The Difference Between Cell Site Location and GPS Location in CDRs

Published on May 8, 2026

Contact Mark CV Download

Cell-Site vs GPS Labels in CDR Evidence: What People Get Wrong and Why It Matters

People tend to treat a carrier call detail record (CDR) “location” like a GPS dot. They also mix up tower coordinates with phone coordinates.

A CDR usually tells you which cell tower and which sector handled a call, text, or data event, and some carriers also provide separate network based location estimates. Neither one automatically equals the phone’s GPS position.

Training materials used in law enforcement cell analysis warn “THIS IS NOT GPS,” note that coordinates are not exact, and say it is not advisable to plot them as the phone’s location.

Sector data has its own limit because it only supports that the phone was somewhere inside the serving sector’s service area at the start or end of an event, and that service area can stretch from a small city area to many miles. To avoid overclaiming, you first have to define what the “location” field represents.

What a CDR Location Field Actually Represents (Tower/Sector vs Handset Estimate)

CDR “location” can mean two different things, and I treat them as separate buckets. First, many CDRs record the serving tower and sector for an event.

Analysts then map that event by plotting the tower’s latitude and longitude and showing the sector direction. That map point is the tower, not your phone, even if the report labels it as a phone location.

Second, some carriers provide network based handset location products in separate files or sections, such as Verizon RTT or AT&T NELOS.

Those outputs aim at the phone’s location, but they still produce estimates with uncertainty, and courts may scrutinize them because the carrier uses proprietary methods that you cannot test from the outside.

If you keep these two categories apart, you can explain what a plot shows before you discuss how cell towers interact with phones.

Mechanisms Behind Cell-Site Location in CDRs: Selection, Sectors, and Distance Indicators

A phone does not pick a tower because it sits closest to it. For the start of a call, text, or data session, the phone tends to choose the strongest and best quality signal it can use.

That choice can shift with terrain, buildings, network load, and how the carrier tuned the network, so the same address can land on different sectors at different times.

A sector has an azimuth, which is the direction it faces, and a beamwidth, which is the spread of that direction. The wedge you see on a map shows direction, not a measured boundary of coverage.

Some records add distance style indicators such as Timing Advance, which the network uses to manage timing, not to place the phone on a map. Timing Advance supports a range band, often shown as an arc, and the same training materials warn that it is not GPS and that the coordinates are not exact.

During longer sessions, the network can hand off a connection to a new site for reasons that include movement and network management, so a single handoff does not prove movement. Once you understand these limits, it becomes easier to see why GPS works differently.

Mechanisms Behind GPS Location (and Why It Usually Isn’t in the Carrier CDR)

GPS is a satellite based system. Your device calculates its position by receiving signals from satellites, and it can do that without cellular service or internet access if it has satellite visibility.

Because the phone computes GPS on the device, the best GPS style evidence often lives on the device or in app accounts, such as navigation history, fitness logs, photo geotags, or other app records that store latitude and longitude as the phone’s position.

Some carriers offer “precise location” tools, but retention can be short compared with standard CDR retention, so by the time a civil case starts, the carrier may no longer have that higher precision product. That gap forces you to translate these technical differences into what you can claim on a map and in testimony.

Implications for Mapping, Accuracy Claims, and Admissibility Risk

When you only have sector based CDRs, the defensible claim stays narrow: at the event time, the phone connected to a specific sector, so the phone had to be somewhere the sector could serve. That does not “pinpoint” the phone, and sector size can range from small in dense cities to very large in rural areas.

Maps that draw an arbitrary circle or neat arc can mislead because real radio coverage rarely matches simple shapes, and RF engineers often rely on modeling and measurements to estimate coverage.

When you do have carrier derived handset estimates such as RTT or NELOS, you still have estimates, not GPS, and plotting them as exact points can overstate certainty, especially when you cannot audit the carrier method.

Data sessions also create timing and interpretation traps because a logged tower can reflect setup, a short handoff, or a longer dominant connection, and some carriers add “time on tower” fields that change what you can infer. These limits lead to a practical question: what records should you seek for the question you need to answer?

Choosing the Correct Records and the Least-Overstated Method for the Question at Issue

If you need to answer “was the phone generally in this area,” I use tower and sector mapping from CDRs, but I label tower coordinates as tower coordinates and I scale the map so you can see how broad the sector could be. I also avoid language that treats a mapped tower point as the handset location.

If you need to answer “where exactly was the device,” I prioritize device side GPS and app artifacts, and I also consider timely carrier precise location products when the case timing makes that possible.

If you have distance or range to tower data such as Timing Advance, RTT, or NELOS, that can be uncertainty bounded estimate, not GPS, and that uncertainty can be explained in plain language.

When the shape of coverage matters, I look for methods grounded in RF modeling or measurement rather than arbitrary arcs, so your conclusion matches what the data can support.

Contact Mark CV Download