Coded data generation or conversion – Analog to or from digital conversion – Detecting analog signal peak
Reexamination Certificate
1999-03-30
2001-02-27
JeanPierre, Peguy (Department: 2819)
Coded data generation or conversion
Analog to or from digital conversion
Detecting analog signal peak
C367S059000, C330S129000
Reexamination Certificate
active
06195028
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to peak detection, and more specifically to a method of digital peak detection of disk drive servo signals.
BACKGROUND OF THE INVENTION
In disk based magnetic recording systems, data are typically stored in circular tracks. The recorded information of the disk surface is divided into sectors. A servo signal is provided on the disk to direct the actuator motor to find and then follow a track. Disk drives commonly use embedded servo information, wherein a small part of the disk surface is allocated to servo fields. The servo fields are encoded to indicate to the servo control system the position of the read/write head on the disk. The servo fields store information about the sector number, the track number, and how to center the head on a track. The track number is encoded using grey codes. The servo information is recorded onto the disk after the Head Disk Assembly (HDA) has been assembled.
Information is stored on the disk in the form of magnetic transitions. Each bit of information stored on the disk corresponds to one magnetic transition. A read head generates electrical signals corresponding to the magnetic transitions. A “1” may be used to designate the presence of a magnetic transition, and a “0” to designate the lack of a magnetic transition. The read head generates either a positive or negative pulse for each magnetic transition depending on the polarity of the transition. Data are read from the disk by processing transition responses.
A peak detector circuit attached to the read head analyzes the analog signal generated by the read head to identify the presence of pulses which indicate magnetic transitions. One conventional analog peak detection method is to differentiate the signal and detect zero crossings of the signal derivative. The signal derivative is zero for local minimums and local maximums. The amplitude of the signal where the derivative is zero is then compared to a threshold level to identify peak samples. Such peak detection methods typically have a sample comparison window two to three samples wide.
A conventional digital peak detection method is to convert the analog samples to digital samples, and then compare a sample to the previous sample and subsequent sample. If the sample is greater than the previous and subsequent sample then the sample is compared with a threshold level. If the sample exceeds the threshold then the sample is identified as a peak. Similar to the conventional analog peak detection system such digital methods typically have a peak comparison window that is three samples wide. One problem with such detection methods is that noise spikes can cause two or more peaks to be reported within a timing window in which there is only one recorded peak. Such false peaks reduce operating efficiency where they are detected and trigger a second reading of the data. When such errors go undetected they can cause system failures.
Peak detectors typically use a variable gain amplifier (“VGA”) to amplify the read signal so that the pulses approximate a desired amplitude to optimize pulse detection. Conventional analog peak detectors use the charge on a capacitor to control the VGA gain. One capacitor is used for reading servo signals in servo mode and one capacitor is used for reading data signals in data mode. The capacitors are coupled to the VGA control terminal using a multiplexer to switch between the two capacitors. An automatic gain control (AGC) feedback loop is used to control the VGA gain. When a pulse is detected above the AGC target amplitude, a small amount of charge is discharged from the gain control capacitor to reduce the VGA gain. A resistor coupled to the gain control capacitor is used to charge the capacitor. When a series of pulses are below a minimum qualified pulse amplitude the resistor charges the capacitor which increases the VGA gain and thereby increases the pulse amplitude. One problem with these conventional analog peak detectors is that if the read head goes through a damaged data field where none of the pulses exceed the qualified pulse amplitude level then at the end of the data field the gain is generally increased to a higher level then is desired for the subsequent intact data field. This causes delays while the AGC loop recovers.
At the end of a servo field the servo gain control capacitor stores the gain value for the next servo field. Similarly the data gain control capacitor stores the gain for the next data field. During these storage periods the value of the stored gain deteriorates due to leakage of the capacitor.
Thus there is a need for an improved peak detection system that provides enhanced performance and overcomes these and other problems of the prior art.
SUMMARY OF THE INVENTION
The present invention provides a digital peak detection system. The system first converts an analog signal into digital samples. The digital samples are then converted into absolute values. The absolute values are provided to a delay line consisting of a series of one cycle delays. In one embodiment a sequence of comparators are provided, each of which receives one input from a subsequent stage of the delay line. The other input of all of the comparator stages are coupled together to the input of the delay line. With this configuration each new digital sample is compared to the preceding N samples in parallel, where N is the number of active comparator stages. The comparator stages can be activated using control signals to determine the length of the sample comparison window. To provide robust peak detection a comparison window that is larger than the number of samples between recorded peaks can be used. The outputs of the comparator stages are ANDed together so that the AND gate outputs a high signal when the current sample is greater than all of the preceding samples that it is compared against. Here, a sample whose absolute value is larger than that of the N previous samples is referred to as a current preliminary peak sample.
If the absolute value of a current preliminary peak sample is also greater than or equal to those of the succeeding N samples, it is considered a legitimate peak in a window of 2N+1 consecutive samples. This embodiment minimizes the number of comparator stages required by then using the same comparator stages to compare the current preliminary peak sample to each of the subsequent samples in the comparison window. The subsequent samples are compared by monitoring the output of the comparator stages and one active post-comparison stage for each active comparator stage. In this embodiment, only one comparator is used for every two samples in the 2N+1 sample comparison window. The post-comparison stages each comprise an AND gate, a one-cycle delay, and a control circuit to activate the stage. One input of each of the post-comparison stages is coupled to the AND gate output from the comparator stages. The other input of each of these post-comparison stages is coupled to the preceding post-comparison stage through the delay element. The post-comparison stages thereby check that the current preliminary peak sample is greater than or equal to each of the subsequent samples in the comparison window. If the current peak sample is greater than or equal to the other samples in the comparison window then the sample is a legitimate peak in the window and is next compared with a qualifier threshold. If the sample is greater than the qualifier then it is identified as a qualified peak.
A second component of the digital peak detection system comprises a gain error control circuit with a countdown timer. A variable gain amplifier is used to optimize the peak detection processes. The gain of the variable gain amplifier is updated based on the difference between the actual peak sample amplitudes and the desired peak sample amplitudes. Detecting a qualified peak is one condition that triggers an update of the variable gain amplifier. When the sample amplitudes are too small to meet the qualifier threshold, a countdown timer is used to increase the variable
Fredrickson Lisa
Stone Dennis C.
Volz LeRoy
Dempster Shawn B.
Heller III Edward P.
Jean-Pierre Peguy
Olson Jonathan E.
Seagate Technology LLC
LandOfFree
Amplifier gain control method using conditional updating does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Amplifier gain control method using conditional updating, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Amplifier gain control method using conditional updating will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2559210