Proposed Module for Viewing-Angle-Switchable Backlight

be refracted into collim

the objective which mentioned in the introduction, the design ta

rned here. Three main design parameters used to set the final goal for VAS backlight included luminance, FWHM of angular distribution, and the uniformity of entire backlight module, and these parameters were compared with common Notebook backlight requirement. Table 3.1 shows the design target of our VAS backlight. All of the requirements shown in this table are indicated that the luminance must be higher than 3500 nits, and the FWHM of viewing angle distribution confined in +/- 20 degrees, and the uniformity was higher than 70% for both normal and oblique viewing modes.

[Table

As shown in Fig. 3.5, the divergent rays of point source can

ating light by single lens (i.e., collimator) in geometrical optics. If the point light source located on the optical axis of lens as in Fig 3.5 (a), the output light would propagate along the direction of optical axis of this lens. When the point source located off the optical axis with the included angle α, the output rays would travel parallelly with the same angle as Fig 3.5 (b). From this simple idea, the viewing direction adjustable backlight system was possible by placing a lens in front of the light source and provides the directional optical field via different locations of the light sources. However, the lens shape would increase much thickness of entire

21 

(a) (b)

Fig. 3.5 (a) O direction.

After the basic concept of viewing angle control, the original opto-mechanism setup

backlight module in practice. In order to overcome this problem, a designed light control film could used to replace the role of common lens to compose the VAS backlight.

n-axis source provide collimating rays in normal (b) Oblique direction rays via off-axis light source.

included the designed prism layer and three LED light bars as shown the cross-section in Fig 3.6, and Fig 3.7 is the scheme of three-dimensional picture. The normal mode is achieved by a designed prism layer and a LED light bar on the optical axis as the green rays in Fig. 3.6. In addition, the oblique mode is achieved by the same prism layer and an off axis LED light bar as the cyan rays. Because of the limit of size and the difficulty of manufacturing, a commercially available cylindrical Fresnel lens was used to replace the prism layer. In the following sections, the cylindrical Fresnel lens was called “Fresnel lens” in this thesis for convenience. The Fresnel lens played well a role in controlling the rays to provide directional optical fields and the peak value of the oblique direction is about 40 degrees.

Fig. 3.6 Schematic configuration of proposed backlight module

Fig. 3.7 The actual setup of proposed backlight module in 3-D

In this opto-mechanism, two parameters could be adjusted: the distance from the surface of LED light bar to the Fresnel lens and the power radio of these LED light bars. Besides, the detail specification of Fresnel lens and LED light were added in chapter four. In addition, it deserved to be mentioned that the focal length of the Fresnel lens was 5 cm.

The distance (i.e. d) of parameter which mentioned above was the first concern in this design procedure. In the beginning, the main assignment was to find out the suitable distance of normal viewing mode. By this reason, we could simulate this configuration with difference operating distance by optical software (i.e. LightTools).

The simulation results are shown as radar sketch in Fig. 3.8. According to Fig. 3.8, if

the LED light bar is held 5 cm (i.e. d = focal length) away from the Fresnel lens as the blue line, the angular distribution would be too sharp to apply on normal viewing mode. However, if the LED light bar is placed on the one-third of the focal length (i.e.

d/3 = 1.67 cm) as the green line, the FWHM of the angular distribution would be too wide to use as the normal mode.

Fig. 3.8 The angular distribution of normal viewing mode with various distances in radar sketch

Fig. 3.9 The FWHM of angular distribution V.S. various distances in normal viewing mode

23 

Fig. 3.9 shows the relationship between FWHM and the distance d. According to the design parameters in Table 3.1, d/2 was the most suitable distance for our design target.

However, how about the optical performances of oblique viewing mode?

normal mode, and I would discuss in the following.

various distances in oblique de

According to Fig. 3.10 and Fig. 3.11, the higher distance have inadequate angle for ob

We could repeat the similar process by optical software to find out the angular distributions in oblique viewing mode. However, it did not easily make a decision like

Fig. 3.10 The simulation results of radar sketch with

mode. The higher distance had inadequate angle for oblique viewing mo and the FWHM of the angular distribution was too wide to use in lower distance.

lique viewing mode and the FWHM of the angular distribution are too narrow to use as the black, brown, and the green lines. However, if the distance decreases to one-third focal length (i.e. d/3 = 1.67 cm), it will increase the FWHM for normal mode which shown in Fig. 3.8 and had much stray light in other viewing directions.

Therefore, we have a trade-off between the normal and oblique modes. So, we chose

d/2 (i.e. half focal length 2.5 cm) as the distance between the Fresnel lens and LED light bars. After the distance chosen by the simulation results above, the simulation of our opto-mechanism could be started. And the simulation result was shown in Fig.

3.12.

25 

Fig. 3.11 The FWHM V.S. various distances at oblique viewing mode.

Fig. 3.12 The illuminance map of normal viewing mode.

There was a hot zone at the center in this map and the uniformity was lower than 70%.

The lower distances caused the inadequate FWHM.

s a hot spot area appear by using this simple setup as the VAS

Via this way, we could reduce the dimension of screen in Fig 2.4 from 3-D into 1-D.

Similar to the process mentioned in chapter 2, the deriving details were shown below.

Fig. 3.13 1-D linear nd illumination distribution By Fig. 3.12, there wa

backlight, and the poor uniformity of normal and oblique mode was the chief defect in this configuration.

In order to solve the uniformity problem, we could review the photometry example shown in Fig 2.4. The result was equation (2.13) which also means the cos θ law. It meant that the irradiance on a flat surface which illuminated by a small point source was direct proportional to cubic of cosθ. Therefore, the main purpose of this study was to modify the viewing angle of observer in a horizontal direction in VAS backlight.

light source a

According to equation (3.2), the illumination distribution of a point source was proportional to cos²θ, and the hot spot issue was formed by this term. Therefore, we could use this property to increase the uniformity of the VAS backlight. As shown in Fig 3.8, the blue line represents the illumination curve caused by the point source. If the source intensity, Is, proportioned to cos^-2θ, the illumination distribution on the screen would become a straight line which means high uniformity. Based on this property, the uniformity could be improved.

Fig. 3.14 The blue line represented illuminance distribution via point source and the designed luminance intensity was red curve.

The discussion about the concept in Fig. 3.14 was an approximate process.

Therefore, we could set 1-D linear light source in optical software to confirm this approximation was a correct method. According to the result in Fig. 3.15, the ideal and simulation curves were similar. It could believe that this reduction progress was usefully.

27 

Fig. 3.15 The gray line represented the simulation curve which could fit the cos θ approximation

By this idea as motioned above, the light sources could be separated into three directions which as Fig. 3.16 to make the inverse cosine-square distribution introduced in the previous pages. Therefore, we could modify the opto-mechanism which was motioned in Fig. 3.6. In addition, a new opto-mechanism was proposed as Fig. 3.17 and the power ratio of each light bar could be adjusted to improve the uniformity.

Fig. 3.16 The light sources could be separated into three directions based on theoretic result to make the inverse cosine-square distribution, and then the hot spot issue could be fixed

Fig. 3.17 The new opto-mechanical structure of VAS backlight unit and the middle section (CDE) was used for normal viewing mode, and the other cluster (AB, and GF) were used for oblique viewing mode.

29 

In brief, Fig. 3.17 shown the middle section was used for normal viewing mode as shown in the green dotted line. On the other hand, the other two clusters (AB and GF) were used for oblique viewing mode. Then, the inclined light bars (BCEF) were used to increase the luminance around the edge of the partition. In order to fit the cos θ optical field distribution, the power ratio of group (ADG) to group (BCEF) was 3, and the inclined angle of tilt plane with respect to the horizontal was 55°.

Fig. 3.18 A 7 inch prototype was constructed by three sections of LED backlight units with the Fresnel lens

Since we are constrained by the fabrication issue, a 7 inch prototype was constructed by three LED backlight units with the Fresnel lens, as shown in Fig. 3.18.

After confirming the final opto-mechanism, we could check the simulation result of uniformity improvement. The simulation outcomes were shown in Fig. 3.19 and Fig.

3.20. The uniformity of one unit was over 70%, which is better than our design target in Table 3.1. In addition, the FWHM of angular distribution is +/- 20 degree, which meet the requirements of design target [Table 2.1].

Fig. 3.19 The simulation outcomes of normal mode uniformity

Fig. 3.20 The simulation outcomes of angular distribution in normal mode