Transcript WIENER - CERN Indico

Status of WIENER crates
18/12/13
VF
1
NA62 WIENER crates
• 2 “prototypes” crates purchased in 2011:
FE type: power supply LHC 9U, +/-12V 46A, 48V 13.5A, +3.3V 100A, +5V 100A ,+/5V 30A. The crate is Frascati. Perfectly working after a series of
reprogramming 12V -> 7.0V -> 7.2V ->7.5 V. Used as a test crate for all LAV FE
boards.
o TEL62 type: power supply LHC 9U, 48V 27A, +3.3V 100A, +5V 100A, +/-5V 100A.
The crate is in Rome. No problem reported.
8 crates delivered in 2012 (all FE type)
o Reprogramming of +/- 12V to 7.0V, then to 7.2V and at the end to 7.5V
• At the beginning by NA62
• At the end by Elcotron
o
•
18/12/13
VF
2
Failures story. Fan-Trays
Unit Rack
Problem
0133 Spare
Doesn't react to ON/OFF switch (the crate stays OFF).
0135 KTAG
0138
LAV1
18/12/13
Repair
Paid by
« Dommage mécanique sur le NA62
switch ON/OFF »
All lights are ON as soon as power cable is connected ”Réparation sous garantie”
WIENER
to mains. The fan tray doesn't react on switching the
power supply ON or on ON/OFF switch on front
panel of the fan-tray.
Weak blast with some smoke
Wrongly mounted voltage
WIENER
regulator (too short distance
between cooling tab and
chassis)
VF
3
Failures story. Power-Supplies
Unit
0004
Rack
KTAG
Problem
The fan tray connected to this PSU gets all LEDs ON as soon as the
mains power is connected
0005
?
Cannot be controlled via the crate nor the RS232 line. The user wanted
to set the output voltage to a different value and the PSU stopped to
respond to the UEP6000 SW.
0007
LAV3
The value of 3.3V increases continuously from 3.3 up the the security
value 3.5V and then the crate is switched off by the security system.
0007
LAV3
48 V -> 0 V around 3 min with or without the voltmeter
0009
CHOD There is following sequence of the fan-tray display once the power
cord connected to plug : UEP6000:2.22 , CAN6.1 , Plug $ Play error , PS
incompatible , +5V0 low voltage , +12V0 low voltage , +48V0 low
voltage , +3.3V0 low voltage , -12V0 low voltage and it is impossible to
turn it ON.
LAV1 Plug & Play error
0009
0009
PC
farm
18/12/13
Repair
Echange std de la carte
MUH, d'un ventilateur
de module et du switch
I/O vert.
Réparée la carte MUH
(self L3).
Paid by
WIENE
R
Réparé carte MUH self
18µH grillée
NA62
Changé la self 18µH sur
la carte MUH
NA62
NA62
Replacement of
SYSRESET transistor and
18µH inductor
A 3K3 resistor is installed in a free connector to pull-up the SYSRESET
line to 5V. The voltage measured on the SYSRESET line is 0.3V, with or
without the boards inside the crate.
VF
4
L3 (18µH), EEPROM and SYSRESET
18/12/13
VF
5
SYSRESET and
+12V
Backplane lines
18/12/13
VF
6
SYSRESET investigation
3 suspects:
• LAV FE (usually 5 boards in crate) – Mauro Raggi (Frascati)
• TEL62 (usually 1 board in crate) – Franco Spinella, Elena Pedreschi (Pisa)
• WIENER (Andreas Koester)
18/12/13
VF
7
3 reactions :
LAV FE
All the 26 FEE boards that are now sitting at CERN have been
tested for all their functionalities inside the crate which is
sitting in Frascati for at least 30 minutes without any
problem.
TEL62
The SYSRESET part of the board is a replica of TELL1 design.
We had so far a couple of boards in which SYSRESET receivers
were found blown (protection diodes were not mounted),
the last being the one which was plugged in the crate which
developed problems on such line during the dry run.
Alibi ?
3 participants:
WIENER
18/12/13
Standard WIENR power supplies used in NA62 crate.
The only possible reason for the destroyed transistor is
connection of this SYSRESET signal to some other voltage via
electronics plugged into the backplane
VF
8
Analysis of the “Burned Inductor at MUH” Problem (WIENER)
18/12/13
VF
9
SYSRESET investigation: capacitors on backplane
From: Andreas Koester [mailto:[email protected]]
Sent: 26 November 2013 19:40
To: Vincent Bobillier
There is a capacitor drawn
in the schematic (C50),
between SYSRESET and
another voltage rail. We
double-checked: This
capacitor is not assembled,
which is OK. (My thought
was charging this 470uF
capacitor could damage the
transistor at the MUH).
18/12/13
VF
10
SYSRESET investigation: capacitors on backplane
From: Andreas Koester [mailto:[email protected]]
Sent: 29 November 2013 14:59
To: Vincent Bobillier
The SYSRESET is the trace
from left to right (see arrows left
& right, "Bottom Side.jpg"), and
no capacitors there.
Below the SYSRESET there
are the three smaller traces for
+12V/0V/-12V, which do have
capacitors. You can see there is
no connection to the
SYSRESET
18/12/13
VF
11
Crate validation tool and procedure
• Rationales:
o
Functions provided by a crate are rather simple:
• Voltages/currents
• SYSRESET
o
The functions are static (under assumption of absence of an element failure)
• Procedure
o
o
18/12/13
Prepare a simple test tool (Riccardo – beginning of January)
Check all crates without boards
VF
12
3 suspects :
Exculpatory
evidence
LAV FE
All the 26 FEE boards that are now sitting at CERN have
been tested for all their functionalities inside the crate
which is sitting in Frascati for at least 30 minutes without
any problem.
TEL62
The SYSRESET part of the board is a replica of TELL1 design.
We had so far a couple of boards in which SYSRESET
receivers were found blown (protection diodes were not
mounted), the last being the one which was plugged in the
crate which developed problems on such line during the dry
run.
Alibi ?
3 participants:
WIENER Standard WIENR power supplies used in NA62 crate.
The only possible reason for the destroyed transistor is
connection of this SYSRESET signal to some other voltage
via electronics plugged into the backplane
18/12/13
VF
Validation
13
What about TEL62 (personal proposals)?
• Create and keep record of all boards movement from one crate to
•
•
another
Systematically check all boards for unwanted contacts (with record
of observation) – can Pisa group do it?
Abandon SYSRESET circuit and use crate OFF/ON instead
18/12/13
VF
14
Other direction of investigation: electrical network
Paramètres d’enregistrement des tendances :
• Courants, tensions, Puissances, facteurs de puissance, harmoniques
• Durée : 14 jours
• Intervalle de moyennage/min/max : 10 minutes.
19/07 après-midi
(8 événements)
10/07 @ 13:30 et 13:40
23/07 @ 15:00
20/07 @ 11:50 (coupure
générale CERN)
Noir = phase 1-neutre
Rouge = phase 2-neutre
Bleu = phase 3-neutre
A l’exception des creux
de tension signalés sur la
Figure 1 et commentés ciaprès au § 2.1.3, les 3
tensions simples sont en
moyenne à 228 V, avec
un maximum à 242 V
pour V1 (phase R). Ces
valeurs et n’appellent
aucun commentaire
particulier et ne posent
aucun problème.
Figure 1 — enregistrement des tensions simples des 3 phases (valeurs efficaces)
Conclusion
18/12/13
Les perturbations relevées (creux de tension d’une durée de quelques périodes réseau, ou
déformations oscillatoires de l’onde sinusoïdale, ou pics transitoires rapides) ont une cause
externe au Canalis, à chercher sur la distribution amont (… plutôt des manœuvres
exécutées par un automatisme de régulation)
Un certain nombre de problèmes peuvent également être corrélés aux orages survenus sur
15
VF de mesure.
le secteur de Genève durant la période
What else?
• Connection of all 3 grounds:
o
From: Vincent Bobillier
Sent: 13 November 2013 16:22
“During the last phone conference (on October 23), what I learned
already from Andreas Koester is that he recommends not to run the VME
PSU with the different grounds not connected together. Andreas clearly
said that these PSUs are not designed to operate in such a floating
grounds setup. Thus it is recommended to interconnect these grounds
via the backplane jumpers or via the modules inserted in the VME cardcage”.
• Systematical test of the crates after new board
•
•
insertion (the problem manifests itself only at crate
re-start)
Check of grounding issues (in particular in LAV racks)
….
18/12/13
VF
16