Router keeps disconnecting from the internet — different from Wi-Fi drops on individual devices. If the router itself is losing its upstream connection (ISP connection fails), that affects every device on the network simultaneously. If only some devices drop, or if the router admin panel still shows internet connected when devices lose access, that’s a different problem. This fits into the wider topic we cover in our Complete Guide to Fixing Windows, Browser, and Software Errors.
Work out which you’re dealing with: when the internet drops, log into the router admin panel (usually 192.168.1.1 or 192.168.0.1) → look at the WAN status or Internet status page. If it shows “Connected” and shows a valid public IP address — the router has internet, and the problem is between the router and your devices. If it shows “Disconnected,” “No IP,” or “Reconnecting” — the router itself is losing the ISP connection.
ISP line issues — check before touching the router
Intermittent connection drops from the ISP side are frustratingly common and easy to misdiagnose as a router problem. Signs this might be ISP-side: drops happen at similar times each day, drops affect cable or DSL modem even when the router is bypassed, other people in the area report the same problem.
Check outages: most ISPs have a status page or Twitter/X account where they post outages. Also check whether the modem (the box from your ISP) has any error lights — particularly the “Internet” or “Online” light flashing or showing amber/red. If the modem itself is losing connection, the router settings don’t matter.
Log into the router and look at connection logs if available — these often show “PPPoE dropped” or “DHCP renewal failed” with timestamps that might reveal a pattern (same time each day = likely a scheduled ISP maintenance window).
Overheating — extremely common and overlooked
Routers run 24/7 and generate more heat than most people realise. Consumer routers without adequate airflow, sitting in an enclosed cabinet or TV unit, overheat and drop their internet connection as a protective measure. The connection recovers after a few minutes, creating an intermittent pattern that looks like a software or ISP issue.
Touch the top of the router during a drop. If it’s very hot — hot enough to be uncomfortable to hold your hand on — overheating is a real possibility. Move it somewhere with better airflow, oriented so the vents aren’t blocked, and observe whether the drops stop. This is especially relevant in summer months or in rooms without air conditioning.
Firmware update
Log into the router admin panel → look for “Firmware update” in Administration, Advanced, or About sections. Many consumer routers check for updates automatically but don’t install them without manual confirmation. An update that fixes a known disconnection bug could be sitting there uninstalled.
This is worth doing regardless of whether it’s the immediate fix — outdated router firmware is one of the most common sources of mysterious connectivity issues, and manufacturers push fixes for disconnection bugs regularly.
MTU mismatch causing connection drops
If the router’s MTU (Maximum Transmission Unit) setting doesn’t match what the ISP expects, certain large packets get dropped, which can cause connections to drop intermittently — particularly on specific types of content or applications. DSL and PPPoE connections commonly require MTU of 1492 rather than 1500; some ISPs use 1400 or even lower.
Router admin → WAN settings → MTU value. If it’s 1500 on a DSL or PPPoE connection, try 1492. If on cable internet, 1500 is usually correct. Some router interfaces have an “Auto” option that negotiates MTU from the ISP — if available, that’s the safest setting.
DNS settings causing apparent disconnections
DNS server failures produce a symptom that looks like internet disconnection: pages stop loading, apps can’t connect, everything appears offline. But the router’s internet connection is actually fine — only domain name resolution has failed.
When the internet “drops,” try visiting a website by IP address instead of by name: in a browser, type 1.1.1.1 (Cloudflare). If that page loads when regular URLs don’t, DNS is failing, not the internet connection. Fix: router admin → DNS settings → change from ISP-provided DNS to 1.1.1.1 and 8.8.8.8 as the DNS servers. The router then uses reliable public DNS instead of the ISP’s potentially flaky servers.
Interference and wireless backhaul issues (mesh systems)
Mesh routers (Eero, Google Nest WiFi, Orbi, TP-Link Deco) use wireless backhaul — the nodes communicate with each other via Wi-Fi. When the backhaul connection between nodes becomes unstable, the secondary nodes lose their internet connection even if the primary node still has it. Devices connected to secondary nodes experience drops while devices connected to the primary node are fine.
Check the mesh system’s app for node connection status. If secondary nodes show “Offline” or weak backhaul signal — move them closer to the primary node, or use a wired backhaul connection (ethernet between nodes) if the system supports it. Interference from neighbouring 5 GHz Wi-Fi networks can degrade mesh backhaul significantly in apartment buildings.
DHCP lease renewal drops
Routers renew their WAN IP address lease from the ISP periodically. If the DHCP renewal fails — ISP server timeout, brief signal issue — the router briefly loses connectivity during the retry. ISPs using very short DHCP leases (1–4 hours) cause more frequent renewal events, increasing the chance of a failed renewal.
Router admin → WAN status → note the DHCP lease time remaining. If it’s short (under 2 hours) and drops correlate with lease renewal times, ask the ISP whether the lease time can be increased, or set a static WAN IP if the ISP supports and allows it.
For home network connection issues from the device side (rather than the router), the connected but no internet guide covers the Windows-side diagnostics. And for understanding whether it’s really the router or the ISP at fault, our DNS troubleshooting context explains the difference between DNS failures and actual connection drops. Microsoft’s home networking guide covers the diagnostic tools available in Windows 11 that help identify whether the connection drop is at the device, router, or ISP level.
Cable quality and physical connections
The coax or phone line between the wall and the modem is easy to overlook. A damaged cable, a loose F-connector on coax, or water ingress in an outside junction box causes intermittent signal loss that the modem (and router) experiences as a connection drop. Drops that correlate with weather — happening more during rain — are a strong indicator of an outdoor cable or connection issue.
Check all the physical connections you can access. The coax connector from the wall to the modem should be hand-tight. If the cable has visible kinks, sharp bends, or damage, replacing it is worthwhile. For anything outside the house — junction boxes, street-side connections — that’s ISP territory to fix.
Channel bonding and modem issues on cable internet
Cable internet modems use channel bonding — they bond multiple channels together for higher speeds. When channel bonding configuration becomes unstable (upstream power too high or too low, SNR degradation), the modem drops and reacquires channels, causing brief disconnections. The modem’s status page (typically at 192.168.100.1) shows signal levels and upstream/downstream channels.
For upstream power levels: 38–48 dBmV is ideal. Above 50 dBmV indicates the modem is working too hard to transmit and drops are likely. This is a cable plant issue — the ISP needs to adjust signal levels at the tap outside. Report the specific upstream power readings when you call support; technicians respond more quickly when you can quote exact values.
QoS and bandwidth saturation causing apparent drops
When the internet connection is fully saturated — someone downloading large files, a 4K stream running, a cloud backup in progress — routers without proper QoS handling can cause all connections to drop briefly while they manage the overloaded buffer. The connection doesn’t actually disconnect; packets are queued too long and connections time out waiting for responses.
Signs: drops only happen when someone else is using lots of bandwidth. Drops resolve within 30–60 seconds. Router admin → QoS or Traffic Management settings → enable QoS and set your total upload/download bandwidth limits to about 80% of your actual ISP speeds. This prevents buffer bloat from causing the timeouts that look like disconnections.
Rebooting the router — when and how often
Consumer routers, unlike commercial networking equipment, accumulate state over time. Memory fragmentation, stale ARP tables, and accumulated log data can cause instability after months of continuous operation. A reboot clears all of this.
If drops are happening: a reboot is worth trying first before any configuration changes — it costs nothing and fixes a surprising number of issues. If drops return within a day or two after the reboot, the underlying cause is still present; the reboot was temporary relief. If the router is stable for weeks after a reboot and then drops start again, scheduling a monthly reboot (most router admin pages have a scheduled reboot option under Administration) prevents the accumulation problem.
When to involve the ISP
Some router disconnection problems genuinely require ISP intervention: degraded signal at the tap, faulty outside wiring, provisioning errors on the ISP’s systems, or a failing modem that’s ISP-supplied equipment. If the modem itself is provided by the ISP and it’s losing its connection to the ISP’s network, the problem is outside your control.
The most useful thing to do before calling: keep a log of when drops happen (time, duration, what the router admin page showed during the drop) and what the modem’s signal levels look like. This documentation turns a 45-minute call into a 10-minute call because the technician can see a pattern and doesn’t spend time asking you to reproduce it in real-time.
ISP-provisioned modem vs your own
If you’re renting a modem from the ISP and it keeps losing connection: request a replacement. ISP equipment wears out. They won’t always proactively replace it. Asking for a swap is legitimate and they’ll usually send a replacement within a few days. If you own your modem and it’s 4+ years old, it might simply be reaching end of life — modem replacement is often the fix for persistent drops that survive all other troubleshooting.
Ping test to track drops objectively
Rather than waiting to notice a drop manually, run a continuous ping to an external server to capture exactly when disconnections happen and how long they last. Open Command Prompt and run:
ping 8.8.8.8 -tLeave this running. Each successful response shows the latency. Timeouts appear as “Request timed out.” When a drop happens, you’ll see exactly when it started (first timeout) and when it recovered (first response after timeouts). The duration and pattern of timeouts — seconds vs minutes, regular vs random — reveal whether it’s a brief ISP blip or a sustained outage.
Also run ping 192.168.1.1 -t simultaneously in a second window (pinging the router itself). If pings to 8.8.8.8 time out but pings to 192.168.1.1 succeed during a drop, the router’s connection to the ISP failed but the router hardware is working. If both time out simultaneously, the router itself became unreachable — overheating or hardware failure is more likely.
IPv6 prefix delegation drops
ISPs that provide IPv6 use Prefix Delegation to assign IPv6 address blocks to routers. Some ISPs have buggy Prefix Delegation servers that drop and re-assign prefixes frequently, causing brief disconnections specifically for IPv6 traffic. IPv4 traffic on the same connection remains stable.
This is hard to notice because most troubleshooting focuses on IPv4. If the ping test to 8.8.8.8 (IPv4) shows stable connectivity while connectivity to IPv6-only services drops periodically, IPv6 prefix delegation instability is the cause. The fix: router admin → IPv6 settings → disable IPv6 entirely (temporary but immediate fix) or contact the ISP about their Prefix Delegation server stability.
High traffic times and shared infrastructure
Cable internet uses shared bandwidth in a neighbourhood — the available speed is distributed among all users in the cable segment. During peak hours (evenings, weekends), the segment can become congested enough that some connections are shed. Drops that happen consistently in the evening and resolve by midnight are almost certainly this pattern. If this sounds familiar, USB Device Keeps Disconnecting is worth a look.
This isn’t fixable from the home side — it’s a capacity issue on the ISP’s infrastructure. Document the pattern and report it to the ISP as congestion during specific hours rather than as a connectivity problem. ISPs take congestion reports more seriously when they include specific times and ping data, because it helps them identify which cable nodes are oversubscribed. If the pattern persists after reporting, switching to a provider with fibre infrastructure (rather than cable) or DSL may resolve it — fibre and DSL aren’t shared in the same way cable segments are. Our guide on Ethernet Keeps Disconnecting covers an adjacent issue.
