Your powerline adapter reports a gigabit link. It is not delivering one.

If your wired connection is slower than your Wi-Fi, something is wrong — and if the wire runs through powerline adapters, you already know what.

This one is worth writing up because the diagnostic evidence is genuinely misleading. Every indicator says the link is fine.

The measurement that fools everyone

Your PC's network card negotiates its link with whatever it is plugged into. With a powerline adapter, that is the adapter — sitting right there on the wall, connected by a short Ethernet cable.

So the card reports 1 Gbps, full duplex, zero errors. All of which is true, and none of which tells you anything about what happens after the signal leaves the adapter and travels across your house's electrical wiring.

That second leg is the actual link, and it is where the throughput goes. Real-world powerline performance is typically 50–150 Mbps regardless of the "AV1000" or "AV2000" number on the box, and it degrades with distance, circuit topology and electrical noise.

We measured a machine recently that showed a perfect gigabit link and delivered 70 Mbps — while the Wi-Fi on the same machine, to the same router, delivered 160. The cable was less than half the speed of the wireless.

How to prove where the ceiling is

Measure the same thing on both paths. Run a sustained transfer over the wired connection, then unplug it and run the identical test over Wi-Fi. If the numbers differ substantially, the limit is per-link, not upstream. If they land on the same number, the bottleneck is your router or your ISP.

Use parallel streams to rule out TCP effects. A single connection can be limited by latency and window size rather than by capacity. If four simultaneous downloads total roughly the same as one, you are at a hard ceiling. If they total four times as much, your single-stream test was measuring the wrong thing.

Check the error counters. A genuinely faulty cable or port shows up as errors and discards:

Get-NetAdapterStatistics -Name 'Ethernet' |
  Select-Object ReceivedPacketErrors, ReceivedDiscardedPackets, OutboundPacketErrors

Zeros across the board, combined with a link speed far above your throughput, points at something beyond the first hop — which is exactly the powerline signature.

Sanity-check your measurement tool. Test against more than one server. A single slow endpoint will convince you your connection is broken when it is not.

Why powerline degrades over time

Powerline performance is a function of your electrical wiring, and your electrical wiring changes.

• A surge protector or power strip will cripple it. They filter the frequency range powerline uses. Both adapters must go directly into wall outlets — this alone accounts for an enormous share of "it used to be faster" reports.
• New appliances inject noise. LED drivers, phone chargers, motors, space heaters and treadmills all degrade the signal.
• Different circuits or phases force the signal across the breaker panel, which costs a great deal of throughput.

If it was faster before and is slower now, one of those three changed.

What to do instead

Ranked by how close it gets you to the speed you are paying for:

1. A real Ethernet run. Unglamorous, and the only option that reliably delivers gigabit.
2. MoCA over existing coax. If you have cable TV or a cable modem, the coax is already in your walls. MoCA 2.5 is genuinely gigabit-capable, unlike powerline, and it is the best answer when running new cable is impractical.
3. Modern Wi-Fi. A current Wi-Fi 6 card will often beat powerline outright — as it did on the machine above, by more than double.
4. Tuning the powerline. Wall outlets only, same circuit where possible. Worth ten minutes; will not reach gigabit under any circumstances.

A note on upload

One detail worth checking separately: measure your upload speed on both paths. If download differs between paths but upload is identical, the upload limit is upstream of your house — your router or your ISP — not your local link.

That distinction saves a pointless support call. An upload figure around 35 Mbps on a connection advertised as symmetric gigabit usually means the plan is not what it was believed to be.