Real-World Powerline Networking Benchmarks
A practical look at HomePlug AV2 performance across outlets, with iperf3 & ping
TL;DR: Powerline can be acceptable—or terrible—depending on the outlet. In this home, one office outlet delivered ~86/51 Mbit/s (up/down) with low latency, while another office outlet managed only ~12/4 Mbit/s, far worse than Wi-Fi.
Testbed & Methods
These measurements were collected in a single residential home using a TP-Link TL-PA7010 (primary) + TL-WPA7510 (extender) HomePlug AV2 kit and a Netgear Nighthawk access point (AP)/switch. The primary adapter was Ethernet-wired to the AP; the client was usually a Dell Latitude laptop wired to the powerline extender. A Minisforum HM80 mini-PC on 5 GHz Wi-Fi served as the Wi-Fi control in the office, on another floor and across the house from the AP.
Note on hardware: My powerline kit is several years old. The TL-WPA7510 and TL-PA7010 are HomePlug AV2 adapters rated for up to 1000 Mbps (AV1000) and include a single Gigabit Ethernet port. The TL-WPA7510 also provides dual-band AC750 Wi-Fi (300 Mbps on 2.4 GHz and 433 Mbps on 5 GHz). These were mid-range when new, but they are not the fastest options available today. I rediscovered them after they were sitting in a drawer for five years, so I decided to run proper performance tests.
Metrics
- iperf3 forward (client to router): upload capacity
- iperf3 reverse (router to client): download capacity
- Ping: min/avg/max RTT, loss
Commands
iperf3 -c 10.0.0.1 # forward (upload)
iperf3 -c 10.0.0.1 -R # reverse (download)
ping 10.0.0.1 -n 100 # Windows, typical run
Note: iperf is LAN-LAN and reflects the link & LAN stack, not ISP bottlenecks. A wired control (Dell laptop wired to the AP) hit ~750-850 Mbit/s iperf with 1ms ping while Internet speedtests on that same link were ~175/36 Mbit/s, illustrating that the WAN pipe is lower than the LAN capabilities.
Results
Primary question: Should I switch my office from Wi-Fi to Powerline?
The main comparison is Wi-Fi 6 (5 GHz) vs Powerline, both in the office. We report medians to reduce run-to-run noise. These tests were in the same physical location, my office.
Scenario (Office) | Client | iperf3 forward (upload) | iperf3 reverse (download) | Ping avg |
---|---|---|---|---|
Wi-Fi control | Minisforum HM80 | 35.85 Mbit/s | 15.40 Mbit/s | 5.45 ms |
Powerline | Dell laptop | 11.60 Mbit/s | 4.35 Mbit/s | 9.0 ms |
At this outlet, Powerline was ~68% lower on upload and ~72% lower on download than Wi-Fi, with slightly higher latency. |
Outlet sensitivity: same room, different outcomes
Changing the office outlet dramatically changed Powerline results:
Powerline Outlet (Office) | iperf3 forward (up) | iperf3 reverse (down) | Ping avg | Verdict vs Wi-Fi |
---|---|---|---|---|
East wall (Test 6) | 86.3 Mbit/s | 51.3 Mbit/s | ~3 ms | Much faster than Wi-Fi |
South wall (Test 7) | ~29-35 Mbit/s | ~16.8 Mbit/s | ~5 ms | Comparable to slightly worse |
West wall (Test 9) | ~9-12 Mbit/s | ~4.2-4.5 Mbit/s | ~9 ms | Worse than Wi-Fi |
What's possible near the base unit?
When the PL extender is near the primary PL adapter (same room), Powerline throughput was strong:
Scenario | iperf3 forward (↑) | iperf3 reverse (↓) | Ping |
---|---|---|---|
Test 2 | ~141-149 Mbit/s | ~151-157 Mbit/s | avg 2-3 ms |
Test 3 | ~137-146 Mbit/s | ~165-166 Mbit/s | avg ~3 ms |
Wired control (for context)
The laptop wired directly to the Nighthawk achieved ~750-854 Mbit/s iperf, ping ~0-1 ms. Internet speedtests on that same link were ~175/36 Mbit/s due to the WAN cap. This underscores why iperf is the right tool for evaluating Powerline: it measures the LAN link itself, not your ISP.
Discussion: Why does outlet choice matter so much?
HomePlug AV2 uses wideband OFDM and, in its MIMO mode, can transmit across multiple wire pairs (Line-Neutral, Line-Ground, Neutral-Ground). This MIMO capability improves robustness and throughput, but it requires grounded 3-prong outlets to exploit those additional paths; 2-prong connections limit options and may reduce performance.
Also, the familiar "up to 1000 Mbps" rating is a physical-layer maximum under ideal conditions; real application throughput is much lower due to protocol overheads, electrical noise, and home wiring characteristics. Performance varies by outlet, distance, and circuit topology (breakers/phases/hops).
Modern Alternatives
Today's top powerline adapters are much faster and often include Wi-Fi 6. For example, the Devolo Magic 2 Wi-Fi 6 supports speeds up to 2,400 Mbps and offers mesh Wi-Fi with two Gigabit Ethernet ports. Another strong performer is the TP-Link TL-PA9020P KIT, which provides dual Gigabit ports and excellent wired performance for AV2 powerline. For those who need the absolute best wired speeds and have coaxial cabling, MoCA 2.5 adapters like the Actiontec ECB7250 or goCoax MoCA 2.5 deliver up to 2.5 Gbps over coax, far exceeding what powerline can achieve in most homes.
Powerline and MoCA are never plan A. If you can use wired Ethernet, it's the best option.
Key Findings
- Outlet matters. In this home, Powerline ranged from sufficient(165/140 Mbit/s) to poor (12/4 Mbit/s) in the same room.
- Under better conditions, powerline speeds were symmetrical, and under worse conditions download bandwidth was higher.
- With the extender near the base, Powerline delivered ~140-165 Mbit/s. That is much lower than its nominal rating of 1000 Mbps, and Wi-Fi or Ethernet could do much better in the same room.
- Versus Wi-Fi in the office: Powerline was either clearly better or clearly worse, depending on outlet.
- Latency was consistently low (single-digit ms) on good Powerline links and competitive with Wi-Fi.
Recommendations
- Test multiple outlets in the target room; pick the one with the best iperf forward/reverse and low jitter.
- Use grounded, 3-prong wall outlets to enable AV2 MIMO paths (if your hardware supports it) and to reduce attenuation.
- Plug adapters directly into the wall; avoid surge protectors/filtered strips.
- If feasible, relocate the base unit to reduce breaker/phase "hops" to the destination outlet.
- Validate with iperf3 (both directions): run multiple trials and use median values.
No comments:
Post a Comment