E.4 Derivation of the results
38.521-23GPPNRPart 2: Range 2 StandaloneRadio transmission and receptionRelease 17TSUser Equipment (UE) conformance specification
E.4.1 EVM
For EVM create two sets of Z’(f,t)., according to the timing ” 
–W/2 and 
+W/2” using the equalizer coefficients from E.3.3.
Perform the iDFTs on Z’(f,t) in the case of DFT-s-OFDM waveform. The IDFT-decoding preserves the meaning of t but transforms the variable f (representing the allocated sub carriers) into another variable g, covering the same count and representing the demodulated symbols. The samples in the post IDFT domain are called iZ’(g, t). The equivalent ideal samples are called iI(g,t). Those samples of Z’(f,t), carrying the reference symbols (=symbol 2,7,11) are not iDFT processed.
The EVM is the difference between the ideal waveform and the measured and equalized waveform for the allocated RB(s)
,
where
t covers the count of demodulated symbols with the considered modulation scheme being active within the measurement period, (i.e. symbol 0,1,3,4,5,6,8,9,10,12,13 in each slot, 🡪|T|=11)
g covers the count of demodulated symbols with the considered modulation scheme being active within the allocated bandwidth. (|G|=12*
(with: number of allocated resource blocks)).
are the samples of the signal evaluated for the EVM.
is the ideal signal reconstructed by the measurement equipment, and
is the average power of the ideal signal. For normalized modulation symbols 
is equal to 1.
From the acquired samples 2n EVM values can be derived, n values for the timing 
–W/2 and n values for the timing 
+W/2
E.4.2 Averaged EVM
EVM is averaged over all basic EVM measurements.
The averaging comprises n UL slots
where
for PUCCH, PUSCH.
The averaging is done separately for timing¦ 
–W/2 and +W/2 leading to and
is compared against the test requirements.
E.4.3 In-band emissions measurement
The in-band emissions are a measure of the interference falling into the non-allocated resources blocks.
Explanatory Note:
The inband emission measurement is only meaningful with allocated RB(s) next to non-allocated RB. The allocated RB(s) are necessary but not under test. The non allocated RBs are under test. The RB allocation for this test is as follows: The allocated RB(s) are at one end of the channel BW, leaving the other end unallocated. The number of allocated RB(s) is smaller than half of the number of RBs, available in the channel BW. This means that the vicinity of the carrier in the centre is unallocated.
There are 3 types of inband emissions:
1. General
2. IQ image
3. Carrier leakage
Carrier leakage are inband emissions next to the carrier.
IQ image are inband emissions symmetrically (with respect to the carrier) on the other side of the allocated RBs.
General are applied to all unallocated RBs.
For each evaluated RB, the minimum requirement is calculated as the higher of PRB – 30 dB and the power sum of all limit values (General, IQ Image or Carrier leakage) that apply.
In specific the following combinations:
– Power (General)
– Power (General + Carrier leakage)
– Power (General + IQ Image)
1 and 2 is expressed in terms of power in one non allocated RB under test, normalized to the average power of an allocated RB (unit dB).
3 is expressed in terms of power in one non allocated RB, normalized to the power of all allocated RBs. (unit dBc).
This is the reason for two formulas Emissions relative.
Create one set of Y(t,f) per slot according to the timing “
”
For the non-allocated RBs below the in-band emissions are calculated as follows
,
where
the upper formula represents the in band emissions below the allocated frequency block and the lower one the in band emissions above the allocated frequency block.
is a set of 
DFT-s-OFDM symbols with the considered modulation scheme being active within the measurement period,
is the starting frequency offset between the allocated RB and the measured non-allocated RB (e.g. 
for the first upper or 
for the first lower adjacent RB),
and 
are the lower and upper edge of the UL transmission BW configuration,
and 
are the lower and upper edge of the allocated BW,
is the SCS, and
is the frequency domain signal evaluated for in-band emissions as defined in clause E.3.3
The allocated RB power per RB and the total allocated RB power are given by:
The relative in-band emissions, applicable for General and IQ image, are given by:
where
is the number of allocated resource blocks,
and
is the frequency domain samples for the allocated bandwidth, as defined in clause E.3.3.
The relative in-band emissions, applicable for carrier leakage, is given by:
where RBnextDC means: Resource Block next to the carrier.
This can be one RB or one pair of RBs, depending whether the DC carrier is inside an RB or in-between two RBs.
Although an exclusion period may be applicable in the time domain, when evaluating EVM, the inband emissions measurement interval is defined over one complete slot in the time domain.
From the acquired samples n functions for general in band emissions and IQ image inband emissions can be derived. n values or n pairs of carrier leakage inband emissions can be derived. They are compared against different limits.
The in-band emissions are averaged over the n samples (equivalent to 10 UL subframes):
E.4.4 EVM equalizer spectrum flatness
For EVM equalizer spectrum flatness use EC(f) as defined in E.3.3. Note, EC(f) represents equalizer coefficient 
，f is the allocated subcarriers within the transmission bandwidth ((|F|=12*
)
From the acquired samples n functions EC(f) can be derived.
EC(f) is broken down to 2 functions:
Where Range 1 and Range 2 are as defined in Table 6.5.2.4.5-1 for normal condition and Table 6.5.2.4.5-2 for extreme condition
The following peak to peak ripple is calculated:
,which denote the maximum ripple in Range 1
,which denote the maximum ripple in Range 2
,which denote the maximum ripple between the upper side of Range 1 and lower side of Range 2
,which denote the maximum ripple between the upper side of Range 2 and lower side of Range 1
E.4.5 Frequency error and Carrier leakage
See E.3.1.
E.4.6 EVM of Demodulation reference symbols (EVMDMRS)
For the purpose of EVM DMRS, the steps E.2.2 to E.4.2 are repeated 6 times, constituting 6 EVM DMRS sub-periods. The only purpose of the repetition is to cover the longer gross measurement period of EVM DMRS ( time slots) and to derive the FFT window timing per sub-period.
The bigger of the EVM results in one n TS period corresponding to the timing¦ 
–W/2 or 
+W/2 is compared against the limit. (Clause E.4.2) This timing is re-used for EVM DMRS in the equivalent EVM DMRS sub-period.
For EVM the demodulation reference symbols are excluded, while the data symbols are used. For EVMDMRS the data symbols are excluded, while the demodulation references symbols are used. This is illustrated in figure E.4.6-1
Figure E.4.6-1: EVMDMRS measurement points
Re-use the following formula from E.3.3:
Z’(f,t) = MS(f,t) . EC(f)
To calculate EVMDMRS , the data symbol ( t=0,1,3,4,5,6,8,9,10,12,13) in Z’(f,t) are excluded and only the reference symbols (t=2,7,11) is used.
The EVM DMRS is the difference between the ideal waveform and the measured and equalized waveform for the allocated RB(s)
,
where
t covers the count of demodulation reference symbols (i.e. symbols 2,7,11 in each slot, so count=3)
f covers the count of demodulation reference symbols within the allocated bandwidth. (|F|=12*
(with
: number of allocated resource blocks)).
are the samples of the signal evaluated for the EVM DMRS
is the ideal signal reconstructed by the measurement equipment, and
is the average power of the ideal signal. For normalized modulation symbols 
is equal to 1.
n such results are generated per measurement sub-period.
E.4.6.1 1st average for EVM DMRS
EVM DMRS is averaged over all basic EVM DMRS measurements in one sub-period
The averaging comprises n UL slots
The timing is taken from the EVM for the data. 6 of those results are achieved from the samples. In general the timing is not the same for each result.
E.4.6.2 Final average for EVM DMRS
