Transcript Document

QPLL Status – December 2004
Paulo Moreira
People
S. Baron1
H. Furtado1,
R. Haeni2,
A. Marchioro1,
P. Moreira1,
J. Parsons3,
S. dalla Piazza2
and
S. Simion3
1) CERN,
2) Micro Crystal
3) Nevis
Paulo Moreira
http://www.cern.ch/proj-qpll
2
Outline
• QPLL jitter problem
– Crystal activity dips
– Power reduction network
• Circuit
• Layout
• Irradiation tests
• TTCrq
Paulo Moreira
http://www.cern.ch/proj-qpll
3
QPLL Jitter Problem
• August 2004:
– Stefan Simion found that at some frequencies an temperatures
the QPLL jitter was largely exceeding the typical values.
• September 2004
– Both Nevis and CERN worked quite hard at the problem but all
the hypotheses were either rejected or difficult to confirm.
• October 2004
– With the help of Micro Crystal it was possible to confirm that
the problem was due to activity dips in the crystal due to
excessive power driving.
• October/November 2004
– During this period work was done to find a simple and effective
way of reducing the power delivered to the crystal.
Paulo Moreira
http://www.cern.ch/proj-qpll
4
Activity dips
Activity dips are due to
vibration modes that are
mechanically coupled to the
fundamental resonant mode.
– The fundamental mode is a
thickness shear motion while
modes causing activity dips
are not.
– These modes can have
frequencies quite close to the
fundamental mode and are
very dependent on
temperature.
– They can thus interfere with
the fundamental mode
distorting the electrical
characteristics of the crystal
near the resonance.
Paulo Moreira
Excess Jitter Points for crystal xtal MC A
40.084
40.083
40.082
40.081
Frequency [MHz]
•
40.08
40.079
40.078
40.077
40.076
40.075
40.074
-10
0
10
http://www.cern.ch/proj-qpll
20
30
Temperature [C]
40
50
60
70
5
Power reduction network
Excess Jitter Points for crystal xtal super A MC
40.084
40.083
40.082
Frequency [MHz]
40.081
40.08
40.079
40.078
40.077
40.076
R1 = 62 W, R2 = 240 W, C = 10 nF
40.075
40.074
-10
Paulo Moreira
0
10
http://www.cern.ch/proj-qpll
20
30
Temperature [C]
40
50
60
70
6
Power reduction network
• Parasitic capacitance on “N1”:
• Increases the power drive
• Pulls the resonance frequency
• Must be minimized
• As far as we can tell:
• Activity dips are eliminated
• QPLL jitter and centre
frequency remain basically
unchanged.
Paulo Moreira
http://www.cern.ch/proj-qpll
7
Power reduction network - Layout
0805
Paulo Moreira
http://www.cern.ch/proj-qpll
8
Irradiation Tests
• Irradiation tests made in Boston by:
– Sefan Simion and John Parsons
• Proton beam:
– Energy: 160 MeV
– Fluence: 2.3 to 2.5 1013 p/cm2
• QPLL3
• Quartz crystals:
– Micro Crystal
– Conner Winfield
• (Accelerated aging tests being done at CERN)
Paulo Moreira
http://www.cern.ch/proj-qpll
9
Irradiation
cw
Paulo Moreira
http://www.cern.ch/proj-qpll
MC
cw
10
Irradiation
cw
MC
cw
Jitter higher than what is measured normally in the
lab due to experimental conditions and instruments used
Paulo Moreira
http://www.cern.ch/proj-qpll
11
TTCrq
• New TTCrq on the drawing board
– Introduction of the power network
– Pin J2 - 39 will become a +2.5 V power input
• For cards working in a radiation hard environment that can not use
the internal 2.5 V regulator
• Optional 0 W resistor for 100% compatibility with the previous
version
– Optional 100 W internal terminations for the LVDS signals
Paulo Moreira
http://www.cern.ch/proj-qpll
12