AD9545 datasheet(127/157 Pages) AD

127 / 157 page

Data Sheet

AD9545

Rev. A | Page 127 of 157

COMPENSATION METHOD 1

Compensation method 1 employs the open-loop method of

system clock compensation (see the Open-Loop Method

section). As shown in Figure 99, Compensation Method 1 takes

the polynomial coefficients (Cx) and a temperature value (T) to

compute a compensation factor (CF1). Compensation Method 1

differs slightly from the open-loop method by the inclusion of a

digital filter (see the Digital Filter section).

POLYNOMIAL

CALCULATOR

C0 C1 C2 C3 C4 C5

DIGITAL

FILTER

COEFFICIENT

OUTPUT FILTER

CUT OFF

CF1

Figure 99. Compensation Method Number 1

Program the coefficients via Register 0x028A to Register 0x029C.

The temperature value comes from the internal temperature

sensor or a temperature register (see the Temperature Sensor

section for details). The CF1 compensation factor applies to the

designated NCOs and TDCs to correct for frequency variations

of the system clock source. CF1 embodies the factor, 1 +

FFECOMP, shown in Figure 97.

How to Derive Coefficients from Frequency Data

To compensate for the frequency variation of the system clock

source, knowledge of the frequency vs. temperature profile is a

necessity. Generally, the temperature profile of the source is a

table of temperature values with corresponding values of the

source frequency. Table 53 shows a hypothetical example of a

25 MHz source in the two leftmost columns.

Table 53. Example 25 MHz Source Data

Temperature

(°C)

Frequency

(MHz)

FFE (×10−6)

FPE

(×10−6)

−20

24.9757265855

−970.93658

971.88021

−10

24.9745666538

−1017.33385

1018.36987

24.9751062576

−995.74970

996.74220

24.9758628851

−965.48459

966.41766

24.9780535472

−877.85811

878.62943

24.9789426612

−842.29355

843.00361

24.9809631193

−761.47523

762.05552

24.9810049826

−759.80070

760.37843

24.9800392641

−798.42944

799.06744

24.9777091418

−891.63433

892.43005

24.9742106356

−1031.57458

1032.63982

Because the AD9545 implements compensation in terms of

fractional error, it is necessary to convert the frequencies in

Table 53 to FFE values (refer to the System Clock Compensation

Overview section). Table 53 shows the resulting FFE values.

However, Compensation Method 1 generates CF1 as a time error

correction rather than a frequency error correction; therefore,

the user must convert the FFE data to fractional period error

(FPE) data (refer to the System Clock Compensation Overview

section for the conversion formula). Table 53 shows the

resulting FPE values. From the calculated FPE values, find the

coefficients (Cx) that best fit the FPE data, which generally

involves mathematical regression techniques using specialized

mathematical software tools.

Applying such regression techniques to the hypothetical FPE

values in Table 53 yields the following coefficients:

= 1.0046936 × 10−3

= −2.8927765 × 10−6

C2 = −1.9110192 × 10−7

= 2.2493907 × 10−9

= 2.7943570 × 10−11

= −2.4817310 × 10−13

Figure 100 shows the FPE vs. temperature curve resulting from

these coefficients. The trace is the predicted FPE vs.

temperature curve based on the coefficients, and the solid

points are the temperature vs. FPE data from Table 53.

1400

1300

1200

1100

1000

900

800

700

600

–50 –40 –30 –20 –10

10 20 30 40 50 60 70 80 90 100

TEMPERATURE (°C)

Figure 100. FPE vs. Temperature

Given a set of coefficients like C0 to C5, the user must convert

the numeric values to suitable register contents (see the

Programming the Coefficients (Ck) for Compensation Method 1

section for details).

Programming the Coefficients (Ck) for Compensation

Method 1

Compensation Method 1 makes use of six coefficients, Cx,

where the Index x ranges from 0 to 5. The user programs the

constant coefficient, C0, which is independent of the temperature

parameter, via the signed 40-bit constant compensation value

bit field in Register 0x0289 to Register 0x028D. The relationship

between C0 and the value stored in the constant compensation

value (CV) bit field is:

Constant CV = C0/2−45

AD9545 Datasheet(PDF) 127 Page - Analog Devices

AD9545 Datasheet(HTML) 127 Page - Analog Devices

Html Pages

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 ...More

Datasheet Download

Link URL