5.1.2 Beam management sensitivity study of NF based solutions

38.8843GPPRelease 17Study on enhanced test methods for Frequency Range 2 (FR2) User Equipment (UE)TS

5.1.2.1 Simulation assumptions

For NF based solutions, where beam peak search is necessary to perform all applicable test case procedures, an evaluation of UE beam management sensitivity to magnitude/phase variation of the DL signal is needed. Two assumptions are made about the NF based system:

– Beam peak search is performed in the NF (i.e. DNF system); OR

– Beam peak search is first performed in the FF/IFF and test case is executed in the NF (i.e. CFFNF system).

Using the spherical coverage measurement grid assumptions shown in Table 5.1.2.1-1, evaluations were performed of the UE beam management sensitivity in terms of simulated radiated performance metrics for each of the assumptions.

Table 5.1.2.1-1: Beam management sensitivity simulation assumptions

Parameter

Value

Notes

Spherical coverage Measurement Grids baseline assumption

Annex G.1.1 in TR38.810

Antenna array

– 8×2 and 4×1

– Antenna element HPBW: {260/130, 90/90} deg

Element near-field assumption is implementation specific

Simulated DUT

Two antenna arrays are integrated in the UE for the spherical coverage analyses

– Antenna panels are studied with Nz x Ny with Nz>Ny, e.g., 8×2 corresponds to Nz = 8 and Ny = 2

– The implementation loss for the antenna near the front is 0dB less than that for the antenna near the back

– The antenna in the back is on the opposite side of the UE (mirrored around (0,0,0)).

See Figure 5.1.2.1-1 for example positions of two antenna arrays

Beam steering

– In the xy plane, assume 45º beam steering granularity (AZ from -45º to +45º)

– In the xz plane, assume 22.5o beam steering granularity (EL from -90º to 90º)

Offsets

– Various antenna offsets (yoffset, zoffset) beyond 7.5cm in radius (12.5cm max)

– For TRP analysis, model random antenna offsets anywhere within the 30cm spherical QZ

Offset is defined with respect to the center of antenna array

Range Lengths

– 30cm, 20m (more range lengths are not precluded)

– Goal is to eventually determine min. range length and MU for performing spherical coverage tests in DNF

Defined as distance between centre of QZ/positioning axes and measurement probe

Test methodology

– CFFDNF/DNF (while taking path loss offsets into account)

– CFFNF

Sampling grid

Study finer than 7.5deg step size for constant-step size grids

Parametric studies to show convergence for the selected assumption

Figure 5.1.2.1-1 below illustrates example positions of two antenna arrays in the simulated DUT.

Figure 5.1.2.1-1: Simulated DUT antenna assumptions for beam management sensitivity study

5.1.2.2 Simulation results

Table 5.1.2.2-1 below summarizes the results from simulations of beam management sensitivity of a DNF system (i.e. beam peak search is performed in the NF).

Table 5.1.2.2-1: Beam management sensitivity results of a DNF system

Company label

Swept parameters

Beam management performance maximum ∆ relative to reference (dB)

Notes

Beam peak

50% CDF

TRP

Company A

Array: 8×2

Range: {0.2, 0.4, 0.8} m

Offset: {0, 0.05, 0.10} m

HPBW: {90/90}

2.5

Not analysed

Not analysed

There is approximately 2.5 dB of BP error when range length is reduced to 0.2m in presence of the module offset mentioned above. There is also significant perturbation of the CDF curve. CDF statistics start to converge when the range length is at least 4 times the offset.

Company B

Array: 8×2, 4×1

Range: {0.25, 0.3, 0.45, 20} m

Offset: {0.125 in y, 0.125 in z, 0.09 in y & z} m

HPBW: {260/130}

7.0

1.0

TRP analysed separately

The EIRP beam peak (100%-ile EIRP) and direction cannot be measured accurately with the direct NF methodology

Company B

Array: 8×2

Range: 0.2 m

Offset: 0.15 m in x, y, z

HPBW: {260/130}

0.66 dB systematic

0.46 dB RSS’ed

Large uncertainties can be observed for TRP for measurements performed in the NF utilizing the black back box approach

Company C

Array: 4×1

Range: {100, 4.2, 0.9, 0.45, 0.3} m

Offsets: not specified

Full phone model (including the PCB and phone house) has been considered

0.3

0

Not clear whether 0.1 or 0.4

Figure of merits such as EIRP, TRP, and Spherical Coverage are not influenced dramatically from range length

Company B

Array: {4×1, 8×2}

Range: 0.25 m

Offset: {0, 0.125, 0.9} m in y, z

HPBW: {90/90}

4.2

UE selected different beam between NF beam peak direction and FF beam peak direction

Company B

Array: {4×1, 8×2}

Range: 0.25 m

Offset: {0, 0.53, 0.75} m in x, y, z

HPBW: {90/90}

10.4

UE select the same beam in the NF as in the FF more often, we still see concerning trends with the peak EIRP deltas

Company B

Reuse assumptions used by Company A:

Array: 8×2

Range: {0.2, 0.4, 0.8} m

Offset: {0, 0.05, 0.10} m

HPBW: {90/90}

2.5

1.2

Simulations were performed to establish alignment with another company

Table 5.1.2.2-2 below summarizes the results from simulations of beam management sensitivity of a CFFNF system (i.e. beam peak search is first performed in the FF/IFF and test case is executed in the NF).

Table 5.1.2.2-2: Beam management sensitivity results of a CFFNF system

Company and reference

Swept parameters

Beam management performance maximum ∆ relative to reference (dB)

Notes

Beam peak

50% CDF

TRP

Company B (“Black box with transform approach”)

Array: 8×2, 4×1

Range: {0.22 – 0.30} m

Offset: {0, 0.50, 0.10, 0.125} m

Max µ = 0.2

Max σ = 0.3

Not analysed

Not analysed

These results were obtained using a transform-based approach to correct the incurred path loss. Feedback from industry is requested whether to continue efforts in terms of simulations and empirical investigations on this enhanced NF methodology with transform utilizing black-box approach

Company B (“Black & White box with transform approach”)

Array: 8×2, 4×1

Range: {0.22 – 0.30} m

Offset: {0, 0.50, 0.10, 0.125} m

Max µ = 0.1

Max σ = 0.3

Not analysed

TRP analysed separately

These results were obtained using a transform-based approach to correct the incurred path loss. Feedback from industry is requested whether to continue efforts in terms of simulations and empirical investigations on this enhanced NF methodology with transform utilizing the white&black-box approach

Company B (“TRP with compensation for antenna offset”)

Array: 8×2

Range: 0.2 m

Offset: 0.15 m in x, y, z

HPBW: {260/130}

0.02 dB systematic

0.21 dB RSS’ed

These results were obtained using the DNF methodology with declared offset; alternatively, these results could be obtained using a transform based approach to estimate the phase centre offset. With the offset of the antenna array known, e.g., estimated with the enhanced NF methodology introduced in this contribution, very accurate TRP measurements in the NF can be made with a TRP offset compensation approach

Company C

Array: 4×1

Range: {100, 4.2, 0.9, 0.45, 0.3} m

Offsets: not specified

Full phone model (including the PCB and phone house) has been considered

0.3

1.0

0.8

These results were obtained using the DNF methodology. Figure of merits such as EIRP, TRP, and Spherical Coverage are not influenced dramatically from range length

Full phone model (including the PCB and phone house) has been considered