GB2552212A

GB2552212A - Pre-distortion compensation for voltage controlled oscillators

Info

Publication number: GB2552212A
Application number: GB1612261.6A
Authority: GB
Inventors: Alloulah Mohammed; Dawkins Mark; Burdett Alison
Original assignee: Frontier Microsystems Ltd
Current assignee: Frontier Microsystems Ltd
Priority date: 2016-07-14
Filing date: 2016-07-14
Publication date: 2018-01-17
Anticipated expiration: 2036-07-14
Also published as: GB2552212B; GB201612261D0

Abstract

The invention relates to control of a voltage controlled oscillator that may form part of a polar modulator, e.g. in an OFDM transmitter. The invention provides a method of correcting errors in the output of a VCO arising from the non-linearity of the VCO, and also quantisation errors. The invention estimates errors based on a mathematical model, accumulates the estimated errors over a number of input samples, and periodically provides a correction factor to the VCO. The method includes the steps of receiving a phase modulating signal comprising a sequence of samples, each sample indicating a required frequency deviation, mapping the required frequency deviation of each sample to a second allowed frequency deviation value selected from a set of allowed values, and using the second value to control the VCO input. The method also includes the steps of determining a frequency deviation correction factor using a predetermined number of samples, and modifying the VCO input using this factor, which is an estimate of the accumulated discrepancy between the required frequency deviations and the respective second allowed frequency deviation values.

Description

(54) Title of the Invention: Pre-distortion compensation for voltage controlled oscillators Abstract Title: Pre-distortion compensation for a VCO in a two-point polar modulator (57) The invention relates to control of a voltage controlled oscillator that may form part of a polar modulator, e.g. in an OFDM transmitter. The invention provides a method of correcting errors in the output of a VCO arising from the non-linearity of the VCO, and also quantisation errors. The invention estimates errors based on a mathematical model, accumulates the estimated errors over a number of input samples, and periodically provides a correction factor to the VCO. The method includes the steps of receiving a phase modulating signal comprising a sequence of samples, each sample indicating a required frequency deviation, mapping the required frequency deviation of each sample to a second allowed frequency deviation value selected from a set of allowed values, and using the second value to control the VCO input. The method also includes the steps of determining a frequency deviation correction factor using a predetermined number of samples, and modifying the VCO input using this factor, which is an estimate of the accumulated discrepancy between the required frequency deviations and the respective second allowed frequency deviation values.

Figure 6

This print incorporates corrections made under Section 117(1) of the Patents Act 1977.

At least one drawing originally filed was informal and the print reproduced here is taken from a later filed formal copy.

1/14

10 17

Figure 1

2/14

10 17

VCO

stream

Figure 2

3/14

10 17

Figure 3

4/14

10 17

Figure 4

5/14

10 17

Figure 5

6/14

10 17

Figure 6

7/14

10 17

Figure 7

8/14

10 17

Figure 8

9/14

Obtaining a resultant frequency deviation by adding the frequency deviation correction factor to the second allowed frequency deviation value of a current sample.

Determining whether the resultant frequency deviation is inside of a predetermined range

	z-	—\		f-	⁴-\
	If it is inside the		If it is outside the
44	predetermined	46	predetermined
	range I			range	j

IO 45

Using the resultant frequency deviation to set the VCO input value for the current sample

Z- -\

Adding a portion of the frequency deviation correction factor to the required frequency deviation of a current ^sample /Adding a further portion of the frequency deviation correction factor to the required frequency deviation of a \gubsequent sampleJ

Figure 9

10/14

Receiving a phase modulating signal comprising a sequence of samples, each sample having a required frequency deviation;

Mapping the required frequency deviation of each sample to an allowed frequency deviation value, wherein the allowed frequency deviation value is selected from a plurality of allowed frequency deviation values; wherein the mapping comprises determining a second allowed frequency value using a ^nodel of the non-ideal response of the VCO.

Using the second allowed frequency deviation value to control a VCO input;

10 17

Providing a periodic correction to the VCO input by the steps of:

(i) determining a frequency deviation correction factor, the determining being performed using a predetermined number of samples, wherein the frequency deviation control factor is an estimate of an accumulated discrepancy between the required frequency deviations and the respective allowed frequency deviation values. The estimate of the accumulated discrepancy is obtained using an estimate of a periodic error and a non-ideal response of the VCO.

(ii) modifying the VCO input using the frequency deviation correction factor.

Figure 10

11/14

10 17

Figure 11

12/14

10 17

Figure 12

13/14

10 17

Figure 13

14/14

10 17

Figure 14

Intellectual

Property

Office

Application No. GB1612261.6

RTM

Date :19 December 2016

The following terms are registered trade marks and should be read as such wherever they occur in this document:

Wi-Fi (Page 2, 3)

Intellectual Property Office is an operating name of the Patent Office www.gov.uk/ipo

Pre-distortion compensation for Voltage Controlled Oscillators

Field of invention

The present invention relates to pre-distortion compensation for a Voltage Controlled Oscillator (VCO).

Background

Polar modulation is a technique used in a number of Radio Frequency (RF) transmitter architectures. Recently, it has found particular application in devices implementing the EDGE and 3G/UMTS telecommunication standards. Polar modulation is a combination of amplitude and phase modulation, which, instead of using in-phase and quadrature components as in quadrature modulation, uses the components of amplitude and phase. The modulation is performed separately for the amplitude and phase components, with the phase modulation typically performed using a Phase Locked Loop, PLL.

A problem with polar modulation is that the polar signal path must carry wideband phase modulation and tolerates a very small delay mismatch between the separate amplitude and phase modulated signals. A solution to this problem is the use of two point modulation, in which lower frequency components of the phase modulated signal are introduced into the phase lock loop and the higher frequency components are inserted directly into the Voltage Controlled Oscillator, VCO.

Figure 1 is a schematic diagram of a typical arrangement for two point polar modulation. The transmitter (1) comprises a phase locked loop (2), which in turn comprises a phase detector (3), a low pass filter (4), a voltage controlled oscillator VCO (5), a divide-by-n divider (6) and a multimodulus divider MMD (7). The phase modulation signal component (8) and a static RF channel selection signal (9) are summed at a summing unit (10) and the result is fed into a sigma-delta modulator (11). This is in turn fed to the multi-modulus divider MMD (7). This supplies the low frequency components to the PLL. The output of the MMD (7) is fed into the phase detector (3) which compares the output of the MMD with a reference signal (12) generated by a crystal oscillator in the reference signal source (“reference clock”) (13). The phase detector (3) produces an output which is proportional to the difference in phase between its two inputs.

To provide higher frequency components, the phase modulation signal component (8) is fed into a Digital-to-Analogue Converter, DAC (14), which controls the input to a second modulation point (15), this point being an input to the VCO (5). As with the input to the VCO from the low pass filter (4), the signal input to the second modulation point (15) changes the output frequency of the VCO. As the second modulation signal is not applied to the VCO via the low pass filter (4), it is not directly affected by that filter. This arrangement allows phase modulation over a sufficiently wide frequency range. In such polar transmitter architectures however, the design of the direct, second-point modulation port poses some severe design challenges. In particular, the non-linearity of VCO’s presents a problem for phase modulation over a wide frequency range. The polar modulation is completed by adding in the amplitude component (16), via a Digital-toAnalogue Convertor (17), into the amplifier at a third modulation point (18).

Two remedial approaches are treated in prior art, mostly targeting GSM/EDGE systems. First, in a digital PLL two-point modulation variant, elaborate pre-distortion in the digital domain has been proposed in Staszewski et al (“All-Digital PLL and GSM/EDGE Transmitter in 90 nm CMOS”, ISSCC Dig. Tech Papers, pp 316-317, Feb 2005), in order to linearize a digitallycontrolled oscillator (DCO) and its associated gain (KDCO). In Youssef et al (“A low-power GSM/EDGE/WCDMA Polar Transmitter in 65-nm CMOS”, Solid State Circuits, IEEE Journal of, vol. 46, pp3061 - 3074, Dec. 2011), a closed-loop linearized voltage-controlled oscillator (VCO) is used to guarantee an accurate and constant gain (KVCO).

Other solutions have been proposed to cope with the much wider-band OFDM signals used in Wi-Fi (e.g. 20MHz). Gunturi et al., (Principal architectural changes in polar transmitter in DRP design for WLAN, in Communications (NCC), 2013 National Conference on , vol., no., pp.1-5, 15-17 Feb. 2013) propose algorithmic mitigation for phase modulation in a fully digital, DCObased polar Tx architecture. Zheng and H. C. Luong, (A WCDMA/WLAN Digital Polar Transmitter With Low-Noise ADPLL, Wideband PM/AM Modulator, and Linearized PA, in IEEE Journal of Solid-State Circuits, vol. 50, no. 7, pp. 1645-1656, July 2015) perform elaborate DCO-based linearization for OFDM and WCDMA . Ao Ba et al. (“A 1,3nJ/b IEEE 802.11ah Fully Digital Polar Transmitter for loE Applications,” in ISSCC 2016) deal with a reduced bandwidth OFDM (1-2 MHz) through DCO-based polar architecture.

The much wider bandwidth of Wi-Fi OFDM signals coupled with the constraints of practical VCO design, both in terms of supported range (e.g. +/- 50 MHz) and linearity requirements have hitherto hindered the use of direct-modulation, VCO-based polar architectures in Wi-Fi transmitters, despite their potential considerable power consumption savings.

Summary of invention

In general, the invention provides a method of correcting errors in the output of a Voltage Controlled Oscillator, VCO, caused by quantization errors and, optionally, further correcting output errors due to the nonlinearity of the VCO, by estimating the errors based upon a mathematical model, accumulating the estimated errors over a plurality of input samples and periodically providing a correction factor.

Accordingly, in a first aspect of this disclosure, there is provided a method of controlling a voltage controlled oscillator, VCO, comprising receiving a phase modulating signal comprising a sequence of samples, each sample indicating a required frequency deviation, mapping the required frequency deviation of each sample to a second allowed frequency deviation value, wherein the second allowed frequency deviation value is selected from a plurality of allowed frequency deviation values using the second allowed frequency deviation value to control a VCO input and providing a periodic correction to the VCO input.

The second allowed frequency deviation value may be the allowed frequency deviation value closest to the required frequency deviation. Alternatively, optionally, the mapping also prescales the second allowed frequency deviation by a reciprocal gain factor that inverts the nonideal VCO response with respect to an ideal, linear version such that the output VCO frequency deviation delivered is rendered as close to the idealized version as possible.

The periodic correction comprises the steps of (i) determining a frequency deviation correction factor, the determining being performed using a predetermined number of samples and (ii) modifying the VCO input using the frequency deviation correction factor. The frequency deviation control factor is an estimate of an accumulated discrepancy between the required frequency deviations and the respective second allowed frequency deviation values.

In an embodiment the step of determining the frequency deviation correction factor comprises using a model of a non-ideal response of the VCO to produce an estimated VCO output.

In an embodiment, the determining of the frequency deviation correction factor comprises the steps of (a) setting a count to zero, (b) setting an accumulated frequency deviation to zero and (c) for each sample, mapping the required frequency deviation onto a model of a non-ideal response of the VCO to produce an estimated VCO output, determining a difference between output of the required frequency deviation and the estimated VCO, adding the difference to an accumulated frequency deviation value and incrementing the count. If the count is equal to the predetermined number of samples, the method further comprises setting the frequency deviation correction factor equal to a VCO input correction value required to generate a VCO output equal to the accumulated frequency deviation value.

In an embodiment, modifying the VCO input using the frequency deviation correction factor comprises obtaining a resultant frequency deviation by adding the frequency deviation correction factor to the mapped second allowed frequency deviation value of a current sample, determining whether the resultant frequency deviation is inside of a predetermined range. If the resultant frequency deviation is inside the predetermined range, the method further comprises using the resultant frequency deviation to set the VCO input value for the current sample. If the resultant frequency deviation is outside the range, the method further comprises adding a portion of the frequency deviation correction factor to the required frequency deviation of a current sample and adding a further portion of the frequency deviation correction factor to one or more required frequency deviation of subsequent sample(s).

In an embodiment the model of the non-ideal response of the VCO comprises a gain and an offset, and mapping the required frequency deviation of the sample to the model comprises multiplying an index of the required frequency deviation by the gain and adding the offset to determine an estimated VCO output.

In an embodiment the model is a piecewise linear model which comprises a plurality of gains, each of the plurality of gains corresponding to a different set of indices of the required frequency deviations.

In an embodiment the model of the non-ideal response of the VCO comprises a vector, wherein each component of the vector is an estimated VCO output value corresponding to one of the plurality of allowed frequency deviation indices.

In an embodiment the model of the non-ideal response of the VCO is obtained by a calibrating unit by comparing outputs from the VCO with the required frequency deviations obtained from the input calibrating data stream samples.

In an embodiment the modulation data stream is an orthogonal frequency division multiplexed data stream.

In a second aspect there is provided an apparatus for controlling a voltage controlled oscillator, VCO, the apparatus comprising a first connector for receiving a data stream, a second connector for receiving a calibration signal, a third connector for sending a control signal to a VCO, a processor and a memory. The apparatus is configured to receive at the first connector a phase modulating signal comprising a sequence of samples, each sample having a required frequency deviation, map the required frequency deviation of each sample to a second allowed frequency deviation value, wherein the second allowed frequency deviation value is selected from a plurality of allowed frequency deviation values, and use the second allowed frequency deviation value to control a VCO input and provide a periodic correction to the VCO input by the steps of:

(i) determining a frequency deviation correction factor, the determining being performed using a predetermined number of samples;

(ii) modifying the VCO input using the frequency deviation correction factor, wherein the frequency deviation control factor is an estimate of an accumulated discrepancy between the required frequency deviations and the respective second allowed frequency deviation values.

In an embodiment, the apparatus is configured to determine the frequency deviation correction factor by using a model of a non-ideal response of the VCO to produce an estimated VCO output.

In an embodiment, the apparatus is further configured to determine the frequency deviation correction by:

(a) setting a count to zero;

(b) setting an accumulated frequency deviation to zero; and (c) for each sample:

mapping the required frequency deviation onto a model of a non-ideal response of the

VCO to produce an estimated VCO output;

determining a difference between the output of the required frequency deviation and the estimated VCO output;

adding the difference to an accumulated frequency deviation value; and incrementing the count; and (d) if the count is equal to the predetermined number of samples, setting the frequency deviation correction factor equal to a VCO input correction value required to generate a VCO output equal to the accumulated frequency deviation value.

According to a third aspect there is provided a transmitter comprising a Voltage Controlled Oscillator and apparatus for controlling a VCO according to the first aspect.

In addition to the above described embodiments, the invention may be implemented in the following embodiments, described in the numbered paragraphs below:

1. An apparatus for controlling a voltage controlled oscillator, VCO, the apparatus comprising a phase modulating mapping unit, an error calculating unit, an accumulating unit and a multiplexing unit, wherein:

the phase modulating mapping unit is configured to:

receive a phase modulating data stream comprising a sequence of samples, each sample having a required frequency deviation; and for each sample:

map the required frequency deviation of the sample to a second allowed value, wherein the second allowed value is selected from a plurality of allowed frequency deviation values;

the error calculating unit is configured to:

receive the phase modulating data stream; and map the required frequency deviation of the sample onto a model of a non-ideal response of the VCO to produce an estimated VCO output; and determine a difference between the required frequency deviation and the estimated VCO output;

the accumulating unit is configured to:

receive the difference between the required frequency deviation and the estimated VCO output;

add the difference to an accumulated frequency deviation;

increment a count; and if the count is equal to a predetermined number of samples:

forward the accumulated frequency deviation to a multiplexing unit;

set the count to zero; and set the accumulated frequency deviation to zero; and the multiplexing unit is configured to:

receive the accumulated frequency deviation; and inject the accumulated frequency deviation into an input of the VCO.

2. An apparatus according to numbered paragraph 1, wherein the multiplexer is configured to obtain a resultant frequency deviation by adding the accumulated frequency deviation to 15 the required frequency deviation of a current sample;

determine whether the resultant frequency deviation is inside of a predetermined range; and if it is inside of the predetermined range:

use the resultant frequency deviation to set the VCO input value for the 20 current sample;

if it is outside of the range:

add a portion of the accumulated frequency deviation to the required frequency deviation of a current sample;

add a further portion of the accumulated frequency deviation to the 25 required frequency deviation of at least one subsequent sample.

3. An apparatus according to numbered paragraph 2 or numbered paragraph 3, wherein the model of the non-ideal response of the VCO comprises a gain and an offset, and mapping the required frequency deviation of the sample to the model comprises multiplying the required frequency deviation by the gain and adding the offset to determine an estimated VCO output.

4. An apparatus according to any of numbered paragraphs 1 to 3, wherein the model comprises a plurality of gains, each of the plurality of gains corresponding to a different set of required frequency deviations.

5. An apparatus according to any of numbered paragraphs 1 to 4, wherein the model of the non-ideal response of the VCO comprises a vector, wherein each component of the vector is an estimated VCO output value corresponding to one of the plurality of allowed frequency deviation values.

6. An apparatus according any of numbered paragraphs 1 to 5, wherein the model of the non-ideal response of the VCO is obtained by a calibrating unit, the calibrating unit being configured to compare output values of a VCO controlled by the apparatus with required frequency deviations values of an incoming data stream..

7. An apparatus according any of numbered paragraphs 1 to 6, further configured to receive an orthogonal frequency division multiplexed data stream.

Brief description of the figures

Figure 1 is a schematic diagram of a two point modulator according to the prior art;

Figure 2 is a schematic diagram of a control apparatus for a VCO according to the prior art;

Figure 3 is a graphical representation of a typical non-ideal response of a VCO, illustrating the effects of signal quantization and differential non-linearity;

Figure 4 is a graphical representation of a typical non-ideal response of a VCO, illustrating how the errors are most pronounced at the extremes of the range;

Figure 5 is a graphical representation of a typical non-ideal response of a VCO, illustrating the asymmetry of the range;

Figure 6 is a schematic diagram illustrating the components of a VCO controller according to an embodiment.

Figure 7 is a flow chart illustrating a method of control of a voltage controlled oscillator according to an embodiment;

Figure 8 is a flow chart illustrating a method of determining a frequency deviation correction factor according to an embodiment;

Figure 9 is a flow chart illustrating the steps of a method of implementing the modification of a VCO input using a frequency deviation correction factor according to an embodiment;

Figure 10 is a flow chart illustrating a method of control of a voltage controlled oscillator 10 according to another embodiment;

Figure 11 is a flow chart illustrating the steps of a method for determining a total compensated frequency deviation value;

Figure 12 is a schematic diagram of the functional components of a VCO controller according to one embodiment;

Figure 13 is a schematic diagram of the functional components of a VCO controller according to another embodiment; and

Figure 14 is a schematic diagram of the physical components of a VCO controller according to an embodiment.

Detailed description

Embodiments will now be described more fully hereinafter with reference to the accompanying drawings, in which certain embodiments are shown. However, other embodiments in many different forms are possible within the scope of the present disclosure. Rather, the following embodiments are provided by way of example so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

The present disclosure provides a method and apparatus for providing correction to the phase errors caused by the limited range and non-linearity of a Voltage Controlled Oscillator (VCO). This is performed by statistical pre-processing of the phase modulation stream fed to the VCO.

In an embodiment, the incoming phase modulating data stream is mapped onto the VCO’s nonideal response The statistical preprocessing provides a correction for the control voltage of the VCO.

Figure 2 is a schematic diagram of a control arrangement for a VCO. The data stream (19) is mapped at a phase modulation map (20) to control words which control the VCO (22). Typically the data stream is a phase modulating stream for the second point of a two point modulator. The required frequency deviation of the modulating signal is mapped onto a control word. Each control word is a digital representation of the voltage needed to produce the required frequency deviation from the VCO. The control word is forwarded to a digital-to-analogue converter (21), which provides the control voltage for the VCO.

The phase output from a VCO over a transmission time slot T_T is given by:

0(t) = 2π f^^T[g₁(_t)₊g_2(t)] d£ (1) where gi(t) and g₂(t) are first and second modulation point components respectively. These are provided to the VCO in terms of control words, which are selected from a set of allowed values corresponding to a set of allowed frequency deviations.

The control words do not however provide an exact representation of the incoming frequency deviation requirements. Quantization of the signal is a direct result of the digitization, along with distortions due to non-linearity of the VCO and the asymmetry of the range. Figures 3, 4 and 5 are graphical representations of typical VCO responses. Figure 3 is a graph of frequency step against control word, illustrating the effects of signal quantization and the non-uniformity of steps. The graph (25) plots the frequency step in kilo Hertz (kHz) (23) against control word (24). Figure 4 is a graph (28) of integral non-linearity (26) against control word (27) over the control range, illustrating how the errors are most pronounced at the extremes of the range. Figure 5 is a graph (31) which plots VCO frequency (29) against control word (30) illustrating the asymmetry of the range.

These errors and distortions in the input and response result in inaccuracies in the phase output of the VCO, which may be represented mathematically as:

0'(t) = 2π/θ^Γτ[#_ιω + g_2(t) + e(t)] dt (2) where e(t) is the error in frequency deviation and cp’(t) is the distorted phase output. Integrating the above expression and separating the terms gives:

0'(t) = 0(t) + f^e(t)dt (3)

0'(t) = 0(t) + 0_e(t) (4) where <p_e(t)is the phase error.

The phase errors which are introduced in the manner described above can be cumulative and cause a significant error in the output when they build up. The mechanism of the present invention is to estimate the error on each sample and to sum the cumulative error over a plurality of samples. A correction factor is then added into the VCO over one or more samples to compensate for the period over which the error has accumulated. In an embodiment, the errors compensated are those caused by quantization of the input samples. In another embodiment, the non-linearity of the VCO response is also compensated.

The error caused by the quantization is referred to as the periodic error and the compensation for the non-linearity is referred to as the reciprocal gain pre-scaling. The reciprocal gain pre15 scaling is estimated by dividing the ideal response of VCO by the non-ideal response to produce a reciprocal gain (equation 5):

_ . , . Ideal response

Reciprocal qain = -:Non-ideal response (5)

The reciprocal gain pre-scaling is determined by multiplying the reciprocal gain predistortion by the input frequency deviation (equation 6)

Reciprocal gain prescaling = reciprocal gain x input frequency deviation (6)

The total compensation is then determined by adding the reciprocal gain pre-scaling to the periodic error (equation 7).

Total compensation = Reciprocal gain pre-scaling+periodic error (7)

In an embodiment, frequency deviation correction is performed over a correction cycle 25 comprising a predetermined number of samples of the modulating signal. For each sample, the sample value represents a frequency deviation to be converted to a voltage for use in control of the VCO. A correction factor is then added into VCO control words over one or more samples to compensate for the period over which the error has accumulated.

A target use for a non-ideal VCO is in Orthogonal Frequency Division Multiplexed (OFDM) systems. An OFDM signal comprises a plurality of subcarriers modulated separately and then combined together by an inverse Fourier Transform. A typical data stream for such an OFDM system, wherein the subcarriers are modulated by a vector d = [di, d₂, ...d_K] may be represented as:

sd)=^=,d_le²^^t (8) where s(t) is the time domain signal and vector d = [d_1; d₂, ...d_K] represents the modulation of each of the subcarriers.

A phase error as described above will result in a time varying multiplicative effect in the complex time domain, which may be represented as:

s'(t) = s(t~)e^^e^ (9) where <p_e(t) represents the phase error.

In order to determine the effect of the time-varying error, the OFDM signal can be demodulated. The demodulation process may be represented as follows.

(10) s(t) = ^Ef=i d_t f₀ ^Te^27l^i-fk)t_e^_e(t) dt (11) s(t) = ^Ef=id_t f_Q ^Te^27Tj^~^^t+lo^Te^dt] _{dt (12)}

The last integral term of equation 12 represents the error integral which the present invention seeks to minimize. It is readily observable that the impact of phase error accumulation in OFDM is far-reaching, affecting all data subcarriers within one symbol duration T and beyond throughout a transmission time slot T_T.

Figure 6 is a schematic diagram illustrating the components of a VCO controller according to an embodiment. The data stream (19) for modulating the VCO is passed to the phase modulator (23). The phase modulator comprises the phase mapping (20), an error calculator (24), an accumulator (25) and a summing unit (26). The data stream is a phase modulating signal comprising a sequence of samples, each sample indicating a required frequency deviation. The data stream is sent to the phase mapping unit (20), where it is mapped onto a second allowed frequency deviation and optionally pre-scaled by a reciprocal gain factor. The second allowed frequency deviation value is selected from a plurality of allowed frequency deviation values. This value is then sent through the summing unit (26) to the digital-to-analogue converter (21) and thence to the VCO (22). This provides the conventional control of the VCO.

In parallel, in an embodiment, the data stream is passed to the error calculator unit (24). The purpose of the error calculating unit is to provide a periodic correction to the VCO input. The error calculator unit comprises a model of the non-ideal response of the VCO, which is characterised prior to use of the error calculating unit. In an embodiment, the model is preprogrammed as a permanent set of values after one initial calibration at start-up time. In another embodiment, a continuous real-time calibration of the VCO takes place to maintain an accurate model. However, this is expensive to implement and very power hungry. Hence in order to overcome this problem, in an embodiment, the non-ideal response comprises a scalar model comprising a scalar gain and an offset value. In another embodiment, the non-ideal response comprises a plurality of scalar gain values, each value being associated with a different part of the frequency deviation range. In yet another embodiment, the non-ideal response comprises a vector value, with a frequency deviation error value associated with each of the plurality of allowed frequency deviation values, effectively each of the vector components being associated with a VCO control word.

Once the sample has been mapped onto the non-ideal response, the error is sent to the accumulator (25), where it is summed with the current accumulator value. The correction required is then periodically introduced into input of the VCO.

Figure 7 is a flow chart illustrating the steps of a method according to an embodiment. The method comprises:

(1) Receiving (27) a phase modulating signal comprising a sequence of samples, each sample having a required frequency deviation;

(2) Mapping (28) the required frequency deviation of each sample to a second allowed frequency deviation value and optionally pre-scaling by a reciprocal gain factor, wherein the second allowed frequency deviation value is selected from a plurality of allowed frequency deviation values;

(3) Using the second allowed frequency deviation value to control a VCO input (29);

(4) Providing (30) a periodic correction to the VCO input by the steps of:

(a) determining (31) a frequency deviation correction factor, the determining being performed using a predetermined number of samples;

(b) modifying (32) the VCO input using the frequency deviation correction factor.

Figure 8 is a flow chart illustrating the steps of a method of determining the frequency deviation 10 correction factor according to an embodiment. The method comprises:

(1) Setting a count to zero (33);

(2) Setting an accumulated frequency deviation correction factor to zero (34); and

For each sample (35):

(3) Mapping the required frequency deviation (36) onto a model of a non-ideal response of the VCO to produce an estimated VCO output;

(4) Determining a difference (37) between output the required frequency deviation and the estimated VCO;

(5) Adding the difference (38) to an accumulated frequency deviation value;

(6) Incrementing the count (39); and determining if the count is equal to the predetermined number of samples (40), and if so:

(7) Setting (41) the frequency deviation correction factor equal to a VCO input correction value required to generate a VCO output equal to the accumulated frequency deviation value.

If the count is not equal to the predetermined number of samples, steps 36 to 39 are repeated for the next sample.

In an embodiment, the correction to the output of the VCO is implemented by altering the input value of the control word. Optionally, the VCO input correction factor is added to the VCO control word for the sample after the predetermined number of samples has been reached. Optionally, the control word associated with second allowed frequency deviation to the sum of the frequency deviation for the sample and the frequency deviation correction factor is used. If however the VCO input correction factor is too large to be added to the control word for the sample, another approach is needed. Optionally a technique known as clip and distribute is implemented. This comprises implementing only part of the correction factor in one sample and then implementing any residual correction on subsequent samples. In an embodiment, a modulo operation is used to provide the frequency deviation correction instead of clip and distribute. Clip and distribute is however, more flexible in that a choice can be made as to where and how much frequency deviation discrepancy is released onto adjacent samples.

Figure 9 is a flow chart illustrating the steps of a method of implementing the modification of the

VCO input using the frequency deviation correction factor. The method comprises:

(1) Obtaining (42) a resultant frequency deviation by adding the frequency deviation correction factor to the second allowed frequency deviation value of a current sample;

(2) Determining (43) whether the resultant frequency deviation is inside of a predetermined range; and

If it is inside (44) the predetermined range:

Using (45) the resultant frequency deviation to set the VCO input value for the current sample; and

If it is outside (46) the predetermined range:

(a) Adding (47) a portion of the frequency deviation correction factor to the required frequency deviation of a current sample; and (b) Adding (48) a further portion of the frequency deviation correction factor to the required frequency deviation of one or more subsequent sample(s).

In an embodiment as large a portion as possible of the VCO correction factor may be added to the first available control word and the maximum portion of any remaining portion to the following control word, so as to inject the VCO correction factor as soon as possible. In an alternative embodiment, it may be divided equally over a number of control words. The person skilled in the art will appreciate that there are many alternative implementations and the invention is not limited to any one way of implementing the VCO input correction.

In an embodiment, the model used for the non-ideal response is a linear response comprising a gain value and an offset. This arrangement is suitable if a low processing time overhead is required and/or low processing power consumption for the calibration electronics. However, if the transmitter can tolerate longer calibration latency and the necessary power consumption, then a vector VCO response is measured. In an embodiment, a frequency response per control word is measured. This provides a comprehensive pre-distortion and residual perturbation that is matched to the VCO as closely as possible. This results in more accurate modulation and is suitable for systems such as 64-QAM. However the long term statistics are equivalent for both scalar and vector based approaches. This means that at the expense of some instantaneous EVM degradations, a simple linear scalar model with light-weight calibration will equally stabilize the PM transmission stream similar to the heavy-weight vector-based calibration. In a further embodiment, a compromise between the two approaches is used, in which a piecewise linear model with a plurality of linear portions having respective gains is used for a section of the characteristic to provide the error estimate.

The long-term statistics of the scalar- and vector-based models are equivalent. This is an example of a result of the central limit theorem, in this case acting in relation to the fine-grained, per-sample and accumulated VCO error estimates.

While the instantaneous fine-grained estimate will have a distribution heavily dependent on the VCO response curve (hence harder to accurately approximate), the distribution of the binned version (i.e. accumulated) will tend to a Gaussian distribution the longer the binning interval (and is therefore less critically dependent on the VCO response curve). This may be demonstrated by modelling the instantaneous frequency deviation errors of data-bearing OFDM symbols as independent and identically distributed (i.i.d) random variables X_b ..., X_b (usually data payload is i.i.d by construction). If the binning interval is designated as b, then the average X_b random variable is approximately Normally distributed, with a mean that is consistent with the mean of the true fine-grained error μ_βΓΓ0Γ and a standard deviation that equals σ/yfb, where σ is the standard deviation of the true fine-grained error. That is, by virtue of binning, the standard deviation of the new statistic is made less sensitive to σ of the original fine-grained error.

In an embodiment, in addition to quantization errors, the non-linear response of the VCO is also corrected. Figure 10 is a flow chart illustrating the steps of a method according to this embodiment. The steps of this method are the same as those of the method of Figure 7, except for steps (2) and (4)(a):

1) Receiving (27) a phase modulating signal comprising a sequence of samples, each sample having a required frequency deviation;

(2) Mapping (49) the required frequency deviation of each sample to a second allowed frequency deviation value and optionally pre-scaling by a reciprocal gain factor, wherein the second allowed frequency deviation value is selected from a plurality of allowed frequency deviation values, wherein the mapping comprises determining a second allowed frequency value using a model of the non-ideal response of the VCO;

(4) Providing (30) a periodic correction to the VCO input by the steps of:

(a) determining (50) a frequency deviation correction factor, the determining being performed using a predetermined number of samples;

Figure 11 is a flow chart illustrating the steps of a method for determining a total compensated value:

1) Estimating (51) a periodic error for a group of samples;

2) Using (52) a reciprocal gain, the reciprocal gain being an ideal response of a VCO divided by a non-ideal response, to estimate reciprocal gain pre-scaling of input frequency deviation;

3) Summing (53) the reciprocal gain pre-scaling and the periodic error to provide a total compensated value.

Figure 12 is a schematic diagram of a VCO controller according to an embodiment. The controller is configured to receive a phase modulating data stream (19). The controller comprises a phase modulator (23), which further comprises a mapping unit (20), an error calculating unit (24), an accumulator (25) and a summing unit (26). A digital-to-analogue convertor (21) is provided to supply a voltage signal to a VCO (22). The controller further comprises a calibration unit (54). During calibration, the phase modulating data stream is forwarded to the calibration unit, which compares the modulating stream with an output (55) from the VCO in order to produce a model of the non-ideal response of the VCO. In an embodiment the calibration unit operates to continually update this model in real time. The calibration unit may produce a scalar based model, a vector based model or the compromise described above. The gain values from the calibration unit are forwarded to the error calculation unit (24), which, in an embodiment, implements steps 27 to 31 of the method of Figure 7. A frequency deviation correction value is forward and stored in the accumulator unit (25). In an embodiment, the correction factor is added to the VCO input by means of a multiplexer (56), which outputs zero values (57) except at the end of a correction cycle, when the correction value is output. The correction cycle is controlled by a clock signal (58).

Figure 13 is a modified version of Figure 12, which enables the implementation of the embodiment of Figure 10. The components are the same, except that an additional connection (59) is provided from the calibration unit (54) to the mapping unit (60). The PM mapping unit (60) is adapted to select the second allowed frequency based on both the periodic error and the non-linear response of the VCO. In an embodiment, the calculation error unit (61) is similarly adapted.

Figure 14 is a schematic diagram showing the physical components of an apparatus for a VCO controller (62) according to an embodiment. There is provided a first connector (63) for receiving a data stream, a second connector for receiving a calibration signal (64), a third connector (65) for sending a control signal to a VCO, a memory (66) for storing data and at least one processor (67) for implementing instructions. In an embodiment the apparatus is configured to receive at the first connector a phase modulating signal comprising a sequence of samples, each sample having a required frequency deviation, map the required frequency deviation of each sample to a second allowed frequency deviation value and optionally pre-scale by a reciprocal gain factor, wherein the second allowed frequency deviation value is selected from a plurality of allowed frequency deviation values, use the second allowed frequency deviation value to control a VCO input and provide a periodic correction to the VCO input by the steps of:

(a) determining a frequency deviation correction factor, the determining being performed using a predetermined number of samples;

(b) modifying the VCO input using the frequency deviation correction factor,

In an embodiment the apparatus is further configured to determine the frequency deviation correction by:

(a) setting a count to zero;

mapping the required frequency deviation onto a model of a non-ideal response of the VCO to produce an estimated VCO output;

determining a difference between output the required frequency deviation 15 and the estimated VCO;

The present disclosure has mainly been described above with reference to a few embodiments. However, as is readily appreciated by a person skilled in the art, other embodiments than the ones disclosed above are equally possible within the scope of the present disclosure, as defined by the appended claims.

Claims

Claims

1. A method of controlling a voltage controlled oscillator, VCO, comprising:

receiving a phase modulating signal comprising a sequence of samples, each sample indicating a required frequency deviation;

5 mapping the required frequency deviation of each sample to a second allowed frequency deviation value optionally pre-scaled by a reciprocal gain factor, wherein the second allowed frequency deviation value is selected from a plurality of allowed frequency deviation values;

using the second allowed frequency deviation value to control a VCO input; and

10 providing a periodic correction to the VCO input by the steps of:

(i) determining a frequency deviation correction factor, the determining being performed using a predetermined number of samples;

(ii) modifying the VCO input using the frequency deviation correction factor, wherein the frequency deviation control factor is an estimate of an accumulated

15 discrepancy between the required frequency deviations and the respective second allowed frequency deviation values.
2. A method according to claim 1, wherein the step of determining the frequency deviation correction factor comprises using a model of a non-ideal response of the VCO to produce an

20 estimated VCO output.
3. A method according to claim 2, wherein the mapping pre-scales the second allowed frequency deviation by a reciprocal gain factor comprising a reciprocal of the non-ideal VCO response with respect to an ideal response of the VCO.

25
4. A method according to any preceding claim, wherein determining the frequency deviation correction factor comprises:

(a) setting a count to zero;

(b) setting an accumulated frequency deviation to zero; and (c) for each sample:

mapping the required frequency deviation onto a model of a non-ideal response of the VCO to produce an estimated VCO output;
5 determining a difference between output of the required frequency deviation and the estimated VCO;

adding the difference to an accumulated frequency deviation value; and incrementing the count;

(d) if the count is equal to the predetermined number of samples, setting the

10 frequency deviation correction factor equal to a VCO input correction value required to generate a VCO output equal to the accumulated frequency deviation value.

5. A method according to any preceding claim, wherein modifying the VCO input using the 15 frequency deviation correction factor comprises:

obtaining a resultant frequency deviation by adding the frequency deviation correction factor to the second allowed frequency deviation value of a current sample;

determining whether the resultant frequency deviation is inside of a predetermined range;and

20 if it is inside the predetermined range:

using the resultant frequency deviation to set the VCO input value for the current sample; and if it is outside the range:

adding a portion of the frequency deviation correction factor to the required 25 frequency deviation of a current sample: and adding a further portion of the frequency deviation correction factor to the required frequency deviation of at least one subsequent sample.
6. A method according to any of claims 2 to 5, wherein the model of the non-ideal response 5 of the VCO comprises a gain and an offset, and mapping the required frequency deviation of the sample to the model comprises multiplying the required frequency deviation by the gain and adding the offset to determine an estimated VCO output.
7. A method according to claim 6, wherein the model is a piecewise linear model which 10 comprises a plurality of model portions having respective gains, each of the plurality of model portions corresponding to a different set of required frequency deviations.
8. A method according to any of claims 2 to 5, wherein the model of the non-ideal response of the VCO comprises a vector, wherein each component of the vector is an estimated VCO

15 output value corresponding to one of the plurality of allowed frequency deviation values.
9. A method according to any of claims 2 to 8, comprising the step of obtaining the model of the non-ideal response of the VCO by a calibrating unit by comparing outputs from the VCO with the required frequency deviations obtained from a calibrating data stream of samples.
10. A method according to any preceding claim, wherein the data stream is an orthogonal frequency division multiplexed data stream.
11. An apparatus for controlling a voltage controlled oscillator, VCO, the apparatus 25 comprising a first connector for receiving a data stream, a second connector for receiving a calibration signal, a third connector for sending a control signal to a VCO, a processor and a memory, the apparatus being configured to:

receive at the first connector a phase modulating signal comprising a sequence of samples, each sample indicating a required frequency deviation;

map the required frequency deviation of each sample to a second allowed frequency deviation value optionally pre-scaled by a reciprocal gain factor,

5 wherein the second allowed frequency deviation value is selected from a plurality of allowed frequency deviation values;

use the second allowed frequency deviation value to control a VCO input; and provide a periodic correction to the VCO input by the steps of:

(i) determining a frequency deviation correction factor, the determining

10 being performed using a predetermined number of samples;

(ii) modifying the VCO input using the frequency deviation correction factor, wherein the frequency deviation control factor is an estimate of an accumulated discrepancy between the required frequency deviations and the respective second

15 allowed frequency deviation values.
12. An apparatus according to claim 11, configured to determine the frequency deviation correction factor by using a model of a non-ideal response of the VCO to produce an estimated VCO output.
13. An apparatus according to claim 12, wherein the apparatus is further configured to prescale the second allowed frequency deviation by a reciprocal gain factor comprising a reciprocal of the non-ideal VCO response with respect to an ideal response of the VCO.

25
14. An apparatus according to any of claims 11 to 13, further configured to determine the frequency deviation correction by:

(a) setting a count to zero;

(b) setting an accumulated frequency deviation to zero; and (c) for each sample:

mapping the required frequency deviation onto a model of a non-ideal response of the VCO to produce an estimated VCO output;

5 determining a difference between output of the required frequency deviation and the estimated VCO;

adding the difference to an accumulated frequency deviation value; and incrementing the count;

(d) if the count is equal to the predetermined number of samples, setting the

10 frequency deviation correction factor equal to a VCO input correction value required to generate a VCO output equal to the accumulated frequency deviation value.
15. An apparatus according to any of claims 11 to 14, wherein the apparatus is further 15 configured to obtain a resultant frequency deviation by adding the accumulated frequency deviation to the required frequency deviation of a current sample;

determine whether the resultant frequency deviation is inside of a predetermined range; and

20 if it is inside of the predetermined range:

use the resultant frequency deviation to set the VCO input value for the current sample;

if it is outside of the range:

add a portion of the accumulated frequency deviation to the required 25 frequency deviation of a current sample;

add a further portion of the accumulated frequency deviation to the required frequency deviation of at least one subsequent sample.
16. An apparatus according to any of claims 11 to 15, wherein the model of the non-ideal 5 response of the VCO comprises a gain and an offset, and mapping the required frequency deviation of the sample to the model comprises multiplying the required frequency deviation by the gain and adding the offset to determine an estimated VCO output.
17. An apparatus according to any of claims 12 to 16, wherein the model comprises a 10 piecewise linear model which comprises a plurality of model portions having respective gains, each of the plurality of model portions corresponding to a different set of required frequency deviations.
18. An apparatus according to any of claims 12 to 16, wherein the model of the non-ideal 15 response of the VCO comprises a vector, wherein each component of the vector is an estimated

VCO output value corresponding to one of the plurality of allowed frequency deviation values.
19. An apparatus according to any of claims 12 to 16, further comprising a calibration unit configured to derive the model by characterizing the non-ideal response of the VCO by
20 comparing outputs from the VCO with the required frequency deviations obtained from a calibrating data stream of samples.

20. An apparatus according to any of claims 11 to 19, further configured to receive an orthogonal frequency division multiplexed data stream.
21. A two point modulator comprising a Voltage Controlled Oscillator and apparatus for controlling a voltage controlled oscillator, VCO, according to any of claims 11 to 20.
22. A transmitter comprising a Voltage Controlled Oscillator for generating a signal for transmission and an apparatus for controlling a voltage controlled oscillator, VCO, according to any of claims 11 to 20.

5
23. A transmitter according to claim 22 wherein the transmitter is configured to transmit an

Orthogonal Frequency Division Multiplexed signal.
24. An apparatus for controlling a Voltage Controlled Oscillator substantially as herein described with reference to accompanying Figures 6 to 14.
25. A method for controlling a Voltage Controlled Oscillator substantially as herein described with reference to accompanying Figures 6 to 14.

Intellectual

Property

Office

Application No: Claims searched:

GB1612261.6

1-23