Transcript Issues with CSC`s - Indico

Overview of What Was Done
Last Week…
• Worked with Alex M. to measure ALCT
turn-on curves
– See next slides…
• Obtained values of currents for all onchamber electronics components at point
5
– See later slides…
– Would like to finalize values to send to Valeri
to implement “out of range” alarms in DCS…
28 Jan 2009
G. Rakness (UCLA)
1/15
Overview (cont.)…
• Upgraded online software to make the flagging of
configuration errors depend on the “problem_mask.”
– If (configuration error != configuration error expected) then output
is red
– (Used to be: if (configuration error) then output is red…)
• Turned on full set of CSC’s underground today…
– One OPFC+Maraton not working, to be addressed by Petr
tomorrow…
– Checked firmware on entire system… After Misha’s work, much
firmware will need to be loaded…
• Daily CSC commissioning phone meeting started (again)
– +1-510-665-5437… Meeting ID = 2175
– 17h30 CERN time
– There is also a DCS phone meeting at 17h00, ID = 5566…
• Any interest in a public google calendar for CSC events?
– I already have one, it is just a matter of making it public…
28 Jan 2009
G. Rakness (UCLA)
2/15
ALCT turn-on curves
Executive summary:
•
Features in scope traces understood
•
No “smoking gun” seen in traces to be a likely cause of blowing
fuses on the ALCT
•
At the commissioning phone meeting Monday 26 January 2009,
it was decided to power up the system as soon as we could…
More scope traces and explanations can be found at…
http://www.phys.ufl.edu/~madorsky/alctv/curves-explained.doc
28 Jan 2009
G. Rakness (UCLA)
3/15
1.8V current on ALCT as ALCT is
powered on…
 Blue = 1.8V current
Several features seen as the
ALCT powers up…
1
1) Large current spike
3
5
2) 600msec of small current
draw
3) 200msec of 500mA (full)
current draw
2
4
4) ~30msec of small current
draw
5) Full current draw
Features similar on 3.3V line…
To be explained on subsequent
slides...
28 Jan 2009
G. Rakness (UCLA)
4/15
Powering-On the Peripheral Crate
Recall: there is a specific sequence to the
voltages which are enabled when we turn ON
the peripheral crate…
• Enable power to 1.5V… pause 150ms
• For (TMB/DMB pair 1 to 9)
–
–
•
•
•
Enable pair… pause 150ms
N.B. When DMB powers on, it enables the power to
the ALCT…
Additional pause 150ms
Enable Crate Controller… pause 100ms
Enable MPC/CCB
–
Clock on
28 Jan 2009
G. Rakness (UCLA)
5/15
Feature 1) Current spike during voltage ramp
 Powering on the DMB enables voltage to the ALCT. This charges up
the bypass capacitors on ALCT (and capacitance of floating scope…?)
ME+1/2/15
Yellow: 1.8V voltage (inverted)
Blue: 1.8V current (50mV/div =
5A/div)… peak at ~30A
Time scale = 200sec/div
 From this trace, I2T = 0.18
A2-sec at ALCT power-on…
 Rating of ALCT fuse (A.
Madorsky) = 1.6 A2-sec
 The inrush of current at
ALCT power-up is probably
not causing fuses to blow…
(similar for 3.3V line)
28 Jan 2009
G. Rakness (UCLA)
6/15
Features 2) and 3) Time between spike and large
current
Feature 2) The FPGA is idle until the CCB is powered on.
Feature 3) When the clock is powered on, the FPGA programs…
ME+1/3/28
Yellow: 3.3V voltage
Blue: clock at ALCT
 The time between ALCT
power-on and the clock
depends on where the DMB
is located within the crate…
(The noise on 3.3V voltage
curve [coinciding with clock]
is due to the ungrounded
scope.)
28 Jan 2009
G. Rakness (UCLA)
7/15
Feature 2) (detail) Hard resets seen as each
TMB/DMB pair is powered on…
•
FPGA hard-resets are seen as each TMB+DMB
pair is powered on…
ME+1/2/15
Yellow – 3.3V power
Blue – hard reset
Magenta – INIT
Green – DONE
Note: all resets come when
there is no clock, so they
have no effect…
28 Jan 2009
G. Rakness (UCLA)
8/15
Features 4) and 5): after FPGA programmed,
~30msec current dip followed by full current
•
Dip in ALCT core current from ALCT FPGA reprogramming due
to hard reset sent from CCB after its FPGA is programmed
ME+1/2/15
Yellow – 3.3V power
Blue – hard reset
Magenta – INIT
Green – DONE
Note: This is an intentional
feature of the CCB
firmware…
28 Jan 2009
G. Rakness (UCLA)
9/15
All hard reset and clock features go away if
DMB fully controls power-on…
ME+1/3/28
Blue: 3.3V current
Note: regular FPGA power
up and configuration. The
first ~30 ms of low power
consumption is
configuration time, then
normal operation. No
strange artifacts as seen
with CRB turn on.
Is it worth the effort to
rework the power-on
sequence to achieve this?
Probably not…
28 Jan 2009
G. Rakness (UCLA)
10/15
Distributions of Current for OnChamber Components
Goal: to determine values beyond which the
Detector Control System (slow controls) should
raise an alarm
28 Jan 2009
G. Rakness (UCLA)
11/15
On-chamber Currents  ALCT
Data from LVDB read through DMB using XMAS monitoring (J. Liu)
 Values consistent with values in DCS database (X. Yang)
ALCT288
ALCT384
Red lines indicate (update to) proposed high/low alarm set points for DCS…
28 Jan 2009
Rakness (UCLA)
(1.8V current hasG.been
lowered from 0.4A  0.3A)
12/15
On-chamber Currents  ALCT and CFEB
ALCT672
CFEB
… if current limits for ALCT are OK, I will send to V. Sytnik to
implement after this meeting…
28 Jan 2009
G. Rakness (UCLA)
13/15
13 Chambers Outside of ALCT Limits
• 4 cases: ALCT's which have been or will be replaced
– either from a blown fuse or some other problem...
• 4 cases: exhibit the “smoking gun” of an EEPROM losing its
program, i.e., a bit going from 1  0
– 1.8V current ~0.1A
• 4 cases: looks like a "bad read"
– read 0 on both 1.8V and 3.3V line
– I believe we should not error on these, but read the values again... OR,
we need to read the ALCT ADC as well as the LVDB currents...
• 1 case: not known
– Low (non-zero) values on both 3.3V and 1.8V lines
– Maybe we need to read the ADC as well as the LVDB currents...
Oftentimes, when it is bad on one line, it is bad on the other:
• ALCT288: 5 chambers bad... 3 bad on both 3.3V and 1.8V criteria
• ALCT384: 4 chambers bad... 2 bad on both 3.3V and 1.8V criteria
• ALCT672: 4 chambers bad... 2 bad on both 3.3V and 1.8V criteria
28 Jan 2009
G. Rakness (UCLA)
14/15
To do
• Continue towards getting a cosmic ray run
going at point 5
• Attempt to measure time-of-flight in beamhalo or CRAFT data…
• Upcoming talks:
– CSC Trigger Primitives at L1 Trigger Online
Software Review (Thursday 5 Feb.)
– RAT results at RPC Detector Performance
Group meeting (Tuesday 3 Feb.)
28 Jan 2009
G. Rakness (UCLA)
15/15
Backup
28 Jan 2009
G. Rakness (UCLA)
16/15
Probe Measurement
Locations
(ALCT288 schematic)
Curves measured with floating scope…
= voltage measurement drop points
( current)… Current curve
behavior seen by measuring
voltage drop across (fuse+10m
resistor) at power-supply input
with 20MHz bandwidth filter
= voltage measurement point…
Voltage curves measured at powersupply input relative to ground of
chamber (full bandwidth)
28 Jan 2009
G. Rakness (UCLA)
17/15
3.3V current when power-on the
DMB (ALCT)
ME+1/3/28
 3.3V current spike with CRB
turn on
Yellow: 3.3V voltage (inverted)
Blue: 3.3V current (50mv/div =
5A/div)
Similar behavior as on
the 1.8V line…
28 Jan 2009
G. Rakness (UCLA)
18/15