KIVEKÄS et al.: BANDWIDTH, SAR, AND EFFICIENCY OF INTERNAL MOBILE PHONE ANTENNAS

77

Fig. 5. Distribution profiles of maximum 10-g average SARs in watts per kilogram in head at (a) 900 MHz

4.5 W kg and (b) 1800 MHz

7 mm.

9.4 W kg . Distance from head to phone

Fig. 6.

Examples of the measured unaveraged SAR in watts per kilogram under the chassis area on the phantom surface at (a) 900 and (b) 1800 MHz. Chassis

101 mm. Distance from head to phone

1 W.

7 mm.

length

2) With Hand: The effect of the chassis length on the SARs

decreases the efficiency roughly by 2 to 4 dB when  is in-

and radiation efficiency was studied in the presence of the hand

creased from 21 to 81 mm, and approximately 5 dB with longer

7 mm . Similar general behavior in the

models as well

chassis. The decrease due to hand2 is approximately 3 dB for

61 mm and approximately 4 dB otherwise. At 1800 MHz,

SARs in the head was observed with either one of the hand

the decreases due to hand1 and hand2 are approximately 3 and

models than without hand. Only the results with hand2 are il-

2 dB, respectively.

lustrated with figures (Fig. 7). As expected [19], lower SARs in

the head are obtained when a hand is present, because it absorbs

some of the power. Owing to the shape and location of hand1

C. Performance Comparison With Different Antenna

close to the chassis [see Fig. 2(a)], the SARs in the head are

Orientations

generally lower with hand1 than with hand2, and the average

The effect of the antenna orientation with different chassis

SARs in hand1 are higher than in hand2. The values of the 10-g

lengths was investigated by studying the performances of the

average SARs in hand2 are very low and nearly constant with

antenna elements of Fig. 1 with different feed and short-cir-

all chassis lengths (Fig. 7) due to their location in palm, which

cuit positions (Figs. 8­11). Locations marked with 1 (Loc1)

is relatively far from the chassis. The high unaveraged SARs in

are the configurations studied earlier in this section. The effect

hand2 at 1800 MHz are explained by their location in the fingers

of different resonant frequencies caused by different feed and

near the chassis edge [see Fig. 2(b)].

short-circuit positions and chassis lengths has not been removed

Also, the trend in the radiation efficiency at 900 and at

from the results.

1800 MHz, with either hand1 or hand2 included in the simula-

At 900 MHz, the impedance bandwidths (Fig. 8), as well as

tion, is rather similar to that obtained without hand. Compared

the SAR and radiation efficiency characteristics (see Figs. 4(a)

to the radiation efficiencies without hand, at 900 MHz hand1