T-Mobile claims it didn’t lie about 4G coverage, says FCC measured wrong

T-Mobile says the Federal Communications Commission screwed up 4G measurements in a report that accused the carrier of exaggerating its mobile coverage. The FCC report "incorrectly implies, based on a flawed verification process, that we overstated coverage," T-Mobile said in an FCC filing Monday.

The FCC staff report, issued in December, found that Verizon, T-Mobile, and US Cellular exaggerated their 4G coverage in official filings. As the FCC said, "Overstating mobile broadband coverage misleads the public and can misallocate our limited universal service funds."

The FCC relies on carriers' submissions to determine which parts of the country receive government funding to expand broadband access. The disputed submissions are among those the FCC is using to distribute up to $4.5 billion in Mobility Fund money over the next 10 years.

The dispute over whether T-Mobile exaggerated its coverage won't affect any punishment because the FCC already said it wouldn't punish the carriers in any way, even though making false filings violates federal law. The FCC said that its investigation "did not find a sufficiently clear violation" of data-collection requirements to warrant any penalties.

However, figuring out who's right in this dispute could affect the distribution of government funds and whether the FCC or T-Mobile need to make changes to their measurement processes.

It wasn't only the FCC that found T-Mobile exaggerated coverage. Small rural carriers conducted their own speed tests, finding that T-Mobile and Verizon exaggerated their coverage, which spurred the FCC to investigate.

T-Mobile’s claims

T-Mobile says that "when the FCC collected speed-test data to evaluate T-Mobile's maps, it failed to follow commonly accepted coverage testing procedures as well as its own MF-II [Mobility Fund Phase II] instructions."

T-Mobile's filing alleged:

• The sampling was extremely limited and not statistically sound, using only a single testing point for each grid
• In contravention of the FCC's own requirement to measure outdoor coverage, staff conducted drive tests using handsets mounted inside cars, which is considered in-vehicle coverage
• Testers failed to recognize or remedy the fact that their device was locked onto a 3G signal while attempting to test 4G speeds
• The Staff Report contains obvious measurement errors

On the first point, T-Mobile wrote that the FCC's "outdoor testing consisted of only seven test locations in the two states" and that "these stationary tests were completed at the locations only once."

"This is insufficient from a statistical basis to form conclusions regarding specific locations and is wholly inadequate to form the basis for conclusions about coverage," T-Mobile said.

Second, the FCC measurement on phones mounted inside a car "is inconsistent with the [MF-II] requirement to measure outdoor coverage," T-Mobile said.

"Measuring in-vehicle coverage, rather than the FCC-prescribed outdoor coverage, likely resulted in approximately 6dB of signal attenuation due to blockage from the vehicle," T-Mobile wrote. "If in-vehicle coverage maps incorporating in-car penetration losses were compared to the Staff's drive testing, the two data sets would be much more closely aligned."

The FCC report said that commission staff analyzed both in-vehicle and outdoor coverage. "At selected locations along the drive test routes, the agent stopped the vehicle, turned off all other phones (including any modem built-in to the test vehicle), and configured the test handset to perform tests outside of the vehicle," the FCC staff report said.

The FCC report also said that commission staff "follow[ed] standardized test procedures" for both in-motion drive testing and stationary testing.

T-Mobile says phones were locked on 3G

On its third point, T-Mobile said that some "of the so-called 'failed' tests apparently occurred when the testing device was inadvertently operating on 3G technology even though the phone was in an area with 4G LTE coverage." T-Mobile said it suspects "the test phone may have been inadvertently locked" on 3G even when 4G signals were available at testing locations.

T-Mobile continued:

For example, in western Montana and northwest North Dakota, staff's testing on April 1, 2019 showed 109 failing samples on 3G technologies in areas where 4G LTE was available. This constitutes nearly one-third of the failures in Montana. In T-Mobile's experience, when a device is stuck in 3G mode, performing speed tests with short idle intervals between active mode tests (which appears to be exactly what the staff was doing) may fail to reacquire 4G LTE even though strong 4G LTE coverage is available. This issue is understood within the drive-testing community, and normally test operators will guard against it before conducting a speed test by making sure that the device is manually locked onto the correct mode (in this case, 4G LTE). FCC staff appears not to have attempted this standard remedy.

Finally, T-Mobile described the "obvious measurement errors" that it says it found in the FCC report.

"For example, there were more than 60 test results in Montana where the reported RSRP value was less than -150dBm, often with clearly unrealistic values such as -222 or -218dBm, while also often observing multi-megabit download speeds," T-Mobile wrote. "These were not locations where T-Mobile lacked coverage, but rather test points where the measured speeds were clearly not correlated with the measured signal strength. There were more than 130 similar results in Alabama. Furthermore, a large percentage of 'MF-II failed' tests—i.e. those tests showing less than the MF-II required download speed of 5Mbps—were affected by those erroneous test points."

FCC said carrier tests didn’t match real-world experience

We contacted the FCC about T-Mobile's claims today and will update this article if we get a response.

The FCC report in December summarized its findings as follows:

Only 62.3 percent of staff drive tests achieved at least the minimum download speed predicted by the coverage maps—with US Cellular achieving that speed in only 45.0 percent of such tests, T-Mobile in 63.2 percent of tests, and Verizon in 64.3 percent of tests. Similarly, staff stationary tests showed that each provider achieved sufficient download speeds meeting the minimum cell edge probability in fewer than half of all test locations (20 of 42 locations). In addition, staff was unable to obtain any 4G LTE signal for 38 percent of drive tests on US Cellular's network, 21.3 percent of drive tests on T-Mobile's network, and 16.2 percent of drive tests on Verizon's network, despite each provider reporting coverage in the relevant area.

An FCC official said at the time that maps submitted by carriers were based on industry-standard propagation models and that the FCC's own tests made it clear that those industry models do not reflect on-the-ground experience. The FCC staff report recommended an audit of mobile-broadband coverage maps submitted by carriers and said the FCC should develop a better system for verifying the accuracy of carrier-submitted data.

Verizon and US Cellular do not appear to have contested the FCC report so far, as the FCC docket on the matter doesn't contain any filings from those two carriers.