EE 435. Lecture 26. Data Converters. Data Converter Characterization

Transcription

1 EE 435 Lecture 26 Data Converters Data Converter Characterization

2 . Review from last lecture. Data Converter Architectures n DAC R-2R (4-bits) R R R R V OUT 2R 2R 2R 2R R d 3 d 2 d 1 d 0 V REF By superposition: d V =V d +V d +V d +V d = V V 3 4 k 4-k OUT REF 3 REF 2 REF 1 REF 0 REF 4-k REF k k=0 2 k=1 2 d

3 . Review from last lecture. Performance Characterization of Data Converters Static characteristics Resolution Least Significant Bit (LSB) Offset and Gain Errors Absolute Accuracy Relative Accuracy Integral Nonlinearity (INL) Differential Nonlinearity (DNL) Monotonicity (DAC) Missing Codes (ADC) Low-f Spurious Free Dynamic Range (SFDR) Low-f Total Harmonic Distortion (THD) Effective Number of Bits (ENOB) Power Dissipation

4 . Review from last lecture. Performance Characterization of Data Converters Dynamic characteristics Conversion Time or Conversion Rate (ADC) Settling time or Clock Rate (DAC) Sampling Time Uncertainty (aperture uncertainty or aperture jitter) Dynamic Range Spurious Free Dynamic Range (SFDR) Total Harmonic Distortion (THD) Signal to Noise Ratio (SNR) Signal to Noise and Distortion Ratio (SNDR) Sparkle Characteristics Effective Number of Bits (ENOB)

5 . Review from last lecture. Performance Characterization of Data Converters What is meant by low frequency? Operation at frequencies so low that further decreases in frequency cause no further changes in a parameter of interest Low frequency operation is often termed Pseudostatic operation

6 . Review from last lecture. Low-frequency or Pseudo-Static Performance Paramater Pseudo-Static Region f

7 Performance Characterization of Data Converters Static characteristics Resolution Least Significant Bit (LSB) Offset and Gain Errors Absolute Accuracy Relative Accuracy Integral Nonlinearity (INL) Differential Nonlinearity (DNL) Monotonicity (DAC) Missing Codes (ADC) Low-f Spurious Free Dynamic Range (SFDR) Low-f Total Harmonic Distortion (THD) Effective Number of Bits (ENOB) Power Dissipation

8 Performance Characterization Resolution Number of distinct analog levels in an ADC Number of digital output codes in A/D In most cases this is a power of 2 If a converter can resolve 2 n levels, then we term it an n-bit converter 2 n analog outputs for an n-bit DAC 2 n -1 transition points for an n-bit ADC Resolution is often determined by architecture and thus not measured Effective resolution can be defined and measured If N levels can be resolved for an DAC then logn n EQ= log2 If N-1 transition points in an ADC, then logn n EQ= log2

9 Performance Characterization Least Significant Bit n Assume N = 2 Generally Defined by Manufacturer to be X LSB = /N Effective Value of LSB can be Measured For DAC: X LSB is equal to the maximum increment in the output for a single bit change in the Boolean input For ADC: X LSB is equal to the maximum distance between two adjacent transition points

10 Offset Performance Characterization For DAC the offset is (assuming 0 is ideal value of (<0, 0>) (<0,, 0>) - absolute X OUT 0,...,0 X LSB - in LSB Offset C 0 C 1 C 2 C 3 C 4 C 5 C 6 C 7 (If ideal value of (<0, 0>) 0, offset is shift from ideal value at <0, 0>)

11 Offset Performance Characterization (for DAC) Offset C 0 C 1 C 2 C 3 C 4 C 5 C 6 C 7 Offset strongly (totally) dependent upon performance at a single point Probably more useful to define relative to a fit of the data

12 Offset Performance Characterization (for DAC) Offset C 0 C 1 C 2 C 3 C 4 C 5 C 6 C 7 Offset relative to fit of data

13 Offset Performance Characterization For ADC the offset is X T1 -X LSB X X X T1 LSB LSB C 7 (assuming X LSB is the ideal first transition point) - absolute - in LSB C 6 C 5 C 4 C 3 C 2 C 1 X LSB C 0 X T1 X OFFSET (If ideal first transition point is not X LSB, offset is shift from ideal)

14 Performance Characterization Offset For ADC the offset is C 7 C 6 C 5 C 4 C 3 C 2 C 1 X LSB C 0 X T1 X OFFSET Offset strongly (totally) dependent upon performance at a single point Probably more useful to define relative to a fit of the data

15 Performance Characterization Offset For ADC the offset is C 7 C 6 C 5 C 4 C 3 Fit Line C 2 C 1 C 0 X LSB X OFFSET Offset relative to fit of data

16 Performance Characterization Gain and Gain Error For DAC Ideal Output Actual Output Gain Error C 0 C 1 C 2 C 3 C 4 C 5 C 6 C 7 Gain error determined after offset is subtracted from output

17 Performance Characterization Gain and Gain Error For ADC C 7 C 6 Actual Output Ideal Output C 5 C 4 C 3 Gain Error C 2 C 1 C 0 X LSB Gain error determined after offset is subtracted from output

18 Performance Characterization Gain and Offset Errors Fit line would give better indicator of error in gain but less practical to obtain in test Gain and Offset errors of little concern in many applications Performance often nearly independent of gain and offset errors Can be trimmed in field if gain or offset errors exist.

19 Integral Nonlinearity (DAC) Nonideal DAC C 0 C 1 C 2 C 3 C 4 C 5 C 6 C 7

20 Integral Nonlinearity (DAC) Nonideal DAC End Point Fit LIne X OF (k) C 0 C 1 C 2 C 3 C 4 C 5 C 6 C 7 XOF X m= N-1 -X 0 OUT OUT N-1 k = mk+ XOUT 0

21 Integral Nonlinearity (DAC) Nonideal DAC (k)-x OF (k) C 0 C 1 C 2 C 3 C 4 C 5 C 6 C 7 k OUT OF INL = X k -X k INL= max 0 k N-1 INL k

22 Integral Nonlinearity (DAC) Nonideal DAC INL C 0 C 1 C 2 C 3 C 4 C 5 C 6 C 7

23 Integral Nonlinearity (DAC) Nonideal DAC INL often expressed in LSB INL = X k INL= max OUT OF 0 k N-1 X k - X LSB INL k k INL C 0 C 1 C 2 C 3 C 4 C 5 C 6 C 7 INL is often the most important parameter of a DAC INL 0 and INL N-1 are 0 (by definition) There are N-2 elements in the set of INL k that are of concern INL is almost always nominally 0 (i.e. designers try to make it 0) INL is a random variable at the design stage INL k is a random variable for 0<k<N-1 INL k and INL k+j are almost always correlated for all k,j (not incl 0, N-1) Fit Line is a random variable INL is the N-2 order statistic of a set of N-2 correlated random variables

24 Integral Nonlinearity (DAC) Nonideal DAC INL C 0 C 1 C 2 C 3 C 4 C 5 C 6 C 7 At design stage, INL characterized by standard deviation of the random variable Closed-form expressions for INL almost never exist because PDF of order statistics of correlated random variables is extremely complicated Simulation of INL very time consuming if n is very large (large sample size required to establish reasonable level of confidence) Model parameters become random variables Process parameters affect multiple model parameters causing model parameter correlation Simulation times can become very large INL can be readily measured in laboratory but often dominates test costs because of number of measurements needed when n is large Expected value of INL k at k=(n-1)/2 is largest for many architectures Major effort in DAC design is in obtaining acceptable yield!

25 ENOB of DAC Nonideal DAC INL C 0 C 1 C 2 C 3 C 4 C 5 C 6 C 7 Concept of Equivalent Number of Bits (ENOB) is to assess performance of an actual DAC to that of an ideal DAC at an equivalent resolution level Several different definitions of ENOB exist for a DAC Here will define ENOB as determined by the actual INL performance Will use subscript to define this ENOB, e.g. ENOB INL

26 ENOB INL of DAC Nonideal DAC Premise: A good DAC is often designed so that the INL is equal to ½ LSB. Thus will assume that if an n-bit DAC has an INL of ½ LSB that the ENOB INL =n. INL C 0 C 1 C 2 C 3 C 4 C 5 C 6 C 7 Thus define the effective number of bits, n EFF by the expression INL V REF neff neff where INL is in volts 1 n EFF VREF ENOBINL log2 1 INL Thus, if an n-bit DAC has an INL of ½ LSB n VREF 2 V LSB n 1 ENOBINL log2 1 log2 1 log2 2 1 INL VLSB 2 n

27 End of Lecture 26