Network synchronization measurements in the time domain (part 2)

Ok, what does interference look like and what causes it?

Network time-synchronization is obtained from boundary clocks and the central master clock (also known as the grandmaster clock) and a time protocol link (e.g. PTP1588). The result is a complicated structure of network components and several time protocol links. If links or network components are faulty, the network will become unsynchronized in the time domain. This also has a definite impact on frequency, because it is the inverse value of time. As a result, unsynchronized networks display suspicious behavior in the time and frequency domains.

Assume four 5G NR networks are deployed, using band n77.

Each of the 5G NR networks modulates thousands of subcarriers, transmitting synchronization and user data. In the real world, intermodulation from non-linear components in the signal chain come into play. Non-linearities can occur anywhere, such as in amplifiers in active antenna systems and for any number of reasons, such as manufacturing errors in RF connection antenna elements or because a rusty fence is too close to the antenna. Instead listing of all the possible intermodulation causes, we have to be prepared for the fact that intermodulation can occur everywhere.

How to prepare for intermodulation?

Intermodulation products can be calculated with mathematical formulas and the intermodulation principle has been known for decades. In the end, intermodulation must be taken into consideration during network planning, either by arranging carriers so that intermodulation products do not overlap with the uplink spectrum in FDD networks or by time synchronizing networks. Spectrum is getting crowded and carriers can no longer be arranged to avoid intermodulation. This means networks have to be time-synchronized in an absolute time-domain. Regulators in different countries around the world require network operators to use specific uplink and downlink slot structures in the time domain when operating TDD networks and align their network with the UTC second. The UTC second is unique around the world and provides a perfect baseline for time synchronization. After having networks synchronized to the UTC second and giving them uniform uplink and downlink slot structures, intermodulation impact is kept to an absolute minimum. Sending and receiving time slots are aligned and the weakest path (uplink) remains clear from intermodulation when the base station is transmitting at high power (downlink).

How to measure time synchronization to the UTC second?

Measuring network synchronization to the UTC second requires extremely accurate measurements. Country-specific regulations, deployment scenarios and network configurations demand that the deviation from the GNSS-derived UTC second must not exceed 130 ns. As a consequence, measurement equipment must have a mandatory accuracy of << 60 ns. In addition, over-the-air measurements performed in the field that require perfect RF conditions to avoid falsification from multipath propagation. The measurement receiver must detect 5G NR carriers, decode PCIs, SSBs and radio frames to uniquely determine the radio frame transmission time from gNodeB. The R&S®TSMA6 or R&S®TSME6 network scanning receivers can do all this.

The instrument’s host TAE measurement software tools (either R&S®ROMES4 for the R&S®TSMA6 or R&S®TSME6 or QualiPoc Android for the R&S®TSMA6) guide the user through the measurement process.

First, the equipment has to be calibrated to ensure the highest possible accuracy. Every ns counts, making it necessary to configure the cable length, the cable type and the distance to the gNodeB. Preconditions must be met for the second step, including a free view of the sky and perfect 5G NR RF conditions. The best place for performing such measurements are where line-of-sight conditions in the main lobe of the gNodeB antenna have high SINR values.

Once preconditions are met, the equipment delivers the UTC time and the radio frame offset to the UTC second. In the example below, the radio frame #124 was received at 08:21:39,080 (UTC), with a delay of 70.8 ns from the expected radio frame transmission. The frame number (SFN) continuously cycles from 0 to 1023, as you will see during measurements.

In addition to the UTC second offset, several quality parameters are also available. The reference time error (of 17.5 ns in the example below) and the sigma/confidence interval (of 17.4 ns in the example below) provide continuous feedback about current measurement accuracy, depending on the GNSS signal quality and the 5G NR RF conditions. Accuracy of 17 ns is very high, when in light of the time synchronization guidelines that mention synchronization accuracy of 130 ns to 1500 ns.

As mentioned above, time and frequency errors are related. Frequency errors can be detected by measuring the center frequency of the 5G NR synchronization block down to the Hz level (!) and frequency errors (expressed in ppm) are calculated. 3GPP clearly specifies the limits of the max. allowed frequency error.

Network synchronization measurements over a longer period of time help detect drifts over time. The frequency and time offsets can also be visualized in graphs to evaluate time-dependent behavior.

Having all this available lets network operators, field technicians and regulators easily check whether their network complies with regional network synchronization and frequency accuracy guidelines.