> For the complete documentation index, see [llms.txt](https://docs.premium-positioning.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.premium-positioning.com/learn/testing-a-correction-service.md).

# Testing a correction service

How to test an RTK correction service yourself: check coverage at your own sites, measure accuracy against a known point, log availability over several days, and read the result.

A spec sheet gives one accuracy figure in good conditions. A test shows what a service delivers where you work, including on the days that are not good. Test any service, RTK Premium included, on coverage, accuracy and availability, and use the same receiver, antenna and settings for every service you compare.

### Coverage at your sites

A coverage map shows where a service is offered, not where a fix holds. Connect at the sites you actually work, including the edges of your area, and confirm that corrections arrive and the receiver reaches a fixed solution. A connection that opens but delivers no corrections counts as no coverage, so watch the correction age as well as the connection status.

### Accuracy against a known point

1. Choose a survey control mark with published coordinates, in the reference frame the service delivers or one you can transform to.
2. Set up over the mark and measure the antenna height.
3. Connect to the service, wait for a fixed solution, and log one position per second for at least 15 minutes.
4. Compare the logged positions with the published coordinates in east, north and up.
5. Repeat on different days and at different times of day. Ionospheric conditions vary, so one session gives an indication, not a verdict.

To compare two services, run them at the same time on two identical receivers fed from one antenna through a splitter, so both see the same sky. Where that is not possible, alternate sessions on the same mark with the same setup.

### What to record

* Time to first fix: how long the receiver takes from connection to a fixed solution.
* Fix availability: the share of the session spent in a fixed solution.
* Trueness: the average offset of the positions from the known coordinate, in east, north and up.
* Precision: how widely the positions scatter around their own average.
* Correction age: a few seconds is normal, and more than 5 seconds points to a struggling data link.

### Availability over time

A single session shows one moment. Availability shows over days and weeks: leave a receiver or a logging client connected for several days, record every interruption and its length, and compare the result with the provider's published service level. [How to compare correction services](/learn/how-to-compare-correction-services.md) lists what to ask about that commitment.

### Reading the result

* A large average offset with little scatter is a bias, not noise. It most often comes from a difference in reference frame or epoch between the service and the known coordinate. See [Reference frames and constellations](/learn/reference-frames-and-constellations.md).
* Scatter shows how repeatable a service is. Two services can hold the same average and still differ in how far single positions wander.
* A fixed status alone does not prove the position. A receiver can report fixed with a wrong solution, which [Inconsistent RTK accuracy on site](/troubleshooting/inconsistent-rtk-accuracy-on-site.md) explains.
* One mark stands for one location. Your sites differ in sky view, multipath and distance to the network's stations, so test where you work. [Accuracy expectations](/learn/accuracy-expectations.md) covers the conditions that affect the result.

### Testing on RTKTEST.com

Where no control mark is at hand, RTKTEST.com, the open benchmark Premium Positioning runs, compares up to three services against a surveyed point in one 15-minute session. See [RTKTEST](/learn/rtktest.md).

### Testing RTK Premium

The 14-day free trial gives full access to the RTK Premium network on your own receiver. [Start a free trial](https://www.premium-positioning.com/start-14-day-free-trial?utm_source=docs\&utm_medium=referral\&utm_campaign=docs\&utm_content=testing-a-correction-service), then connect with the settings in [Connection settings](/set-up-and-connect/connection-settings.md).

### FAQ

How long should an RTK accuracy test run? Log one position per second for at least 15 minutes in a fixed solution, and repeat the test on different days and at different times of day.

Can I test RTK Premium on my own receiver before buying? You can. The 14-day free trial gives full access to the RTK Premium network on your own receiver.

How do I compare two correction services fairly? Run them at the same time on the same mark, ideally on two identical receivers fed from one antenna, with the same settings for both. RTKTEST.com does this on a surveyed point for up to three services at once.

Why does a service show a steady offset from the known point? A large average offset with little scatter is a bias, most often a difference in reference frame or epoch. See [Reference frames and constellations](/learn/reference-frames-and-constellations.md).

What correction age is normal? A few seconds is normal. An age that climbs above 5 seconds points to a struggling data link rather than to the correction service.

Last updated: 15 October 2026