LM9812 datasheet(28/37 Pages) NSC

28 / 37 page

http://www.national.com

This PGA is static - it does not change at the pixel rate. The Pixel-

Rate Shading Multiplier is used to eliminate pixel-to-pixel gain

errors (typically sensor PRNU and shading errors).

1.2.2 Static Offset DACs

The Static Offset DACs remove DC offsets generated by the sen-

sor and the LM9812’s analog signal chain. The DACs should be

set during calibration to the lowest value that still results in an

ADC output code > 0 for all the pixels. Each LSB of the offset

DAC is typically 4.2 ADC LSBs, providing a total offset adjust-

ment range of ±130 ADC LSBs. The equation for the offset DAC

is:

Like the PGA, the Offset DAC is static - it does not change at the

pixel rate. The Pixel Rate Offset Subtractor is used to eliminate

pixel-to-pixel offset errors.

1.2.3 ADC

The 10 bit ADC quantizes the output from the analog chain and

passes the result to the digital section for pixel rate error correc-

tion. The ADC’s input range is equal to VREF IN/C, where C is the

gain constant compensating for all sources of gain error through-

out the entire analog signal path.

1.3 DIGITAL SIGNAL PATH

1.3.1 Pixel-Rate Offset Subtractor

The output of the ADC is fed to the Pixel-Rate Offset Subtractor.

Each pixel of image data may have a different offset error. If bit 0

of Register 9 is set to a 0 and a byte representing the offset error

code is fed to the Pixel-Rate Offset Subtractor through the CD0-

CD9 databus, the offset for each pixel can be subtracted from the

ADC output code at the pixel conversion rate. If this feature is not

desired, the Offset Subtractor can be set to a fixed value by writ-

ing the desired value to Register 6 and setting bit 0 of Register 9

to a 1.

The offset subtractor input is 8 bits wide, input on CD0-CD7, for a

correction range of 0 to 255. Information on bits CD8 and CD9 is

ignored.

1.3.2 Pixel-Rate Shading Multiplier

The Pixel Rate Shading Multiplier follows the Offset Subtractor.

This stage compensates for nonuniformities between individual

pixels (shading error). Each pixel of image data may have a differ-

ent gain error. If bit 1 of Register 9 is set to a 0 and a word repre-

senting the gain correction coefficient is fed to the Pixel-Rate

Shading Multiplier through the CD0-CD9 databus, the gain for

each pixel is changed at the pixel rate to eliminate pixel-to-pixel

gain errors. If this feature is not desired, the Shading Multiplier

can be set to a fixed value by writing the desired gain value to

Registers 7 and 8 and setting bit 1 of Register 9 to a 1.

The equation for calculating the gain of the multiplier is given in

Equation 5. The Pixel Rate Shading Multiplier has three different

gain ranges as shown in Table 1. For systems where the ratio of

the strongest pixel to the weakest pixel is less than 1.5, the multi-

plier gain range setting of 1:1.5 provides very accurate control of

the gain of each pixel. Systems with larger variation can use gain

ranges of 1:2 or 1:3.

When the 8 bit Coefficient Bus Width is selected (Register 25, bit

6=1), the coefficient data is only 8 bits wide, supplied on CD0-

CD7. This reduces the accuracy of the gain correction, but allows

the use of an 8 bit path to store coefficient data, instead of a 10 or

16 bit wide path.

1.4 SENSOR CLOCK GENERATION

The LM9812 generates all of the clock signals required to directly

drive most commercial linear CCDs and some CIS - no external

clock buffers are necessary. Most linear CCDs designed for scan-

ner applications require 0 to 5V signal swings into 20 to 500pF

input loading. Series resistors are typically inserted between the

driver and the CCD to control slew rate and isolate the driver from

the large load capacitances. The values of these resistors are

given in the CCD’s datasheet.

1.4.1 TR1 and TR2

The LM9812 supports one or two TR (transfer, or shift) pulses as

shown in Diagrams 8 and 9. This pulse is used to transfer the

contents of each pixel’s photodiode to the CCD’s serial shift regis-

ter for clocking out of the CCD. Configuration Registers 16, 17,

and 18 control the TR1, TR2, and TR guardband pulse widths,

while Register 24, bit 3 determines whether the LM9812 gener-

ates TR1 only or both TR1 and TR2 at the beginning of a line.

The polarities of the TR pulses are determined by Register 26,

bits 2 and 3. If not needed, one or both TR pulses can be dis-

abled by setting Register 25, bits 3 and 4 to the appropriate value.

Offset ADC LSBs

()

4.2 DAC Code

()

Equation 3: Offset DAC

Offset Subtractor Out

ADC Output

Offset Error Code

–

Equation 4: Offset Subtractor

Multiplier Gain Range

1:1.5

(33%)

1:2.0

(50%)

1:3.0

(67%)

Minimum Gain

(V/V, Multiplier Input = 0)

Maximum Gain

(V/V, Multiplier Input = 1024)

1.5

2.0

3.0

N (used in Equation 5)

2048

1024

512

Bits D3, D2 of Register 9

1, 0

0, 1

0,0

Table 1:Shading Multiplier: 10 bit Coefficient Bus Width

Multiplier Gain Range

1:1.5

(33%)

1:2.0

(50%)

1:3.0

(67%)

Minimum Gain

(V/V, Multiplier Input = 0)

Maximum Gain

(V/V, Multiplier Input = 256)

1.5

2.0

3.0

N (used in Equation 5)

512

256

128

Bits D3, D2 of Register 9

1, 0

0, 1

0,0

Table 2:Shading Multiplier: 8 bit Coefficient Bus Width

Multiplier Gain(V/V)

Gain Correction Coefficient

------------------------------------------------------------------------

Equation 5: Shading Multiplier Gain

LM9812 датащи(PDF) 28 Page - National Semiconductor (TI)

LM9812 датащи(HTML) 28 Page - National Semiconductor (TI)

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

датащи скачать

ссылки URL