Features
Part of Fixed wireless: questions about the unit, path and hub
Fixed wireless trials: what to log for two weeks and how to read it
A two week fixed wireless trial, not someone else's review: what to measure hourly and per minute, and what each pattern in your log means for the link.
Someone else's fixed wireless link is aimed at a different mast, through different trees, on a sector with different neighbors. Their review is a fact about their roof.
This page is the trial that replaces it, and how to read the log it produces.
What to take away
- Two weeks of your own measurements are worth more than any number of ratings, and most providers will allow a trial or a short initial term.
- Log four thingsthroughput both ways, delay, variation in delay, and the moments the link drops. Averages hide all four.
- The patterns in the log are diagnostic. A daily dip, a rainy-day dip and a random dip mean three different things.
What to measure, and how often
| Measure | Frequency | Why it matters |
|---|---|---|
| Download and upload, separately | Every hour, automatically | The upload is the direction calls and backups use |
| Round-trip delay to a fixed destination | Every minute | Delay is what makes conversation awkward |
| Variation in that delay | Every minute | Variation, not average, is what a listener hears |
| Drops, with timestamps | Continuously | Four short drops a day matter more than a lower average |
| Signal readings from the radio | Every hour | This is what separates a radio problem from a network one |
| Weather | Daily note | It is the variable you cannot control and must account for |
Run the measuring device wired to the router. A wireless measurement measures
What to log for two weeks
- Download and upload, separatelyevery hour
- Round-trip delay to fixed destinationevery minute
- Variation in that delayevery minute
- Drops, with timestampscontinuously
- Signal readings from the radioevery hour
- Weatherdaily note
Reading the patterns
A dip at the same time every weekday. The sector is congested. This is
Reading the log patterns
Dip same time every weekday?
sector congestion; ask sector questions
check next pattern
about how many customers the provider has put on it, and it gets worse as they sell more. Ask the sector questions before you commit.
A dip during rain, recovering after. Rain absorbs these frequencies. If the drop is small, it is a fact of the technology. If the link fails in ordinary rain, the margin on the path is too thin, which usually means alignment, mounting, or a sight line that is worse than the survey admitted.
A gradual decline over weeks. Something is moving or growing. A mount that has loosened, an antenna that has drifted, or foliage coming into leaf.
Random short drops with good signal. The radio path is fine and the fault is elsewhere: the equipment inside, the power feed, or the provider's network.
Good throughput, poor calls. Delay and variation are the problem, not capacity. VoIP systems covers what a call actually needs.
The parts a review can tell you
Not the link. The company. Whether an engineer turns up, how long a
realignment takes, whether the bill matches the contract, and how a cancellation goes. Those are the same for every customer and worth reading about. Anything describing speed is describing a roof you cannot see.
Ask the provider for a reference on the same mast rather than in the same town. Sector is the unit that matters, and a customer three streets away may be on a different one.
Ending the trial honestly
Write the verdict down before you renegotiate, and write it against the
Ending the trial honestly
- Write the verdict before renegotiating
- Write it against the requirement, not the advertisement
- Sentence onewhat the link delivered at busy hour
- Sentence twowhat it did in worst weather
- If either sentence is uncomfortable, choose a second path
If either sentence is uncomfortable, the answer is a second path rather than a better trial. Holding one trial's verdict against another provider's is the comparison provider comparisons sets out.
Federal preparedness guidance on power outages covers the failure that no radio survey addresses and that every trial should include at least once.
Common questions
Will a provider let us trial before committing?
Many will offer a short initial term or a cancellation window instead. Ask for one in writing; a refusal is itself a data point.
How long is long enough?
Two weeks including a weekend and, ideally, a wet day. A month is better if your year has a peak coming.
The installer says the signal is good. Is that enough?
It is one reading at one moment in fair weather. Signal quality and delivered capacity are related but not the same, and neither predicts the busy hour.
Should we test the backup path too?
Yes, and by unplugging the primary during working hours. A failover nobody has watched is a plan, not a backup.
