LM98725CCMTSLASHNOPB datasheet(39/145 Pages) TI1

39 / 145 page

LM98725

www.ti.com

SNAS474H – APRIL 2009 – REVISED MARCH 2015

7.4.12.7 General White Loop Operation

The White Peak value is determined by a moving window average done throughout the Active/White pixels

region of the line (Refer to LM98725 Typical Line Timing and Pixel Gain Regions). The size of the moving

window is set at Page 2, Register 4, Bits[2:0]. 4, 8, 16 or 32 (default) pixels are averaged together at a time and

compared with the previous Peak White value. If the new White Average is greater than the previous Peak White

value, then the Peak White value is set equal to the White Average. The window is then shifted over to the next

group of pixels (one window width) and a new average is calculated. This process is repeated until the

Active/White Pixels End value is reached. The final Peak White value is then compared with the White Target

value (Page 2, Register 5). The gain settings are then adjusted to make the Peak White value closer to the White

Target value.

7.4.12.8 White Loop Modes

•

White Full Calibration (Cold start or Pre-Job) - In this mode, the gain settings are completely reset to

default at the beginning of the loop. The first change or gain decision is whether the CDS/SH gain setting will

be 1x or 2x. Subsequent decisions are made to adjust the PGA setting to the provide the Peak White value

that is closest to the White Target value.

•

White Fine Tuning (Page Gap Calibration) - In this mode, the previous gain settings are used as the

starting point. The CDS/ SH gain setting will not be changed. The PGA gain will be increased or decreased

by one step at a time, to fine tune the gain. This mode is normally used where only minor changes in the

white level are expected.

7.4.12.9 Bimodal (Automatic) Correction

Due to the architecture of the ADC used in the LM98725, it is possible that a small offset in the resulting data will

be present between the Even and Odd pixels processed by the device. This offset is referred to as a Bimodal

distribution in the data. It can occur because the signals being procesed go through two different paths to

achieve the required throughput. Each of these paths can give a slightly different signal offset, resulting in the

bimodal distribution in output data, for a given fixed input voltage.

When the automatic Black Level (offset) correction circuit is engaged, a portion of that circuitry automatically

corrects the bimodal distribution. The bimodal correction applied is stored at Page 1, Registers 0x1Ch, 0x1Dh,

and 0x1Eh. The stored value is divided by two. Half is added to the Even pixels, and half is subtracted from the

Odd pixels.

Submit Documentation Feedback

Product Folder Links: LM98725

LM98725CCMT/NOPB Datasheet(PDF) 39 Page - Texas Instruments