Backup mobile mode held a 500 Mbps service near 24 Mbps

A useful diagnosis starts by separating the service, access equipment, router, local link and client device. In this record, both the desktop Ethernet test and WiFi test stayed near 25 Mbps after a self-install, which pointed away from a single client device. That observation mattered because it reduced the number of plausible explanations and kept the work tied to measurements or account information that could actually be checked.

The next step was proportionate to that evidence: used video guidance to correct the Telstra modem installation and stop traffic relying on the mobile-backup path. No unrecorded hardware, carrier policy, plan feature or building condition has been added. Where the source stopped at an appointment or escalation, this page also stops there rather than presenting a later repair as complete.

• Write down the expected service state or plan rate before changing anything.
• Test one known path at a time and record the device, connection type and result.
• Check whether the carrier sees the account and access equipment online.
• Change only the layer supported by the evidence, then repeat the same test.

After the recorded work, the phone over WiFi and desktop over Ethernet both measured about 500 Mbps, while reported game latency fell from 40 ms to 18 ms. The comparison is useful because it connects the intervention to an observable outcome. It should not be read as a promise that another property, plan or device will produce the same number or turnaround time.

Backup-mode behaviour depends on the modem and service state. A slow result should be verified before changing settings or removing any SIM. Before resetting equipment, reinstalling a system or changing a service, preserve any credentials and data that may be needed for recovery. If the evidence points upstream, collect concise test results and escalate them instead of repeatedly replacing local equipment.

A final verification should mirror the original complaint. Use the same client where possible, note whether the link is wired or wireless, and record the time. For intermittent faults, one successful test is evidence of capability, not proof of long-term stability, so continued observation may still be appropriate.

For a similar issue, the practical takeaway is simple: establish a baseline, isolate one layer, keep the test conditions consistent, and describe both the confirmed result and what remains unverified. That makes the next technical or carrier conversation faster and avoids turning a plausible theory into an unsupported claim.