WO1994017520A1

WO1994017520A1 - System for correcting transducer tangential skew in a dual actuator disk drive system

Info

Publication number: WO1994017520A1
Application number: PCT/US1994/000712
Authority: WO
Inventors: Kurt Anderson; Aaron Wilson
Original assignee: Conner Peripherals, Inc.
Priority date: 1993-01-22
Filing date: 1994-01-18
Publication date: 1994-08-04
Also published as: HK1008366A1; ATE178156T1; JPH08505976A; DE69417403T2; EP0685102A1; US5523901A; DE69417403D1; EP0685102B1; EP0685102A4

Abstract

An apparatus in a dual actuator disk drive system having a nonpackwriting actuator (9) and a packwriting actuator (8) for controlling the sensing of address marks by the transducers (13, 23, 24, 26, 28 and 30) associated with the nonpacking actuator after a switching operation between any two of the transducers (13, 23, 24, 26, 28 and 30) associated with the nonpackwriter actuator (9). The apparatus comprising an address mark means (48) for generating an address mark search signal to start an address mark search when an address mark is expected to be read by the presently selected transducer (13, 23, 24, 26, 28 or 30) associated with said nonpackwriter actuator (9) and an adjusting means (52, 49) connected to said address mark means (48) for altering the time when said address mark means (48) will generate said address mark search signal such that the newly selected transducer (13, 23, 24, 26, 28 or 30) will read the next address mark to occur on the track being read by the newly selected transducer (13, 23, 24, 26, 28 or 30) after said newly transducer (13, 23, 24, 26, 28 or 30) has been selected.

Description

SYSTEM FOR CORRECTING TRANSDUCER TANGENTIAL SKEW IN A DUAL ACTUATOR DISK DRIVE SYSTEM

A portion of the disclosure of this patent document contains material which is subject to copyright protection. The copyright owner has no objection to the facsimile reproduction by any one of the patent documents or the patent disclosure as it appears in the United States Patent & Trademark Office patent file or records, but otherwise reserves all copyright rights whatsoever.

CROSS-REFERENCE TO RELATED APPLICATIONS DISK DRIVE SYSTEM USING MULTIPLE EMBEDDED QUADRATURE SERVO FIELDS, Serial No. 386,504, filed July 27, 1989, assigned to the assignee of the present application;

ADAPTIVE READ EQUALIZER WITH SHUTOFF MODE FOR DISK DRIVES, Serial No. 559,899, filed July 30, 1990, assigned to the assignee of the present application;

DISK DRIVE SYSTEM EMPLOYING ADAPTIVE READ/WRITE CHANNEL CONTROLS AND METHOD OF USING SAME, Serial No. 420,371, filed October 12, 1989, assigned to the assignee of the present application;

HIGH PERFORMANCE DISK DRIVE ARCHITECTURE, Serial No. 07,612,427, filed November 9, 1990, assigned to the assignee of the present application; and

MULTIPLE ACTUATOR DISK DRIVE, Serial No. 07,614,853, filed November 9, 1990, assigned to the assignee of the present application.

Each of these Related Applications is hereby incorporated by reference. BACKGROUND OF THE INVENTION

Field of the Invention

The present invention is generally related to hard disk drives and disk drive electronic control systems. In particular, the present invention relates to a system for tangential skew correction of the transducers for reading and writing data from the disk in a dual actuator disk drive system.

Description of the Related Art

A dual actuator disk drive system is shown in Figure 1 and is comprised of disk assembly 7 and actuators 8 and 9. Transducer 14 on actuator 8 and transducer 13 on actuator 9 read and write information to and from the disk on disk assembly 7. Further it is desirable for each of the transducers 13 and 14 to read data regardless which of the transducers recorded the data.

Figure 2 shows in more detail the dual actuator disk drive system of Figure 1. Actuator 8 has a head arm assembly 15 where each arm of the head arm assembly 15 carries a transducer to communicate with one of the surfaces of a disk in disk assembly 7. The head arm assembly 15 is connected to a motor 11 for placing a given transducer over a desired track on one of the disk surfaces of disk assembly 7. Disk assembly 7 is comprised of a plurality of disks 10, 16 and 18, shaft 19 and a motor 20. Actuator 8 employs transducers 14 and 22 to interact with top and bottom surfaces of disk 10 respectively, transducers 24 and 26 to interact with the top and bottom surfaces of disk 16 respectively and transducers 28 and 30 to interact with the top and bottom of disk 18 respectively. Actuator 9 is comprised of a motor 12, a head arm assembly 17 and transducers 13, 23, 25, 27, 29 and 31. In a dual actuator disk drive system, one of the actuators is used as a packwriter to record on each disk surface system operational information such as servo data, sector marks, address marks and like information. Actuator 8 is used as the packwriter to record the system information on the surfaces of disks 10, 16 and 18. Thus there is no radial or tangential skew problem associated with transducers 14, 22, 24, 26, 28 and 30 because their physical location with respect to each other and with respect to the disk surface remains a constant.

The spatial relationship between transducers 14, 22, 24, 26, 28 and 30 of actuator 8 will be different than the spatial relationships between actuators 13, 23, 25, 27, 29 and 31 of actuator 9. There is no physical means to align the transducers on the head arm assembly of actuator 9 to accurately duplicate the spatial relationship of the transducers on actuator 8. The radial and tangential skew between the transducers on the head arm assemble of the nonpackwriter actuator does cause a delay in obtaining proper transducer placement over a desired track during a transducer switching operation between transducers associated with the nonpackwriter actuator.

Assuming that transducer 13 is used as the reference transducer, it can be seen from Figure 3 that transducers 23, 25, 27, 29 and 31 can be radially offset either to the right or left of reference transducer 13. Such an offset to the right or left of transducers 23, 25, 27, 29 and 31 may cause these transducers to be located over different tracks than the reference transducer 13. Of course, a transducer may in fact be in alignment with the reference transducer 13 as shown with respect to transducer 31. Therefore, transducer 31 will be over the same track as transducer 13 which still may not be the correct track because the spatial relationship between transducers 13 and 31 may be different than the spatial relationship between packwriter transducers 14 and 30. Thus a track correction may have to be made for transducers 23, 25, 27, 29 and 31 to place the transducer over the correct track before data may be read or written properly.

Figure 4 uses transducer 13 as a reference and shows that transducers 23, 25, 27, 29 and 31 may be tangentially offset to the front or to the rear of reference transducer 13. Where the transducer is offset to the front of reference transducer 13, the address mark will appear earlier than the address mark at the reference transducer 13. Where the transducer is set to the rear of reference transducer 13, the address mark will occur later than the address mark at the reference transducer 13.

In the disk drive system using a sectored embedded servo, control circuitry is synchronized to data read from the desired track on the disk surface. The system expects the occurrence of servo data, sector marks and address marks to occur at specific times and sets timers to trigger circuitry to search for the occurrence of servo data, sector marks and address marks.

If reference transducer 13 is being used and the system switches to transducer 23, the system would still expect the address mark to occur at the time that the address mark would have occurred for transducer 13. Therefore the start of the address mark search by the system can begin after the address mark has passed beneath transducer 23 since transducer 23 is physical ahead of transducer 13. The address mark will therefore be missed and the system would have to go through the entire synchronization procedure before the system could ascertain the track location of transducer 23. SUMMARY OF THE INVENTION

It is therefore an object of the present invention to establish a system to compensate for the tangential offset of the transducers on the nonpackwriting actuator such that the next address mark appearing on the data track after a transducer switching operation will be properly read by the newly selected transducer, thereby increasing the overall efficiency of the disk drive system.

The present invention develops a reference table of the time offset between designated transducer and all of the other transducers of the nonpackwriting actuator in the dual drive actuator disk drive system. The table is stored in memory and is used in each transducer switching operation of the nonpackwriting actuator. During a transducer switching operation, means calculate the correction factor for the newly selected transducer from the table stored in memory and adjust the time when the start of the next index mark search should occur, as established by the prior transducer, such that the time for the start of the address mark search will be that time for properly starting the address mark search for the newly selected transducer. This system will cause the newly selected transducer to sense the first address mark to occur after the transducer switching operation. As a result of obtaining the address mark information, the transducer can then be placed over the correct track such that a data transfer operation may proceed.

BRIEF DESCRIPTION OF THE DRAWINGS The invention will be described with respect to the particular embodiments thereof and reference will be made to the drawings, in which: FIGURE 1 is a high level diagram of a dual actuator disk drive system;

FIGURE 2 is a diagram showing the relationships of the various transducers associated with each of the actuators with a disk surfaces of the disk drive system;

FIGURE 3 is a diagram of the skew, along the X axis, to the left or right of a reference transducer of the other transducers associated with the nonpackwriting actuator in a dual actuator disk drive system;

FIGURE 4 is a diagram showing the skew, along the Z axis, ahead or to the rear of a reference transducer of the other transducers associated with the nonpackwriting actuator of a dual actuator disk drive system;

FIGURE 5 is a logic diagram for tangential skew correction during a transducer switching operation for the nonpackwriting actuator in a dual actuator disk drive system; and

FIGURE 6 is a table describing the tangential skew correction factors stored in memory to be used in conjunction with the logic of Figure 5.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

Referring to Figure 5, the logic associated with establishing the timing for sensing an address mark in a dual actuator disk drive system is shown.

A disk drive system normally has associated with it a central processing unit (CPU) 35 to control the operation of the disk drive system. CPU 35 will have a processor 40 for controlling such operations as the track seeking, transducer switching, track following, data reading and writing, and transferring data to and from the disk drive system to another processor or memory. CPU 35 has an associated memory unit 42 to store the various control program to be used by processor 40 in controlling the operation of the disk drive system. The system employs self-synchronizing techniques and uses encoded data in specific areas on the disk to obtain data relevant to system control functions and procedures. Once such type of data is an address mark which is used to determine the identity of the data being read by a transducer in the system. To locate the address mark, processor 40 will establish the time for the start of the search for an address mark in terms of a count value of counter 46. Counter 46 is stepped by clock pluses appearing on line 43. Comparator 48 compares the count value of counter 46 on line 47 and the count value for the for initiating an address mark search appearing on line 44. When a comparison occurs, an address mark search signal is generated on line 51 to signal the start of an address mark search by the system. The system (by logic not herein shown) will open a window during which time the data being read by the selected transducer is expected to be an address mark.

This invention incorporates into the foregoing, subtractor 52 to subtract the tangential skew correction factor for the presently selected transducer on line 38 from the tangential skew correction factor for the transducer to be selected on line 39. The results of the subtraction is provided on line 45 as the transducer switching correction factor to be used in the transducer switching procedure. Adder 49 adds the transducer switching correction factor on line 45 to the anticipated start address mark search count on line 44 to generate a new start address mark search count on line 50 to be used by comparator 48. It should be noted that the transducer switching correction factor is only used to search for the first address mark after a transducer switching operation. After the first address mark has been sensed, processor 40 will thereinafter generate the proper next start address mark search count in accordance with the normal address mark search procedure. More specifically, referring to Figure 6, a table is stored within memory of the tangential skew correction factors for transducer switching operations between reference transducer 13 and transducers 23, 25, 27, 29 and 31 of actuator 9 The usable recording surface of the disk is divided into four zones and a correction factor is established for each of the zones. It has been found that the arc path transversed by the transducers on the packwriting actuator is different from the arc path transversed by transducers of the non-packwriting actuator. As a result of the differences in the arc path followed by the transducer associated with both actuators for interacting with the same disc surface, the tangential skew correction factor is not a constant. To mitigate this problem, the tracks have been divided into four zones and the tangential skew correction factor is established in each zone with the reference transducer located over the center track of the zone. During a transducer switching operation, the processor 40 will read the correction factor for the present transducer selected and the correction factor for the transducer to be selected. The correction factor of the present selected transducer is subtracted from the value of the newly selected transducer and this value becomes the value of the correction factor which appears on line 45 of Figure 5 and sent to adder 49. This procedure allows the correction factor between any two transducers in any zone to be determined while only storing the correction factor for switching between the reference transducer and each of the other transducers for each of the zones. The amount of data to be stored in support of this invention is thereby minimized.

For example, if reference transducer 13 is selected and it is desired to switch to transducer 27 and the reference transducer 13 is presently within zone 2, then correction factor C23 will be used. Since the reference transducer is the transducer to be switched from, the correction value for that transducer is deemed to be zero and, therefore, the effective correction factor for that transducer switching operation is the value C23 as shown in Figure 6.

Next assume that transducer 31 is selected and it is desired perform a transducer switching operation to select transducer 23. Further assume that transducer 31 is presently within zone 3. The processor 40 recovers from memory 42 the value of C35 for transducer 31 and C32 for transducer 23. The transducer switching correction factor for this transducer switching operation will be C35 minus C31. It should further be realized that the correction factors can be of a positive or negative number as they are adjusting the count in a forward or backward direction. By viewing Figure 4, transducer 31, being to the rear of the reference transducer 13, would have a positive correction factor which would cause the start address mark search signal on line 51 to occur at a later time than had reference actuator 13 been used. In similar fashion, actuator 23 is ahead of reference actuator 13 and its correction factor would be a negative correction factor which will cause the start address mark search signal on line 51 to occur at an earlier time than the address mark signal which would have been generated for reference transducer 13. The correction factor for the transducer switching operation from transducer 31 to transducer 23 would mean that transducer 23 is in front of transducer 31 and, therefore, a negative value should result to cause the start of an address mark search to occur at an earlier time than would have been predicted for the start of an address mark search when using transducer 31. Therefore the subtracting from the newly assigned transducer 23, the value of the present selected transducer 31 will yield a negative number which will cause the start of the address mark search for transducer 23 to occur at an earlier time than the start of the address mark search that would have occurred for transducer 31.

Once the address mark has been properly read by transducer 23 then a correction can be made to bring transducer 23 to the desired track. Since the address mark identifies the track presently being read, the system can adjust transducer 23 to the correct track such that a data operations may be initiated.

The correction factors as shown in Figure 6 are obtained by means of the circuitry shown in Figure 5. During the manufacture of the dual actuator disk drive system, a microprogram is run to determine the values of the correction factors. Since there is no simple way, without additional hardware, of measuring the time difference directly within the system, a trial and error system is employed. The trial and error system will be described for a transducer switching operation from transducer 13 to transducer 23 in zone 1 for exemplary purposes. The same procedure is performed for the transducer switching operation for each transduce with regard to the reference transducer for each zone on the disk. The procedure calls for a transducer switching operation to be performed between transducer 13 and 23 and the procedure notes if an address mark was sensed by transducer 23 at the expected time. If no address mark was sensed during the address window, then a correction factor is added in a forward direction and the switch operation is again attempted. The method will alternate by adding a forward or backward correction factor to the initial setting, increasing the value of the correction factor each time used, until an address mark is first obtained. The system then incrementally increases the correction factor in the direction in which the address mark was first sensed until the address mark is lost. This indicates a maximum correction factor. The correction factor is then incrementally decreased during transducer switching operations until the address mark is again lost and this determines the minimum correction factor for the transducer switching operation. The maximum value and minimum value are averaged together such that the correction factor would be in the middle of the maximum and minimum correction values. This value is then stored as the tangential skew correction factor C11 in the table in memory. It should be noted that the process of determining the minimum and maximum values are repeated a number of times to ensure the correctness of the minimum and maximum values .

In the preferred embodiment, the method for generating the tangential skew correction factors is performed by a microprogram stored within memory 42. Further, the method for generating the new address mark search count is performed by a microprogram stored within memory 42 to be used by processor 40 during a transducer switching operation. The invention as described heretofore uses known digital software implementation and the microprogram was assembled using Motorola 68C11 Assembler, series 5.0. The microprogram for carrying out the method for establishing the tangential skew correction factors as shown in Figure 6 is provided in Appendix A. The microcode for establishing the new corrected address mark search count appearing on line 50 of comparator 48 is provided in Appendix B. The programs of Appendix A and Appendix B carry out the functions as described within the specification. It should be understood that the method can be embodied in other microprograms using other programming languages and assemblers. The microprogram may be permanently stored in a read only memory such that the microprograms become a permanent part of the overall operating system. It is also possible for the transducer switching system to be embodied in hardware logic using the correction factor stored in memory and a state machine to provide the necessary timing. While the invention has been particularly shown and described with reference to the preferred embodiments therefore, it will be understood by those skilled in the art that changes in form and detail may be made therein without departing from the spirit and scope of the invention. Given, the above disclosure of general concepts and specific embodiments, the scope of the protection sought is defined by the following.

APPENDIX A

Conner Peripherals , Inc . 1992

5459

5460

5461 ; =================================================

5462 ; Head switch timing offsets test

5463 ;

5464 ; Entry: Head iin B. >max = calibrate all heads

5465 ; we're already at the desired cylinder

5466 ; =================================================

5467 961E 17 HSWTST: TBA ;end head = start head

5468 961F C1 09 CHPB #M4 XHD ;check the head

5469 9621 2303 BLS HSWTSO ;if (head is invalid) then

5470 9623 5F CLRB ; start at head 0

5471 9624 8609 LDAA #MAXHD ; end at the last one

5472 9626 HSWTSO: PSHD ;save the head limits

5473 9626 37 PSHB

5474 9627 36 PSHA

5475 9628 EWDH

5476 9628 CC 96 B7 LDD #MSD.HS1 ;print the

5477 9628 BD CD 72 JSR HSG.WT ;header.

5478 962E CC 0000 LCD #REFHEAD ;set the test head

5479 9631 B71091 STAA REF.HD ;to the reference head and

5480 9634 BD 8855 JSR SETH ;give it an offset of zero.

5481 9637 PULD ;restore the head limits

5482 9637 32 PULA

5483 9638 33 PULB

5484 9639 EHDM

5485 HSWTS6: ;repeat

5486 9639 C1 00 CHPB #REFHEAD ; compare current head to reference head

5487 963B 2768 BEQ HSWTS8 ; if (this is not the reference head) then

5488 9630 PSHD ; save the heads

5489 9630 37 PSHB

5490 963E 36 PSHA

5491 963F EHDH

5492 963F BD 96 C4 JSR HSTEST ; run the test.

5493 9642 275F BEQ HSWTS7 ; if (it failed) then

5494 9644 FE 1007 HSWTSZ: IDX RWS.LNK ; repeat

5495 9647 8C 90 CC CPX #RS.IDLE ; check for idle

5496 964A 26 F8 BNE HSWTSZ ; until (r/w idle)

5497 964C CC 96 AC LDD #MSD.HSO ; print the retrying

5498 964F BD CD 72 JSR MSG.WT ; message

5499 9652 PULD ; retry

5500 9652 32 PULA

5501 9653 33 PULB

5502 9654 ENDM

5503 9654 PSHD ; this

5504 9654 37 PSHB

5505 9655 36 PSHA

5506 9656 ENOM

5507 9656 BD 96 C4 JSR HSTEST ; head.

5508 9659 2748 BEQ HSWTS7 ; if (it failed again) then

5509 9658 FE 1007 HSWTSY: LDX RWS.LNK ; repeat

5510 965E 8C 90 CC CPX #RS.10LE ; check for idle

5511 9661 26 F8 BNE HSWTSY ; until (r/w idle)

5512 9663 PULD ; head

5513 9663 32 PULA

5514 9664 33 PULB

5515 9665 ENDM

5516 9665 PSHD ; to B

5517 9665 37 PSHB

5518 9666 36 PSHA

5519 9667 ENDM

5520 9667 5A DECB ; one less

5521 9668 2A 02 BPL HSWTSA ; if (it underflowed) then

5522 966A C609 LDAB #MAXHD ; wrap it around

5523 966C F71091 HSWTSA: STAB REF.HD ; make it the reference head

5524 966F CC 96 AC LDD #MSD.HSO ; print the retrying

5525 9672 BD CD 72 JSR MSG.WT ; message

5526 9675 PULD play

5527 9675 32 PULA

5528 9676 33 PULB

5529 9677 ENOM

5530 9677 PSHD ; it again.

5531 9677 37 PSHB

5532 9678 36 PSHA

5533 9679 ENDM 5534 9679 BD 96 C4 JSR HSTEST ; Sam.

5535 967C PSHD , save the result

5536 967C 37 PSHB

5537 967D 36 PSHA

5538 967E EMDM

5539 967E 8600 LDAA #REFNEAD ; restore the

5540 9680 B7 10 91 STAA REF.HD reference head

5541 9683 PULD ; restore the result

5542 9683 32 PULA

5543 9684 33 PULB

5544 9685 ENDM

5545 9685 40 TSTA ; check it

5546 9686 27 1B BEQ HSWTS7 ; if (it failed) then

5547 9688 FE 10 07 HSWTSX LDX RWS.LNK ; repeat

5548 968B 8C 9D CC CPX ITRS.IDLE ; check for Idle

5549 968E 26 F8 BNE HSWTSX ; until (r/w idle)

5550 9690 BD CD 72 JSR HSG.WT ; print the failure message

5551 9693 CE 04 00 LDX #ACT.HLT*256 ; set the

5552 9696 8600 LDAA #STHLTH ; health

5553 9698 SMDCMD ; bit.

5554 9698 12 A0 40 FC A?: BRSET CSTATS,CMD.OUT,A? ;wait for output empty

5555 969C DF A2 STX COATAH

5556 969E 97 A1 STAA CCMD

5557 96A0 CE 10 00 LDX #RAMADR

5558 96A3 ENDH

5559 96A3 HSWTS7: PULD ; restore the heads

5560 96A3 32 PULA

5561 96A4 33 PULB

5562 96A5 EHDH

5563 96A5 5C KSUTS8: IUCB ; next head please

5564 96A6 11 CBA ; compare limit to current head

5565 96A7 24 90 BHS HSWTS6 ;until (limit head < current head)

5566 96A9 7E 93 D1 HSWTS9: JMP TEND ;get outta Dodge

5567

5568 96AC 72 65 74 72 79 MSD.HS0 DB 'retrying',CR,LF,0

5569 96B7 00 0A 43 79 6C MSD.HS1 DB CR,LF,'Cyl ',6,<CMD.CYL,2,<CMD.CYL+1,CR,LF,0

5570

5571 ; ===========================================

5572 ; Head switch timing offsets adaptive test

5573 ;

5574 ; Entry : head n B

5575 ; we're already at the desired cylinder

5576 ;

5577 ; Exit: pointer to error string in 0, 0 s no error, 2 is valid

5578 ; if no error: the optimal value is chosen and set in Niwot.

5579 ; including EEROM

5580 ; if error: head is set back to value it was before test

5581 ;

5582 ; ===========================================

5583 96C4 F7 10 17 ISTEST: STAB CMD.MD ;save the head

5584 96C7 37 PSMB ;stack it, too

5585 96C8 BO 8879 JSR GETH ;get this head's current offset

5586 96CB 37 PSHB ;stack it with the head

5587 96CC CC 98 00 LDD #MSD.HS2 ;prιnt which head

5S88 96CF BD CD 72 JSR MSG.WT ;we're doing

5589 9602 32 PULA ;restore the current offset

5590 9603 36 PSHA ;save it again

5591 9604 CE 807F LDX #807FH ;stack initial max and

5592 9607 3C PSHX ;mιn values

5593 ;

5594 ; loop out from the current offset, looking for a range of values

5595 ; that sort of work

5596 ;

5597 9608 F610 17 LDAB CMD.MD ;head back to B

5598 H STES1: ;repeat

5599 96DB CE 00 14 LDX #20 ; 20 seeks max

5600 96DE BD 98 11 JSR HSWTIM ; run the timing test

5601 96E1 27 0B BED HSTESB ; if (we couldn't do 20 seeks) then

5602 96E3 30 TSX ; subtract the original

5603 96E4 A0 02 SUBA 2,X ; offset from this one

5604 96E6 2E 01 BGT HSTESA ; if (we're 0 or negative from the

5605 96E8 4A PECA ; reduce it by one

5606 96E9 40 X STESA: NEGA ; now switch directions

5607 96EA AB 02 ADDA 2,x ; and add back the original

5608 96EC 20 1E BRA HSTESX ; else if (we did all 20 seeks) then 5609 96EE 36 HSTESB: PSHA ; try

5610 96EF 80 02 SUBA #2 ; again

5611 96F1 CE 00 14 LDX #20 ; at

5612 96F4 BD 98 11 JSR HSWTIH ; two

5613 96F7 32 PULA ; less

56K 96F8 27 04 BEQ HSTESC ; if (we couldn't do 20 seeks) then

5615 96FA 88 02 ADDA #2 ; move away from the failure

5616 96FC 20 0E BRA HSTESX ; else if (we did all 20 seeks) then

5617 96FE 36 HSTESC: PSHA ; try

5618 96FF 88 02 ADDA #2 ; again

5619 9701 CE 00 14 LDX #20 ; at

5620 9704 BD 98 11 JSR HSWTIH ; two

5621 9707 32 PULA ; more

5622 9708 27 02 BEQ HSTESX ; if (we couldn't do 20 seeks) then

5623 970A 80 02 SUBA #2 ; move away from the fai lure

5624 970C 8C 00 00 HSTESX: CPX #0 ; check the result

5625 970F 27 0E BEQ HSTES2 ;until (we find a range that works)

5626 9711 81 78 CHPA #120 ;

5627 9713 2C 04 BGE HSTESD ;or (we've reached the upper limit)

5628 9715 81 88 CMPA #-120 ;

5629 9717 2E C2 BGT HSTES1 ;or (we've reached the lower limit)

5630 ;if (we went too far) then

5631 9719 CC 97 B4 HSTESD: LDD #MSD.HSZ ; point D to failure message

5632 971C 7E 97 A5 JMP HSTES8 ;else if (we found a workable value) then

5633 ;

563* ; Continυe working down from 1

5635 ;

5636 HSTES2: ; repeat

5637 971F CE 00 0A LDX #10 ; 10 seeks max

5638 9722 BO 98 11 JSR HSWTIH ; run the timing test

5639 9725 2706 BEQ HSTESS ; if (we apparently reached a limit) then

56-0 9727 CE 00 0A LDX #10 ; try again

5641 972A BD 9811 JSR HSWTIH ; to be sure

5642 9720 22 0A HSTESS: BHI HSTES4 ; until (we find a non-usable value)

5643 972F 4A DECA ;

5644 9730 81 88 CMPA #-120 ;

5645 9732 2C EB BGE HSTES2 ; or (we go too far)

5646 ; if (we went too far) then

5647 9734 CC 97 C7 LDD #MSD.HSL ; point 0 to failure message

5648 9737 206C BRA HSTES8 ; else if (we found a non-working value) then

5649 9739 4C HSTES4: INCA ; last offset that worked to A

5650 973A PSHD ; stack the unadjusted offset - head

5651 973A 37 PSHB

5652 973B 36 PSHA

5653 973C ENDM

5654 973C F6 1091 LOAB REF.HD ; reference head

5655 973F BO 8879 JSR GETH ; offset to B.

5656 9742 30 TSX ; adjust our lower limit

5657 9743 EB 00 ADDS 0,X ; by this much

5658 9745 E703 STAB 3,X ; save it

5659 9747 F710 A7 STAB PARM1+1 ; print

5660 974A CC 9806 LDD #HS0.HS4 ; the adjusted

5661 9740 BD CD 72 JSR HSG.UT ; result

5662 9750 PULD ; restore the unadjusted offset & head

5663 9750 32 PULA

5664 9751 33 PULB

5665 9752 ENDM

5666 ;

5667 ; Now work up from here +5 to the upper limit ...

5668 ;

5669 9752 88 05 ADDA #5 ; go up 5 for hysterisis

5670 HSTES5: ; repeat

5671 9754 CE 00 0A LDX #10 ; 10 seeks max

5672 9757 BD 98 11 JSR HSWTIH ; run the timing test

5673 975A 2706 BEQ HSTESZ ; if (we apparently reached a limit) then

5674 975C CE 00 OA LDX #10 ; try again

5675 975F BD 98 11 JSR HSWTIH ; to be sure

5676 9762 22 0A HSTESZ: BHI HSTES6 ; until (we find a non-working value)

5677 9764 4C INCA ;

5678 9765 81 78 CHPA #120 ;

5679 9767 2F EB BLE HSTES5 ; or (we go too far)

5680 ; if (we went too far) then

5681 9769 CC 97 D8 LDD #MSD.HSW ; point 0 to failure message

5682 976C 20 37 BRA HSTES8 ; else if (we found the upper limit) then

5683 976E 4A HSTES6 DECA ; back up one to the last offset that worked 5684 976F PSHD ; save the unadjusted offset and head

5685 976F 37 PSHB

5686 9770 36 PSHA

5687 9771 ENDM

5688 9771 F6 10 91 LDAB REF.HD ; reference head

5689 9774 BD 8879 JSR GETH ; offset to B.

5690 9777 30 TSX ; adjust our lower limit

5691 9778 E8 00 ADDB 0,x ; by this much

5692 977A E702 STAB 2,X ; save it

5693 977C F7 10 A7 STAB PARM1+1 ; print

5694 977F CC 98 06 LOO #MSD.HS4 ; the

5695 9782 BD CD 72 JSR MSG.WT ; result

5696 9785 PULD ; restore the head

5697 9785 32 PULA

5698 9786 33 PULB

5699 9787 ENDM

5700 ;

5701 ; calculate the midway point, set the head to this value

5702 ;

5703 9787 30 TSX ; upper limi t

5704 9788 A6 00 LDAA 0,X ; to A

5705 978A A0 01 SUBA 1 ,X ; subtract the lower l imi t

5706 978C 81 06 CHPA #6 ; check for at least a range of 6

5707 978E 2C 05 BGE HSTES7 ; if (range < 6) then

5708 9790 CC 97 EF LDD -ΗSD.HSR ; point D to failure message

5709 9793 20 10 BRA ; else if (range >= 6) then

5710 9795 44 HSTES7: LSRA ; divide by 2

5711 9796 AB 01 ADDA 1,X ; now add back to lower limit

5712 9798 B7 10 A7 STAA PARHH1 ; save the result

5713 979B BD 8855 JSR SETH ; set the head to the result

5714 979E CC 98 OC LDD #HSD.HS5 ; print

5715 97A1 BD CD 72 JSR ; the result

5716 97A4 4F CLRA ; no errors

5717 97A5 38 HSTES8: PULX ;clean off the stack

5718 97A6 38 PULX ;initial offset & head to B

5719 97A7 40 TSTA ;check the result

5720 97A8 27 09 BEQ ;if (we failed) then

5721 97AA PSHD ; save error pointer

5722 97AA 37 PSHB

5723 97AB 36 PSHA

5724 97AC ENDH

5725 97AC 8F XGOX ; set the head back

5726 97AD BD 88 55 JSR ;to it's original offset

5727 97B0 PULD ;restore the error pointer

5728 97B0 32 PULA

5729 97B1 33 PULB

5730 97B2 ENOM

5731 97B2 40 TSTA ; validate Z

5732 97B3 39 HSTES9: RTS

5733

5734 97B4 46 61 696C 65 MSD.HSZ DB 'Failed - No zero',CR,LF,0

5735 97C7 46 61 696C 65 HSO.HSL DB 'Failed - No lower',CR,LF,0

5736 97DB 46 61 696C 65 HSD.HSU DB 'Failed - No upper',CR,LF,0

5737 97EF 46 61 696C 65 HSO.HSR DB 'Failed - range',CR,LF,0

5738

5739

5740

5741 9800 48 06 173A 20 MSD.HS2 DB 'H',6,<CMD,HD, ': ,

742 9806 02 A7 20 20 20 HSD.HS4 DB '0

5 2,<PARH1+1,' ',0

5743 980C 02 A700 0A 00 HSD.HS5 DB 2,<PARM1+1,CR,LF,0

5744

5745 ;

5746 ; HSWTIM ... head switch timing test

5747 ;

5748 ; Entry: offset to use in A

5749 ; head to test in B

5750 ; max # of seeks to do in X

5751 ;

5752 ; Exit: A and B are preserved

5753 ; X contains nuiύer of seeks not performed due to excessive errors 5754 ; Z reflects X's value

5755 ;

5756 9811 37 HSWTIM: PSHB ;save the head

5757 9812 36 PSHA ;and the offset

5758 9813 3C PSHX ;and the seek limit 5759 9814 BD 88 55 JSR SETH ;set this head to this offset

5760 HSWTIO ;repeat

5761 9817 30 TSX ; test head

5762 9818 E603 LDAB 3,X ; to B

5763 981A F7 10 17 STAB CMD . HD ; point to it

5764 9810 BD 9840 JSR SEK.CM ; seek there

5765 9820 36 PSHA ; save the error

5766 9821 F6 1091 HSWTIZ LDAB REF.HD ; seek to

5767 9824 F7 10 17 STAB CMD.HD ; the reference

5768 9827 BD 9840 JSR SEK.CM ; head

5769 982A 30 TSX ; add

5770 982B AA 00 ORAA 0,X ; the

5771 9820 A7 00 STAA 0,X ; errors.

5772 982F 32 PULA ; get the errors

5773 9830 81 01 HSWTI1: CHPA #1 ; check it

5774 9832 22 05 BHI HSUTI2 ; until (we get > 1 error)

5775 9834 38 PULX ;

5776 9835 09 DEX ;

5777 9836 3C PSHX ;

5778 9837 26 DE BNE HSWTI0 ;or (we've done all the seeks)

5779 9839 38 HSWTI2. PULX ;restore the seek counter

5780 983A 32 PULA ;restore the offset

5781 983B 33 PULB ;restore the head

5782 983C 8C 00 00 CPX #0 ;validate Z

5783 983F 39 RTS

5784

5785

5786

5787 9840 CE 0064 SEK.CM: LDX #100 ;loop count to X

5788 9843 18 CE 90 CC SEKCMO: LOY #RS.IDLE ;required r/w state to Y

5789 9847 09 DEX ;decrement loop count

5790 9848 2776 BEQ SEKCH9 ;while (not timed out)

5791 984A 12 2B 04 0A BRSET ER.FLAG.SPN.ER,SEKCMA ;and (spin error)

5792 ; or

5793 984E 13 81 02 06 BRCLR RU.CTL.ALOURT,SEKCMA ; (we're not allowing wri

5794 9852 18 BC 1007 CPY RUS.LNK ; or

5795 9856 27 0C BEQ SEKCH1 ; (r/w is not ready) then

5796 9858 3C SEKCMA: PSHX ; save the loop count

5797 9859 CC 00 64 LDD #100 ; wait

5798 985C BD CA F5 JSR DELAY ; 100 mS

5799 985F 38 PULX ; restore the loop count

5800 9860 86 10 LDAA #10H ; busy error to A

5801 9862 20 DF BRA SEKCHO ;end do

5802

5803 9864 CE 10 00 SEKCM1 : LDX #RAMADR

5804 9867 EC 17 LCD <CMO.KD,X ;get orig par-ins back

5805 9869 ED 42 STD <TGT.HD,X ;save to target

5806 9868 5F CLRB

5807 986C E7 01 STAB <ERROR,X

5808

5809 986E CE FF FF LDX #-1

5810 9871 09 SEKCM2: DEX

5811 9872 27 09 BEQ SEKCM3 ;b if timeout-force seek

5812 9874 FC 10 05 LDD ACT.LNK

5813 9877 1A 83 9C 4E CPO #TRAK10 ; track fol lowing?

5814 987B 26 F4 BNE SEKCM2 ; B if no

5815

5816 9870 14 A0 01 SEKCM3: BSET CSTATS.SK.MODE ;set seek mode to host

5817 9880 7F 10 71 CLR SK.TIME

5818 9883 FC 10 15 LDD CHO.CYL

5819 9886 FO 10 40 STD TGT.CYL ;xfr new cyl addr

5820 9889 CE 10 00 LDX #RAMADR

5821 988C CC B5 E6 LDD #SEEK20

5822 988F ID 98 10 BCLR <R.FLAG,X,DID.SEK ;reset seek retry

5823 9892 FD 10 05 STD ACT.LNK ;issue seek and

5824 9895 14 2B 01 BSET ER.FLAG.SK.MODE

5825

5826 9898 4F CLRA ;assume no errors

5827 SEKCH4: ;repeat

5828 9899 13 2B 04 08 BRCLR ER.FLAG,SPN.ER,SEKCM6 ; if (spin error) then

5829 9890 4C INCA ; count it

5830 989E 26 01 BNE SEKCM5 ; if (overflow) then

5831 98A0 4A DECA ; uncount it

5832 98A1 12 2B 04 FC SEKCM5: BRSET ER.FLAG.SPN.ER,* ; wait for it to go away

5833 98A5 1228 01 FO SEKCM6: BRSET ER.FLAG,SK.MODE,SEKCM4 ;until (seek complete) 5834 98A9 FE 10 05 LDX ACT.LNK ;

5835 98AC 8C 9C 4E CPX #TRAK10 ;

5836 98AF 26 E8 BNE SEKCM4 ; and (we're track following)

5837 98B1 1381 02 E4 BRCLR RU.CTL.ALOURT,SEKCM4 ;and (we're allowing write)

5838 98B5 CE 10 00 LDX #RAMADR ;point to ram base

5839 98B8 IF 9B 10 04 BRCLR <R.FLAG,X,DID.SEK.SEKCM9 ;if (we had to retry the seek) then

5840 98BC 4C INCA ; count as an error

5841 98BD 26 01 BNE SEKCM9 ; if (overflow) then

5842 988F 4A OECA ; uncount it

5843 98C0 40 SEKCH9: TSTA

5844 98C1 39 RTS

5845

APPENDIX B

Conner Peripherals , Inc . 1992

10968 ; =========================================================

10969 ; Actuator state 2 - initiate seek to TGT.CYL, TGT.HD

10970 ; =========================================================

10971 ; Note - as of 1-8-91 state 2 has 80 cycles of margin

10972 ;Note - changed health bit setting in state 9. smk 1/16

10973 ;Changes to states 5,6,7,8 to improve access time JS 4-13

10974 ;

10975 B5E5 3B SEEK2E: RTI

10976 B5E6 12 2B 04 FB SEEK20: BRSET ER.FLAG,SPN.ER,SEEK2E ;B spin error

10977 B5EA 14 2B 01 BSET ER.FLAG.SK.MCOE ;Set seek mode

10978 BSED 86 10 LDAA #10H

10979 85EF B7 10 75 STAA PK.DET ;Seek mode peak detector value 10980

10981 ;Select target head value in PORTD and adjust expected sector time 10982 [01] IFTRUE TWOACT

10983 B5F2 FE 10 8F LDX HSWPTR

10984 [01] ELSE

10985 LDX #HSWTCH

10986 [00] ENOIF

10987 B5F5 F6 10 34 LDAB HEAD ;get current hd #

10988 B5F8 C4 OF ANDB #HD.0+HD.1+H0.2+HD.3 ; remove flag bits

10989 B5FA 3A ABX

10990 B5FB A6 00 LDAA 0,X ;0ld head offset

10991 101] IFTRUE TUOACT

10992 B5FD FE 10 8F LDX HSWPTR

10993 [01] ELSE

10994 LDX # HSWTCH

10995 [00] ENOIF

10996 B600 F6 10 42 LDAB TGT.HD ;New head address

10997 B603 C4 OF ANDB #HD.0+HD.1+HD.2+HD.3

10998 B605 F7 10 34 STAB HEAD ;Save new head number 10999 B608 07 08 STAB PORTD ;Head select lines 11000 S60A 3A ABX

11001 B60B A0 00 SUBA 0,X ;Calc (old-new) offset 11002 B600 40 NEGA

11003 B60E 16 TAB ;sign extend...

11004 860F 48 LSLA

11005 B610 86 00 LDAA #0 ;CLRA clears carry 11006 B612 82 00 SBCA #0

11007 B614 03 1A ADDD TOC3

11008 8616 DO 1A STD TOC3 ;Adjust expected sector timing 11009 B618 C3 00 6E ADDD #PWR.TM

11010 B61B DO 16 STD TOC1 ;Adjust reference time 11011

11012 [01] IFTRUE TWOACT

11013 [02] IFTRUE ACTB

11014 B61D CC B6 24 LDD #SK202

11015 B620 F0 10 05 STD ACT.LNK

11016 B623 RTI

11017 B624 SK202:

11018 [01] ENDIF

11019 [00] ENDIF

11020

Claims

What is claimed is: 1. In a dual actuator disk drive system, each said actuator having a plurality of transducers, said system employing a transducer switching procedure for switching between a first presently selected transducer and a second to be selected transducer, an apparatus, associated with the nonpackwriting actuator, for controlling the sensing of an address mark by said second transducer, said apparatus comprising:

address mark means for generating an address mark search signal to start an address mark search when an address mark is expected to be read by said first transducers; and

adjusting means connected to said address mark means for altering the time when said address mark means will generate said address mark search signal such that said second transducer will read the next address mark to occur on the track being read by said second transducer after said second transducer is selected.

2. The apparatus of Claim 1 wherein said adjusting means further comprises;

storage means for storing a tangential skew correction factor for each said transducer associated with said nonpackwriting actuator, said tangential skew correction factor correcting for the tangential skew between one of said transducers established as a reference transducer and each of said remaining transducers; calculating means connected to said storage means for generating a transducer switching correction factor from said first transducer's tangential skew correction factor and said second transducer's tangential skew correction factor; and

correction means connected to said calculating means and said address mark means to adjust, by said transducer correction factor, the time when said address mark means generates the next said address mark search signal.

3. The apparatus of Claim 2 wherein said calculating means further comprises;

subtracting means for generating said transducer switching correction factor by subtracting said first transducer's tangential skew correction factor from said second transducer's tangential skew correction factor.

4. The apparatus of Claim 1 wherein said dual actuator disk drive system has a system clock and said address mark means comprises;

counter means step by said system clock;

time means generating the value of the count of said counter when the next address mark search signal is to be generated;

comparator means connected to said counter means and said time means for generating said address mark search signal when said count generated by said time means equals the count of said counter.

5. The apparatus of Claim 4 wherein said adjusting means further comprises;

storage means for storing a tangential skew correction factor for each said transducer associated with said nonpackwriting actuator, said tangential skew correction factor correcting for the tangential skew between one of said transducers established as a reference transducer and each of said remaining transducers;

calculating means connected to said storage means for generating a transducer switching correction factor from said first transducer's tangential skew correction factor and said second transducer's tangential skew correction factor; and

correction means connected to said calculating means and said address mark means to adjust, by said transducer correction factor, said count value when said address mark means generates the next said address mark search signal.

6. The apparatus of Claim 5 wherein said calculating means further comprises;

7. The apparatus of Claim 5 wherein said calculating means further comprises;

adder means connected to said time means and said compare means for adding said count value of said time means to said transducer switching correction factor and inputting said resulting summation to said compare means in place of said count value generated by said time means.

8. The apparatus of Claim 7 wherein said calculating means further comprises;

adder means connected to said subtracting means, said time means and said compare means for adding said count value of said time means to said transducer switching correction factor of said subtracting means and inputting said resulting summation to said comparator means in place of said count value generated by said time means.

9. In a dual actuator disk drive system, each said actuator having a plurality of transducers, said system employing a transducer switching procedure for switching between a first presently selected transducer and a second to be selected transducer, a method, associated with the nonpackwriting actuator, for controlling the start of an address mark search to enable the next address mark after the completion of a transducer switching procedure to be read by said second transducer, said method comprising the steps of:

generating a time value for when an address mark search should be initiated to enable the next address mark to be read by said first transducers; and altering said time value for when an address mark search should be initiated such that said second transducer will read the next address mark to occur on the track being read by said second transducer after said second transducer is selected.

10. The method of Claim 9 wherein said altering step further comprises the steps of;

storing a tangential skew correction factor for each said transducer associated with said nonpackwriting actuator, said tangential skew correction factor correcting for the tangential skew between one of said transducers established as a reference transducer and each of said remaining transducers;

calculating a transducer switching correction factor from said first transducer's tangential skew correction factor and said second transducer's tangential skew correction factor; and

correcting said time value by said transducer switching correction factor.

11. The method of Claim 10 wherein said calculating step includes the step of;

subtracting said first transducer's tangential skew correction factor from said second transducer's tangential skew correction factor to generate said transducer switching correction factor.

12. The method of Claim 9 wherein said dual actuator disk drive system has a system clock and said generating step further comprises the steps of;

stepping a counter by said system clock;

establishing the count of said counter for said time value;

generating an address mark search signal when said time value count equals the count of said counter.

13. The method of Claim 12 wherein said altering step further comprises the steps of;

correcting said time value by said transducer switching correction factor.

14. The method of Claim 13 wherein said calculating step includes the step of;

15. The method of Claim 13 wherein said calculating step further comprises the step of;

adding said time value count to said transducer switching correction factor;

substituting said resulting summation for said time value count in said comparing step.

16. The method of Claim 15 wherein said calculating step further comprises the step of;

adding said time value count to said transducer switching correction factor;

17. The method for generating a tangential skew correction factor for each transducer associated with the nonpackwriting actuator in a dual actuator disk drive system, said method comprising the steps of:

a) selecting one of said transducers as a reference transducer;

b) determining the first maximum value that can be added to the address mark detection means within said system such that a second transducer cannot read the next address mark that occurs on the track being read by said second transduce after a transducer switching procedure is completed from said reference transduce to said second transducer;

c) determining the second maximum value that can be subtracted from the address mark detection means within said system such that a second transducer cannot read the next address mark that occurs on the track being read by said second transduce after a transducer switching procedure is completed from said reference transduce to said second transducer;

d) averaging said first maximum value and said second maximum value to form said tangential skew correction factor for said second transducer;

e) storing said tangential correction factor in memory;

f) repeating steps b,c,d and e for each transducer associated with said nonpackwriter actuator:

g) setting the value of tangential correction factor for the reference transducer to zero; and

h) storing said tangential correction factor for said reference transducer in memory.

18. The method of Claim 17 including the steps, of; i) dividing the data tracks on said disks within said system into zones; and

j) repeating steps a, b, c, d, e, f, g and h for each of said Zones.