Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Self-learning RRO cancel by feed forward

IP.com Disclosure Number: IPCOM000014808D
Original Publication Date: 2001-Dec-29
Included in the Prior Art Database: 2003-Jun-20
Document File: 2 page(s) / 86K

Publishing Venue

IBM

Abstract

For HDD's servo control, RRO is a large factor to degrade the servo stability. The technical method described here is one of the technique to reduce RRO using feed forward servo control. There are three steps to cancel RRO by self learning and feed forward as followed. STEP-1: generate RRO information (calculate RRO components) STEP-2: update rro_table with dumping (RRO learning process) STEP-3: feed forward for coming sector (output process) In each steps, following calculation is done.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 100% of the total text.

Page 1 of 2

Self-learning RRO cancel by feed forward

For HDD's servo control, RRO is a large factor to degrade the servo
stability. The technical method described here is one of the
technique to reduce RRO using feed forward servo control.

There are three steps to cancel RRO by self learning and feed
forward as followed.

STEP-1: generate RRO information (calculate RRO components)
STEP-2: update rro_table with dumping (RRO learning process)
STEP-3: feed forward for coming sector (output process)

In each steps, following calculation is done.

N : servo sector number of 1 revolution
PES : Y(i) = CurrentCylPos(i) - TargetCylPos
rro_table: X(i) is rro_table of i'th servo sector
Kdf : feed forward gain

(STEP-1) generate RRO information

Y0(i) = Y(i) <--- position
Y1(i) = Y0(i) - Y0(i-1) <--- velocity
Y2(i) = Y1(i) - Y1(i-1) <--- acceleration
Y3(i) = Y2(i) - Y2(i-1) <--- deviation of 4th order
Y4(i) = Y3(i) - Y3(i-1) <--- deviation of 5th order
rro_input(i) = Y0(i) + Y1(i) + Y2(i) + Y3(i) + Y4(i)

(STEP-2) update rro_table with dumping
delta = rro_input(i) - X(i-N)
X(i) = X(i-N) + epsilon
where epsilon is determined by delta's value

(STEP-3) feed forward for coming sector
rro_val(i) = Kdf * { X(i-N+1) + X(i-N+2) }
dacout(i) = dacout(i) + rro_val(i)

By using this method, RRO can be removed and servo stability can be
improved.

Following figuure shows the effect of this technique.

1

Page 2 of 2

2

[This page contains 2 pictures or other non-text objects]