Transcript Software - Indico

Measurements of Voltage Oscillations
after a Fast Power Abort in Sector 67
during the Christmas stop 2010-11
Emmanuele Ravaioli – TE/MPE
Thanks to A. Verweij, F. Formenti
24 November 2010
Main goals of the measurements
1.
Understanding the behavior of the circuit during the power-converter shut
down, with particular attention to the oscillations of the aperture voltages,
and their difference.
In each sector, about 60% of the dipoles present an unexpected behavior: during
the PC shut-down, U_QSO reaches values ~5 times larger than the other magnets
and than the values calculated by Pspice simulations.
2. Assessing the voltage across the switches (after the installation of the
snubber capacitor).
The proposed measurements are required because the oldQPS and newQPS have
a sampling frequency that is too low (200 Hz and ~1 kHz) to properly analyse the
transitory.
It is important to understand the unexpected behavior of U_QSO, since this issue
concerns about 60% of the magnets and the analysis of the present data shows that
it is not related with the electrical or physical position of the magnets, nor with the
magnet and cable manufacturers, nor with the chronology of installation.
Emmanuele Ravaioli
2
Scheme of the measurement system
0.3 km
:100
PC
:100
SW8
Arc 67
USB
PC
A
40 m
A
A
A
A
Card
NI USB-6251-BNC
1 MS/s
Isolation
Transformer
PC
Wi-Fi
Isolation
Transformer
VDSL
Wi-Fi
2.7 km
SW1
Differential
Probe
Card
NI USB-6251-BNC
1 MS/s
USB
USB
Wi-Fi
A
3 Crates hosting
2x3 Isolation amplifiers
AD210
Differential
Probes
Card
NI USB-6251-BNC
1 MS/s
25 m
Point 7
:100
Point 6
VDSL
Wi-Fi
PC
Wi-Fi
Isolation
Transformer
VDSL
Wi-Fi
Safety issues
•
•
•
•
•
•
•
No magnet will be touched; the measurements will be performed accessing to
the QPS cabling (only QPS & ElQA experts will touch the cabling).
The NI card measuring the magnets will be isolated from the magnets
themselves by means of insulator amplifiers.
Fuses have been installed in the magnet patches, to protect the IFX cable in
case of a short circuit inside the patch.
The three measurement systems will be fully floating; this will be assured by
insulator transformers.
The full commissioning of the circuit up to 6 kA and an ElQA test are required
before the tests. It is therefore important to install the equipment before the
ElQA test already planned for the circuit (after January 20th ?).
No ElQA test is required after the measurements (Nuria Catalan confirms),
provided the items are uninstalled with the proper carefulness.
All the equipment will be fully tested for 1.9 kV before the installation in the
tunnel because it will have to sustain the ElQA test.
Emmanuele Ravaioli
4
Status
•
All the required material is already available.
•
In the past weeks, the material has been tested:
•
•
•
•
Patches for connection to the QPS cables of the magnets (fuses added to improve
safety)
NI Cards
Insulator amplifiers & Crates
Wi-Fi modules (tested in the tunnel)
•
Yesterday (17 November 2010) the first test of acquisition with the software-based
triggering have been successfully performed.
•
A full test of the acquisition system is foreseen in the PowerHall after the campaign for
the snubber qualification.
Persons involved
System architecture: E. Ravaioli; G. J. Coelingh; A. Rijllart
Hardware preparation & testing: E. Ravaioli; F. Formenti; Fabienne Boisier; Gaelle Dib
Installation in the tunnel: E. Ravaioli; Mathieu Favre
Software: Odd Oyvind Andreassen
Other contributions: F. Formenti; K. Dahlerup-Petersen; N. Catalan Lasheras; B. Panev; R. Denz;
G. D’Angelo; H. Thiesen; M. Solfaroli; A. Verweij
Emmanuele Ravaioli
5
Planned measurements
Signal:
Triggering
Time frame
( t=0 at triggering)
Frequency of
acquisition
Voltage range
Vout of PC
Software
0s – 5s
500 kS/s
-160 V – +160 V
Software
0s – 5s
150 kS/s
-10 V – +5 V
Software
0s – 5s
150 kS/s
-10 V – +5 V
Software
0s – 5s
150 kS/s
-10 V – +5 V
U_RES switch pt6
Software
0s – 5s
500 kS/s
+0 V – +900 V
U_RES switch pt7
Software
0s – 5s
1.25 MS/s
+0 V – +900 V
U1 and U2 of magnet
MB.B9R6
U1 and U2 of magnet
MB.A10R6
U1 and U2 of magnet
MB.A11R6
Assembly name
Top assembly name Electrical numbering Distance from IP6
MB.B9R6
LBBRS.9R6
# 153
331.97
MB.A10R6
LBARL.10R6
#3
356.775
MB.A11R6
LBARA.11R6
#4
395.84
Emmanuele Ravaioli
Relative distance
~ 25 m from MB.B10R6
~ 39 m from MB.B10R6
6
Planned tests
A set of tests are planned in Sector 67 in order to clarify the odd behavior observed
in some of the main dipoles.
(Additional info: Analysis of the QSO signals after PC shut down at different current levels - ver2 )
At least four tests will be launched; in each test, the current of the RB circuit will be
raised with a dI/dt = 10 A/s. When I_max is reached (while the current is still
raising), the power converter will be shut down (triggered by the QPS); after 350
ms, the first switch (pt7, middle of the chain) will be opened; after another 250 ms,
the second switch (pt6, close to the power converter) will be opened as well.
The values of I_max in the four tests are: 1 kA, 2 kA, 4 kA, 6 kA (actually a bit below
6 kA, because the PC slows down at the end of the ramp).
If there is sufficient time possibly also tests with I_max = 3 kA and 5 kA can be
performed. Another set of tests with dI/dt = 0 A/s (no ramping up) would also be
interesting, at the same current levels. NB: Every test above 2 kA requires a sensible
time due to the cooling down of the switches.
(Additional info: \\cern.ch\dfs\Users\e\eravaiol\Public\Planning of the measurements in the tunnel
during Christmas stop 2010-2011\ )
Emmanuele Ravaioli
7
Schedule of the tests
Activities
Time
Comment
Test of the equipment items
Oct-Nov 2010
NI cards, isolation amplifiers, Wi-Fi modules, …
Test of the acquisition software
Nov 2010
Installation and testing of the equipment in the PowerHall
Installation of the equipment in Sector 67
Testing of the equipment in Sector 67
Dec 2010
after 7 Jan 2011
after 7 Jan 2011
Commissioning up to 6 kA and ElQA test of the circuit
after 20 Jan 2011
Tests from the Control Room
after 20 Jan 2011
During the night
Tests to be performed (FPA via QPS)
Duration
Comment
First check test (Ramp to 1kA, 10 A/s)
15 min
Test + PM buffer receiving
Preliminary data analysis (to check every signal is fine)
Ramp to 2 kA, 10 A/s
Ramp to 4 kA, 10 A/s
Ramp to 6 kA, 10 A/s
30 min
30 min
1h
1.5 h
(NI cards, oldQPS, newQPS)
Test + PM buffer receiving
Test + PM buffer receiving + Switch cooling
Test + PM buffer receiving + Switch cooling
Ramp to 3 kA, 10 A/s
Ramp to 5 kA, 10 A/s
Ramp to 1 kA, 0 A/s
Ramp to 2 kA, 0 A/s
45 min
1.25 h
15 min
30 min
Test + PM buffer receiving + Switch cooling
Test + PM buffer receiving + Switch cooling
Test + PM buffer receiving
Test + PM buffer receiving
Ramp to 4 kA, 0 A/s
Ramp to 6 kA, 0 A/s
1h
1.5 h
Test + PM buffer receiving + Switch cooling
Test + PM buffer receiving + Switch cooling
Total time
9h
Emmanuele Ravaioli - TE-MPE
After snubber campaign is finished
Before commissioning and ElQA!
Acquisition of the signals from a power supply
8
Annex
Emmanuele Ravaioli
9
Magnet choice
The three magnets have been selected based on the behavior they showed during past tests:
Magnets 3 and 153 showed a U_QSO signal during the PC shut-down transitory at 2 kA bigger than
that at 6 kA.
Magnet 4 showed a regular behavior, but has been selected to highlight the difference as compared
to the other two magnets.
Besides, the three magnets are physically close each other, and this will simplify the operations and
reduce the number of the required insulator transformers.
Emmanuele Ravaioli
10
Presentation of the issue (1)
Since the oscillations in the QSO signals (difference between voltage across the
apertures of each magnet) are caused by a voltage wave travelling throughout the
magnet string, one would expect that larger voltage waves cause larger QSO signals.
On the contrary, during the PC shut down (before switch openings), some magnets
present bigger values of the QSO signal at lower current (2 kA), when the voltage
wave is slightly smaller.
Magnet 001  Blue
Magnet 154  Red
Magnet 001  Blue
Magnet 154  Red
Emmanuele Ravaioli
11
Presentation of the issue (2)
Looking closer to the results, one can notice that not every magnet of the string
exhibits the same behavior: about half of the magnets show a smaller QSO signal at 2
kA, as one would expect, while the other half presents values 5-6 times larger.
In the plots below one can observe the measurements of the same experiment
shown in the previous slide: on the left only the normally-behaving magnets have
been plotted; no the right only the oddly-behaving ones.
Magnet 001  Blue
Magnet 154  Red
Magnet 001  Blue
Magnet 154  Red
Emmanuele Ravaioli
12
Max QSO before switch opening – 2 kA
The PSpice model simulates well only the behavior of the ‘normal magnets’
Emmanuele Ravaioli
13
Max QSO before switch opening – 6 kA
The PSpice model simulates well only the behavior of the ‘normal magnets’
Emmanuele Ravaioli
14