Last updated: December 2, 2008
This section is not applicable to GSM.
The test set can be used to perform receiver sensitivity testing on GPRS mobiles in a number of ways. This section provides a comparison of the BER and BLER measurements, and looks at the different testing methods and configurations that can be used to verify the performance of a mobile's receiver. The section contains:
ETSI standards referred to in this section are the Release 99 versions of 4.14, 5.02, 5.03, 5.05 and 11.10. The ETSI GSM 11.10 standard has now been superseded by 3GPP 51.010.
Traditionally, GSM mobile receiver sensitivity testing is done using bit error rate (BER) testing. In the case of GPRS and EGPRS, the standards refer to the use of Block Error Rate (BLER) rather than BER.
BLER is similar to BER, but the resolution is at a block, rather than a bit level. A block consists of four radio bursts. The number of data bits in a block depends on the channel-coding scheme that is in use. There are four different channel coding schemes (CS-1 through CS-4) with varying levels of error protection used within GPRS. Similarly for EGPRS, there are nine different modulation and coding schemes (MCS-1 through MCS-9) using different modulation combined with varying levels of error protection. The reference sensitivity for mobiles varies depending on the coding scheme that is in use and the power class of the mobile (that is, the maximum power that it transmits.)
In a BER measurement, the method typically involves transmitting a sequence of bits (usually a pseudo-random binary sequence (PRBS)) from the test set to the mobile. This data is transmitted at a low power level (usually somewhere close to the reference sensitivity level of the mobile's receiver.) The mobile receives and re-transmits (typically at a much high power level to avoid introducing further bit errors) all the received data back to the test set. The test set compares the received data with the transmitted data and calculates the bit error rate. There are different BER tests defined within the standards for GSM, depending on whether channel coding is considered or not.
For a BLER measurement, the data is transmitted in a similar way to the BER measurement, but the computation is based on the number of blocks received in error by the mobile compared to the number of blocks sent. A block consists of four bursts. Each block has a Block Check Sequence (BCS) associated with it. If the BCS is bad, the block is counted as being in error.
In normal operation, the mobile receives the blocks and, if the Medium Access Control (MAC) header polling bit is set, sends back an ACK (Acknowledged) or NACK (Not Acknowledged) response, depending on whether the block was properly received. The transmitting device (that is, the base station) can record the relative number of negatively acknowledged blocks to the total number of blocks acknowledged by the mobile. This data can be used to calculate the BLER as a ratio. The advantage of this method is that a network can monitor BLER performance in normal use.
It is also possible for a mobile to loop back all the data received to a test set and for the test set to do the BLER calculation in a similar way to a traditional BER test. The test set allows you to do this (see Block Error Measurement ).
The Release 99 GSM 5.05 standard defines the reference sensitivity of GPRS mobiles. The following table shows the reference sensitivity input levels for GPRS mobiles in the various bands under static conditions.
The BLER in each case must be less than or equal to 10%.
|
GSM450, GSM480, GSM750, GSM850, PGSM (GSM900), DCS1800, PCS1900, RGSM, T-GSM810 |
Reference level under static conditions (dBm) |
|---|---|
| -104 | |
| -104 | |
| -104 | |
| -101 | |
| -104 | |
| -104 | |
| -104 | |
| -101.5 | |
| -98 | |
| -96 | |
| -93 | |
| -90.5 | |
| -86 |
The input levels provided in the above table are referenced to a normal PGSM mobile (MS), and must be corrected by the following values for other MS:
The mobile is also required to receive data when there is power in adjacent timeslots with respect to the following two situations:
There are different configurations available in the test set for testing GPRS and EGPRS mobiles. Some of these are available in Active Cell operating mode and others are GPRS BCH+PDTCH and EGPRS BCH+PDTCH Test Mode operating modes. All of these configurations allow transmitter tests to be performed. Only the following allow receiver tests to be performed:
Further details of each of these configurations are provided in the following sections.
Five of the connection types available in Active Cell operating mode are BLER, ETSI Type A, ETSI Type B (Unacknowledged), ETSI Type B (Acknowledged), and Switched Radio Block Loopback. The ETSI test modes A and B are defined in GSM 04.14:
In Active Cell operating mode both BER measurement (see GPRS Bit Error Measurement ), an SRB BER measurement (see EGPRS SRB Bit Error Measurement ), and a test set calculated BLER measurement (see Block Error Measurement ) are provided. However, it is not possible to run any of these measurements simultaneously.
In the case of a BER measurement, all of the GPRS coding schemes include a Frame Check Sequence (FCS). In normal operation, if a mobile (or the test set) receives a block for which any of the bursts fail this check, the entire block is discarded. The BER measurement allows you to specify how bad blocks are interpreted. Either of the following two settings can be used:
The BLER connection type is an Agilent proprietary method that makes use of standard signaling messages. This connection method sends GPRS Mobility Management (GMM) Information messages to the mobile with the polling bit set in the MAC header. This forces the mobile to respond with ACK/NACK messages, which allows the test set to calculate BLER (see Block Error Measurement ). Note that the default BLER connection mode does not send a PRBS on the downlink, so results may be different from those when compared with an equivalent BER test or when the test set calculates the BLER (where a PRBS is used by default).
In GPRS Test Mode BCH+PDTCH, a BER measurement can be made in a similar way as the method used with the ETSI test mode B (Unacknowledged) connection type. The primary difference is that the mobile must use some proprietary method for looping back the data from the downlink timeslot of interest. Only Layer 1 signaling is provided in this mode. For this reason it is not possible to make an ACK/NACK based BLER measurement in this test mode, although the test set calculated BLER measurement is still available.
EGPRS Test Mode BCH+PDTCH can be used to make BLER measurements in the same manner as GPRS Test Mode BCH+PDTCH.
There are some additional features that you can configure which improve the reliability of these connection methods with different manufacturers' mobiles.
For example, when using the BLER connection type, by default a message is sent to the mobile in every block. This can cause some mobiles to be overrun, so to counteract this the polling rate can be slowed down (see CALL:FUNCtion:DATA:BLER:POLLing:INTerval ). Note that this additionally slows down the rate that the mobile transmits at, so if transmitter measurements are being made in parallel, these are also slowed down because there are fewer bursts available to make measurements on.
Another situation that can occur when using the BLER connection type is that mobiles can respond to the GMM information message in different ways. The GMM information elements are all optional. The test set transmits none of these elements, effectively sending a header with an empty message. In some cases this causes the mobile to respond differently from what could be expected. To overcome this problem, the Logical Link Control (LLC) Frame Check Sequence (FCS) can be deliberately corrupted (see CALL:FUNCtion:DATA:BLER:LLC:FCSequence ). This causes the mobile to discard the LLC blocks, preventing the messages from reaching the GMM layer in the mobile. In the meantime, the RLC/MAC layer keeps responding to the polls and transmits ACK/NACK messages, allowing the test set to compute BLER (and make transmitter measurements).
When a TBF is established in all of the active cell data connection types, the test set uses a relative starting frame number by default. If you have not yet implemented this mandatory part of the standards, options have been added to use an absolute starting frame number or to start immediately (see CALL:FUNCtion:DATA:FRAMe:STARt ).
The following additional features are also available:
In the BLER mode data connection, it is possible to transmit a PRBS or other specific payload patterns. This is done by selecting the corrupt LLC FCS option (see CALL:FUNCtion:DATA:BLER:LLC:FCSequence ) then selecting the payload pattern you want using CALL:FUNCtion:DATA:PAYLoad:PATTern:BLER . Note that if you want to switch between the corrupt and valid FCS messages, it is recommended that you should end the current data connection and then restart it.
In all test set operating modes, it is possible to switch on the power in unused timeslots using the CALL:(PDTCH|PDTChannel):PREDuction:UBURst . The power level for all the unused downlink timeslots can be set to either Power Reduction Level 1 (PRL1) or Power Reduction Level 2 (PRL2). This allows adjacent channel rejection tests to be done. Note that all unused timeslots are set to the same level; that is Off, PRL1 or PRL2.
As mentioned previously, there are two configurations that need to be tested. Using the test set, it is possible to set up power differences between adjacent active downlink timeslots.
The test set uses power control mode A for setting the downlink power. The standards specify a maximum difference of 10 dB between active timeslots. The test set provides even greater flexibility by allowing you to specify a maximum difference of up to 25 dB (although the source accuracy is currently only specified up to a difference of 20 dB). This allows greater dynamic range in GPRS testing. See CALL:(PDTCH|PDTChannel):PREDuction:LEVel<1|2> .
Because an indication of the power difference is included in the RLC/MAC header, there may be some cases where mobiles are confused by the use of differences greater than 10 dB.
The reason power control mode B has not been implemented in the test application (which allows a difference of up to 30 dB) is that this requires fixed allocation mode. Only dynamic allocation mode has been implemented in the test application. Fixed Allocation is available in the lab application.
The multislot configuration can be set up in any of the test set's operating modes/ data connection types (ETSI Type A, ETSI Type B, BLER and BCH+PDTCH test mode), to any one of the supported multislot configurations, that is 1x1, 2x1, 3x1, 4x1, 2x2 or 3x2.
In each case, there are two power levels (PRL1 and PRL2) that can be set on the downlink for the active timeslots (see CALL:(PDTCH|PDTChannel):PREDuction:BURSt<[1]|2|3|4|5> ). PRL1 and PRL2 are the levels below the P0 level at which bursts can be transmitted. (The P0 level is the amount by which the PDTCHs are reduced below the BCH level when the PDTCHs and BCH are on different ARFCN (Absolute Radio Frequency Channel Numbers). The P0 level can differ from the BCH level by up to 30 dB. According to the standard, the P0 level must be set to an even number. Each used burst can be set to either PRL1 or PRL2.The P0 level can be set independently (see CALL:(PDTCH|PDTChannel):PZERo:LEVel ). PRL1 and PRL2 are both set relative to P0. To achieve the maximum level difference between active timeslots, set PRL1 to 0 dB and PRL2 to 25 dB, or set PRL1 to 25 dB and PRL2 to 0 dB. Assign PRL1 to the first burst and PRL2 to the second burst. If you want more details on power reduction levels, including a figure which illustrates GPRS power reduction levels, refer to Downlink PDTCH Power Control .
It is possible to perform multislot BLER testing on the test set with downlinks at different power levels. In BLER mode, the count of ACK/NACK messages is not currently associated to individual timeslots. For normal use this is acceptable, as the standards specify that BLER must be within limits for the multislot configuration as a whole.
If you want to attribute BLER to a single timeslot with other timeslots at higher levels in a multislot downlink, then the following type of procedure can be used. Note that this procedure assumes that your mobile's multislot performance is of an acceptable standard.
The procedure is to first verify that the mobile has a BLER of zero when the downlink timeslots are at different powers, but all received timeslots are above the reference sensitivity level. The levels of all timeslots are then reduced, but the relative difference between them is maintained. When performing reference sensitivity level testing, only one of the timeslots should be close to or below the reference sensitivity level. In this case, any BLER is likely to be attributable to this lower power timeslot. To verify this, perform a BER measurement in either ETSI test mode B or in the GPRS Test Mode BCH+PDTCH (see Example of Multi-slot Receiver BER Testing ). You should also ensure that the power is switched off in unused timeslots (see CALL:(PDTCH|PDTChannel):PREDuction:UBURst ) to remove any possibility of power in unused bursts affecting the test results.
Set the Connection Type to BLER.
Set up the multislot configuration (for example, 2x1) in the required band.
Make sure the BCH and the PDTCH's are on different ARFCN's. (If they are the same, then the standard requires all downlink timeslots to be transmitted at the same power level - namely that of the BCH.)
Set the Cell Power to a suitably high level well above the reference sensitivity level of the mobile (for example, -80 dBm).
Set P0 to a suitable level (for example, 0 dB).
Set PRL1 and PRL2 to appropriate levels (for example, 0 and 6).
Assign the first burst to PRL1 and the second to PRL2 (which means the first timeslot is transmitted at -80 dBm and the second at -86 dBm.)
Set the BLER measurement to single trigger mode, and specify the number of blocks to measure, for example 100. See SETup:BLERror .
Check that the result is zero. If not, adjust the parameters in steps 4 through 6 above until the result is zero.
Change the P0 level to reduce both the downlink levels (for example 14 dB, causing the first burst to now be transmitted at -94 dBm and the second at -100 dBm.)
In GPRS BCH+PDTCH Test Mode or ETSI Test Mode B, the first downlink burst to be looped back can be specified. By default this is the first active timeslot, but you can change this if required.
For example, to measure BER on a second received timeslot that is at a lower power level than the first, perform the following procedure:
Set the operating mode to GPRS BCH+PDTCH Test Mode or set the operating mode to Active Cell and the connection type to ETSI Test Mode B. If you want to use GPRS BCH+PDTCH Test Mode, set the mobile into an appropriate loopback mode.
Ensure that the BCH and PDTCH's are on different ARFCN's. (If they are the same, then the standard requires all downlink timeslots to be transmitted at the same power level - namely that of the BCH.)
Set the cell power and P0 level to the values you want, for example -70 dBm and 14 dB.
Set the multislot configuration to 2x1.
Set the PRL1 level to the value you want, for example 0 dB. (The level transmitted is cell power - P0 - PRL1, in this example, -84 dBm).
Set the PRL2 level to a higher value, for example 10 dB. (The level transmitted is cell power - P0 - PRL2, in this example, -94 dBm).
Assign PRL1 to the first burst that you are interested in.
Assign PRL2 to the second burst.
Set the first burst to loop back to 2 (that is, the second received downlink.)