Overprint Errors of Prints (2)

Place a plate with a very accurate registration mark on the press and measure the overlay error. In the Cartesian coordinate system, the overlay error is represented by the abscissa and the error occurrence rate is represented by the ordinate, resulting in the curve shown in Figure 8-2. This curve is very similar to the bell curve of the Gaussian normal distribution law, indicating that the overprint accuracy characterized by the overlay error is a statistical data with a normal distribution.

Figure 8-2 illustrates the normal market curve of overprinting errors. During the printing process, there is a large overprinting error. For example, prints larger than ±0.03 mm are few, there is no overprinting error, and 100% of registered prints are not actually printed. In most prints, there is a medium overprint error. Therefore, as long as the overprinting error meets the normal distribution, the normal distribution law can be used to evaluate the printing overlay accuracy.

Figure 8-2 Normal distribution curve of overlay error

The normal distribution is represented by the standard error, and its value can be calculated by using the difference between the actual measured overprint error and the ideal overprint. Since the ideal overprint does not exist, the actual overprint error is represented by the average value obtained from the measured values.

Specifically, when the first color is printed, a thin line is printed. This thin line is a measurement mark. When the second color is printed, another fine line mark with a certain distance from the first color mark money is printed. If there is no overlay error on the printed sheet, the distance between the two thin line markers on the printed sheet is always consistent, which achieves an ideal overprint. In fact, there is always an error in the printing process. There is no ideal overprinting situation. Therefore, the distance between the two thin line markers always changes, so that the distance between the thin line markers can be measured on many printed sheets. Averaged, that is:

x=1/nnd=1xd (8-1)

Where n is the number of printed sheets and Xd is the distance between the two thin line markers measured on each sheet. When the mean value is close to the ideal overprint value, the standard error of the normal distribution can be expressed by the following formula:

S=1/nnd=1(xd-xi)2 (8-2)

The xi in the formula is the ideal overprint value.

Experiments have shown that 68.26% of overprint errors obtained by actual measurement are within the range of single standard error, 95.45% are within double standard error range, 99.73% are within triple standard error range, ie

The actual error within ±1s is 68.26%;

The actual error within ±2s is 95.45%;

The actual error in the ±3s range is 99.73%.

In the printing process, in terms of register, if the standard error is set at 0.015mm, the ±1s limit means that the 68.26% overprint error of all the sheets is only 0.015mm; the ±2s limit means all the sheets 95.45% of overprint errors are less than or equal to 0.03mm; and 99.73% to 95.45% of the printed sheets, that is, 4.28% overprint of all sheets is 0.03-0.045mm. The rest of 0.27% of the sheets have an overlay error of more than 0.045mm.

If more than two factors of overlay error are encountered, then each overprint error is regulated to have a normal distribution property, and they are added together to produce a new normal distribution. This new normal distribution is There is a new standard error, which can be calculated using the following formula:

S2a=S21+S22+S23+...S2n (8-3)

Sa is the new standard error of the new normal distribution, S1, S2, S3......... Sn is the standard error of a single normal distribution.

If the error of the continuation paper of the known printing press is small, ten mm, and the paper transfer error is 0.03 mm, the overprint error should be

S=(0.052+0.032)1/2=0.00341/2=0.058mm