INTRODUCTION
The discs with vanes of rotary fertilizer spreaders throw the fertilizer by the action of centrifugal force. These discs may be flat or coneshaped. A coneshaped disc has an angle greater than 0° between its own surface and the horizontal plane, while the angle between the surface of a flat disc and the horizontal plane is 0°. Therefore, the particles discharge from a coneshaped disc with an upward component of velocity. As this result, the particles are thrown a larger distance (Srivastava et al., 1993; Hofstee et al., 1999).
The revolution speeds of discs used on the rotary fertilizer spreaders range mostly between 540 min^{1} and 1000 min^{1}. Olieslagers et al. (1996), using a twin flat disc fertilizer spreader reported that as the revolution speed of disc was increased, the distribution width increased. Aphale et al. (2003), using spreader with a flat disc with different material, notified that the mean distribution distances increased with an increase in the revolution speed of disc (from 540 to 810 min^{1}). Dintwa et al. (2004) reported that the revolution speed of disc (280, 516, 752 and 910 min^{1}) had the effect on fertilizer distribution pattern in singleflat disc spreader. Carman (1991) determined that the distribution uniformity increased as the revolution speed of disc with a cone angle of 2° was increased. But, the effect of different cone angles along with different revolution speeds of disc on fertilizer distribution uniformity is unknown.
The objective of this study was to determine the effect of disc cone angle on fertilizer distribution uniformity and the best combination between the cone angle and the revolution speed with different flow rates for triple superphosphate.
MATERIALS AND METHODS
A domestic mounted singledisc rotary fertilizer spreader mounted on a tractor was used in the experiments. The fertilizer used in the experiments was triple superphosphate. Bulk density, moisture content and repose angle of this fertilizer were 1057 kg m^{3}, 3.51% and 35.50°, respectively. The sieve analysis and mass median diameter of this fertilizer are shown in Table 1.
The experiments were performed with a flat disc and two conical disc, three
revolution speeds of disc and three flow rates of fertilizer. The cone angles
of the discs used in the experiments were 0° (flat disc), 10° and 20°.
The dimensions of these discs and the angular position of circular orifice with
respect to travel direction are shown in Fig. 1.

Fig. 1: 
Dimensions of discs and positions of orifices 
The discs were run at the revolution speeds of 405, 540 and 810 min^{1}
in the experiments. The height of the discs from the top of the collection trays
was 600 mm. The ground speed was 4.44 km h^{1} in all tests. The experiments
were conducted according to ASAE S341.2, Procedure for Measuring Distribution
Uniformity and Calibrating Granular Broadcast Spreaders (ASAE Standards, 1998).
The circular orifices with diameters of 30, 40 and 50 mm at the bottom of the conical fertilizer hopper were used for obtaining the different flow rates. The flow rates obtained from these circular orifices with diameters of 30, 40 and 50 mm were 8.85, 22.69 and 43.27 kg min^{1}, respectively.
The heights and lengths of
type vanes used in the experiments were 65 mm and 120 mm, respectively. The
six of this vane type were mounted radially on each disc. The distance between
the center of disc and the inner end of vanes and the average feeding radius
of the fertilizer on the disc were 127 and 80 mm, respectively (Fig.
1).
Collection trays were 500 mm wide, 500 mm long and were subdivided into compartments with 10 cm wide, 10 cm long and 10 cm high made of cardboard to decrease the possibility of particles ricocheting out of trays.
Collection trays were placed adjacent to each other in a line perpendicular to direction of travel of the spreader. Thus, the swath width of 26 m was obtained. One collection tray on each side of centerline was removed to allow the tractor wheels to pass through the collection trays. The amounts of fertilizer, at the locations removed the collection trays, were interpolated from the adjacent collection trays (Parish, 2003).
The fertilizer collected in each tray was weighted using a precision balance with an accuracy of 0.01 g. The coefficient of variation (CV) of fertilizer weight data were calculated based on the sidebyside circuitous mode of operation, according to ASAE S341.2 (ASAE standards, 1998).
The preliminary tests were conducted to determine the angular positions of the orifices obtaining the optimum pattern performance with flat disc at the revolution speed of 540 min^{1} for each orifice. According to the preliminary tests, the angle α between the center of orifice and the travel direction for the orifice diameters of 30, 40 and 50 mm were 5°, 10° and 20° counterclockwise, respectively. All experiments were performed with these angular positions.
The treatments were arranged factorially in three replications including three cone angles of disc, three revolution speeds of disc for three flow rates of fertilizer. The analysis of variance (ANOVA) was applied to evaluate the treatment effects for each orifice diameter separately. Duncan’s multiple range test procedure was used to compare the mean values of CV obtained from each combination.
RESULTS AND DISCUSSION
Table 2 presents skewing, CV and minimum and maximum values.
Skewing value is expressed as the percentage of material on the left side of
the pattern centerline compared with the percentage of material on the right
side of the pattern centerline. The CV, widely used to characterize the quality
of distribution uniformity in overlapped pattern, defined as the standard deviation
divided by the mean. The minimum and maximum values are expressed as a percentage
of the mean rate in the overlapped pattern. Onal (1995) reported that the difference
between amount of fertilizer on left side of the pattern centerline and amount
of fertilizer on right side of the pattern centerline (skewing) should be ≤4%.
Table 2: 
Summary of pattern data^{[a]} 

^{[a] }Each data point shown is the mean of three
replications.^{ [b] }Skewing in the overlapped pattern is expressed
as the percentage of material on the left side of the pattern centerline
compared with the percentage of material on the right side of the pattern
centerline. ^{[c] }The coefficient of variation (CV) is the minimum
value (best) in the overlapped pattern. ^{[d] }Means followed by
the same letter in each group are not significantly different at level of
probability specified, according to Duncan’s Multiple Range Test. ^{[e]
}Minimum and maximum points in the overlapped pattern (at effective
swath width) are expressed as a percentage of the mean rate 
According to Parish (2002), a high skewing value cause a high CV and increase
the differences between minimum and maximum values. In an acceptable overlapped
pattern, the value of CV should be ≤20%, minimum and maximum values should
be ≥80% and ≤120%, respectively (Parish, 2001). Each data point in Table
2 is the mean of each combination between the cone angle of disc and the
revolution speed of disc.
The results of ANOVA for each orifice diameter indicated that there were significant differences (p<0.001) among both cone angles and revolution speeds of disc. The values of CV from the interaction between the cone angle and the revolution speed of disc were significantly different from each other. These differences were statistically significant at the 1% (p<0.01) level of probability for the orifice diameter of 30 and 40 mm and at the 5% (p<0.05) level of probability for the orifice diameter of 50 mm (Table 2).
As increased the revolution speed of disc, the values of obtained CV decreased for all the cone angle of disc, therefore the fertilizer distribution uniformity became better (Table 2). For a given the revolution speed of disc, the CV obtained from overlapped pattern increased with increasing cone angle of the disc, which indicated that the uniformity of fertilizer distribution was getting worse. These cases were the same at all orifice diameters (30, 40 and 50 mm).
For all the cases, the values of obtained CV changed from 9 to 43%. The minimum value (best) of CV was obtained at the combination between the cone angle of 0° (flat disc) and the revolution speed of 810 min^{1} with diameter orifice of 30 mm. The maximum value (worst) of CV was obtained at the combination between the cone angle of 20° and the revolution speed of 405 min^{1} with diameter orifice of 50 mm.
Table 2 shows, at the same time, the values of CV obtained
from the same combination increased with an increase in orifice diameter (therefore
increasing flow rate onto disc).

Fig. 2: 
The distribution patterns belonging to the orifice diameter
of 30 mm at the disc revolution speed of 405, 540 and 810 min^{1}.
The center of the disc passed midway between the centers of collection trays
26 and 27. Each line in the graphs is the mean of three replications 
Parish (2002) and Yildirim and Kara (2003) also reported that the fertilizer
distribution uniformity was deteriorated with increasing flow rate onto disc.
All the values of CV obtained at the cone angle of 0° and 10° were found to be <20% for the disc revolution speed of 540 min^{1}, used generally in practice, except for the combination between the cone angle of 10° and the diameter orifice of 50 mm. For the disc cone angle of 20°, all the values of CV were found to be >20% at all the revolution speed of disc.
Figure 24 showed the maximum and minimum
points became closer to 100% when the revolution speed of disc was increased
(405, 540 and 810 min^{1}).

Fig. 3: 
The distribution patterns belonging to the orifice diameter
of 40 mm at the disc revolution speed of 405, 540 and 810 min^{1}.
The center of the disc passed midway between the centers of collection trays
26 and 27. Each line in the graphs is the mean of three replications 
On the other hand, increasing the cone angle (0°flat disc, 10° and
20°) caused the patterns to skew more to the left. This case is undesirable
for uniformity of fertilizer distribution. According to Grift (2000), an acceptable
fertilizer distribution pattern should be similar to the normal distribution
(also called Gaussian distribution). As the minimum and maximum values became
closer to 100% and the skewing values became closer to 50%, the distribution
patterns are similar to normal distribution.

Fig. 4: 
The distribution patterns belonging to the orifice diameter
of 50 mm at the disc revolution speed of 405, 540 and 810 min^{1}.
The center of the disc passed midway between the centers of collection trays
26 and 27. Each line in the graphs is the mean of three replications 
CONCLUSIONS
The data from this study demonstrated that the cone angle of disc along with
the revolution speed of disc had an effect on the fertilizer distribution uniformity.
Increasing the cone angle of disc deteriorated the fertilizer distribution uniformity.
On the other hand, as the revolution speed of disc increased, the values of
obtained CV decreased. The best distribution uniformity was obtained from the
combination between the disc cone angle of 0° (flat disc) and the revolution
speed of 810 min^{1} for all orifices.
The values of CV obtained from all the experiments varied from 9 to 43%. The
values of CV for the cone angle of 0° and 10° with the revolution speed
of 540 min^{1} were <20%, except for the combination between the
cone angle of 10° and the diameter orifice of 50 mm. Therefore, at disc
revolution speed of 540 min^{1}, used widely in practice, manufacturers
should be taken into consideration that the angle of disc has to be <10°
to obtain an acceptable fertilizer distribution uniformity.