1. Field of the Invention
The present invention relates generally to RF impedance measurement and in particular to RF impedance measurements using two point voltage sampling without a phase detector. Some embodiments also relate to adjusting an impedance matching network after the measurement.
2. Description of Related Art
Mobile handsets such as cellular phones are being manufactured using higher levels of integration and use in broader frequency band coverage. As a result, the performance limits of embedded antenna technology are being stretched. Variations in load impedance at the antenna due to environmental changes such as the position at which the phone is held, the frequency band being used and other contributors create a mismatch or increased voltage standing wave ratio (VSWR) at the antenna port. In addition, the body effects of a head or hand near the antenna contribute to capacitive loading which also results in an impedance mismatch. This can lead to a shift in antenna center frequency and an increased VSWR mismatch. In addition to reception problems, any mismatch will further result in a reduction in power radiated from the antenna.
The antenna impedance Zant can change, as previously noted, due to a change in the physical environment surrounding the antenna. The impedance at the input of the tunable matching network 30 is monitored by periodically measuring the amplitude of the RF voltage at the input and output of the directional coupled 28 using respective peak detectors 32A and 32B. The phase relationship between the two detected voltages is measured using a phase detector 34. The peak voltage measurements and the phase measurement are then provided to a processing device 36 such as a digital signal processor to compute the impedance. In the event the measured impedance differs from the target impedance due to a change, by way of example, in the antenna characteristics, the processor adjusts the tunable matching network 30 as needed to return to the target impedance.
The above-described approach requires an impedance sensing section which is separate from the impedance matching section. In addition, a phase detector is used. A phase detector having good accuracy and low current is difficult to achieve over the 690 Mhz to 2690 Mhz range of interest in many cell phone applications. As will become apparent to those skilled in the art upon a reading of the following Detailed Description of the Invention together with the drawings, an RF impedance improved detection scheme is disclosed which does not rely upon a phase detector and which does not require a sensing element separate from the matching network.
Referring again to the drawings,
The adaptive matching network module 38 initially transforms the impedance of the antenna 24 to a target impedance which may be, by way of example, a 50Ω real impedance. Environmental fluctuations may cause the impedance of antenna 24 to change so that the matching network is no longer optimal. As will be described, the adaptive matching network module 38 monitors the impedance of the matched network and, if the impedance varies from the target value, will adjust the matching network so that the impedance is returned to the target value.
The exemplary matching network used in module 38 is a pi type network which includes a series connected inductor Lsense and a pair of shunt connected capacitor arrays C1 and C2 disposed on either side of the inductor. The capacitor arrays each include an array 44A and an array 44B of RF-MEMS (micro-electromechanical system) capacitive switches Cn to C1a. The capacitive switches are preferably disposed in a binary weighted manner, with there being five capacitive switches connected in parallel, with the relative capacitive values being C, 2 C, 4 C, 8 C and 16 C. The five capacitive switches are individually enabled and disabled to provide a total capacitance ranging from C to 31 C in increments of C. As is well known, high voltage switching circuitry (not depicted) is used to control the state of each of the five switches. Lsense has a typical inductance of 2 to 8 nano-Henries, with the value of C of the capacitive switches being 0.5 to 4.0 pF. Each capacitor bank further includes a small (Co<0.125 pF) switched capacitor which is periodically connected in parallel with each of the MEMS capacitive switches 44A and 44B. The smaller the value of Co, the greater the voltage detection accuracy required of the RF detectors employed as peak detectors 52A and 52B to be described. A dither clock present on line 50 is used to control the states of switches 48A and 48B which operate to switch capacitors Co in circuit and out of circuit. The frequency of the dither clock is determined by the required response time of the RF impedance measurement, which may be as low as a 100 Hz or up to around 1 MHz. Preferably the dither frequency is not so high as to introduce spikes on the RF sensing lines.
In addition to forming part of the impedance matching, inductor Lsense also functions as part of the impedance sensor. A pair of peak voltage detectors 52A and 52B are connected to detect respective voltages V1 and V2 at opposite ends of inductor Lsense. The voltages are periodically sensed when the switched capacitors 46A and 46B are connected in circuit by switches 48A and 48B and then sensed a second time when the capacitors are switched out of circuit. As will be explained, these four voltage measurements permit the impedance looking into the matching network to be determined. In the event that measured impedance is out of range, the matching network is adjusted by way of capacitor switches 44A and 44B to bring the impedance back into range. A control block 54 provides various control functions, including the production of the dither clock on line 50, control of the peak detectors 52A and 52B, the computation of the actual network impedance and the re-adjustment of the adaptive matching network to bring the impedance back into range.
Note that
Z
L
=R
L
+jX
L (1)
Voltages V1 and V2 are measured using respective peak detectors 52A and 52B when switch 48B is opened based upon the polarity of the dither clock on line 50 so that dither capacitor 46B (Co) is out of circuit. Thus, it can be seen from the
The value of ZL is preferably determined using signal processing circuitry disposed within control unit 54. The phase angle φ is expressed as follows:
Cos φ=−0.5[Xdp2(Vr12−Vr22)+XS2]/(XS*Xdp*Vr1) (2)
where,
Once the phase angle is known, the reactive component XL and real component RL of the impedance ZL can be calculated as follows:
X
L=(XS/2)[(Vr12−1)/(Vr12+1−2Vr1 cos φ)−1] (3)
and
R
L
=[X
S
2/(Vr12+1−2Vr1 cos φ)−XL2]1/2 (4)
Assuming that the value of ZL has moved away from the target value because, for example, of changes in the antenna environment, the signal processor in the control unit 54 will proceed to alter the matching characteristics in the matching network. As will be described in greater detail, this is carried out by changing the value(s) of capacitors 44A and 44B.
A change in the antenna load impedance is simulated in the timing diagrams at a time T1=25 μs. Prior to that time, it can be seen from
Z
L1
=R
L1
+jX
L1=100−j35.4 Ω (5)
The change in antenna impedance could be caused, by way of example, by a change in the antenna environment such as adjusting the manner in which a cell phone is held. As can be seen in
Z
L2
=R
L2
+jX
L2=50−j17.7 Ω (6)
If it is assumed that the target impedance is reflected by equation (5) above, the control unit 54 will then precede to alter the matching network by way of the MEMS 44A and 44B so that the matched impedance has returned to the target impedance. One approach for adjusting the matching network will now be described. As will be seen, only the change in matching network capacitance to arrive at the target values is needed and not the actual final value of that capacitance.
As was previously shown by equations (3) and (4), the values for RL and XL represent the respective real and imaginary components of the measured impedance. Using these values, the needed change in value of matching network capacitances 44A and 44B, the MEMS capacitor arrays, is determine using a signal processor or the like. A chart of the complex impedance plane is shown in
Initially, assume that that the matching network is at the optimum value to transform the present antenna impedance Zant to the optimum value in this example of in this example of 50+j0Ω purely real resistance. This condition is represented on the
In order to return the altered impedance to the target impedance at point A, it is usually necessary to adjust both the value of capacitances C1 and C2 of the matching network. First, the value of C2 is changed by ΔC2 to provide a new value of ZL, referred to here as ZLnew. By adding a parallel reactance, the impedance moves along an arc 72A of a constant admittance circle from point B to point C. The distance and direction of the movement is a function of size of the change ΔC2 and the polarity. In the present example, the polarity is positive (C2 is to be increased). The magnitude of ΔC2 is determined so that point C is at a location in the complex plane such that, when the fixed value inductor Lsense of impedance XS is added in series, the combined, new value of impedance will fall on the constant admittance circle 69 of 20 milli-Siemens. That value at point D is the sum of ZLnew plus XS. At this point, a value of C1 of the matching network is then produced which provides a reactance X1 which is of a magnitude sufficient to move the impedance ZLnew plus XS to close to a pure resistance of 50Ω as represented by point A. Since the MEMS cap arrays 44A and 44B that make up the majority of respective capacitances C1 and C2 have only a finite number of possible values, the final impedance value may differ somewhat from the ideal value of 50Ω.
In order to carry out the above transformation, one approach is to first determine the change in capacitance C2 to move from point B to point C of
ΔX2=XS[XSXL+RL2+XL2−(RL(−XSRL50RL2+50XL2))1/2]/[(XL+XS)2+RL(RL−50)] (7)
where
XS is the impedance of the inductor Isense;
XL is the measured reactive component of ZL per equation (3);
RL is the measured resistive component of ZL per equation (4); and
the value 50 is target impedance in ohms.
Thus, the needed change to the present value of C2 in order to move from point B to point C of
ΔC
2−1/(ωΔX2) (8)
where ω is the radial frequency 2nf.
The new value of C1 needed to shift the full combined impedance (matching network+Zant) from point D back to point A is then determined. The equation for calculating the impedance X1 provided by the new value of C1 is set forth below. Variables Rn and Xn, to be defined later, are used to simplify the following equation for X1.
X
1=5[10Xn+(−100Rn2+2RnXn2+2Rn3)1/2]/(Rn−50) (9)
where,
Rn is a variable determined by equation (11) below; and
Xn is a variable determined by equation (12) below.
The new value of C1 is then as follows:
C
1=−1/(ωX1) (10)
where ω is the radial frequency 2nf.
The values of variables Rn and Xn used in equation (9) are as follows:
R
n=(ΔX22RL)/[RL2+(ΔX2+XL)2] (11)
and
X
n
=X
S
+[R
L
2
ΔX
2
+X
L
2
ΔX
2
+X
L
ΔX
2
2]/[RL2+(ΔX2+XL)2] (12 )
where
XS is the reactance of inductor Lsense;
ΔX2 is the reactance of C2 per equation (7); and
RL and XL are the real and imaginary parts of ZL per equations (3) and (4).
Thus, once the new value of C1 of the matching network has been provided per equation (10), the impedance looking into the matching network on the C1 side will have returned to point A of
Note that MEMS switched capacitors 44A and 44B if
As can be seen in
As previously noted, the impedance matching networks of
The RF detectors are implemented in both the
Thus, various embodiments of an adaptive impedance network and associated circuitry have been disclosed. Although these embodiments have been described in some detail, certain changes can be made by those skilled in the art without departing from the spirit and scope of the present invention as defined by the appended claims.