Receiver sensitivity testing

Posted in Applications

An important factor of GPS testing is that of receiver sensitivity. The main areas of test are acquisition sensitivity and tracking sensitivity. Typically, the RF power level received by an antenna on the ground will be between -125dBm and -150dBm depending on environmental factors.

To generate very low RF power levels in this range, it is required to use external passive attenuators to reduce the LabSat output. In this way, the signal level is reduced to the required range with minimal added noise. Actual values of attenuation should be determined by the user to suit the equipment under test but as a guide, two in-line 20dB attenuators (40dB total) will give an RF power range of around -125 to -155dBm.

Although a user recorded RF signal can be used for testing in this way, the recommended method is to use a computer generated simulation file created using the SatGen software. This is because the SatGen created file will contain a 'pure' GPS signal with constant signal to noise ratio. A user recorded scenario will contain additional noise that was present during recording and will also have constantly changing signal to noise levels which make comparison difficult.

sensitivity-testing

Image: Examples of RF power level output using a SatGen created GPS signal. The -85dBm to -115dBm range corresponds to the standard LabSat output range.

By adjusting the attenuation slider during replay, the RF power output level can be reduced from -85dBm down to -115dBm.

However, because the -85 to -115dBm range is above the background noise level, the GPS signal is always visible to the GPS receiver and so the measured C/NO dBHz level will have little correlation to the slider attenuation. Some reduction in C/NO will be observed when reducing the LabSat RF level but it is not a linear reduction.

Adding 40dB of external attenuation to the LabSat output will reduce the RF power to a range of around -125dBm to -155dBm. This range is in line with the RF level seen by a GPS antenna when outdoors and falls below the level of background noise. By reducing the signal in this way a more linear control of C/NO is possible.

sensitivity-table

The table above shows C/No values measured by a UBLOX GPS engine using various external attenuators on the LabSat output. For each external attenuator value, the LabSat RF level is changed in 5dB steps.

The table shows that the linearity of signal to noise control improves as the external attenuation is increased. However, with external attenuation much above 40dB, the useable slider range is reduced.

U-centre

The screenshot above shows an example output from the UBLOX U-Center software. The GPGSV NMEA message is switched on to show C/NO level for each satellite.

A UBLOX TIM-LA device was tested using the following setup:

  • LabSat
  • Static scenario generated using SatGen
  • 2 x 20dBm attenuator (Minicircuits VAT-20W2)

The UBLOX data sheet quotes, for the TIM-LA, -138dBm for acquisition sensitivity and -146dBm tracking sensitivity. Using the setup above, the following measurements were made:

  UBLOX Datasheet Normal Mode LabSat Calculated
Aquisition -138 dBm -138 dBm
Tracking -146 dBm -147 dBm

The LabSat calculated value was obtained by using a baseline measurement of the un-attenuated LabSat output using a spectrum analyser. The measured value was summed with the 40dB external attenuator and the slider value to give estimated RF power.

As can be seen, with a relatively crude calibration using a spectrum analyser, a good approximation of receiver sensitivity is possible. In practice, it is recommended that the user would use a dedicated RF power meter obtain a more accurate reading to account for tolerances in cable and attenuator performance.

Customer Example

Engineers at HuaXun Technology use LabSat to test the sensitivity of their GNSS chipsets because of its ability to use realistic scenarios from every part of the globe.

“The variety of scenarios give us a complete solution. We don’t have to travel half way round the world to be able to test there, just like we don’t have to actually use a sports car (with the added cost for both car and fuel) to do pilot testing with high dynamics - it only needs recording once and then we can use it as often as we like ... "

Liang Rong Zhou, Test Engineer

Read the full case study here.