Page tree
Skip to end of metadata
Go to start of metadata


Description: EVM measurement with reference signal input
Library: 3GPPFDD 10-99, Measurement

Name Description Default Unit Type Range
StartSym start symbol 2560   int [0, ∞)
SymBurstLen number of symbols within burst to be measured 2560   int [1, 10000]
SampPerSym number of samples per symbol 16   int [1, ∞)
SymDelayBound upper bound of delay detection, in symbol, -1 for no detection -1   int [-1, ∞)†
NumBursts number of bursts to be measured 1   int [1, ∞)††
MeasType type of measurement: EVM_rms, EVM_peak, EVM_95th_percentile EVM_rms   enum  
SymbolRate symbol rate 3840000 Hz real (0, ∞)†††
EVMValue EVM value expression options: EVM_Ratio, EVM_Percent EVM_Ratio   enum  
† The model fulfills the synchronization (detects the delay of test signals, and aligns them with the reference signals) inside this boundary. If set to -1, synchronization will not be applied.
†† EVM results are determined over multiple bursts; NumBursts indicates the number of bursts to be measured and averaged.
††† SymbolRate is used to calculate the frequency offset; the default value 3840000 Hz is the symbol rate of 3GPP.

Pin Inputs




Signal Type



signals to be measured for EVM




reference signals for EVM measurement


  1. This subnetwork is used to accomplish the EVM measurement with numeric signals in baseband. The schematic for this subnetwork is shown in the following figure.

    WCDMA3G_EVM_WithRef Schematic
  2. EVM measurements are used to evaluate the modulation accuracy of modulators.
    Typically, the measurement is calculated at the symbol times within one burst. Z(k) is the complex vector produced by observing the real transmitter at the optimal phase of symbol k. S(k) is the reference (ideal) signal of symbol k sampled at the same phase as that of Z(k). The transmitter model is


    W = edr + jda accounts for both a frequency offset giving da radians per symbol phase rotation and an amplitude change of dr nepers per symbol
    C0 is a constant origin offset representing quadrature modulator imbalance
    C1 is a complex constant representing the arbitrary phase and output power of the transmitter
    E(k) is the residual vector error on sample S(k), and the value range of k is K which is [0,L−1]. By setting the parameter StartSym, designers can select which symbol the simulation starts with (Sth symbol). By setting the parameter SymBurstLen, designers can select the length of the burst to be measured (L).

    The error vector E(k) is measured and calculated for each instance k.

    The sum square vector error for each component is calculated over one burst. The relative RMS vector error is defined as

    The symbol EVM at symbol k is defined as

    which is the vector error length relative to the root average energy of the burst.
    C0, C1 and W are used to minimize RMS EVM per burst, then calculate the individual vector errors E(k) on each symbol. The symbol timing phase of the receiver output samples calculate the vector error to give the lowest value for the RMS EVM; this phase is called the optimal phase.
    • RMS EVM (MeasType=EVM_rms) for one burst is defined as

      RMS EVM should be measured by averaging over multiple bursts.
    • Peak EVM (MeasType = EVM_peak) is the peak error deviation within a burst (that is, the maximum of E(k)) measured at each symbol interval.
      Peak EVM should be measured by averaging over multiple bursts.
    • 95th percentile (MeasType = EVM_95th_percentile) is the point where 95% of the individual EVM (EVM(k)), measured at each symbol interval, is below that point. That is, only 5% of the symbols are allowed to have an EVM exceeding the 95th-percentile point.
      The 95th percentile should be measured by averaging over multiple bursts.

  1. ETSI SMG2 EDGE Tdoc 370r1/99, Modulation accuracy for EDGE MS and BTS, Paris, France, August 24-27, 1999.
  • No labels