Significantly Reduces Distortion in RF Chain (ADCs, DACs, Sampling Circuitry, Amplifiers or Complete RF Chain)

LinComp significantly improves the performance of analog-to-digital converters (ADCs), digital-to-analog converters (DACs), sample-and-hold circuitry, buffer amplifiers, and RF power amplifiers (e.g., Solid-State Power Amplifiers (SSPAs) and Traveling Wave Tube High-Power Amplifiers (TWT HPAs) ), or the combination of these devices in an RF signal chain (as shown in Figure 1-1 for a complete RF receiver chain and in Figure 1-2 for a complete RF transmit chain). Linearity errors cause harmonic distortion and intermodulation distortion which can limit the performance of state-of-the-art electronic systems, such as radar systems, digital transceivers for wireless communications, laboratory test equipment, medical imaging, and audio and video compression. Of particular interest is the ability to pre-compensate the transmit signal chain (especially the output power amplifier) to significantly reduce the need to lower the level of the RF power amplification to meet distortion specifications. This reduces the power rating and therefore the size, cost, and power consumption of the output power amplifier in transmitter systems. Reducing errors in digital-to-analog converters, analog-to-digital converters, sample-and-hold circuitry, and buffer and power amplifiers can significantly improve the performance of the critical conversion process.

Figure 1-1. LinComp digital signal processing can be used to improve the dynamic range of a combination of devices in an RF receive
chain by 24 dB (4 bits).

Figure 1-2. LinComp digital signal processing can be used to pre-compensate a combination of devices in an RF transmit chain
by 24 dB (4 bits).

LinComp Performance will Always Exceed State-of-the-Art

The LinComp architecture will always exceed state-of-the-art because it can easily be upgraded as new, more powerful ADCs, DACs, and amplifier products become available. By eliminating much of the analog electronics, the high-performance LinComp can significantly reduce the size, power, and cost of RF transceivers, radar systems, and cellular communications systems by performing more of the processing digitally in reconfigurable software. The size, power, and cost improvements afforded in the receiver by the significant improvement in performance of LinComp more than outweigh its additional signal processing requirements.

Dramatic Improvement in Dynamic Range

Figures 1-3 and 1-4 illustrate the measured Fourier output spectrum of a 12-bit, 340 MHz sample rate ADC direct IF sampling at 250 MHz with and without the LinComp processing, respectively. In this case, the LinComp technology reduces the harmonic distortion by approximately 18 dB, providing 79 dB spurious free dynamic range.

Figure 1-3. FFT Spectrum of 12-bit ADC without LinComp (sample rate fs = 340 MHz) with
test frequency ft = 249.08 MHz. Aliasing/imaging spur at 260.92 MHz is virtually eliminated and
does not limit the resolution.

Figure 1-4. FFT Spectrum of 12-bit ADC with LinComp (sample rate fs = 340 MHz)
with test frequency ft = 249.08 MHz. LinComp improves the SFDR to 79 dB (improved from 61 dB).

Beyond State-of-the-Art Performance Enables Accurate Conversion or Synthesis at High IF or RF

The LinComp technology is effective at reducing harmonic and intermodulation distortion by up to 24 dB at very high frequencies. This enables direct sampling or direct digital synthesis at high IF or RF frequency. As shown in Figure 1-5, the LinComp technology improves the dynamic range by up to 4 bits and increases the RF or IF frequency by a factor of 2 to 3. For example, a state-of-the-art 12-bit ADC may provide direct sampling of signals at 130 MHz IF with 12-bit dynamic range. LinComp can improve the performance of this ADC to provide direct sampling up to 370 MHz IF with full 12-bit dynamic range. LinComp can also improve the performance of this ADC to provide 16-bit dynamic range (up to 130 MHz IF).

Figure 1-5. The dynamic range of typical state-of-the-art ADCs degrades as input frequency is increased (i.e., direct
sampling at IF or RF) beyond the third or fourth Nyquist zone (i.e., input frequencies greater than 3 f_s / 2 to 2 f_s ; one
Nyquist zone has bandwidth f_s /2 ). By correcting linearity errors that occur when directly sampling IF or RF, the V-Corp
LinComp processing increases the IF or RF frequency by a factor of 2 - 3 (e.g., for a typical ADC, the IF/RF frequency is
increased from 3 f_s / 2 to 7 f_s / 2 while still maintaining the full dynamic range).

LinComp is well-suited for the V-Corp wideband, high-resolution Advanced Filter Bank (AFB) Analog-to-Digital or Digital-to-Analog converter architecture, but it can also be used to improve the dynamic range of other ADCs, DACs, sample-and-hold circuitry, and amplifiers.

Reduces (Eventually Eliminates) RF Electronics in Digital Transceiver Systems

LinComp enables very accurate conversion or synthesis of wideband data at high frequencies without necessitating the use of gigasample-per-second sample rates or bulky, inaccurate RF mixing electronics. Moving the IF frequency higher significantly reduces the image rejection filter requirements and complexity of the RF-to-IF downconversion circuitry, improving the performance, easing the task of single-chip integration, and reducing cost. Eventually, converter performance will improve sufficiently to allow conversion of signals directly at the antenna element, completely eliminating the RF downconversion electronics. LinComp greatly accelerates this progress. The size, power, and cost improvements afforded in the receiver by the significant improvement in performance afforded by LinComp more than outweigh its modest additional signal processing requirements. The LinComp technology therefore reduces the size, power, and cost of transceivers by eliminating much of the RF electronics and reducing the digital signal processing requirements (by reducing the data rate from GHz speeds to MHz speeds). Also, by performing more of the processing digitally in reconfigurable software LinComp improves system versatility and flexibility.

LinComp Overcomes the Critical Conversion Bottleneck

Performance of a conversion system is typically quantified by two parameters, speed (in samples per second) and resolution (in bits). Designers face the challenge of trading off the resolution of the conversion with its speed. Manufacturers typically limit harmonic distortion spurs introduced by the converter to values on the order of the quantization error. This distortion typically limits the dynamic range of the of the converter to approximately 6n where n is the number of bits in the converter (i.e., spurious harmonics are approximately 6n dB below the input signal). However, transceiver applications require dynamic range far in excess of this value. LinComp can be used to effectively increase the analog bandwidth of the converter, enabling direct sampling or synthesis of high IF signals.

LinComp is Ideally-Suited for High-Performance Transmitters, Receivers,
RF Communications, and Radar

Achieving high-performance data conversion is currently the limiting factor in the cost, size, and power consumption of many electronic systems. For example, improving DAC and ADC performance in radio frequency (RF) transceivers and radar systems can:

• · Eliminate much of the analog electronics by performing more of the processing digitally
• · Enable simultaneous support of multiple analog and digital signaling and protocols
• · Seamlessly support next generation standards with software updates
• · Reduce power, mass, volume and cost

Figure 1-6. LinComp digital signal processing can be used to improve the dynamic range of state-of-the-art electronics for both
transmission and reception by at least 24 dB.

Traditional Gain/Phase Calibration Does Not Improve Dynamic Range

Many transceiver systems and test equipment use digital filtering to compensate for gain and phase drift across frequency. A pseudo-random noise signal is periodically injected into the system and the output is re-calibrated for constant gain and phase performance. Since this method uses a linear filtering operation, it does not correct nonlinear distortion and therefore does not improve the dynamic range. LinComp performs full gain and phase calibration as well as correction of nonlinear distortion.

Digital Signal Processing Enhances Performance

The LinComp technology is derived from techniques developed for adaptive antenna array processing used to null interfering signals and enhance the desired signal. Linearity errors are treated as interference and their distortion profile can be used to effectively null the errors in the digital output. The approach requires less computational processing and hardware than competing compensation approaches while providing greater reduction in harmonic and intermodulation distortion.

Algorithm Description

The linearity profile of the device to be compensated is measured by injecting known signals into the device and measuring the resulting harmonic and intermodulation distortion. The device is measured across a wide range of amplitudes and frequencies (including super-Nyquist frequencies for direct IF sampling or synthesis applications) to yield a broadband model. An efficient auto-calibration procedure is implemented by using pseudo-random comb signals to quickly generate a broadband model.

The algorithm uses the linearity model to predict the distortion which is then nulled in the output. To simplify the computational requirements, only the dominant errors (e.g., second and third order components) are chosen to be nulled in the output. The algorithm is dynamic since it nulls errors for inputs across a wide range of frequencies (unlike static look-up table techniques); the algorithm uses digital filters with a response that varies across frequency instead of using slope estimation (as used in phase-plane compensation to account for frequency-dependent errors.

Generalized Polynomial Linearity Error Model

The LinComp processor generates out-of-phase distortion products which are used to cancel the errors. Note that LinComp is a general linearity error compensator that accurately nulls harmonic distortion and intermodulation distortion for a wide range of electronic devices, and it corrects errors for arbitrary wideband signals as well as sinusoidal signals.

Figure 1-7. Generalized polynomial linearity error model used in the LinComp processor to null distortion.

Calculation of LinComp Coefficients is Straightforward

The LinComp coefficients can be determined, for example, by injecting sinusoidal signals into the device to be modeled and measuring the amplitude and phase shift of the harmonic distortion spurs in the device output. The amplitude and phase shift of the harmonic distortion spurs are directly related to the LinComp coefficients, so their calculation is straightforward.

Specific harmonics can be chosen to be compensated while others are ignored by assuming the amplitudes of harmonics to be ignored are zero.

Efficient Implementation of LinComp Processing

The LinComp digital signal processing can be implemented efficiently with finite impulse response (FIR) filters, multipliers, and adders. In practice, distortion products beyond the fourth or fifth order are usually negligible and can be ignored, so only the first few distortion products need to be nulled. Digital signal processing with relatively low order FIR filters (less than length 32) can be used to implement algorithm. The LinComp processor can use parameters that vary with input frequency if the device linearity error profile varies with frequency. The frequency band may be partitioned into smaller bands using a digital filter bank and each band may processed with separate compensation algorithms if the linearity error profile changes significantly as a function of frequency.

Efficient and Accurate Automatic Calibration

The LinComp processor is calibrated by injecting known test signals into the device to characterize its linearity error distortion profile. The device is measured across a wide range of amplitude and frequency (including super-Nyquist frequencies for direct IF synthesis applications) to yield a broadband model. An efficient auto-calibration procedure can be implemented by using pseudo-random comb signals to quickly generate a broadband model.

Since the linearity error profile of many electronic devices is relatively constant, re-calibration of the LinComp compensator can be performed infrequently (e.g., when the system hardware is reconfigured). Calibration is computationally efficient, only requiring a few clock cycles.

LinComp Outperforms Traditional Compensation

LinComp Technology is Not Sensitive to Slope Estimates
and is Suitable for Super-Nyquist Compensation

The V-Corp proprietary LinComp approach requires less hardware than phase-plane compensation and provides up to 24 dB or more reduction in harmonic distortion, does not require slope estimates, and is capable of super-Nyquist error compensation (i.e., direct synthesis of high IF data).

Traditional Static Compensation Provides Limited Performance Improvement

One technique for converter compensation is simply a static look-up table to correct the digital values. This method sometimes provides up to 10 dB reduction in harmonic distortion. Researchers have realized that this type of correction improves the dynamic range of the converter only near the calibration frequency.

Traditional Dynamic Compensation is Sensitive to Slope Estimates
and is Not Suitable for Super-Nyquist Frequencies

Phase-plane compensation is a dynamic approach in that it accounts for errors that are a function of both amplitude and frequency. Like static compensation, a lookup table is used to correct the digital values, but in this case, the lookup table is indexed by the current digital value of the converter and the estimated slope of the output (to account for frequency). This method accomplishes all that static compensation does but yields improved performance for its ability to errors that are a function of frequency. This method can sometimes provide 10-15 dB reduction in harmonic distortion.2 This method is more hardware-intensive than static compensation since it needs to estimate the slope of the signal; inaccurate slope estimates significantly degrade the performance. In addition, this method is not suitable for super-Nyquist input frequencies (signals above the Nyquist frequency) due to the ambiguity in the slope. Super-Nyquist compensation is necessary in transmitter applications that use intermediate frequency (IF) synthesis.

Wide Range of Military and Commercial Applications of the LinComp Solution

Applications of high-resolution, high-speed data conversion include:

• · Smart radios for wireless communications (cellular and satellite)
• · Radar systems
• · Medical imaging systems such as Ultrasound, MRI, CT Scan, PET Scan, and X-Ray
• · Sonar systems
• · General test equipment such as oscilloscopes, spectrum analyzers, network analyzers
• · Special test equipment
• · Wide bandwidth modems

Enables Software Radio

One particularly attractive application of LinComp is the Software Radio, which can accommodate two or more RF modulation standards simultaneously by performing tuning and demodulation on the digital data in software. The Software Radio can seamlessly integrate new standards as they arise. The Software Radio promises lower power, smaller size, and lower cost by processing 50 or more channels in software instead of dedicated hardware. For military applications, the Software Radio is capable of understanding many different signalling protocols with a compact, low-power transceiver. For cellular telephone applications, the Software Radio allows for universal coverage without necessitating worldwide agreement on a single standard (which, due to politics and competition, has proven to be impossible) since it can understand signals from many different types of cellular telephones.

Enhances Adaptive Antenna Array Processing and Digital Beamforming

Space Division Multiple Access (SDMA) is used to improve communications capacity, reduce jamming and interference, and improve security by using spatial processing for intelligent beam steering (for both transmission and reception). Current systems are limited by converter bandwidth and precision; adaptive algorithms require wide dynamic range to accurately null interfering signals. LinComp overcomes the critical bandwidth and precision limitations.

Home | Products | Services | About V-Corp | News

V Corp®, AFB®, LinComp®, and the double-bar logo are registered trademarks of V Corp Technologies, Inc.
© 2005 V Corp Technologies, Inc.