Trivetek — Home network measurements | trivetek.com
Hold test conditions constant and log idle and loaded behaviour so runs can be compared.
Fix the conditions before the first run: same device, same server, same test duration, no background sync or streaming. Note the time of day, because ISP congestion and neighbourhood load change results across the day.
Take a wired baseline with three runs at least two minutes apart. Then repeat over Wi-Fi in the same room as the router and again where the problem appears. Comparing like with like is what makes the difference meaningful.
On narrow screens, swipe or scroll the plate sideways.
Add a loaded run: start a large download and measure latency and packet loss during it. If idle latency is fine but loaded latency rises by hundreds of milliseconds, queueing rather than raw line speed is the bottleneck.
The FCC fixed-broadband methodology uses three concurrent TCP connections for download and upload throughput measurements, so a single-stream test can under-report a fast line. Use a multi-connection test for throughput and a separate latency test for timing.
The FCC thirteenth report describes idle-latency tests using approximately 2,000 packets per hour, which shows that latency sampling is a deliberate measurement rather than a by-product of a download.
Further reading