Checkout After Repair; Flt Documentation - IBM System/360 2050 Maintenance Manual

Processing unit
Table of Contents

Advertisement

3.
If
any input is bad, follow its inputs until
a point is found with all inputs good. This
point is in the failing net, and the card or
some other point in the net is bad.
c.
If
there are no G/F points listed and no cards
listed, or if none of the G/F points agree
with Scopex, refer to the logic and scope the
dead entries.
Checkout After Repair
After effecting an FLT repair, rerun all FLT's to
be sure of the fix.
Use storage ripple (all ones) to
put all good parity in storage (FLT's use both good
and bad parity), and neutralize all FLT controls.
FLT DOCUMENTATION
This section describes documentation presently
available for the Model 50 fault locating tests.
MS Mode Hardcore: Documentation consists of a
program listing (FLTOl) of the tests as they are in
storage. This includes the op code, data, function,
and IAR value. There is also a table provided for
SAR and/or IAR failures.
MS Mode ROS Bit Tests: Documentation consists of
a listing (FLT04) providing samples of the first two
tests of record 1 and an explanation of how to deter-
mine the starting address of a failing test. This in-
cludes op code, data, function, and IAR and SDR
values.
ROS Mode Hardcore: Documentation consists of a
program listing (FL T02) of the tests as they are in
storage. This includes data, function, and IAR
values.
ROS Mode Zero-Cycle Tests: The zero-cycle test
documentation is shown in Figure 37. The first
column is the test number in hex. This is the num-
ber displayed in SDR bits 20-31 on a termination.
The second column contains two characters
which show the condition the trigger should attain
for the test. The first character indicates whether
the trigger should be set (S) or reset (R) by the test,
and the second character tells whether the scope
point given should be an up (1) or a down (0) level.
The only combinations that appear are RO and Sl.
The third column gives the pin at which the
value (0 or 1) should be observed; the fourth column
provides a logic page reference to the trigger being
tested, and the fifth column gives the name of the
trigger.
Test
Desired
Logic
Trigger
No.
Result
Test Point
Page
Name
0120
RO
01 A-C4M4D12
RR071 AA4
R Reg 28
012E
Sl
012F
RO
0130
RO
01A-C4M4B12
RR071 AB4
R Reg 29
0131
SI
0132
RO
0133
RO
01A-C4M4JIO
RR071 AC4
R Reg 30
0134
SI
0135
RO
0136
RO
01 A-C4M4GOB
RR071 AD4
R Reg 31
0137
Sl
0138
RO
0139
RO
01 A-C3H2D06
RROOI AB4
M Reg 0
013A
Sl
0138
RO
013C
RO
01 A-C3H2D05
RMOOl AC4
M Reg 1
0130
Sl
013E
RO
Example: Stop at 133 indicates R register (30) does not reset off.
Stop
at
134 indicates R register {30) does not set on.
Stop at 135 indicates R register (30) does not reset off ofter setting on.
FIGURE 37. ZERO-CYCLE TEST DOCUMENTATION
ROS Mode One-Cycle Tests (Scopex): The one-cycle
test documentation (Scopex) is an automatically pro-
duced FLT document that allows conversion from
the pattern displayed on the console lights to the
suspect cards or nets. Scopex consists of a series
of lists, each referring to one particular test. Lists
are separated from the identification number of the
test by asterisks. Each list consists of several
lines of information, each line referring to a parti-
cular pin or net in the machine. A typical list is
shown in Figure 38. When a test sequence stops,
the hex number displayed in bytes two and three is
the segment number and test number within the
segment.
The term "tree" includes all nets encountered
when tracing back from a trigger and continuing
until each path reaches either another trigger or
another entry to the combinational logic; e.g., 1/0
lines, console switches, etc. Entries to the trees
are identified by the word "entry", if they are
triggers which can be set to either state by the
scan-in logic.
If
entry points are always at the
same state at the conclusion of scan-in, they are
"dead" entries.
Each line of the list refers to a net in the ma-
chines which,
if
it does not assume the value which
the input pattern is trying to force it to, will cause
the test to fail. These nets are given in the fourth
column (net) of the list. The first line (Ref-AA) in
each list corresponds to the output point, the next
lines to the nets feeding this output point, and so on
back through the logic to the entries. The pin at
which each net can be observed is shown in the third
column (Pin). Generally, only the nets having pins
are printed on the list because the failure of a net
internal to the card will nearly al ways propagate and
give the wrong value on a pin. The second column
Maintenance Features
(3/71)
61

Advertisement

Table of Contents
loading

Table of Contents