Title: Test Tolerance analysis for TS 34.122 Test case 8.2.2.6.1

Source: CATT

1 Introduction

The test case in section 8.2.2.6.1 of 34.122, UTRA to E-UTRA TDD cell reselection: E-UTRA is of higher priority, has not been completed. The time periods is not aligned with 25.123. The measurement uncertainty and test tolerance is missed. This document describes the process to derive the Test Tolerances. The calculations are provided in the accompanying spreadsheet.

2 Test case in TS 34.122

The test conditions are defined in the following extract from TS34.122 v9.6.0, adapted according to TS25.123 v9.6.0 to re-order the time periods and avoid a separate step forcing the UE to cell 2 in the initialisation phase.

The following mapping was used:

• T1: Reselect to Cell 1 (UTRA) becomes T3

• T2: Change identity of Cell 2 (E-UTRA) becomes T1

• T3: Reselect to Cell 2 (E-UTRA), check within defined time becomes T2

which gives the resulting sequence, assumed in the rest of this document:

• T1: Change identity of Cell 2 (E-UTRA)

• T2: Reselect to Cell 2 (E-UTRA), check within defined time

• T3: Reselect to Cell 1 (UTRA)

In the current test case, the S_{prioritysearch1} is missed, it should be added as 24dB, i.e. -79dBm/1.28MHz to let UE using lower rate to search high priority inter-RAT cell. Additionally, the S_{prioritysearch2} = 0dB and Ssearch _E-UTRA is added also, but they are not verdict in the test case.

8.2.2.6.1 UTRA to E-UTRA TDD cell reselection: E-UTRA is of higher priority

……

<<Some clauses skipped>>

8.2.2.6.1.4.2.1 Initial conditions

This test scenario comprised of 1UTRA TDD serving cell ,and 1E-UTRA TDD cell to be re-selected. Test parameters are given in table 8.2.2.6.1.4.2.1-1, 8.2.2.6.1.4.2.1-2 and 8.2.2.6.1.4.2.1-3.The test consists of three successive time periods, with time duration of T1,T2 and T3 respectively. UTRA cell 1 is already identified by the UE prior to the start of the test, At T1 the UE is camped on to cell1. Cell2 is of higher priority than cell1.

<<Table 8.2.2.6.1.4.2.1-1 skipped>>

Table 8.2.2.6.1.4.2.1-2: Cell specific test parameters for cell re-selection UTRA TDD to E-UTRA TDD test case (cell 1)

Parameter	Unit	Cell 1 (UTRA)
Timeslot Number		0			DwPTS
		T1	T2	T3	T1	T2	T3
UTRA RF Channel Number (Note1)		Channel 1
PCCPCH_Ec/Ior	dB	-3	-3	-3
DwPCH_Ec/Ior	dB				0	0	0
OCNS_Ec/Ior	dB	-3	-3	-3
	dB	11	11	11	11	11	11
	dBm/1.28 MHz	-80
PCCPCH RSCP	dBm	-72	-72	-72	n.a.
Propagation Condition		AWGN
Qrxlevmin	dBm	-103
Qoffset1s,n	dB	C1, C2: 0
Qhyst1s	dB	0
Threshx, high (Note2)	dB	46 (-94dBm)
Sprioritysearch1	dB	24 (-79dBm)
Sprioritysearch2	dB	0
SsearchE-UTRA	dB	Not send
Note1: In the case of multi-frequency cell, the UTRA RF Channel Number is the primary frequency’s channel number. Note2: This refers to the value of Threshx, high which is included in UTRA system information, and is a threshold for the E-UTRA target cell

Table 8.2.2.6.1.4.2.1-3: Cell specific test parameters for cell re-selection UTRA TDD to E-UTRA TDD test case (cell 2)

Parameter	Unit	Cell 2
		T1	T2	T3
E-UTRA RF Channel Number		2
BWchannel	MHz	10
PBCH_RA	dB	0	0	0
PBCH_RB	dB
PSS_RB	dB
SSS_RB	dB
PCFICH_PA	dB
PHICH_PA	dB
PHICH_PB	dB
PDCCH_PA	dB
PDCCH_PB	dB
PDSCH_PA	dB
PDSCH_PB	dB
OCNG_RANote 1	dB
OCNG_RBNote 1	dB
Qrxlevmin	dBm/15kHz	-140	-140	-140
	dBm/15kHz	-98
RSRP	dBm/15kHz	-inf	-87	-101
	dB	-inf	11	-3
Snonintrasearch	dB	Not sent
Threshserving, low	dB	46 (-94dBm)
Threshx, low (Note2)	dB	24 (-79dBm)
Propagation Condition		AWGN
Note 1: OCNG shall be used such that cell is fully allocated and a constant total transmitted power spectral density is achieved for all OFDM symbols. Note2: This refers to the value of Threshx, low which is included in E-UTRA system information, and is a threshold for the UTRA target cell

……

3 Discussion

In the test case, there are two cells, cell 1 (UTRA cell) and cell 2 (E-UTRA cell) with higher priority. The test case consists of three successive time periods, T1, T2 and T3. During T1, cell 2 shall be powered off and changed scrambling code, to let the cell is unknown to UE. UE selects cell 1. At starting T2 cell 2 becomes stronger than threshold. The UE is expected to detect cell 2 and reselect to cell 2 due to with higher priority. During T3, UE selects cell 1 due to signal of cell 2 is below threshold. The value of Srxlev for each cell versus time, together with the thresholds are shown in figure 1.

Figure 1 Srxlev in cell reselection process

3.1 Selection criterion for UTRA

The selection criterion is satisfied according to the equation from TS 25.304 v9.6.0 clause 5.2.3.1.2:

Srxlev = Q_rxlevmeas – (Q_rxlevmin + Q_{rxlevminoffset}) – Pcompensation

The values in this equation are derived as follows:

• Srxlev is evaluated for each cell at T1, T2 and T3

• Q_rxlevmeas is the measured cell Rx level value (P-CCPCH RSCP) as given in tables 8.2.2.6.1.4.2.1-2.

• Q_rxlevmin for the UTRA cell is -103dBm/1.28MHz as given in tables 8.2.2.6.1.4.2.1-2.

• Q_{rxlevminoffset} is only taken into account as a result of a periodic search for a higher priority PLMN while camped normally in a VPLMN, and is therefore not applicable here.

• Pcompensation = 0 according to the table in TS 25.304 v9.6.0 clause 5.2.3.1.2, provided (UE_TXPWR_MAX_RACH – P_MAX) is negative. UE_TXPWR_MAX_RACH must therefore be lower than P_MAX.

The equation therefore reduces to:

Srxlev = Q_rxlevmeas – Q_rxlevmin

which is drawn for each cell at T1, T2 and T3 in the diagram above. For example, Cell 1 at T1, T2 and T3 has P-CCPCH RSCP = -72dBm/1.28MHz, so Srxlev = -72 – (-103) = 31dB.

3.2 Selection criterion for E-UTRA

For Cell 2 (E-UTRA) the selection criterion is satisfied according to the equation from TS 36.304 v8.9.0 clause 5.2.3.2:

The values in the equation for Cell 2 E-UTRA Srxlev are derived as follows:

• Srxlev is evaluated at T1, T2 and T3

• Qrxlevmeas is the Measured cell RX level value (RSRP) as given in table 8.2.2.6.1.4.2.1-3

• Qrxlevmin for the E-UTRA cell is -140dBm as given in table 8.2.2.6.1.4.2.1-3

• Qrxlevminoffset is only taken into account as a result of a periodic search for a higher priority PLMN while camped normally in a VPLMN, and is therefore not applicable here.

• Pcompensation =0 according to the table in TS 36.304 v8.7.0 clause 5.2.3.2, provided (P_EMAX – P_PowerClass) is negative. P_EMAX must therefore be lower than P_PowerClass.

The equation therefore reduces to:

Srxlev = Qrxlevmeas – (Qrxlevmin) which is drawn for each cell at T1, T2 and T3 in the diagram above.

Cell 2 (E-UTRA) at T1 has RSRP -infinity so Srxlev = -infinity – (-140) = -infinity.

Cell 2 (E-UTRA) at T2 has RSRP -87dBm so Srxlev = -87 – (-140) = 53.

Cell 2 (E-UTRA) at T3 has RSRP -101dBm so Srxlev = -101 – (-140) = 39.

As the test must work for a Rel-8 UE, the approach taken has been to refer to Rel-8 core specifications. Rel-8 of TS 36.304 does not use Squal.

3.3 Thresholds

The thresholds for UTRA are defined in TS 25.304 v9.6.0 clause 5.2.6.1.5:

Thresh_{x, high}

A Srxlev based threshold used by the UE for cell reselection towards a higher absolute priority layer when absolute priorities are applied. A threshold is defined for each absolute priority layer.

S_{prioritysearch1}

This threshold is used in the measurement rules for cell re-selection when absolute priorities are used. It specifies the value of Srxlev in the serving cell controlling the rate of inter-frequency and inter-RAT measurements.

The thresholds for E-UTRA are defined in TS 36.304 v8.9.0 clause 5.2.4.7:

Thresh_{x, low}

This specifies the threshold used in reselection towards frequency X priority from a higher priority frequency. Each frequency of E-UTRAN and UTRAN, each group of GERAN frequencies, each band class of CDMA2000 HRPD and CDMA2000 1xRTT might have a specific threshold.

Thresh_{serving, low}

This specifies the threshold for serving frequency used in reselection evaluation towards lower priority E-UTRAN frequency or RAT.

3.4 Test case verdict

The design of the test case relies on the UE being able to measure RSRP and P-CCPCH RSCP during T1, T2, and T3, and correctly compare it to the signal thresholds. In this test case, only signal level threshold is used.

A detailed analysis of the test case is given in section 5 of this document.

4 Choice and values of uncertainties to be specified

The SS provides two cells on different frequencies with AWGN. We propose to control the following parameters:

• AWGN absolute power on cell 1 frequency, I_oc ± 0.7dB

• AWGN absolute power on cell 2 frequency, N_oc ± 0.7dB

• Ratio of cell 1 signal / AWGN, Î_or / I_oc ± 0.3dB

• Ratio of cell 1 code level / I_or , Ec / I_or ± 0.1dB

• Ratio of cell 2 signal / AWGN, Ê_s / N_oc ± 0.3dB

This choice forms a minimum set, so the superposition principle can be applied if necessary. For these test cases, the signals are not faded, and the values for Ê_s / N_oc and Î_or / I_oc are chosen to be similar to equivalent parameters in LTE and W-CDMA. The absolute levels of N_oc and I_oc are specified as ±0.7dB, similar to the uncertainty for other absolute power values such as RefSens. The Ec / I_or is specified as ± 0.1dB in TS34.122.

5 Calculation of Test Tolerances

General approach

The general approach is given in the steps below:

1. Copy the originally specified key parameters from the core requirements

2. Where relevant, calculate derived parameters from the core requirements

3. Define uncertainties for a minimum set of parameters

4. Define controlled parameters (critical to the test verdict), calculate sensitivity factors and uncertainty

5. Determine which original or derived parameters to offset (apply Test Tolerances to) and by how much

6. Recalculate original or derived parameters including Test Tolerances

7. Check that the controlled parameters meet requirements to get the correct test verdict

Each step is explained below, and the calculations are given in the accompanying spreadsheet.

a) Original specified key parameters

The key parameters are copied from Table 8.2.2.6.1.4.2.1-2 and Table 8.2.2.6.1.4.2.1-3 in TS 34.122. The key parameters are selected as the minimum set to define the cell power levels, and which are subject to a test system uncertainty which may affect the verdict of the test. Some signalled parameters S_{prioritysearch1}, Thresh_{x, high}, Thresh_{serving, low}, and Thresh_{x, low} are also copied, although these are not subject to uncertainty.

The key parameters appear in section a) of the accompanying spreadsheet. The layout for Cell 1 and Cell 2 are similar to Table 8.2.2.6.1.4.2.1-2 and Table 8.2.2.6.1.4.2.1-3 in TS 34.122, but the time period is rearrangement and the AWGN is given a separate set of columns for each frequency. This allows the spreadsheet calculations to be done in a consistent way.

b) Derived parameters

A number of derived parameters are calculated, using the base information in a). The reason for deriving each additional parameter is given in the “Comment” column of section b) in the accompanying spreadsheet. The UE compares Srxlev values to thresholds, so the RSRP and RSCP values need to be converted to Srxlev.

In this test case, the UE compares Srxlev values to thresholds, so a further step is done to calculate the difference between Srxlev and the thresholds:

• (Cell 1 Srxlev – Thresh_x_low)

• (Cell 1 Srxlev – Sprioritysearch1)

• (Cell 2 Srxlev – Thresh_serving_low)

• (Cell 2 Srxlev – Thresh_x_high)

c) Uncertainties

The choice of uncertainties is covered in section 4 of this document. They appear in section c) of the accompanying spreadsheet.

d) Controlled parameters critical to verdict

A diagram giving the Cell 1 and Cell 2 Srxlev values during T1, T2 and T3 is provided in section 3 of this document. It also includes the relevant thresholds. This is the most appropriate view to analyse the critical parameters for this test.

From 34.122 clause 8.2.2.6.1, Cell 2 is of higher priority than Cell 1.

1. Before the start of T1 the UE has selected Cell 1 (low priority)

2. During T1, Cell 1 is the only cell. The relevant conditions are:

• The low priority Cell 1 must be good enough. It seems that none of the thresholds Thresh_{x, high}, Thresh_{x, low} or Thresh_{serving, low} apply when there is only one cell.

• Cell 2 must be undetectable

• S_{prioritysearch1} is specified for Cell 1, which is the serving cell during T1. Cell 1 has a margin of at least 6dB above S_{prioritysearch1} and therefore the UE will only search for inter-frequency and inter-RAT layers of higher priority with low rate, according to TS 25.123 clause 4.2.2.5a.

• Cell 1 must meet the P-CCPCH RSCP and DwPCH side conditions (level, Ec/Io) for a detectable cell in TS 25.123 clause 8.1A.2.2.1. Although it does not seem clear that these values are applicable for the UE during reselection, it seems reasonable to adopt the side conditions for the UTRA cell.

3. During T2, the UE is required to reselect to Cell 2 (high priority). The conditions for the reselection towards higher priority are:

• The high priority Cell 2 must be good enough, expressed as Srxlev ≥ Thresh_{x, high}, with a margin of at least 6dB (from TS 25.123 clause 4.2.2.5a).

• Thresh_{x, low} is not applicable during T2 (only applies when reselecting towards frequency X from a higher priority frequency)

• Thresh_{serving, low} is not applicable during T2 (only applies when reselecting towards lower priority frequency)

• S_{prioritysearch1} is specified for Cell 1, which is the serving cell at the start of T2. The UE will only search for inter-frequency and inter-RAT layers of higher priority, according to TS 25.123 clause 4.2.2.5a, provided Cell 1 exceeds the S_{prioritysearch1} threshold by a margin of at least 6dB. This is required to ensure the UE searches every T_{higher_priority_search} = (60 * N_layers) seconds as specified in TS 25.123 clause 4.2.2.

• Cell 1 must meet the P-CCPCH RSCP and DwPCH side conditions (level, Ec/Io) for a detectable cell.

• Cell 2 must meet the RSRP and SCH side conditions (level, Es/Iot) in TS 25.123 clause 4.2.2.5a.

4. During T3, the UE is required to reselect to Cell 1 (low priority). The conditions for the reselection towards lower priority are:

• Thresh_{x, high} is not applicable during T3 (only applies when reselecting towards higher priority Cell 2)

• The low priority Cell 1 must be good enough, expressed as Srxlev ≥ Thresh_{x, low} with a margin of at least 6dB. Although it does not seem clear that these values are applicable for the UE during reselection (see applicability statement at start of clause 9.1), it seems reasonable to adopt the side conditions for the UTRA cell. The side condition for Io has not been taken into account as the required margin for reselection is specified without any condition on Io.

• The high priority Cell 2 must be bad enough, expressed as Srxlev ≤ Thresh_{serving, low} interpreting TS 25.123 clause 4.2.2.5a to require a margin of at least 6dB.

• Cell 1 must meet the P-CCPCH RSCP and DwPCH side conditions (level, Ec/Io) for a detectable cell.

• Cell 2 must meet RSRP and SCH side conditions (level, Es/Iot) in TS 25.123 clause 4.2.2.5a.

The critical conditions are related to the 9 controlled parameters listed in section d) of the accompanying spreadsheet. They have been derived by study of the test case diagram and by careful reading of the relevant clauses in TS 25.123 and TS 36.133. The reason for each parameter being critical to the test verdict is given briefly in the “Comment” column of section d) in the accompanying spreadsheet. Information about the value to be achieved is given later in the “Comment” column of section g) in the spreadsheet.

Having identified the parameters critical to the test verdict which need to be controlled, we now need to consider how they are affected by the parameters which can be set by the test equipment. The sensitivity factors are the ratio (effect on a critical parameter y / a test equipment uncertainty x), and are usually in dB/dB. In this RRM test case for most parameters there is a simple one-to-one relationship between the parameters that can be set by the test equipment, and their effect on parameters determining the test verdict. The sensitivity factors are therefore derived by inspection as one or zero.

For example, an error of 1dB in the absolute AWGN level Noc would cause 1dB error in the Cell 2 RSRP, so the sensitivity factor is 1.000 during T2 and T3. It would also cause 1dB error in the derived parameter (Cell 2 Srxlev - Thresh_x_high), so the sensitivity factor is also 1.000 during T2 and T3. However the same error of 1dB in the absolute AWGN level Noc would cause no change to Cell 2 Es/Iot, because all other powers are specified relative to Noc, so the sensitivity factor is zero.

In some cases, the sensitivity factor is an intermediate value. For example, the Cell 1 Ior/Ioc has an effect on P-CCPCH_Ec/Io which depends on ratios of the powers making up the total. In such cases a sensitivity factor value between 0 and 1 results. It is important to calculate these correctly to obtain the overall uncertainty.

For example,

• The effect of Cell 1 Ior/Ioc uncertainty on Cell 1 P-CCPCH_Ec/Io is x 0.074.

These factors can also be derived intuitively, by considering the relative powers in step b). For example, Ioc forms 7.4% / (7.4%+92.6%) of the total power Io, which is 0.074. A change in the Cell 1 Ior/Ioc has very little effect on the ratio of P-CCPCH power to overall power, because the Ior makes up most of the overall power.

Having filled in the matrix of sensitivity factors, the accompanying spreadsheet calculates the overall uncertainty for each controlled parameter, taking into account the uncertainties and sensitivity factors for each parameter that can be set by the test equipment. This process follows the superposition principle. More details and explanation can be found in section 4 of TS 36.903. Uncertainties are calculated separately for T1, T2 and T3.

The normal procedure of combining uncorrelated uncertainties root-sum-square is followed.

e) Determine parameters to offset

• During T1, T2 and T3, the uncertainties do not take Cell 1 P-CCPCH Ec/Io, DwPCH Ec/Io or P-CCPCH RSCP power outside the allowed range, therefore no offsets are required.

• During T2, Cell 1 can maintain the same conditions as during T1. The UE will only search for inter-RAT layers of higher priority, according to TS 25.123 clause 4.2.2.5a, provided Cell 1 exceeds the S_{prioritysearch1} threshold with a margin of 6dB. Because S_{prioritysearch1} is set to 24dB, and the uncertainties do not reduce the margin below 6dB, therefore no offsets are required.

• During T2, Es/Iot  -4 dB side conditions in TS 25.123 is easily met by Cell 2 due to Es/Iot equal to 11dB, and the RSRP power is easily met, therefore no offsets are required.

• During T2 the UE is required to reselect to E-UTRA Cell 2 (high priority). The Srxlev of high priority Cell 2 is set to be above Thresh_{x, high} with a margin of 7dB, and the uncertainties do not reduce the margin to less than the assumed requirement of  6 dB interpreting TS 25.123 clause 4.2.2.5a, therefore no offsets are required.

• During T3, no offsets are required for Cell 1 due to Thresh_{x, low} for Cell 1 is the same value as S_{prioritysearch1}.

• During T3, Es/Iot  -4 dB side condition in TS 25.123 is just met by Cell 2 with Es/Iot equal to -3dB after considering test uncertainties, and the RSRP power is also met easily. Therefore no offsets are required.

• During T3, the UE is required to reselect to UTRA Cell 1 (low priority). The Srxlev of high priority Cell 2 is set to be below Thresh_{serving, low} with a margin of 7dB, and the uncertainties do not reduce the margin to less than the core requirement of ≤ -6 dB in TS 25.123 clause 4.2.2.5a, therefore no offsets are required.

The check that all other controlled parameters meet their required range is done in step g). In theory it is possible for steps e) to g) to be iterative, or possibly even steps c) to g) to be iterative, but this test case can be done without such iteration.

f) Parameters modified by Test Tolerances

Based on the decision in e), the set of parameters in a) and b) is reproduced in section f) of the accompanying spreadsheet, but this time modified by the Test Tolerances (applied offsets). In this case the offsets are all zero.

Re-derived parameters are calculated using the same methods as were used in step b).

g) Check controlled parameters Min/Max

Using a format similar to that in step d), the nominal value of each controlled parameter is recalculated.

The minimum and maximum values, due to variability from uncertainties, of controlled parameters is then calculated and compared against the requirements (Required margin relative to threshold, Es/Iot range, RSRP level..). It is not necessary to calculate all parameters during each time interval T1, T2 and T3, so a selection is made of those critical to the test verdict. The critical requirement for each parameter is given briefly in the “Comment” column of section g) in the accompanying spreadsheet. The cases closest to limit (in these test cases, all limits are one-sided) are identified by turquoise cells in the spreadsheet. If all requirements are met, then the exercise is complete.

It can be seen that with the uncertainty values and Test Tolerances proposed, all the requirements are met.