by Andrew J. Shepherd
Sprint 4G Rollout Updates
Monday, July 16, 2012 - 1:40 AM MDT
As Sprint LTE 1900 has become live and discoverable in numerous markets over this past weekend, some of our readers, especially those who are using Android 4.0 ICS based ROMs, have expressed concern at the seemingly low signal levels that they have encountered. For example, see this screenshot from an HTC EVO 4G LTE (under Settings > About > Network):
Note the -102 dBm signal level. If this were measuring CDMA1X or EV-DO, then, yes, -102 dBm would be nearing the margin of usable signal. But -102 dBm is actually relatively healthy LTE signal level. To understand why, we need to learn the differences between two types of signal measurement: Received Signal Strength Indicator (RSSI) and Reference Signal Received Power (RSRP).
First, an LTE downlink is divided into subcarriers. A 5 MHz bandwidth downlink, which is the configuration that Sprint is deploying, contains 300 subcarriers. And of those subcarriers, one in three carry LTE reference signals. In other words, of the 300 subcarriers, 100 transmit periodic reference signals.
To illustrate, I captured this power vs frequency sweep with a spectrum analyzer. The LTE downlink graph comes from a Sprint site in the Kansas City area in late April, well before Sprint stopped blocking devices from live LTE sites. So, the sector depicted here exhibits no data traffic; it is transmitting only the periodic reference signals on 100 subcarriers, which you can clearly count in the graph:
Now, RSSI is the more traditional metric that has long been used to display signal strength for GSM, CDMA1X, etc., and it integrates all of the RF power within the channel passband. In other words, for LTE, RSSI measurement bandwidth is all active subcarriers. If we take the above RF sweep of a Sprint 5 MHz bandwidth downlink, RSSI measures the RF power effectively of what is highlighted in yellow:
RSRP, on the other hand, is an LTE specific metric that averages the RF power in all of the reference signals in the passband. Remember those aforementioned and depicted 100 subcarriers that contain reference signals? To calculate RSRP, the power in each one of those subcarriers is averaged. As such, RSRP measurement bandwidth is the equivalent of only a single subcarrier. And using our graph once more, RSRP measures the RF power effectively of what is highlighted in red:
Since the logarithmic ratio of 100 subcarriers to one subcarrier is 20 dB (e.g. 10 × log 100 = 20), RSSI tends to measure about 20 dB higher than does RSRP. Or, to put it another way, RSRP measures about 20 dB lower than what we are accustomed to observing for a given signal level. Thus, that superficially weak -102 dBm RSRP signal level that we saw previously would actually be roughly -82 dBm if it were converted to RSSI.
To conclude, here are a few takeaways about RSSI and RSRP as signal strength measurement techniques for LTE:
- RSSI varies with LTE downlink bandwidth. For example, even if all other factors were equal, VZW 10 MHz LTE bandwidth RSSI would measure 3 dB greater than would Sprint 5 MHz LTE bandwidth RSSI. But that does not actually translate to stronger signal to the end user.
- RSSI varies with LTE subcarrier activity -- the greater the data transfer activity, the higher the RSSI. But, again, that does not actually translate to stronger signal to the end user.
- RSRP does a better job of measuring signal power from a specific sector while potentially excluding noise and interference from other sectors.
- RSRP levels for usable signal typically range from about -75 dBm close in to an LTE cell site to -120 dBm at the edge of LTE coverage.
Sources: 3GPP, author's graphs