|
|||||||||
Tutorial on Designing Delta-Sigma Modulators: Part 2
Tutorial on Designing Delta-Sigma Modulators: Part 2 The transition from second-generation (2G) to third-generation (3G) wireless cellular systems requires multi-standard adaptability in a wireless receiver. An important answer to this request is the use of delta-sigma modulators for IF-to-baseband conversion, which will satisfy the dynamic range requirements for aggressive digital signal processing, and at the same time, add adaptability and programmability to the characteristics of a RF receiver. This articles addresses the issues of how to design effective delta-sigma modulators for wireless receivers that need to support GSM, DECT, and wideband CDMA (W-CDMA) operation. A single-loop third-order modulator topology suitable for low power and high integration multi-standard receiver is described.1 The trade-offs in the modulator design are also presented and explained. In Part 1 of this article, we compared today's typical receiver architectures to the delta-sigma modulator approach. We'll also provide a look at the different types of delta-sigma modulators available and the key elements that make up a modulator designs. Now in Part 2, we'll further the discussion by looking at circuit implementation issues as well as the key sub-elements that make up a delta-sigma converter design. We'll also provide some simulation results as well as future design work that needs to be conducted in this space. Circuit Implementation The three integrators and the comparator are controlled by two-phase, non-overlapping clocks, ck1 and ck2; an early clock phase ck1e to strobe the latched comparator; two delayed clock phases, ck1d and ck2d; and complements of the delayed phases, ck1db and ck2db. The Integrators
During ck1, the input voltage is sampling onto CSampling. During ck2, charge proportional to the difference between V_i and DAC reference voltage, V_ref, is transferred from CSampling to CIntegrating. The sampling and integrating capacitances are shown in Figure 8. The sizing of these switches obeys to the analysis pattern described in chapter 4 of Reference 1, where NMOS switch's W/L ratio is at least 20/1, and the minimum PMOS switch's W/L ratio is 40/1. The delayed clock phases ck1d and ck2d reduce signal-dependent charge injection from switches S_3 and S_4 onto CSampling and CIntegrating. S_5 resets the integrating capacitors during power-on or the transformation from one standard to another. In addition, it also serves as clipper for the integrator output voltages at high swing, which is often referred to as the integrator reset approach for modulator stability. The second integrator is shown in Figure 9. The additional feedback branch, which creates the local resonator in the case of W-CDMA, is controlled by the enable signal, en_wcdma and ck1d, through a MOS AND gate.
The third integrator is shown in Figure 10. Its schematic is similar to that of the first integrator. The sampling and integrating capacitances are shown in the same figure.
Understanding the OTA Gain boosting techniques13 are employed in the design of OTA in order to acquire high gain and high bandwidth at the same time. The basic principle is that a DC gain enhancement can be acquired without affecting the gain and phase characteristics of the operational amp at high frequencies. In order to meet the settling requirements imposed by a clock rate over 100 MHz while maintaining a low power consumption, a telescopic amplifier topology (Figure 11) is chosen for the main amplifier because of its good frequency response and relatively low bias current budget. The switched capacitor common-mode feedback (CMFB) circuit is used to track the output common-mode voltage level. The value of tail current, I_tail, is 590 μA for the main amplifier in the first stage. To save power, the value of I_tail can be scaled down by a ratio of 1/2 and 1/4 for the OTAs in the second and third stage, respectively.
The auxiliary amplifiers are designed using the folded-cascade topology, which aims to provide sufficient DC gain enhancement. The CMFB loop is built by connecting the output nodes to gates of two large transistors, which are working in the linear region. The biasing plan is the following: (1) Individual biasing network is used for each main amplifier to provide sufficient isolation; (2) The auxiliary amplifiers in three integrators share the same biasing network in order to save power. The simulated UGBW and phase margin (PM) of the OTA are 480 MHz and 56 degree, respectively. The simulated dc gain is 83 dB. The idea of the foregoing gain boosting scheme is that DC gain should be determined mainly by the folded-cascade auxiliary amplifiers, while the telescopic main amplifier is optimized for high bandwidth. The direct undesirable result of this configuration is a lower DC gain. To alleviate this problem, a preamplifier can be added in front of the main amplifier14 as illustrated in Figure 12. A DC gain as high as 96 dB has been reported using this amplifier topology.15. However, this amplifier inevitably consumes more power than the previous one (extra 2 to 3 mW at a 2.5 V supply).
Single-Bit Quantizer
The comparator shown in Figure 13 is in the reset mode when the clock phase ck1e is low, and the outputs are set to V_DD by M_9 and M_10. When ck1e goes high, the comparator enters the regenerative mode, and M_3-M_8 form a positive feedback loop in order to amplify the input difference, which is sensed by M_1 and M_2, to a full-scale rail-to-rail output. Once the comparator makes a decision, the cross-coupled transistors will shut down all connections between V_DD and V_SS, which will cause null dc power consumption. Simulation Results
Figure 15 shows the simulated FFT of the modulator for the W-CDMA standard. An IF input signal at 100-kHz offset from the IF center frequency (138.24 MHz) is added. The spectrum indicates that thermal noise floor is higher than the quantization noise floor for frequencies up to 800 kHz. The thermal noise floor is roughly flat up to 2 MHz, which is half the signal bandwidth for W-CDMA standard.
Figure 16 shows the simulated FFT of the modulator for the DECT standard. An IF input signal at 100-kHz offset from the IF center frequency (110.59 MHz) is driving the modulator. The spectrum indicates that thermal noise floor is higher than the quantization noise floor for frequencies up to 300 kHz. The thermal noise floor is roughly flat up to 576 kHz, which is half the signal bandwidth for DECT standard. A noticeable harmonic at 300 kHz (third-order) limits the SNDR performance.
As Table 3 shows, the dynamic range performance of this single-loop modulator is worse than that reported in Reference 16, where a cascaded (MASH) 2-2 topology was used at a price of, however, much higher power consumption (65 mW in total for GSM/DECT applications with a 3.3 V supply). Having demonstrated the post-layout simulation results, it is instructive to mention the chip-level test plan, since the receiver is already in the process of fabrication (a standard 0.35-μm, 2.5-V, DP5M CMOS process in this case). During the test, a fully differential sinusoidal input signal (i.e., the IF signal) will be off-chip bandpass filtered before entering into the modulator. The digital output will be acquired with a test board, and the acquired data stream will then be transferred to a workstation for spectral analysis. The clock for the modulator will be generated from an off-chip RF pulse generator and brought onto chip through a bond-pad and then amplified to a 2.5 V rail-to-rail swing. Although there are band-gap references available on-chip, a clean off-chip voltage reference should be used for the modulator measurement, for the purpose of diminishing the effect of distortion due to an unclean reference. Future Work References
About the Author
|
Home | Feedback | Register | Site Map |
All material on this site Copyright © 2017 Design And Reuse S.A. All rights reserved. |