WITE Revision 1.17 Release Notes

New Features

• PRML Comparator Error Test based on Viterbi algorithm is now implemented on the DSP array. The test now runs at least 10 times faster than in previous releases.

• New functionality implemented to support S-312MP Spinstand.

• New Signal-to-Noise Ratio test is implemented. Unlike the existing test, a signal doesn't have to be erased to determine the SNR. (See: G.Main, T.D.Howell, "Determining a signal to noise ratio for an arbitrary data sequence by a time domain analysis", IEEE Trans.of Magnetic,v.29,n.6,Nov 1993, pp.3999-4001).

• Band Erase operation in Production sequence was optimized so that if enabled, Band Erase will be performed for all heads at the same time, thus avoiding multiple 'seek offset' operations.

• More softbuttons were added to the Operator Panel allowing an operator to run more tests.

• The "Use Last Profile" button was added to the Reshaping Program. Pressing this button sets the "Database Name" text box to the full name of the latest result database used by WITE.

Bugs Fixed

• Sequenced Amplitude Margin test with Viterbi Histogram method sometimes caused "DSP time-out" message and system crash.

• During the Comparator Error Test, with 'PCTR' pattern (or any other pattern dealing with chip digital channel) and write option 'PATTERN', writing through PRML chip frequently failed. This caused huge numbers of comparator errors.

• PW measurement produced wrong results if all of the following conditions were present: (1) the first tested head was bad, (2) "Context" flag was enabled, and (3) the run time limits didn't reject the head.

• An attempt to run Frequency Test with frequencies higher than 160 Mflux on RWA1632/RWA1601 with High Frequency Pattern Generator caused an error "DPRT: ambiguous write".

• Head amplifier TDA5152 always divided the write data by four (both internal and external Flip/Flops were active).

• Comparator Error Rate test didn't recognize EPR4 patterns as PRML patterns that led to "Non PRML Pattern" error message.
• The meaning of Amplification Level settings was always reversed. If UPDATE.DAT had 'Amplevel "High"', the UP6 was initialized to use level "Low" and vice versa.

• NLTS pseudorandom test used the filter from "Control"/"Filter" menu, not from test setup.

• Opening the equalizer menu under "Control"/"Equalizer" always reset equalizer taps from the database.

• "Calibration"/"ADC delay" menu allowed users to specify a delay much bigger than PRML clock period. Unfortunately, such a delay is not allowed by hardware, and it might actually disable PRML clocks completely. Now, an additional validation procedure keeps the delay value within the acceptable range.

• On RWA1632 machine, Pulse Profile test resulted in the error message "DSP time-out" if RPM was less then 4000 and ReadGate = Full Track.

• If the "Cancel" button on Error Message box was pressed when no abortable operation was running, then all subsequent error messages were suppressed. Now the "Cancel" button appears on the Error Message box only if there is a test running by WITE.

• If the absolute path for pattern file was specified in "Configure"/"System" menu, then an attempt to load pattern editor caused an error message.

• If extended (*.EPT) pattern file was corrupted, then an attempt to load this file caused a system crash.

• During installation of WPRML, W747 and WGMA software under Windows 95 the following message was displayed: "Warning - Can not copy file DDEML.DL_ since the destination file is already in use".