The second RL11 controller

The second controller came with a label:

The drive does not show "ready" when connected to this controller. So let's go to work on it..

RLZID1 test failures

The next test is RLZI.

This test fails with:

CZRLI HRD ERR  00101 ON UNIT 00 TST 015 SUB 002 PC: 037702

DIFF OF 1 SEEK TEST
OPER: SEEK 
           FROM 002 DIFF 002 SGN 0 HD 0
RESULT:        DRV STATE IS 5 SB 4
BUS ADD=174400 DRV=0
           RLCS    RLDA    RLBA    RLMP    CYL   HD 
OP INIT = 000106  000401  000000  000000
OP DONE = 000304  000003  000000  000235   002  0

ERR HLT

A second try also failed:

CZRLI HRD ERR  00101 ON UNIT 00 TST 015 SUB 003 PC: 037702

DIFF OF 1 SEEK TEST
OPER: SEEK 
           FROM 000 DIFF 002 SGN 1 HD 1
RESULT:        DRV STATE IS 5 SB 4
BUS ADD=174400 DRV=0
           RLCS    RLDA    RLBA    RLMP    CYL   HD 
OP INIT = 000106  000425  000000  000000
OP DONE = 000304  000003  000000  000335   000  1

ERR HLT

And a third:

CZRLI HRD ERR  00101 ON UNIT 00 TST 015 SUB 002 PC: 037702

DIFF OF 1 SEEK TEST
OPER: SEEK 
           FROM 002 DIFF 002 SGN 0 HD 0
RESULT:        DRV STATE IS 5 SB 4
BUS ADD=174400 DRV=0
           RLCS    RLDA    RLBA    RLMP    CYL   HD 
OP INIT = 000106  000401  000000  000000
OP DONE = 000304  000003  000000  000235   002  0

ERR HLT

The description for test15 reads:

Test 15, part 1: read a header, seek with a difference of 1 on head 0, then get status and check the state is 4. Failing that points at the drive command shift register, a dropped bit in the difference register, or the state ROM.

Part 2 waits around 20ms and expects state 5, listing the difference register, the count logic, the seek ROM, the DC servo and the tacho as suspects. The whole sequence is then repeated with the opposite sign and on head 1.

For this test it is useful to look at the RL02 command word which has the following format:

The RL02 command word. Bits 7 to 15 carry the cylinder address difference, bit 4 selects the head, bit 2 the direction, and a seek is therefore expressed as a delta plus a sign rather than an absolute cylinder.

As can be seen the seek operation works on deltas; a seek is done by sending a delta cylinder address and a sign (upwards, downwards) to the drive. Looking at the first and third error I think we see that the test expects the drive to be on cyl 2, and it sends a command with a delta of +2. This should make the drive arrive at cyl 4, and it appears that it doesn't.

The RESULT line shows that we got drive status 5 while we expected drive status 4. The drive status register looks like this:

The drive status word: errors in bits 15 to 09, drive status in 07 to 03, and the drive state in the bottom three bits STC, STB and STA.

The message apparently talks about state, not status, so those bits have the following values:

What those three bits mean. State 4 is seek-track counting, which the test expects; state 5, the one actually returned, is seek-linear mode with lock on.

So, state 5 would be "Seek-linear mode (lock on)", and state 4 would be "Seek-Track Counting". Reading time..