Patent Application
20080129562
The present invention relates to semiconductor integrated circuits, and more particularly relates to a semiconductor integrated circuit having the function of measuring jitter in an EFM (eight to fourteen modulation) signal used in a CD (compact disc) device and the like.
In optical disk devices, jitter in a data signal read from an optical disk is generally measured by the following two methods. One is a data-to-clock jitter measurement method used in DVD (digital versatile disc) devices and the like, and the other is a jitter measurement method called a 3 T or 22 T method, which is used in CD devices and the like.
A data-to-clock measurement circuit is incorporated into an LSI (large scale integrated circuit) for the learning function of the LSI. In DVD devices, jitter measurement for delivery inspection can thus be performed using the data-to-clock measurement circuit. In a CD device, a circuit for measuring jitter is not incorporated into an LSI. This is because, in the first place, jitter measuring function is unnecessary for LSIs. Jitter measurement in delivery inspection of a CD device is performed using a jitter meter which is commercially manufactured and sold as a measuring instrument.
However, in the conventional method that uses jitter meters to perform delivery inspection of CD devices, a jitter meter must be provided for each fabrication line, which causes the delivery inspection cost to increase. It is thus required that the LSI itself measure jitter without using a jitter meter. Nevertheless, if the function of the above-described jitter meter is just incorporated into LSIs without modification, jitter measurement results vary from LSI to LSI due to fabrication variations in the LSIs, and it is thus difficult to use the measurement results in the delivery inspection.
In view of the above problem, it is an object of the present invention to realize semiconductor integrated circuits capable of highly precise jitter measurement in which variation due to variation among the individual circuits does not occur.
In order to solve the above-described problem, an inventive semiconductor integrated circuit includes: a slicer for binarizing an input signal to generate a data signal; a T/V converter for outputting a voltage corresponding to the data length of an input signal; a multiplexer for selecting either the data signal or a reference signal as the input signal to the T/V converter; an A/D converter for converting the output voltage of the T/V converter to digital data; a processor for measuring jitter in the input signal to the T/V converter in accordance with the digital data; and a correction section for comparing a predetermined voltage with the output voltage of the T/V converter produced where the reference signal has been selected by the multiplexer, and correcting output characteristics of the T/V converter in accordance with results of the comparison.
In the inventive semiconductor integrated circuit, the reference signal is selected by the multiplexer as the input signal to the T/V converter, and the correction section makes a comparison between the predetermined voltage and the output voltage of the TN converter produced as a result of the input of the reference signal. Based on the comparison results, the output characteristics of the T/V converter are corrected. Consequently, highly accurate jitter measurement, in which variation due to variation among the individual circuits does not occur, is realized.
The correction section preferably adjusts a gain and offset of the T/V converter.
The processor preferably measures the jitter in the input signal to the TN converter in accordance with the digital data that falls within a predetermined range.
The processor preferably calculates an average value of the data lengths of input signals to the T/V converter in accordance with the digital data to calculate a deviation of the average value from an ideal value, and the semiconductor integrated circuit preferably includes a slice level correction section for correcting the slice level of the slicer in accordance with the deviation calculated by the processor.
The semiconductor integrated circuit preferably includes an amplifier for amplifying the output voltage of the T/V converter, wherein the A/D converter preferably converts the voltage amplified by the amplifier to the digital data.
The processor preferably calculates a deviation of a gain of the T/V converter from an ideal gain in accordance with first digital data and second digital data which are output from the A/D converter, and corrects the digital data in accordance with the calculated deviation, the first digital data being output where a first reference signal having a first data length has been selected by the multiplexer, the second digital data being output where a second reference signal having a second data length has been selected by the multiplexer.
The processor preferably subtracts a variance of the jitter obtained where the reference signal has been selected as the input signal to the T/V converter from a variance of the jitter obtained where the data signal has been selected as the input signal to the T/V converter.
Function in the semiconductor integrated circuit other than the function of jitter measurement is preferably shared by at least one of the slicer, the A/D converter, and the processor.
As described above, according to the present invention, the semiconductor integrated circuits having the jitter measuring function achieve highly precise jitter measurement in which variation caused by variation among the individual circuits does not occur. Therefore, jitter meters do not have to be used any more in delivery inspection, and the delivery inspection cost and hence the fabrication cost of the semiconductor integrated circuits are reduced.
Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings.
Since the T/V converter 12 has variation caused during the fabrication of the LSI, the gain of the T/V converter 12, which is defined by a change in conversion voltage with respect to a change in the length of input data, varies from LSI to LSI. In addition, an absolute voltage that is output when data having a proper data length is input, that is, the offset, also varies from LSI to LSI. When the processor 14 performs arithmetic processing for obtaining a standard deviation or the like, the former variation becomes variation in the standard deviation itself, and thus directly and adversely affects the jitter measurement results. The latter variation becomes variation in the center voltage in the distribution of the voltage obtained by the T/V conversion. If the center voltage varies greatly, the output voltage of the T/V converter 12 falls outside the input range of the A/D converter 13, which may lead to erroneous jitter measurement results. Therefore, the correction section 16 corrects the output characteristics of the T/V converter 12 to reduce variation among the LSIs.
Specifically, the output characteristics of the T/V converter 12 are corrected as follows. First, the multiplexer 15 inputs the reference signal to the T/V converter 12. The reference signal is a jitter-free data signal having a predetermined data length, that is, a proper data signal. The reference signal may be externally provided, or may be generated inside the LSI. The correction section 16 compares a predetermined voltage with the output voltage of the T/V converter 12, to which the reference signal has been input, and gives feedback to the T/V converter 12 according to the comparison results. More specifically, the correction section 16 adjusts the gain of the T/V converter 12 so that the output voltage of the T/V converter 12 is equal to the predetermined voltage. Consequently, variation in the gain of the T/V converter 12 from LSI to LSI is reduced.
Furthermore, the correction section 16 performs offset adjustments so that the output voltage of the T/V converter 12 is in the vicinity of the center of the input range of the A/D converter 13. As a result, the output voltage of the T/V converter 12 falls within the input range of the A/D converter 13, whereby accurate jitter measurement results are obtained.
As already described, the T/V converter 12 has the function of selecting a data signal having a specific data length. Nevertheless, when jitter is relatively large, a distinction between the specific data length and the other lengths cannot be made clearly, and thus a data signal whose data length is other than the specific data length may be selected. For example, when the specific data length is 3 T, a data signal having a data length of 2 T or 4 T may be selected. If such a data signal having a data length other than the specific data length is contained in the output of the T/V converter 12, the reliability of the final jitter measurement results decreases. Thus, limitations are placed on digital data to be processed by the processor 14, so that the processor 14 only processes data within the predetermined range. Specifically, in the case where the specific data length is 3 T, the limitations are placed so that only digital data whose data length is close to 3 T, e.g., in the range from 2.5 T to 3.5 T, is processed. Then, the accuracy of jitter measurement results is increased further.
As the precision of the correction by the correction section 16 is improved, variation in the output characteristics of the T/V converter 12 from LSI to LSI is reduced. However, the improvement in the precision of the correction requires the circuit size of the correction section 16 to be increased. Furthermore, the predetermined voltage that is compared with the output voltage of the T/V converter 12, and the offset of a comparator (not shown) which performs actual comparison operation vary from LSI to LSI. Thus, the occurrence of some variation in jitter measurement from T/V converter 12 to T/V converter 12 is unavoidable. It is therefore desired that the correction section 16 make coarse adjustments, and the processor 14 make fine adjustments, so that the jitter measurement precision is improved.
To be specific, the processor 14 receives digital data V1 output when a first reference signal whose data length is 3 T is supplied to the T/V converter 12, and digital data V2 output when a second reference signal whose data length is 2.5 T is supplied to the T/V converter 12. The gain of the T/V converter 12 at this time is expressed by (V1−V2)/(3 T−2.5 T). Assume that the ideal value of the digital data related to the first reference signal is V10, and the ideal value of the digital data related to the second reference signal is V20. In this case, in the processor 14, input digital data may be multiplied by (V10−V20)/(V1−V2). Then, the gain error in the T/V converter 12 is corrected by the processor 14, whereby the accuracy of the jitter measurement results is increased further.
In an LSI incorporating jitter measuring function, noise from other circuits in the LSI is sometimes applied to the T/V converter 12 and the A/D converter 13. Such noise applied from the other circuits is measured as jitter to cause an error in the jitter measurement results. The following measures are thus preferably taken. The processor 14 calculates the variance of jitter observed with the reference signal being input. Since the reference signal does not contain jitter, the variance obtained at this time is mainly due to the noise. The processor 14 stores this variance and subtracts the stored variance from the variance of jitter observed by the input of an ordinary data signal, whereby the jitter measurement error caused by the noise from the other circuits in the LSI is cancelled.
As described above, according to this embodiment, in the LSIs having the jitter measuring function, fabrication variations among the LSIs are reduced. As a result, accurate jitter measurement is realized irrespective of variations among the individual LSIs.
It should be noted that the correction section 16 may be configured as hardware or may be performed as software processing by using a DSP (digital signal processor) or the like. Also, each of the slicer 11, the A/D converter 13, and the processor 14 may be shared with other function in the LSI or the like by using a time sharing system, whereby the layout area of the LSI is reduced.
In some of the EFM signals, which are input signals having a specific data length, the specific data length deviates from a proper data length. A processor 14 detects jitter in input signals by using a certain detection window. A signal whose data length deviates falls outside the detection window and the jitter in that signal is not measured. That is, the jitter of the jitter-measurement-target data signal is not measured, which might result in erroneous jitter measurement results. Therefore, the slice level of the slicer 11 is corrected by the slice level correction section 17 to reduce the deviation of the data length of the data signal.
Specifically, the processor 14 calculates the average value of each of the data lengths in the input signal and outputs a deviation of the average value from its ideal value. Based on this deviation, the slice level correction section 17 adjusts the slice level of the slicer 11. To be more specific, the slice level correction section 17 gives feedback to the slicer 11 until the deviation disappears.
As described above, according to this embodiment, in the semiconductor integrated circuit having the jitter measuring function, adjustments are performed so that the average value of the distribution of input signal data lengths is its ideal value. As a result, precise jitter measurement is realized.
In the present invention, a jitter to be measured is defined as a data-length-independent absolute value, more specifically, as a value obtained by dividing a deviation of an observed data length from its ideal data length by a predetermined value (a reference data length). According to this definition, a jitter of Ins in a data signal whose data length is 3 T is the same in magnitude as a jitter of Ins in a data signal whose data length is 11 T, for example.
In the 11 T method, as compared with the 3 T method, an integration time for integrating a sawtooth waveform in a T/V converter 12 is longer, and hence the output voltage is larger. In the 11 T method, therefore, the gain of the T/V converter 12 must be lowered as compared with the 3 T method, so that the output voltage of the T/V converter 12 falls within the input range of the A/D converter 13. However, the lowered gain causes reduction in jitter, such that jitter that is different in magnitude from the actual jitter is measured, resulting in decrease in the jitter measurement accuracy. The processor 14 can compensate for the jitter reduction, but in order to allow the processor 14 to make the compensation, the A/D converter 13 is required to output highly accurate digital data. Improving the accuracy of the A/D converter 13, however, leads to increase in costs and is thus not desirable. As shown in
The amplifier 18 amplifies the output voltage of the T/V converter 12, and the A/D converter 13 outputs digital data corresponding to the amplified voltage. That is, the jitter reduced by the T/V converter 12 is amplified to the original magnitude by the amplifier 18, and then the amplified voltage is supplied to the A/D converter 13. Consequently, in the semiconductor integrated circuit having the jitter measuring function, the precision of the jitter measurement by the 11 T method is increased.
In the above-described embodiments of the present invention, as shown in
The semiconductor integrated circuits according to the present invention have the function of performing highly accurate jitter measurement in which variation caused by variation among the individual circuits does not occur, and are thus applicable to LSIs for writable CD devices.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2004-250224 | Aug 2004 | JP | national |
| Filing Document | Filing Date | Country | Kind | 371c Date |
|---|---|---|---|---|
| PCT/JP05/07164 | 4/13/2005 | WO | 00 | 2/28/2007 |
