by Dr. Ronen Holtzman, General Microwave Israel Ltd.

When spectrum efficiency is important, Narrowband FM (NBFM) is used but when better signal quality is required, Wideband FM (WBFM) is used at the expense of greater spectrum usage. The term WBFM is used in applications where the modulation index is equal to or larger than 1. However, in this article, we are going to address applications and techniques for WBFM with modulation indexes much larger than that, going up to 100 and beyond. In such applications spectral efficiency is less important and sometimes large spectral spread is actually desired. The purpose of this article is to present some major applications in the commercial and defense markets. Within this framework, the common techniques of generating WBFM are presented.

Application Example – FM-CW Radars

FM-CW Radars generate a continuous-wave (CW) signal that is typically modulated by a saw-tooth waveform; such a signal is called a Chirp. This signal is then amplified and transmitted. The received signal is amplified, filtered and converted to zero-IF by mixing with the transmitted signal. The basic block diagram of the FM-CW transmitter is shown in Figure 1. The

Figure 2: FM-CW transmit (red) and received (green) signals

received signal is delayed by the time it takes the signal to reach the target and return. Also, the frequency of the received signal is shifted by the Doppler Effect due to the target’s relative velocity. Overall, by the comparison (or mixing) of the transmitted and received signals both the range and the velocity of the target can be extracted. This principle is shown in Figure 2.

The advantage of FM-CW Radar is excellent Signal to Noise Ratio (SNR) and, since it transmits all the time, the simplicity of the information extraction and the ability to detect very close-range targets.

Pulsed radars, for instance, cannot receive the signal while transmitting. The result is a “Shadow time” that prohibits the pulsed radar from detecting very closed-range targets.

The FM-CW radar overcomes this problem and can support very close-range targets. In order to get an accurate reading of a target, the frequency change rate must be very high, so there will be a detectible frequency difference between the transmitted and received signals. Therefore, FM-CW radars use a very wideband FM modulation technique.

Techniques

There are many techniques to generate a WBFM signal: analog based techniques, digitally based techniques and hybrid techniques. In this article the commonly used solutions and hardware are reviewed.

Free Running Voltage Controlled Oscillator

A free running VCO is a device based on an unstable transistor circuit. The frequency of oscillation depends on the resonance frequency set by its equivalent capacitance and inductance. By applying variable bias voltage to a varactor diode, the capacitance is changed and the oscillating frequency is changed accordingly. Simplified schematic is shown in Figure 3.

The VCO is a very low cost method of generating WBFM signals, such as chirp signals. The VCO has some important properties that are common to all frequency sources. The definitions of these properties are detailed below and will be used for the rest of the article:

Frequency Range

This parameter is defined as the lowest and highest frequencies generated by the VCO. A VCO may cover a full octave band.

Settling Time

This parameter is defined as the time it takes the VCO to reach the final frequency within an allowable window. Typical values are 50 ns to a ±10 MHz and 1 µs to a ±4 MHz for a 12 to 18 GHz jump.

Post-Tuning Drift

After a VCO reaches what seems to be its final frequency it may slowly drift until it reaches the real final value. This post-tuning drift may cause an additional few MHz of deviation after a few micro-seconds.

Sensitivity and Maximum Sensitivity Ratio (MSR)

The sensitivity parameter is defined as the “voltage to frequency” transfer function of the VCO and is measured in MHz/Volt. A perfect VCO will have a constant sensitivity throughout its range of operation. Unfortunately there are no ideal VCOs and therefore the sensitivity varies across the VCO frequency range. The Maximum Sensitivity divided by the Minimum Sensitivity is defined as MSR. Using a VCO with poor MSR (>>1) will yield a wide range of problems. Some examples:

Applying a perfect saw-tooth waveform as the tuning voltage will not generate a perfect chirp. Range measurements in Altimeters, for example, will become inaccurate as a result.

The same modulating waveform, for different center frequencies, will result in different frequency spans. This is demonstrated in Figure 4.

Different modulating waveform amplitudes, for a constant offset voltage, will result in different center frequencies. This is demonstrated in Figure 5.

Figure 5: MSR effect on center frequency

Frequency Total Accuracy

This parameter is defined as the maximal frequency error that will be measured after a “voltage to frequency” calibration table has been established. The frequency error is mainly effected by temperature and aging. This is the major drawback of the VCO as a frequency source. A system using a simple VCO as a WBFM generator may end up with a signal that has some deviation in center frequency and also in its span. For example, an EW system may jam the wrong frequency band reducing its effectiveness and its coexistence capabilities.

Frequency Modulation Span

This parameter is defined as the maximal frequency span that VCO may cover when driven by modulating signal. With VCOs there is no actual limit to the span and a VCO can support WBFM starting at its lowest frequency and ending at its highest frequency. For example, a 4 to 8 GHz VCO will support modulation with a span up to 4 GHz. As will be shown later, this is not the case for other devices.

Modulation Frequency Bandwidth

This parameter is defined as the maximal modulation frequency or modulation rate that may be applied to the modulation control pin before the spans droops by more than 3 dB. For example, a VCO is being modulated by a very slow changing control voltage to generate 1 GHz span. The control voltage is then changed to be a fast sine-wave. The frequency of this control voltage is increased until the span starts to become less than 1 GHz. The frequency that causes the span to be 707 MHz is the 3 dB Modulation Bandwidth. A typical value for a VCO would be 250 MHz.

Digitally Tuned Oscillator

Since the VCO requires the user to prepare a look-up-table in order to know what voltage to apply to get the desired output frequency, a more convenient approach would be to have this look-up-table stored within the module. This will allow the user to input a digital command and the pre-calibrated information will be used to generate the correct frequency. Since the transfer function of a VCO is greatly dependent on the temperature, a heater connected to the VCO is used to produce a constant VCO temperature. For supporting frequency ranges of more than an octave, several VCOs may be housed within the same DTO. The basic block diagram of a multi-octave DTO is shown in Figure 6.

Figure 6: Multi-octave DTO conceptual block diagram

Main advantages of the DTO are its multi-octave frequency range and its relatively low price. The main DTO disadvantage is the need for an elaborate calibration process.

When modulating the DTO by the external modulation signal, only one of its internal VCOs is being modulated and therefore the modulation span is limited. The same problems of changing modulation spans and shifting center frequencies with different modulation voltages exist when using a DTO as well.

Frequency Locked Oscillator

To improve the frequency accuracy of a DTO a correction circuit is used. The output signal is sampled and its frequency is measured by an accurate frequency discriminator. The output of the discriminator is used as a feedback to the tuning voltage of the VCO. The VCO is said to be frequency locked and its accuracy is as good as the discriminator’s ability to measure frequency. When commanded to jump to a new frequency, the FLO’s control circuit applies a tuning voltage to the VCO according to its internal look-up-table. This is called a “DTO Mode” since this is exactly what is being done in a DTO. Once the VCO converges to the vicinity of the final frequency, the discriminator reading is connected with a closed loop to the tuning voltage in order to achieve enhanced accuracy. This is called “FLO Mode”. As with the DTO, the FLO output signal may be modulated. For NBFM the module can still be in “FLO Mode” during the modulation and the center frequency accuracy is guaranteed. However, for WBFM the frequency locked loop must be opened (due to the limited BW of the discriminator) and the module works in a “DTO Mode” with reduced accuracy. Usually, for the same frequency range of operation, the FLO is larger and more expensive than a DTO. The basic block diagram of a FLO is shown in Figure 7.

Figure 7: Multi-octave FLO conceptual block diagram

Fast Indirect Synthesizer

A cost effective solution for generating wide-band signals is the indirect synthesizer. With the indirect synthesizer the VCO is phase-locked to a reference oscillator. That is why the indirect synthesizer is also known as a PLL based signal generator. The frequency accuracy of the output signal is the same as the reference signal used to lock the synthesizer, and is several orders of magnitude better than all the previously described solutions. The basic block diagram of an Indirect Synthesizer is shown in Figure 8.

The indirect synthesizer has been widely used in the market for many years and is successfully supporting many non-modulated frequency applications. To add modulation ability to the indirect synthesizer, there are several approaches.

Figure 8: Indirect synthesizer conceptual clock diagram

NBFM techniques

Two major techniques are commonly being used. The first is to inject the modulating voltage directly to the tuning voltage of the VCO. This solution is effective as long as the modulating signal is a relatively higher frequency than the loop BW (also known as AC coupling). Otherwise the loop will be able to detect and remove this modulation. The second technique is to modulate the reference signal to the PLL. This technique is effective as long as the modulating signal is within the loop BW so that the loop will cause the VCO to follow the changing frequency of the modulated reference. Other methods are also being used, such as hybrid methods (two and three point modulation) but they are behind the scope of this article.

WBFM Techniques

Two major techniques are commonly used. The first is to use the PLL in order to jump to the new center frequency, then to keep the tuning voltage to the VCO at a constant value (e.g., by a S/H) and inject the modulating voltage directly to the tuning voltage. This technique is called “DTO Mode” since the loop is open during the modulation and the VCO is actually in free running mode. This technique suffers from all the drawbacks explained previously for the “DTO Mode” in DTOs and in FLOs.

The second technique is to use a “Pure Locked Mode” (PLM). Using PLM the reference signal to the PLL is modulated and the synthesizer is always locked, similar to the NBFM case. This technique is very challenging due to the fact that the loop elements of the PLL need to support extremely high rates of voltage changes (both voltage and frequency). But the advantages of the PLM are quite clear, perfect center frequency and well known modulation spans, without the need for factory or customer calibration. The PLM supports modulation waveforms from DC to high rates (DC coupled).

Product Example

As described above, there are many benefits of using the Indirect Synthesizer technology to generate WBFM especially when using it in PLM. The Model SM6220, offered by General Microwave Israel (GMI), is a 2 to 20 GHz synthesizer that has a very fast settling time; less than 1 micro-second. This settling time is guaranteed for any jump between any two frequencies, including end to end. The SM6220 is also capable of WBFM in PLM with up to a 1 GHz span. The 1 GHz span can be located anywhere within the 2 to 20 GHz range (no “sub-bands”), thus enabling continuous coverage. The 3 dB modulation bandwidth is 10 MHz. A spectrum plot of a 1 GHz WBFM span is shown in Figure 9. This state-of-the-art product is compared to other solutions in Table 1.

Table 1: Various models that supports WBFM – comparison table

Acknowledgments

All the photos and measurements in this article are courtesy of General Microwave Israel (GMI), a KRATOS Company.

Ronen Holtzman is the VP of Engineering in GMI Ltd. He has more than 25 years of experience and is a specialist in RF and microwave components and subsystems.

(22441)

Wideband Frequency Modulation Applications and Techniques

Share this video