often enough to push a marginal line into a red LOS state.
If you must clean it, use a proper one-click cleaner designed for the connector type.
DSL light blinking red vs. no DSL light at all
On a DSL or VDSL connection, the DSL light reports whether the modem has synced with the DSLAM. A DSLAM is a Digital Subscriber Line Access Multiplexer, the equipment at the exchange or street cabinet that your line connects to.
A blinking red DSL light means the modem is trying to sync and failing. The line is present, but the handshake is not completing.
No DSL light at all is a different problem. It usually means the modem is not seeing the phone line, which points to a cabling fault on your side of the master socket.
That distinction changes your next move. A dead DSL light sends you to your filters, your extension sockets, and your master socket. A blinking red DSL light sends you to your provider for a line test.
Phone light, Wi-Fi light, LAN light: the ones you can ignore for now
These three almost never explain a red broadband light, and chasing them wastes time.
The phone light covers VoIP service. The Wi-Fi light covers the wireless radio. The LAN light covers wired devices plugged into the back.
A red Wi-Fi light means the radio is off or has failed. That is a separate problem, and it will not cause the broadband light to go red. Same story with a red phone light on a line that no longer carries voice over copper.
Fix the broadband light first. The others usually sort themselves out.
Provider-Specific Light Behaviour: What Each Gateway Actually Does
Every brand uses its own colour convention, and the differences trip people up. Here is how the major gateways behave.
| Provider or gateway | Light name | Red means | Amber or orange means | | BT Smart Hub | Broadband | No link to exchange or ONT | Connecting or degraded | | Virgin Media Hub | Internet | No link to the CMTS | Booting or ranging | | Sky Broadband Hub | Internet | No link | Sync in progress | | TalkTalk router | Internet | No link | Connecting | | Vodafone router | Internet | No link | Connecting | | Xfinity gateway | Online | No link | Ranging or registering | | Spectrum modem | Online | No link | Acquiring downstream | | AT&T gateway | Broadband | No link | Connecting | | Verizon router | Internet or WAN | No link | Connecting | | CenturyLink ONT | LOS | No optical signal | Not applicable |
Why the same colour means different things
On some gateways, red means the link is down. On others, it means the device is still trying. On a handful, red means the device is stuck in a boot loop.
Read the manual for your exact model. Manufacturer support pages from Netgear, TP-Link, Asus, Arris, Sagemcom, and Hitron all document the light states for their hardware. The sticker on the base usually tells you the model number.
If the manual is lost, the provider's support site has a light-meaning page for every gateway they ship. Search the model number plus "light meanings".
Why colour-blind users struggle with these lights
Roughly 1 in 12 men and 1 in 200 women have some form of colour vision deficiency. Red and green LEDs are the worst possible combination for the most common type.
If you cannot distinguish the colours reliably, use the blink pattern instead. Solid versus blinking is easier to read than red versus green.
The admin page is the better answer. It states the connection status in words, not colours. Log in and check the WAN status directly.
Read the Blink Pattern Before You Touch Anything
Solid red vs. slow blink vs. fast blink vs. cycling colours
The blink rate carries information. Here is what the common patterns usually mean across major gateway brands.
| Pattern | Typical meaning | First action | | Solid red | Link is down, device has stopped trying | Check cables and outage | | Slow blink (about 1 per second) | Attempting to connect or sync | Wait 2 minutes, then reboot | | Fast blink | Authentication or provisioning failure | Check account status | | Cycling colours | Firmware boot or update in progress | Do not unplug | | Red then green, repeating | Marginal signal, intermittent sync | Check signal levels |
Slow blink is the most common and the least alarming. The device is doing its job and simply cannot complete the handshake.
Fast blink is the one people misread. It often looks like activity, but on most gateways it means the device is failing at the login stage rather than the signal stage.
Cycling colours are the pattern you should never interrupt. Pulling power during a firmware write can brick a gateway outright.
What a red light that turns green on its own tells you
A red light that recovers within two or three minutes is usually a transient sync loss. DSL and cable lines drop sync all the time during electrical noise, bad weather, or a brief cabinet power blip.
If it recovers and stays green, note the time anyway. Repeated recoveries at the same time of day point to a pattern rather than a random fault.
That log becomes powerful evidence later. UK providers under the Ofcom automatic compensation scheme, and several US providers under their own service guarantees, respond very differently to a documented pattern than to a single complaint.
Write it down. Date, time, duration, and which light was red. Ten seconds of effort, and it can be worth real money.
When the pattern points to sync, authentication, or provisioning
Three failure stages produce three different red-light patterns. Mixing them up sends you down the wrong path.
Sync failure. The modem cannot hear the provider's signal. On DSL this is the DSL light. On cable it is a failure to lock the downstream channel.
On fibre it is LOS. No amount of router configuration will fix this.
Authentication failure. The physical link is up, but the login failed. Causes include wrong PPPoE credentials, a MAC address lock after swapping hardware, or a suspended account.
Provisioning failure. The device is online, but the provider has not activated the service on their end. This is extremely common on new installs, self-install kits, and after a house move.
If the pattern suggests authentication, check your account status before you touch any hardware. A failed direct debit or an expired card produces a red light that no amount of rebooting will clear.
Check for an ISP Outage First (Two Minutes, Saves Hours)
Postcode and service status checkers worth using
Every major provider runs a public status page. BT, Sky, TalkTalk, Virgin Media, Xfinity, Spectrum, AT&T, and Verizon all publish one, and most update it within minutes of a known fault.
Use the provider's own status page rather than a third-party outage map. Aggregator sites lag by hours and often report stale incidents.
Check the provider's app too. Most now push postcode-level outage alerts before you even think to ask. Turning those notifications on is one of the few genuinely useful things a provider app does.
In the US, the FCC publishes consumer guidance on broadband service, outage reporting, and the broadband consumer labels that providers must now display. Those labels include the speeds you are actually paying for, which matters if you later claim a fault.
Ask the neighbours: a genuinely useful diagnostic
This is the fastest test in the entire process. It takes about ninety seconds.
If your neighbour on the same provider is also down, you have an area fault. Stop troubleshooting, report it, and wait. Nothing you do inside your house will help.
If your neighbours are fine, the problem is on your side of the network. That still includes the drop cable running to your property, which remains the provider's responsibility to repair.
This one question splits the whole decision tree in half. Everything after it depends on the answer.
Planned maintenance windows and copper or fibre upgrade works
Providers run maintenance overnight, and most publish a window, usually between midnight and 6am. A red light inside that window is expected behaviour.
As of 2026, many regions are mid-migration. Copper networks are being retired across the UK, and DOCSIS upgrades are rolling out across North America. Both generate short, scheduled outages that look identical to a fault.
If your red light appeared during a published maintenance window, wait it out. If it persists past the window, escalate with the window reference in hand.
Weather, roadworks, and cable damage as outage causes
Aerial drop wires are vulnerable to wind, ice, and falling branches. Underground cables are vulnerable to excavation, and utility crews cut them more often than anyone likes to admit.
A red light that appears during a storm is almost never your equipment. A red light that appears the same day a crew starts digging on your street is almost certainly a cut cable.
Report it with the timing. Saying "it went red at 9:40am, right after the crew on Elm Street started digging" gets a faster response than "my internet is broken". Specifics beat sentiment every time.
Your Connection Type Changes the Whole Diagnosis
The technology carrying your signal decides which lights matter, which numbers are worth reading, and which fixes are even possible. Here is the short version.
| Connection type | What carries the signal | Light that matters most | Typical failure point | | FTTP (fibre) | Single-mode fibre to an ONT | LOS or PON | Drop fibre or connector | | Cable (DOCSIS) | Coaxial cable to a cable modem | Online or Internet | Signal levels or splitter | | DSL / VDSL | Copper phone pair | DSL | Sync, filters, line length | | Fixed wireless / 5G | Radio from a mast | Signal or 5G | Alignment or congestion | | Satellite | Radio to a satellite | Network or Status | Weather or dish alignment |
Fibre to the premises (FTTP) and the ONT
FTTP means fibre runs all the way to your wall. An ONT terminates it inside your property, and your router plugs into the ONT with Ethernet.
The ONT has its own lights. Power, PON, LOS, and sometimes a status indicator. If LOS is red, nothing downstream matters.
Your router is fine, your cables are fine, and your settings are fine.
GPON optical receive power typically sits between -8 dBm and -27 dBm. Outside that window, the ONT declares loss of signal. ITU-T G.984 is the standard behind GPON, and the optical budget it defines is why a single bad splice can take an entire street offline.
If the PON light is green and the LOS light is dark, the fibre is healthy and your problem is further down the chain. That is your cue to look at the router.
Cable and coax: DOCSIS signal levels and splitters
Cable internet runs over coaxial cable using the DOCSIS standard, which CableLabs maintains. DOCSIS 3.1 and 4.0 are the current generations in active deployment.
Typical healthy values on a cable modem look like this. Downstream power between +15 and -15 dBmV. Downstream SNR above 30 dB.
Upstream power between 35 and 50 dBmV.
A modem that has lost its downstream lock will show a blinking red or amber light. That is a signal problem, not a login problem, and no router setting will fix it.
Splitters are the classic culprit. Every splitter in the chain reduces signal in both directions. An extra splitter added for a TV box or a second room can be enough to push a marginal line over the edge.
DSL and VDSL: filters, master sockets, and line length
DSL runs over the phone line. VDSL runs over it faster, usually from a street cabinet rather than the exchange, which shortens the copper path.
Two numbers matter more than any other. SNR margin is the headroom above the noise floor. Attenuation measures how much signal is lost over the length of the line.
Healthy VDSL2 lines usually show an SNR margin above 6 dB and attenuation under 40 dB. Long rural lines push attenuation past 60 dB, and those lines drop sync constantly no matter what hardware you fit.
Filters matter too. Every phone, set-top box, or alarm panel sharing the line needs a filter, or you need a filtered faceplate on the master socket. Removing the ring wire from extension sockets is a genuinely effective fix on older UK wiring, because that wire acts as an antenna for interference.
Fixed wireless and 5G home broadband: signal, not cable
Fixed wireless access and 5G home broadband use radio instead of wire. There is no line to cut and no connector to reseat.
A red light here usually means the gateway has lost its cell. Causes include mast maintenance, peak-hour congestion, and a blocked or misaligned antenna.
Signal quality is measured in RSRP and SINR. RSRP is reference signal received power, and SINR is signal to interference plus noise ratio. Both are visible in most gateway admin pages.
Moving the gateway two feet can change everything. So can a new building opposite, a growing tree, or a low-emissivity window film that blocks the relevant frequencies.
Satellite and rural long-line setups
Satellite broadband has the longest latency and the highest weather sensitivity of any option. Heavy rain, thick cloud, and snow on the dish all cause dropouts, and none of them are fixable from indoors.
Rural long-line DSL has the opposite problem. The line is physically intact, but it is simply too long to hold a stable high-speed sync. Attenuation and noise win.
In both cases, the red light is telling you the truth. The connection genuinely is not there, and rebooting will not change physics. What you can change is your provider's expectations, so ask for a line test or a service review rather than a replacement router.
Red Lights in Rural, Coastal, and Extreme Weather Locations
Location changes which faults you get and how often. Here is what to expect in different environments.
Long rural lines and high attenuation
Rural DSL lines can run several kilometres from the cabinet. Attenuation climbs above 60 dB, and sync becomes fragile.
Red lights on these lines are usually weather-related. Damp air raises noise, and noise pushes a marginal line over the edge.
A stability profile is the realistic fix. It trades speed for uptime, and on a long line that is almost always the right trade.
If your line is genuinely beyond copper's reach, ask about fixed wireless or satellite. Both have their own trade-offs, but both sidestep the physics problem entirely.
Coastal salt corrosion and humidity
Salt air eats connectors. Coax F-connectors corrode, DSL joints degrade, and outdoor enclosures fill with moisture.
If you live within a few kilometres of the coast, expect more frequent line faults than inland neighbours. Corrosion is not a fault you caused, and it is not one you can fix.
Report corrosion you can see on your own side of the network boundary. A corroded connector at the wall plate is worth replacing. Anything beyond that belongs to the provider.
Extreme heat, freeze-thaw, and storm seasons
Heat causes expansion, and expansion stresses joints. Cable modems in hot lofts drop sync more often in summer, and their internal temperatures climb past safe limits.
Freeze-thaw cycles are worse. Water gets into a joint, freezes, expands, and breaks the connection. The fault appears in winter and vanishes in spring.
Storm seasons produce the most dramatic failures. Aerial drop wires snap, poles lean, and trees fall across cables. These faults are usually obvious and usually affect multiple properties.
Wildfire, flooding, and grid instability
Wildfire damages poles and cables over a wide area, and recovery takes weeks. Flooding destroys underground joints and can take months to fully repair.
Grid instability causes a different pattern. Repeated brownouts leave equipment half-booted and stuck in red states. A UPS on your router and ONT prevents that specific failure mode.
If you live in a region with regular grid problems, a small UPS is a genuinely good investment. It costs less than a single technician visit.
The Home-Side Checks That Fix Most Red Lights
Start here if you have already confirmed there is no area outage and your neighbours are online. These checks cost nothing and fix a large share of red-light cases.
Work through them in order. Do not skip ahead to the factory reset.
Power, cables, and the ports people plug into wrongly
Check the power first, and check it properly. A router running from a failing power brick can boot, blink, and then fail to hold a link. Try a different socket and a different adapter if you have one.
Now the cables. On a fibre setup, the Ethernet cable runs from the ONT LAN port to your router WAN port. Mixing those two up is more common than anyone admits.
Look at the port labels, not the colour of the cable. A blue cable in a blue port means nothing if the port is labelled LAN.
On a cable setup, check the coax connector at both ends. Finger-tight is not tight enough. It should be snug, and the centre pin should be straight, not bent or pushed in.
On a DSL setup, check the RJ11 cable and every filter in the chain. A filter plugged in backwards, or a phone plugged into an unfiltered socket, will kill your sync.
Power-cycle order: ONT first, then router
The order matters more than the act. Here is the sequence that works.
- Unplug the ONT or modem from power. Leave the router plugged in for now.
- Wait a full 60 seconds. This lets the provider's session time out on their end.
- Plug the ONT back in and wait for its PON light to go solid green. That can take two to three minutes.
- Now unplug the router, wait 30 seconds, and plug it back in.
- Give it another two minutes before you judge the result.
If you power both at once, the router can finish booting before the ONT has a link. It then sits in a red state and never retries cleanly.
That single mistake explains a huge share of "I rebooted it and it didn't work" complaints.
Testing at the master socket or test socket
If you have a DSL or VDSL line, the master socket is the single most useful test point in your house. It is the first socket where the provider's line enters, and everything else is extension wiring you own.
Unscrew the faceplate and plug your modem into the test socket behind it. That test socket sits directly on the incoming line and bypasses every extension, every filter, and every bit of internal wiring.
If the DSL light goes green at the test socket, the fault is inside your home. If it stays red, the fault is on the line and your provider needs to act.
This one test tells you who pays for the repair. Run it before you call.
Bypass the router: connect straight to the modem
If your modem and router are separate boxes, connect a laptop directly to the modem with Ethernet. Skip the router entirely.
If the internet works, your router is the problem. If it does not, your modem or the line is the problem.
This test is decisive, and it takes about two minutes. It also removes every Wi-Fi variable from the equation at once.
On an all-in-one gateway you cannot do this. Skip to the admin page section instead.
Wired vs. Wi-Fi testing so you're not chasing the wrong fault
A red broadband light is never a Wi-Fi problem, but a red Wi-Fi light gets confused with it constantly.
If the broadband light is green and your devices still cannot reach the internet, you have a different fault. Check the router's DNS settings, try a public resolver, and test on Ethernet before you assume the line is down.
Aggregate user reports show DNS resolver outages produce symptoms that look exactly like a dead connection. Pages fail to load, apps time out, and the router looks perfectly healthy.
Firmware, admin page login, and WAN status
Outdated firmware causes genuine link failures, though it is lower on the list than people think. Manufacturer release notes from Netgear, Asus, and TP-Link do document WAN dropouts fixed in specific firmware versions.
Log into the admin page and check two things. First, whether the WAN interface has an IP address. Second, whether the connection status says Connected, Disconnected, or Authenticating.
If it says Authenticating, you have a credentials or account problem. If it says Disconnected with no IP, you have a link problem.
When the Fault Is on the Line, Not in Your House
Once you have eliminated your own cabling, the remaining causes live on infrastructure you cannot legally touch. Recognising them helps you ask for the right thing.
Sync failures, SNR margin, attenuation, and optical power
Each technology fails in its own way, and each reports it with different numbers.
| Technology | Key metric | Healthy range | Red light means | | VDSL2 | SNR margin | Above 6 dB | Cannot hold sync | | VDSL2 | Attenuation | Under 40 dB | Line too long or noisy | | GPON fibre | Optical receive power | -8 to -27 dBm | LOS, no usable light | | DOCSIS cable | Downstream SNR | Above 30 dB | Cannot lock channel | | DOCSIS cable | Upstream power | 35 to 50 dBmV | Cannot reach CMTS |
If your SNR margin is hovering at 3 or 4 dB, the line is marginal. It will drop every time it rains or the electrical noise rises. That is a line fault, not a router fault.
If your attenuation is above 60 dB, you are on a long rural loop. No hardware change will fix that, and a stability-focused line profile is usually the realistic outcome.
Bridge taps, ring wire, extension sockets, and old alarm wiring
Copper lines hate extra connections. Every spur, every unused socket, and every old alarm panel adds a reflection that degrades the signal.
The ring wire is the classic UK culprit. It runs to pin 3 on older extension sockets and acts as an antenna for interference. Disconnecting it at each socket genuinely improves sync on many lines.
Bridge taps do the same thing on longer lines. They are a leftover spur from old wiring, and your provider can often detect and remove them remotely.
If your provider's line test shows a bridge tap, that is a real finding. Ask for it to be removed rather than accepting another router swap.
Water ingress, rodent damage, and corroded connectors
Outdoor joints fail with age. Water gets into the enclosure, the copper corrodes, and the line goes noisy before it goes dead.
Rodent damage is more common than people expect. Squirrels chew through drop wires, and mice chew through indoor cabling behind skirting boards.
Corroded F-connectors cause the same problem on coax. Look for green or white residue around the connector, and check whether the cable has been crushed or kinked.
None of these are DIY repairs on the provider's side of the network boundary. Document what you can see and report it.
Street cabinet, exchange, and headend faults you can't touch
A red light with healthy numbers at your end usually means the fault sits further up. Street cabinets lose power, DSLAM cards fail, and CMTS ports go down.
These faults affect groups of properties rather than single homes. That is why the neighbour test is so useful, and why a fault that hits four houses on your street is almost never a coincidence.
Report it with your postcode and the exact time it started. Providers can correlate your report with their own monitoring, and that speeds up the fix.
Router Admin Page Diagnostics: The Numbers That Tell the Truth
The admin page is where you stop guessing. It shows what your equipment actually sees, and that evidence changes how your provider responds.
Where to look for WAN status, IP, DNS, and PPPoE state
Most routers use one of a handful of default addresses: 192.168.0.1, 192.168.1.1, 192.168.1.254, or 10.0.0.1. The address and the admin password are usually printed on a sticker on the base.
Look for a status, internet, or WAN page. You want four things: the WAN IP address, the default gateway, the DNS servers, and the connection state.
A WAN IP starting with 100.64 through 100.127 means you are behind carrier-grade NAT. That is normal on many connections, but it blocks inbound connections and breaks port forwarding.
Reading DSL and cable signal values
On a DSL modem, the DSL status page shows sync rate, SNR margin, attenuation, and error counters. Rising uncorrectable errors alongside a falling SNR margin means the line is degrading.
On a cable modem, the DOCSIS page lists each downstream and upstream channel with its power level and SNR. Look for channels dropping out or power levels drifting toward the edges of the range.
Take a screenshot before you reboot anything. Those values reset on a power cycle, and losing them makes the conversation with support much harder.
Checking logs for authentication, DHCP, and provisioning errors
Router logs are ugly but honest. You are looking for a handful of specific entries.
PPPoE authentication failure means wrong credentials or a suspended account. DHCP timeout means the provider is not answering. Repeated re-registration attempts point to a provisioning problem on their end.
Save the log to a file if the router allows it. Timestamps make your case, especially when you are claiming compensation for a multi-day outage.
Capturing evidence for your ISP ticket
Build a small evidence pack before you call. It takes ten minutes and saves hours.
- Screenshots of the admin status page and signal values
- A photo of the lights on the ONT, router, and any modem
- The exact time the fault started and whether it is constant or intermittent
- Results of the master socket or direct-to-modem test
- Whether neighbours on the same provider are affected
- Your account number and the equipment serial number
Support agents work from scripts. Evidence lets them skip the script.
Reboot, Factory Reset, or Bridge Mode? Pick the Right One
These three actions get used interchangeably, and they do completely different things. Picking the wrong one wipes your configuration or breaks your connection entirely.
Soft reboot vs. power cycle
A soft reboot restarts the software. A power cycle restarts the hardware and clears the volatile state that the software cannot reset.
If your router has a reboot button in the admin page, that is a soft reboot. It fixes stalled sessions about as often as a power cycle, and it is gentler on the device.
A power cycle is the stronger option. Pull the plug, wait a full minute, plug it back in. That minute matters, because it lets the provider's authentication session expire.
If a soft reboot fixes the light, you have a session problem. If only a power cycle fixes it, you may have a failing power supply.
Factory reset: what you lose and when it's worth it
A factory reset wipes everything. Your Wi-Fi name and password, port forwarding rules, static IP reservations, parental controls, and any ISP-specific settings.
On ISP-supplied gateways, that last part is the danger. Some providers push configuration automatically on reconnect, and some do not. If yours does not, you can end up with a gateway that will not authenticate at all.
Reset only when you have changed a setting you cannot remember, or when the admin page is unreachable and a power cycle has failed. Never reset as a first response to a red light.
Bridge mode and double NAT traps
Bridge mode turns your router into a dumb pass-through. It hands the public IP to whatever sits behind it and stops doing routing.
People enable it to fix gaming NAT types or to use their own router. That is valid. The trap is ending up in double NAT, where two devices both perform routing and neither knows about the other.
Symptoms of double NAT include strict NAT types, failed port forwarding, and broken VPN connections. It does not usually cause a red broadband light, but it causes a lot of confusion when troubleshooting one.
MAC cloning, VLAN tagging, and IP passthrough
Some providers lock service to the MAC address of the equipment they supplied. Swap the router and the connection refuses to authenticate, even though everything is physically fine.
MAC cloning copies the old device's address onto the new one. It is a legitimate fix, and most routers support it.
VLAN tagging matters on fibre and some cable networks. If your provider requires a specific VLAN ID on the WAN interface, a router without that setting will never connect. IP passthrough is the equivalent workaround on AT&T gateways.
If you have swapped hardware and the light went red immediately, suspect one of these three before you suspect the line.
What to Tell Your ISP (and What to Have Ready)
Calling support without evidence gets you a script. Calling with evidence gets you a line test. Here is what to have in front of you before you dial.
The exact details that skip the scripted troubleshooting
The agent's first job is to rule out the obvious. Your job is to show them you already did.
Have these ready:
- Your account number and the service address
- The make and model of your router, modem, or ONT
- The equipment serial number from the sticker
- The exact time the red light started
- Whether it is constant or intermittent
- Which lights are red and which are green
- Whether you tested at the master socket or direct to the modem
- Whether neighbours on the same provider are affected
- Any signal values from the admin page
Lead with the test results. Saying "I tested at the master socket and the DSL light is still red" moves you past three script steps instantly.
Asking for a line test or remote diagnostics
A line test checks the copper or fibre from the provider's end. Ask for it by name.
Providers can run a metallic line test on DSL, check optical power on fibre, and query modem signal levels on cable. All of these happen remotely, and all of them produce a result the agent can read.
If the test comes back clear, ask for a stability profile change on DSL. That trades a little speed for a more stable sync, and it often fixes marginal lines.
If the test comes back with a fault, ask for the fault reference number. Write it down. You will need it for compensation claims later.
Booking an engineer and what "no access" fees mean
If remote diagnostics cannot fix it, the provider books an engineer. That visit is usually free when the fault is on their side of the network boundary.
The no-access fee is the trap. If you are not home, or the engineer cannot reach the equipment, some providers charge for the wasted visit. UK charges in this category typically run from around £30 to £100 depending on the provider.
Be there. Make sure the ONT, master socket, and router are all reachable. Move furniture if you have to.
For internal wiring faults, some providers charge for the repair. Ask before you book. If the fault is at the master socket or beyond, it is almost always their cost.
Mistakes That Turn a 10-Minute Fix Into a Week-Long Ordeal
Most red-light cases resolve in under an hour. The ones that drag on for days almost always involve one of these errors.
Reset-happy troubleshooting that wipes your evidence
Every factory reset wipes the logs and signal values a technician would use. The device forgets what went wrong, and so do you.
Reset once, at most, and only after a power cycle has failed. If a reset does not fix it, do not reset again.
The same applies to rebooting. Each reboot clears the error counters. If your line is intermittently dropping, you want those counters intact.
Replacing equipment before you've confirmed the fault
Buying a new router is the most expensive way to learn your line is faulty. It is also extremely common.
Confirm the fault first. Test at the master socket. Bypass the router.
Check whether the ONT shows LOS. Only replace hardware when you have evidence pointing at the hardware.
If you rent your equipment from the provider, swapping it is usually free. Ask for a replacement before you buy anything.
Ignoring the light that actually matters
On a fibre setup, the ONT lights matter more than the router lights. On DSL, the DSL light matters more than the broadband light.
People stare at the router because it is the device they recognise. Meanwhile the ONT is flashing LOS, which tells the whole story.
Learn which lights belong to which box. It takes two minutes and saves hours.
Safety: what not to open, cut, or climb
Never open the ONT, the NID, or any provider-owned enclosure. They contain no user-serviceable parts, and you risk both injury and a repair bill.
Never climb a pole or a mast. Never cut a cable, even a dead-looking one. Never dig near a marked utility line.
Electrical safety matters too. Outdoor enclosures can hold a charge, and grounding faults are a real hazard. If you see scorch marks or smell burning, stop and call the provider immediately.
Fibre laser safety and why you never look into a connector
Fibre carries infrared laser light. You cannot see it. Your blink reflex will not protect you, because there is nothing visible to blink at.
The damage happens before you feel anything. Retinal burns from fibre are permanent.
Never look into the end of a fibre cable. Never look into an unconnected port on an ONT or a patch panel. Keep dust caps on.
If you need to inspect a connector, use a fibre inspection scope, which projects the image onto a screen.
If you suspect eye exposure, seek medical attention immediately. Do not wait for symptoms.
Costs, Fees, and Compensation: What You're Actually Owed
A red light that lasts days is not just an inconvenience. In many regions, it triggers a payment from your provider. Here is what applies where.
Truck roll fees and when they're waived
A truck roll is a technician visit. Providers charge for it when the fault turns out to be your own wiring or equipment.
Published fee schedules from major US providers typically fall between $50 and $150 for a no-fault visit. UK providers commonly charge £30 to £100 for a no-fault visit or a missed appointment.
The fee is usually waived when the fault is on the provider's side. That includes the drop cable, the street cabinet, the exchange, and the ONT. It also includes any fault found at the master socket test point.
Automatic compensation schemes and total loss of service
The UK has a formal scheme. Ofcom's automatic compensation rules require providers to pay when a repair is delayed or an appointment is missed.
Recent published rates have sat at around £9.76 per day for a delayed repair, £30.10 for a missed appointment, and £9.76 for each day a new service is not activated on the promised date. These figures are adjusted annually, so check Ofcom's current rate before you claim.
The US does not have a single national scheme. Some providers offer service credits under their own terms. Ask specifically for a service credit rather than a goodwill gesture, because the former is a defined remedy and the latter is discretionary.
Speed guarantees, minimum guaranteed speeds, and contract exit rights
Many providers now publish a minimum guaranteed speed. If your line consistently falls below it, you may have the right to exit your contract without penalty.
You usually have to document the shortfall over a set period, often 30 days. Run wired speed tests at different times of day and keep the results.
The FCC's broadband consumer labels, available through the FCC, show the speeds providers commit to. Those labels are useful evidence if you need to escalate.
Escalation paths when the first answer is no
Front-line support cannot authorise compensation. They can only follow their script.
Ask for a formal complaint. Get a complaint reference number. In the UK, the provider has eight weeks to resolve it before you can take it to the relevant ombudsman.
In the US, ask for the executive complaints team. Most large providers have one, and it operates outside the standard call centre.
Keep everything in writing where you can. Chat transcripts and emails are far stronger evidence than a phone call you cannot prove happened.
Resilience: Backup Connections, Alarms, and Vulnerable Users
For most people, a red light is an inconvenience. For some, it is a safety issue. Here is how to plan for both.
Who actually needs a backup connection
If you work from home full time, run a business from your property, or depend on a landline for medical reasons, you need a backup. Everyone else can usually wait out an outage.
The cheapest backup is a mobile hotspot on your phone. It costs nothing extra and covers short outages. Data caps and throttling are the limits.
A dedicated 4G or 5G backup router is the next step up. Some gateways support automatic failover, which switches over the moment the primary link drops.
Dual WAN routers with a second ISP are the top tier. That covers you against a provider-wide outage rather than just a line fault.
Telecare, medical alarms, and VoIP landlines
Traditional analogue landlines work during a power cut because the exchange powers the line. VoIP does not. If your phone service runs over broadband, a red light means no calls.
That matters for telecare alarms, pendant alarms, and any medical device that dials out. If you or someone in your household depends on one, tell your provider. Most maintain a priority services register for exactly this reason.
A mobile backup, a charged phone, and a neighbour who checks in are the practical safeguards. Do not assume the alarm will work when the broadband is down.
Priority services and vulnerable customer registers
UK providers maintain priority services registers for customers who need extra support. That includes people who are elderly, disabled, or medically dependent on their connection.
Being on the register changes how the provider handles your fault. You get priority restoration, proactive updates, and sometimes a named contact.
US providers run similar programmes under different names. Ask specifically for the priority or vulnerable customer list. It is not advertised, but it exists.
What to do during a long outage
Charge your phone. Charge a power bank. Let anyone who needs to reach you know the landline is down.
If you have a mobile hotspot, use it for essential tasks only. Video calls will burn through a data cap fast.
Log the outage with timestamps. You will need it for a compensation claim, and providers rarely volunteer the credit without evidence.
When to Change Provider or Technology Entirely
Sometimes the fix is not a repair. Sometimes it is a different connection altogether.
Repeated faults on the same line
If your line has failed more than three times in twelve months, you have a chronic fault. Providers have a process for this, usually called a chronic case review or a repeat fault investigation.
Ask for it by name. A chronic case review looks at the whole line rather than the current fault, and it can result in a full line replacement.
If the provider refuses and the faults continue, you have grounds to exit the contract. Document everything and ask for a deadlock letter if you need to escalate.
End-of-life copper and the copper switch-off
Copper networks are being retired. The UK's PSTN switch-off is well underway as of 2026, and several other countries are running similar programmes.
If your provider has told you your service is moving to full fibre, treat any red light as potentially migration-related. The old line may be decommissioned before the new one is fully active.
Ask for the migration date in writing. If the switch happens before your fibre is live, you will lose service, and knowing the dates gives you leverage for compensation.
When the technology no longer fits
DSL on a long rural line will never deliver modern speeds. Cable in a heavily oversubscribed node will slow down at peak times. Satellite has latency that no amount of hardware can hide.
If your connection keeps failing because of the technology itself, switching is the honest answer. Check what is actually available at your address using the provider's own coverage checker or the regulator's broadband map.
Do not trust a generic availability site. Use the provider's checker, because it reflects the real network record for your property.
What to check before you switch
Run a line check on the new provider before you commit. Confirm the technology, the speed tier, and the contract length.
Check whether the new provider charges for installation or activation. Check whether the equipment is rented or owned. Check the minimum guaranteed speed.
Read the early termination fee on your current contract too. Sometimes staying and fixing the fault is cheaper than switching. Sometimes it is not.
Do the maths before you decide.
Real Scenarios: Red Light After a Storm, a Move, or a Provider Switch
Context changes the diagnosis. Here is how the same red light means different things in different situations.
Blinking red after a power cut or brownout
Power events are the most common trigger. A brownout can leave a router half-booted and stuck in a red state.
Power-cycle in the right order. ONT first, wait for green, then the router. That sequence alone fixes most post-outage red lights.
If it does not recover, check for a blown power adapter. Surges kill power bricks more often than they kill routers, and a replacement adapter is cheap.
Blinking red on a brand-new router or self-install kit
New hardware almost always fails at provisioning or authentication. The physical link is usually fine.
Check that the service is actually active on your account. Check the activation date, which is sometimes later than the delivery date.
If you swapped from a provider-supplied gateway to your own, check for MAC address locking and VLAN requirements. Both will produce a red light on hardware that is working perfectly.
Blinking red after moving house
Moving is the highest-risk scenario for a red light. The service at the old address has to be ceased, and the service at the new address has to be activated.
If the previous occupant left equipment behind, do not use it. Provider gateways are usually locked to the account and the line.
Check that the activation date has passed and that the line is live. In the UK, an Openreach activation can slip by a day or more, and the red light simply reflects that.
Blinking red but Wi-Fi still works
This combination confuses people, but it has a simple explanation. Your Wi-Fi is working because the router's wireless radio is fine. The broadband light is red because the link to the provider is not.
Devices may still connect to the router and even reach each other. They just cannot reach the internet.
Some routers also keep a cached page or a captive portal running, which makes the connection look half-alive. It is not. Test with a wired connection and a fresh browser tab.
Blinking red on a device that's been fine for years
Age-related failures are real. Power supplies degrade, capacitors bulge, and thermal damage accumulates in routers that run hot for years.
Check the power adapter first. Then check whether the device is still supported by your provider. Many older gateways cannot handle current network profiles, especially after a DOCSIS or fibre upgrade.
If the hardware is more than five years old and the line tests clean, a replacement is a reasonable next step. Ask your provider for one before you buy.
Frequently Asked Questions
Is it my router or my ISP?
It depends on which light is red. If the ONT shows LOS, the fault is on the provider's network. If the ONT is green and only the router is red, the problem is between those two boxes or in the router itself.
Test direct to the modem to confirm.
Can I fix a blinking red broadband light myself?
Often, yes. Power-cycle in the right order, reseat every cable, and test at the master socket. Those steps fix many cases.
If the ONT shows LOS or the line tests faulty, you need your provider. Do not open provider-owned equipment.
How long until it comes back?
A transient sync loss recovers in two to three minutes. A confirmed area outage usually clears within a few hours. A line fault requiring a technician typically takes one to three working days.
UK providers under the Ofcom scheme aim for two working days.
Will a new router fix it?
Only if the router is the faulty part. Most red lights come from the line, the ONT, or the provider's network, and a new router will not help. Confirm the fault first with a direct-to-modem test.
If you rent your equipment, ask for a free swap before buying anything.
Should I keep rebooting it?
No. Reboot once in the correct order and wait two full minutes. Repeated reboots wipe the error logs and signal values a technician needs.
If the light stays red after one clean power cycle, the problem is not a stalled session.
Do I need a technician visit?
Only if remote diagnostics cannot fix it. Ask for a line test first. If the test shows a fault on the provider's side, they will usually book an engineer at no cost.
If the fault is in your internal wiring, you may be charged.
The Decision Guide: Your Next Step, Mapped
Here is the whole thing collapsed into one path. Start at the top and follow the branch that matches what you see.
If the ONT or LOS light is red
Stop. Your fibre connection has no signal, and nothing inside your home can fix that.
Power-cycle the ONT once. Wait three minutes. If LOS stays red, call your provider and ask for an optical power test.
Do not open the ONT. Do not unplug the fibre. Do not look into the connector.
If the router broadband light is red but the ONT is green
The fault sits between the ONT and the router, or inside the router itself.
Check the Ethernet cable first. Reseat both ends. Try a different cable if you have one.
Then log into the admin page and check the WAN status. If it says Authenticating, you have an account or credentials problem. If it says Disconnected with no IP, you have a link problem between the two boxes.
If all lights are off or the router will not boot
You have a power problem, not a network problem.
Test the socket with another device. Try a different power adapter if you have a matching one. Check for a bulging capacitor or a burning smell.
If the router is provider-supplied, ask for a free replacement. Do not buy a new one until you have confirmed the power adapter is not the cause.
If it is red, intermittent, and recovers on its own
You have a marginal line. This is the most frustrating pattern because it looks fine when the engineer arrives.
Keep a fault log. Date, time, duration, and weather conditions. Run a wired speed test each time it drops.
Ask your provider for a stability profile change on DSL, or a signal level check on cable. Intermittent faults are diagnosed from patterns, not from single readings.
If neighbours are also down
You have an area fault. Nothing you do indoors will help.
Report it once with your postcode and the start time. Then wait. Rebooting repeatedly achieves nothing and wipes your logs.
Check the provider's status page every few hours. If the fault runs past 48 hours, ask for a service credit.
If you have already replaced the router and it is still red
The router was never the problem. You have a line fault, a provisioning issue, or a provider-side failure.
Stop buying hardware. Ask for a line test, and quote the fault reference from your earlier calls. If the provider keeps closing tickets without fixing anything, escalate to a formal complaint.