Frequency inputs to the TLG
1..15Hz: Commonwealth Edison 60Hz AC line Phase A in the Mac Room divided by 4.
Used for: 68040 Non-Maskable Interrupt (NMI)
IPA (Non-BSTR/Non-MI) reset output enable
2. 1MHz: Phase locked loop multiple of the Commonwealth Edison 60Hz AC line Phase A
Used for: External rate input to all the delay timers
on the GPIP
|MI Reset Delay to TLG||T:RSTDL1||Z:RSTDL1|
|Non BSTR/MI Reset Delay||T:RSTDL2||Z:RSTDL2|
|Clock event $08 Delay||T:RSTDL3||Z:RSTDL3|
|Clock event $1A Delay||T:RSTD1A||Z:RSTD1A|
|Unused delay channel 4||T:RSTDL4||Z:RSTDL4|
|Unused delay channel 5||T:RSTDL5||Z:RSTDL5|
|Unused delay channel 6||T:RSTDL6||Z:RSTDL6|
3. BMIN: BDOT zero crossing from GMPS
Used for: IPB (BSTR) reset output enable from
|BSTR Reset Delay to TLG||B:RSTDLY||Z:RSTDLY|
D33 Clockscope resolution has been changed since early this year
from 15Hz ticks (67ms) to 1MHz timestamps (1 us). The TLG non-maskable
interrupt is still 15Hz, so a TLG tick is still 67ms. Now, however, the
events are timestamped with 1us resolution. Therefore, the BSTR resets
can be seen to occur at roughly 49ms prior to the Non-BSTR resets. A TLG
tick, then, begins epsilon after the Non-BSTR reset time (also known as
$0F time, for the 15 Hz $0F event) and ends at the $0F. This being the
case, BSTR resets on the very first TLG tick of a supercycle will normally
have negative timestamps of 49ms magnitude.
Enabled on any TLG tick with a $12 (BSTR Prepulse), occurs 3ms after
$0F and 12ms before $12.
BSTR resets ($11, $12, $13, $14, $15, $16, $17, $19, $1C) must occur 2ms before BMIN.
BMIN occurs 20ms after $0F. So, T:RSTDLY is triggered by BMIN and
delayed for 64.654ms, making BSTR resets "pop" at 18ms after $0F, which
is 2ms before BMIN. Booster Beam (BB) resets are $13+$14+$15+$16+$17+$19+$1C.
Null cycles are $11s. Pre-pulse cycles are $12s.
NTF Events ($1A and $1B)
The NTF Priority Request Signal comes from the NTF Interlock box
and is active when the NTF patient treatment switch is on in the NTF control
room. This signal is active low (0=true, 1=false) and is received by the
BSSB "anded" with the BSSB NTF Key (gray) and passed on to the TLG as the
The NTFP bit comes from D69 console application. It is active high
(0=false,1=true). When a module is built, BB resets may be tagged with
NTFP. In the Attributes Window on the Modules subpage, there are two lines
that control this function.
|NTFREQ Status||Percentage Usage by NTF||Action Taken|
|OFF||0%||NTF does not take any cycles|
|ON||25%||NTF takes 1 out of every 4 BB cycles|
|ON||50%||NTF takes 1 our of every 2 BB cycles|
|ON||75%||NTF takes 3 our of every 4 BB cycles|
|ON||100%||NTF takes all BB cycles|
Defined as "NTF Beam Permit", this event actually turns the 58
degree magnet "on" meaning it deflects beam to NTF. The rules for generating
$1A have two parts. If, after a BSTR Beam (BB) reset, there are 5 clock
ticks ahead without a BSTR Beam reset, then $1A is issued. It will be made
to occur at least 2ms , but less than 67ms, after the BSTR beam reset.
Once a $1A has been issued, a 5 sec counter begins decrementing. If the
counter terminates, a check is done to "look ahead" 5 clock ticks. If there
is no BB reset within 5 clock ticks, then the $1A is issued.
Defined as "NTF Beam Inhibit", this event turns the 58 degree
magnet "off", meaning it allows beam to travel undeflected past NTF and
down LINAC. If allowed,$1B happens at least one tick prior to any BSTR
beam reset. Two parameters determine whether $1B shall be inhibited. One
is the NTFP bit from D69. The other is the NTF Priority Request signal
from the BSSB which is active low (0=true, 1=false):
|NTF Priority Request signal||T:NTFPRI||Z:NTFPRI|
The truth table for inhibiting $1B is:
|NTFP=0 (false)||NTFP=1 (true)|
|NTFPRI=0 (true)||Inhibit $1B’s|
If the BSSB Master Key is off or any MSTR switches
are taken (LED on) on the Beam Switch Boxes, then NTF has a Continuous
NTF Beam Permit. This means the NTF 58 degree magnet stays on without regard
to event $1B occurring (NTF gets DC Beam).
If the BSSB Master Key is on and the MSTR switches
are all on (LED off) then the NTFP bit controls whether or not NTF takes
BB reset cycles.
Note: NTF has a direct pulse shifter request connection
which allows NTF to take beam on an $11 (BSTR Null) cycle. This is reflected
on TCLK as event $09, "NTF Permitted Request". Therefore, NTF may be getting
beam on $11s even if there are no BB reset cycles taken.
For TLG testing using TCLK, Linac must be notified
to switch over to LCLK. Before switching to LCLK, make sure that a $1A
was issued. LCLK does contain both $1A and $1B. So the NTF group should
use a pulse generator set up to mimic $1A’s every 5 seconds into their
NTF Interlock Module, while leaving the $1B input to the NTF interlock
MI resets ($20, $21,$23, $29,$2A, $2B,$2D, $2E) until
this time occurred at $0F time. At the request of Phil Martin, the MI resets
can now be delayed from $0F to occur closer to or farther ahead of the
BSTR resets. This will depend on whether the value written is less than
or greater than 18ms, respectively.
Security, Privacy, Legal